i

On the non-creole basis for Afro-Caribbean Spanish

John M. Lipski
University of New Mexico {1993}

0. Introduction
The claim has frequently been made by scholars that Africans in the Spanish Caribbean,
deprived of the opportunity for learning the prevailing dialects of Spanish, developed a Spanish-based
creole. The most ambitious theories assert that this Afro-Hispanic language permanently affected other
varieties of Caribbean Spanish. An accompanying claim is that the postulated creole was built upon an
earlier Afro-Portuguese trade jargon, pidgin, or even fully nativized creole, originally formed on the
coast of Africa, and learned in the African slaving stations and on slave ships crossing the Atlantic. The
present study injects a note of caution, after considering a wider corpus of Afro-Hispanic materials than
has heretofore been examined. It is shown that most recurring features of Afro-Hispanic language are
common to second-language learners, and that only two or three features link Afro-Caribbean Spanish
to creole languages. Moreover, all the features in question come from a small group of 19th century
Cuban and Puerto Rican texts. Closer examination of the texts, together with an inquiry into the
demographic shifts of black slaves and laborers in the 19th century Caribbean, suggests that the most
creole-like features of Afro-Caribbean Spanish may have been acquired from speakers of other
Caribbean creole languages, especially Papiamento. In other cases, spontaneous blending of Spanish
and African languages yielded combinations which bear a superficial resemblance to creole structures.
One of the most interesting chapters in the history of Latin American dialect differentiation is the
African contribution. There exists a tantalizing corpus of literary, folkloric and anecdotal testimony on
the earlier speech patterns of Afro-Hispanics, in Spain and Latin America. The greatest obstacle in the
assessment of earlier Afro-Hispanic language is the high level of racial prejudice, exaggeration and
stereotyping which has always surrounded the description of non-white speakers of Spanish, and which
attributes to all of them a wide range of defects and distortions that frequently are no more than an
unrealistic repudiation of this group. One group which did use a distinctly Afro-Hispanic language were
the bozales, a term referring to slaves born and raised in Africa, who spoke European languages only
with difficulty. Bozal language, a halting approximation to Spanish or Portuguese typical of first-
generation immigrants, first arose in the Iberian Peninsula late in the 15th century; the earliest attestations
come from Portugal. Bozal Spanish makes its written appearance in Spain early in the 16th century,
and continues through the middle of the 18th century. Latin American bozal Spanish was first described
by writers like Sor Juana Ins de la Cruz, who imitated the speech of black slaves (brought from Puerto
Rico) in Mexico at the beginning of the 17th century. Other surviving documents from the 17th century
demonstrate the existence of bozal Spanish in Peru, Bolivia and Central America. Few documents
representing Afro-Hispanic speech remain from 18th century Latin America; Cuba and Mexico are
among the regions so represented. Beginning at the turn of the 19th century, the last big surge of slave
trading, spurred by the sugar plantation boom and by increased urbanization of many coastal regions,
resulted in an outpouring of literary representations of bozal Spanish. The geographical distribution of
extant Afro-Hispanic texts mirrors the profile of the African slave trade in Latin America. The 19th
century texts come principally from three regions: Cuba, coastal Peru, and the Buenos
Aires/Montevideo region (cf. Lipski 1986a, 1986b). Only the Cuban texts, however, together with a
handful of Puerto Rican examples, have been offered as evidence of an Afro-Hispanic creole.
Due to a number of influential studies, the prior existence of a Spanish-based creole in the
Caribbean has become widely accepted, and the hypothesis that an earlier Afro-Portuguese creole,
 ii

such as found in Cape Verde, Annobn, and So Tom enjoys nearly as much support. The number of
Afro-Caribbean texts which have been offered as evidence is, however, very small: fewer than a dozen
sources, among the more than one hundred available poems, plays, novels, travel narratives, and
anthropological documents which describe the speech of Africans in Cuba and Puerto Rico. Upon
considering a wider range of data than has previously been used to state the case, a different picture
emerges. The 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 creole. The present study, while not totally rejecting the possibility that
some Africans taken to Latin America may have spoken a Portuguese pidgin, nor that Afro-Hispanic
modalities may have occasionally stabilized, demonstrates the fragility of the evidence upon which such
claims are based. By considering a wider corpus of bozal texts, the idiosyncratic and
non-representative nature of some `key' texts which form the centerpiece for the creole theories is
revealed. In the balance, a small number of unique texts has been given disproportionate importance in
the evaluation of Caribbean bozal Spanish.

1. Bozal phenomena of non-creole origin

The majority of Afro-Hispanic linguistic phenomena can be accounted for without recourse to
controversial theories, and suggest only a gradual approximation to regional varieties of Spanish. The
rudimentary attempts of adult Africans to learn Spanish resemble the Spanish still used non-natively in
several parts of the world. Africans of a particular language background might lend individualizing
touches to this broken Spanish, but in most cases only the lowest common denominators would be
expected. By taking a cross-section of West African language families known to have been represented
among the Africans taken to Spain and then to Spanish America in the first few centuries of the slave
trade, some elementary predictions can be made, all of which are borne out in literary representations of
pidgin or bozal Spanish (cf. Lipski 1986a, 1986b).
Due to substantial grammatical differences between Spanish and West African languages, we
would predict little attention to inflectional endings in bozal Spanish, together with simplification of
pronouns, articles and prepositions, and use of basic strategies like repetition to achieve specific
grammatical purposes. Since bozal Spanish represents either the halting attempts of language learners,
or the minimal communication strategies adopted by captives forced to use a foreign language, one
would expect simple sentences, with minimal subordination or conjunction. A perusal of bozal texts
yields a high percentage of structures which need no special explanation, either in terms of theories
claiming access to universal aspects of language structure, or as regards the eventual creation of a
uniform dialect of `black Spanish.' If such considerations could exhaustively account for Afro-Hispanic
texts, we could simply stop here, by claiming that the situation is similar to that of immigrant groups
elsewhere in the world: first-generation Africans who learned Spanish imperfectly spoke a pidginized or
bozal form of the language, possibly passing it on to their immediate offspring, but subsequent
generations learned Spanish natively, and the ethnic background left no traces other than occasional
vocabulary items. This model adequately accounts for Afro-Hispanic language in Spain, Peru, Mexico,
and the Ro Plata area.

2. Portuguese elements in early bozal Spanish

More difficult to tease out of the reconstruction of Africanized Spanish is the Portuguese
contribution, in particular the input of stable Portuguese-based pidgins or creoles. During the first two
 iii

centuries of the slave trade, Spain acquired the majority of its slaves from Portuguese traders, and due
to the nature of the Portuguese slaving empire, some of the Africans had acquired a Portuguese pidgin
before being transferred to other regions. The Portuguese maintained feitorias or slave depots in
Angola, So Tom, Fernando Poo, Cape Verde, Annobn and later Brazil, in addition to supplying
some of the market from Africans already resident in southern Portugal. Pidgin Portuguese sprang up as
a coastal African lingua franca (cf. Naro 1978), at one time stretching from Senegal, around the Cape of
Good Hope, along the coast of India and reaching as far as Hong Kong and Indonesia; this type of
speech may also have been used by slaves who spoke mutually unintelligible African languages, although
the claims that such populations were deliberately chosen to minimize uprisings have been overstated.
A comparison of texts purporting to represent early pidgin Portuguese reveals some consistent
features, nearly all of which made their way into Portuguese-based creoles in Africa, as well as some
Latin American creoles. Some recurring tendencies include:

(a) use of (a)mi as first person singular subject pronoun;

(b) confusion of ser and estar, eventually giving rise to the hybrid form sa (and occasionally
santar) for all persons and tenses;
(c) use of vos as the second-person subject pronoun;
(d) use of bai from the third person singular form of ir in Portuguese, as the verb for `to go.'

The earliest Afro-Hispanic texts generally follow the same patterns as the Portuguese examples and at
first the direct imitation of Portuguese writers is evident. Once literary `black Spanish' became
established in Golden Age drama, the linguistic characteristics move sharply away from pidgin
Portuguese, and acquire traits typical of Spanish `foreigner talk,' together with considerable phonetic
deformation. After 1550, use of (a)mi as subject pronoun rapidly disappears (Lipski 1991), as does
use of bai/vai for `to go.' Unstable gender and number assignment remain, as do incorrectly conjugated
verb forms, although use of the uninflected infinitive becomes increasingly rare. Confusion of ser and
estar is still found from time to time, augmented by use of sar, and loss of the copula occurs
sporadically.

3. Portuguese pidgin as the basis for Latin American bozal Spanish?

By all indications, any pidgin Portuguese component in bozal Spanish had all but disappeared
by the time of the first appearance of Afro-Hispanic language in Latin America. The 17th century texts
from Mexico,1 Peru and Puerto Rico, as well as the 18th century Mexican and Cuban bozal specimens,
show no Portuguese features. Most 19th century Latin American examples, including the extensive
corpus from Peru, Argentina and Uruguay, are similarly nondistinctive in their reproduction of Africans'
approximations to regional varieties of Spanish, and do not point to recurring features which cannot be
explained as spontaneous independent developments or as natural learners' errors.
Although most bozal Spanish texts from Latin America do not exhibit strong resemblances with
Afro-Iberian creoles, several scholars noted that certain texts from 19th century Cuba and Puerto Rico
showed striking similarities with Papiamento, Cape Verde creole and Palenquero, which could not be
attributed to mere chance or predicted from the pidginized Spanish of earlier examples. Among the
more striking parallels are verb forms based on the particle ta plus a stem derived from the Spanish
infinitive, and use of an undifferented third person pronoun for masculine and feminine referents. A
number of other traits of Caribbean bozal Spanish, to be surveyed below, were also cited as evidence
of creole origins for Afro-Hispanic pidgin.
 iv

One of the earliest investigators to link Caribbean bozal Spanish to an earlier Afro-Lusitanian
pidgin was Wagner (1949: 101):

I negri "bozales", cio frescamente importati, sapevano pi o meno il negro-portuguese

come si parlava sulla costa occidentale dell'Africa, e lo cambiarono in un primo tempo,
come naturale, con uno spagnolo rudimentale, che rassomigliava molto alle lingue
creole; usavano la terza persona del singolare dei verbi come forma generale ...
scambiavano i generi ... e non sapevano separare bene le parti del discorso ...

Wagner's comments lay fallow until the development of more ambitious theories of creolization and
monogenesis. The first attempt to document the presence of an Afro-Lusitanian pidgin among bozales
in Latin America was Granda (1970)'s analysis of the observations of Sandoval (1956: 94). The latter,
a (Spanish-born) Peruvian priest resident in Cartagena de Indias, remarked in 1627 that African slaves
from So Tom spoke `con la comunicacin que con tan brbaras naciones han tenido el tiempo que
han residido en San Thom, las entienden casi todas con un gnero de lenguaje muy corrupto y
revesado de la portuguesa que llaman lengua de San Thom ...' The reference to some sort of
Portuguese-based pidgin or creole is clear, but the implication that slaves from other regions also
acquired an Afro-Lusitanian pidgin is not, since Sandoval's quote continues: `... al modo que ahora
nosotros entendemos y hablamos con todo gnero de negros y naciones con nuestra lengua espaola
corrupta, como comnmente la hablan todos los negros.' Although Granda interprets Sandoval's
observations to mean that an Afro-Lusitanian pidgin formed a substrate for ALL Afro-Hispanic language
in Cartagena (and by extrapolation, elsewhere in the Spanish American colonies), the final sentence
seems to indicate the opposite, namely that even Africans speaking the `lengua de San Thom'
eventually acquired bozal Spanish. In any event, by the end of the 18th century, when the critical
Caribbean bozal texts appear, Spain was acquiring few slaves from the Portuguese depot on So
Tom.
The perceived similarities between Caribbean bozal Spanish and Afro-Iberian creoles gave rise
to the claim that an Afro-Lusitanian creole once existed in ALL of Latin America, or at least in the
Caribbean region. The same similarities were also used to bolster `monogenetic' theories of creole
formation, which postulate that an original Portuguese-based maritime pidgin or lingua franca was
relexified and coalesced to form creole dialects of English and French in the Caribbean and Africa;
Spanish and Portuguese in Asia; and Dutch in the West Indies, Guyana and possibly South Africa (cf.
Naro 1978; Thompson 1961; Whinnom 1965, and the references therein). The importance of
reconstructing Afro-colonial Spanish thus rises enormously, for far-ranging issues are at stake. In its
most radical form, this theory claims that a SINGLE creole underlay virtually all Afro-Hispanic speech
over a period of more than three centuries, and therefore was more important than the strictly African
element in determining the characteristics of bozal Spanish and its possible repercussions in general
Latin American Spanish. This hypothesis is clearly stated, e.g. by Granda (1976: 5-6):

Los esclavos negros establecidos en diferentes reas de la Amrica espaola desde el

siglo XVI al XIX manejaron, primeramente junto a sus hablas africanas aborgenes y
posteriormente con carcter exclusivo o al menos dominante, un cdigo lingstico
criollo. Esta modalidad de lenguas fue evolucionando ... hacia el espaol subestandard
de las diferentes zonas hispanoamericanas en que el fenmeno se produjo, a travs de
un continuum pos-criollo ...
 v

This creole in turn had its origins in an Afro-Lusitanian pidgin, developed in West Africa (e.g. Granda
1976: 8):

Las modalidades del criollo desarrollado y empleado en las diferentes zonas

hispanoamericanas de poblacin negra derivaron, gentica y por lo tanto
estructuralmente, del ... protodiasistema criollo portugus de Africa que constituy la
base de la cual, por diferentes procesos de relexificacin ... se originaron aqullas.

Although Granda (1976) includes data from Afro-Hispanic groups in Colombia and elsewhere in the
Caribbean region, he concentrates his remarks on similarities involving Colombian Palenquero,
Papiamento, and purported bozal creoles in 19th century Puerto Rico (cf. also Granda 1968) and Cuba
(cf. also Granda 1971, Perl 1982, 1985, 1987, 1989b). Megenney (1984, 1985a) draws together the
triad Palenquero/Papiamento/Afro-Cuban creole to claim a common origin in an Afro-Lusitanian pidgin.
Perl (1989a) also provides comparisons which tend to support this view. Alvarez Nazario (1974)
supports the hypothesis of a Portuguese pidgin infrastructure for Afro-Hispanic language in both Spain
and Latin America, although conceding (p. 128) that bozal Spanish, especially in Spain, lacks the
characteristic preverbal particles found in Afro-Portuguese creoles.

4. Was Puerto Rican bozal Spanish ever a stable creole?

Although Palenquero and Papiamento are undisputed creoles, the corpus of bozal Spanish from
the Caribbean and elsewhere in Latin America is not as unequivocal as regards the former existence of a
stable creole, rather than a rudimentary pidgin which arose as new arrivals from Africa entered the
speech communities.
The existence of a prior stable Afro-Hispanic creole in Puerto Rico is based on a literal handful
of texts analyzed by Alvarez Nazario (1974), principally the skit `La juega de gallos o el negro bozal'
(PR-5). Alvarez Nazario demonstrated many parallels between the Puerto Rican texts and
Afro-Iberian language from other regions and time periods, although his characterization of Puerto Rican
bozal language as a `criollo afroespaol' may refer to a non-native pidgin, rather than to a nativized
creole. That the latter might indeed have existed in Puerto Rico was first claimed by Granda (1968),
who notes (p. 194, fn. 4): `... es fcil demostrar el carcter igualmente "criollo" de la modalidad
lingstica puertorriquea ...' From this point forward the claim that an Afro-Hispanic creole was once
spoken in Puerto Rico has never been seriously challenged, despite the fact that the case rests on such a
small corpus. Among later studies of `Caribbean bozal Spanish,' little attention has been paid to a
possible Afro-Hispanic creole in Puerto Rico, with the latter region usually lumped together with the
more extensive Afro-Cuban corpus.
The creoloid structures in the Afro-Puerto Rican texts just mentioned are subject to alternative
explanations, to be presented below. There is, however, an important text (not included in Alvarez
Nazario's analysis), which suggests much less consistent bozal language in Puerto Rico. In 1884, the
poet/playwright Ramn Mndez Quiones (1847-1889) wrote the play `?Pobre Sinda!,' which was
never published. This neo-Romantic drama is set in Puerto Rico, in the `poca de la odiosa
esclavitud--1864.' Among the characters is a `esclavo viejo, congo,' who in the definitive version of the
manuscript speaks in normal, even sublime, Spanish, as he delivers his impassioned denunciations of the
cruelties of slavery. In unpublished notes, discovered by Girn (1991: 399-411), Mndez Quiones
gives his reasons for not having this African-born slave use bozal language: `Hablando en su jerga, no
convencera, y en los momentos ms patticos no hara sentir, produciendo con sus exclamaciones de
 vi

dolor la hilaridad del pblico' (Girn 1991: 400). To demonstrate his point, the author adds examples
of several scenes which he had originally written in bozal Spanish, `y de las cuales prescind por los
conceptos antes expresados.' These fragments (PR-7) show great similarity with bozal texts from
elsewhere in Latin America, but contain few of the creoloid traits which form the cornerstone of
Caribbean bozal-creole theories. For example, there is no consistent alternative verb system; correctly
conjugated Spanish verbs alternate with inappropriate forms, and no auxiliary particles signal tense or
aspect. Similarly, nouns and adjectives are sometimes inflected for gender and number, while in other
cases there are lapses of agreement.

5. Putative creole origins of Cuban bozal Spanish

The existence of a former Afro-Hispanic creole in Cuba has been forcefully asserted by a
number of investigators. Wagner (1949: 158) stated that `poche e confuse sono le notizie che abbiamo
sul negro-spagnolo parlato una volta a Cuba dai negri bozales ... ma sono sufficienti per dimostrare che
si trattava anche in questo caso di una lingua creola del tipo del papiamento ...' Wagner's case was
based on the poem `Yo bota lan garaf' (C-10) and the `Dilogo' between a negro criollo and a bozal
(C-7). In a later and more comprehensive analysis, Granda (1971: 483) offered the claim that `...
Cuba ha posedo y posee an entre su poblacin negra rastros y manifestaciones lingsticas "criollas" ...
unindose as al "papiamento:, al "palenquero" ... y a las manifestaciones puertorriqueas en la
formacin de un "corpus" dialectal "criollo" de superestrato espaol ...' To prove his case, Granda
made ample use of El monte by Lydia Cabrera (C-20), originally published in 1954, an anthropological
text containing data on religious beliefs among Afro-Cubans, and which includes extensive imitations of
bozal Spanish. Granda unquestioningly accepts the accuracy of Cabrera's imitations, given her high
reputation in other linguistic and folkloric matters, and suggests that such language, `caracterizadores de
una estructura "criolla" de lengua, persistan en el "registro" hablado de negros cubanos ... como
continuacin de la modalidad lingstica adoptada por generaciones anteriores de esclavos ...' In this
article, Granda did not explicitly link the putative Afro-Cuban creole to the monogenetic Portuguese
pidgin hypothesis, but this claim was eventually made in Granda (1976). Lapesa (1980: 560) believes
that `las postreras supervivencias del criollo espaol parecen ser el habla "bozal" que se usaba entre
negros de Puerto Rico en el siglo pasado y todava entre los de Cuba a mediados del actual ... ' Using
phonological data from Afro-Cuban texts, Sosa (1974) also offers the claim that an Afro-Hispanic
creole once existed in Cuba, and affected popular varieties of Cuban Spanish.
Cabrera's extensive writings, particularly El monte, have figured importantly in most subsequent
writings on a putative Afro-Cuban creole. Otheguy (1973) adds to the list of creoloid traits mentioned
by Granda, and claims that Cabrera's work demonstrates the prior existence of an Afro-Hispanic creole
in the Caribbean. Perl (1982) also refers to El monte, as well as to the brief bozal fragments from
Miguel Barnet's Autobiografa de un cimarrn (C-14). Perl (1982: 424) asserts that `... the Cuban
"habla bozal" was no idiolectally determined jargon of the Blacks in the 19th century but a social variety
of Spanish comparable with other varieties of Spanish- and Portuguese-based creoles.' With respect to
a possible extra-territorial origin, Perl (1982: 423) suggests that `... especially the morphosyntactic
features of the "habla bozal" are very suitable for demonstrating the relations to other Iberian-based
creoles and the embedding of the "habla bozal" within the Creoles and the "intermediate varieties" in the
Caribbean area.' This line of approach is extended in Perl (1985, 1987).
Perl (1984, 1985) strengthens his claim that an Afro-Portuguese creole was once spoken in
Cuba, based on two sources. The first is the `baile portugus,' recorded by Garca Herrera (1972) in
the predominantly Afro-Cuban neighborhood La Guinea in the central Cuban village of Lajas. This
 vii

song is of undetermined origin, although oral tradition links the song to the `portugueses de Africa'
(Garca Herrera 1972: 160). Garca Herrera (1972) postulates that the reference is to the
Angola/Congo region, given the documented existence of slaves from that zone in central Cuba, and the
preservation of some lexical items from the languages of this region. Perl (1985: 195) claims that this
song is `un ejemplo tpico para una forma lingstica criolla ms antigua,' although Perl (1984: 56) states
that the text in question is simply `una cancin que contiene voces como p. ej. "gayina" o "vol" que se
puede explicar como palabras ibricas (espaolas o portuguesas).' The text in question is (Garca
Herrera 1972: 162):

Ti ti ti ti
ngongo ti ngur
ca gayina np vol
mangu mangu
makina ma ntu ...

This song is composed of Bantu elements, and rather than demonstrating the prior existence of an
Afro-Portuguese creole, is more typical of the incorporation of Portuguese lexical items in a broad
spectrum of African languages from the Congo Basin.
Perl (1984, 1989b) also quotes from the anti-slavery treatise of Buxton (1839: 122), who in
turn cites a British ship captain who intercepted a Portuguese slave ship headed from Cape Verde to
Cuba. The captain claimed that `two of the Africans ... speak Portuguese ...,' which Perl (1984: 56)
interprets as meaning that `los esclavos que llegaron a Cuba tambin tenan conocimientos del
portugus/portugus criollo.'
Ziegler (1981) assumes axiomatically that Afro-Cuban bozal Spanish constituted a definable
creole, and attempted to write a grammar of this putative creole. Ziegler believes that Cuban bozal
creole resulted from 15th century Portuguese, with later accretions from several West African
languages, from nonstandard Spanish dialects, and from Jamaican creole English, carried by Jamaicans
arriving in Havana in the 18th century. It is known, for example, that during the British occupation of
Havana in 1763, tens of thousands of slaves were quickly imported into Cuba by the British (Knight
1970: 7). The minimal traces of Portuguese in surviving bozal texts is, according to Ziegler, due to
sustained contact with non-creole Cuban Spanish. Megenney (1984, 1985a) adopts Ziegler's evidence,
and groups `Afro-Cuban creole' together with Palenquero and Papiamento in a comparative analysis of
Portuguese-influenced Latin American creoles. Valkhoff (1966: 116) states, without further discussion,
that the only surviving Spanish-based creoles are `Malayo-Spanish' of the Philippines (i.e. Chabacano),
`Negro-Spanish of Cuba,' and Papiamento. Holm (1989: 305-9) is more cautious, speaking only of
`restructured Spanish' in the Caribbean, and noting that while there is ample evidence of a Spanish
pidgin in 19th century Cuba, it is not clear that a true creole developed. Speaking of the possibility for
creolization of Spanish in Cuba, Reinecke (1937: 269) noted that `conditions, one would assume, were
eminently favorable for the formation of a Cuban Spanish creole dialect,' although admitting (p. 271)
that `the jargon [i.e. the rudimentary speech of first-generation bozales: JML] was there, but there is no
indication that it took definite shape.'
An objective look at the Cuban bozal corpus reveals far fewer similarities with acknowledged
Afro-Iberian creoles than might be supposed by the research cited above. One of the earliest surviving
bozal texts from Cuba is an anonymous 18th century canto de cabildo (C-6):

Dond jachero
 viii

pa un palo.
Palo ta duro.
jacha no cotta.
Palo ta brabbo.
?qu son ese?
Si palo so jocuma,
yo so quiebrajacha.
Bamo be quie pue ma.
Tu jabla y no conose.
Tamb ta brabbo.

In addition to exhibiting many vernacular Cuban phonetic traits, including gemination of obstruents
following the reduction of syllable-final liquids, this text gives the first hint of what was to be a
commonly-recurring feature of Cuban bozal Spanish, the use of so(n) as an uninflected copula. In this
example, no creoloid features appear.
The first explicit mention of Cuban bozal Spanish comes in the catechism Explicacin de la
doctrina cristiana acomodada a la capacidad de los negros bozales by the Spanish priest Nicols
Duque de Estrada, written in 1797 (C-55). The author describes bozal speech (Lavia 1989: 67) as
`aquel lenguaje de q. usan ellos sin casos, sin tpos., sin conjunciones, sin concordancias, sin orden ...'
This work contains a few fragments of bozal language, none of which suggest a creole resembling
Papiamento:

pa nuetro ta seno cielo (Lavia 1989: 75)

yo soi un pobre esclavo, yo tiene dos gallinas no ms, gente tiene suelto su cochino, cochino
come mi gallina. Yo ya no tiene con que comprar tabaco ni nada ... ?yo va andando en cueros?
(Lavia 1989: 119)

Many observers (e.g. Martnez Gordo 1982, Valds Bernal 1978)) have based claims on the
creole status of Afro-Cuban speech on the remarks of the Cuban lexicographer Esteban Pichardo
(C-57), which accord to 19th century Cuban bozal speech the status of a separate linguistic variety,
compared with the French creole of Haiti (Pichardo 1985):

... este lenguaje es comun e idntico en los Negros, sean de la Nacin que fuesen, y
que se conservan eternamente, a mnos que hayan venido mui nios: es un Castellano
desfigurado, chapurrado, sin concordancia, nmero, declinacin ni conjugacin, sin R
fuerte, S ni D final, frecuentemente trocadas la Ll por la , la E por la I, la G por la V
&; en fin, una jerga ms confusa mientras ms reciente la inmigracin; pero que se deja
entender de cualquiera Espaol fuera de algunas palabras comunes a todos, que
necesitan de traduccin. Para formarse una ligera idea de esto, vertiremos una
respuesta de las mnos difciles: "yo mi ama Frasico Mandinga, neglito
reburujaoro, crabo musuamo o Mingu, de la Cribaner, branco como carabon,
sua como nan gato, poco poco mir ot, cribi papele toro ri toro ri, Frasico dale
dinele, non gurbia dinele, e laja cabesa, e bebe guariente, e coje la cuelo, guanta
qui guanta"...

Pichardo's imitation of this `special' language, however, does not show the characteristics of a stable
 ix

creole, but rather a rudimentary pidgin such as might arise spontaneously in any environment where
bozal slaves were rapidly acquiring Spanish. Pichardo's text bears considerable resemblance to bozal
texts from 16th-17th century Spain, as well as from elsewhere in 19th century Latin America, but the
similarities with Afro-Iberian creoles are minimal.2
Another widely-cited example of Cuban bozal Spanish, an anonymous mid-19th century text
(C-7) quoted by Bachiller y Morales (1883), bears even fewer resemblances with Papiamento or other
Afro-Iberian creoles:

Ah, si ot no lo cubr,
si ot tova no fu,
?pa que buca que beb?
?Con qu ot lo va pag?
Cuando ot lo cubra, anj,
antonsi ma qui ti muere
beb ot como ot quiere,
como ot como dan gana,
y durm ot una semana
ma que lan tempo si piere.

Bachiller y Morales notes: `no es posible confundir un lenguaje [i.e. the speech of criollo blacks: JML]
con el otro: la supresin de letras, la conversin de otras, no es peculiar de todo negro ...' Indeed, this
example is far less removed from non-African Spanish than Pichardo's text, despite the fact that both
represent the same time period. The example cited by Bachiller y Morales (published in Matanzas
earlier in the 19th century) contains only two elements which cannot be analyzed as simply imperfectly
pronounced Spanish minus a few connecting words: the element tempo instead of tiempo, contains a
non-diphthongized root homologous with Portuguese and Papiamento.3
With the new perspective of a possible Afro-Lusitanian origin, the focus of Afro-Hispanic
studies shifts away from the search for direct African- American links to the postulate of an intermediate
pan- Hispanic creole stage. This intermediate language, through contact with European Spanish
following the abolition of slavery, gradually came to resemble regional Latin American Spanish more and
more, while perhaps transferring some of its own characteristics to the Spanish spoken by descendents
of Europeans. In this vein, Perl (1985) proposes a model for the `decreolization' of Cuban bozal
Spanish, while Megenney (1985) suggests that even `acrolectal' varieties of Caribbean Spanish bear the
imprint of the early Afro-Hispanic/Afro-Lusitanian creole.

6. Opposing views

Although only a handful of texts have previously been used in the discussion of Cuban bozal
Spanish, the available corpus is much wider; a representative sample has been included in the Appendix.
The majority of these texts show few creoloid features, but rather conform to the pidgin/foreigner-talk
patterns expected of Africans learning the rudiments of Spanish as a second language. The remaining
creoloid traits may be attributable to a direct infusion of creoles from elsewhere in the Caribbean, as will
be suggested below. Put in this perspective, the case for a previously existing Afro-Hispanic creole in
Cuba is shown to rest heavily on a single text, Cabrera's El monte (although by extension more of
Cabrera's extensive writings might also be included), together with ambiguous remarks such as those of
Pichardo. The Afro-Lusitanian theory is based on tenuous evidence of Portuguese participation in the
 x

18th-19th century slave trade to Cuba, together with similarities between the putative Cuban bozal
creole and the acknowledged creoles Papiamento and Palenquero, for which Afro-Portuguese roots
may be more uncontroversially established. For Puerto Rico, the case is even more precarious, for only
two literary texts (for which a possible imitation of Cuban models cannot be entirely excluded) establish
claims of a stable Afro-Hispanic creole, and the parallels with Papiamento and Palenquero are therefore
even more limited.
In view of these considerations, not all investigators have accepted the notion that any stable
Afro-Hispanic creole was ever spoken in the Caribbean. Goodman (1987) casts doubt on the notion
that many African slaves arriving in Spanish America had acquired a Portuguese pidgin. He observes
that in any given slave depot in West Africa, a regional African language usually prevailed as a lingua
franca, while the brief passage to the New World provided neither opportunity nor motive for acquiring
a Portuguese-based pidgin. Only slaves brought from So Tom and later Angola were exceptional in
possessing knowledge of an Afro-Lusitanian pidgin. Goodman provides alternative explanations for
Portuguese elements found in Latin American creoles. Laurence (1974) notes that the proportion of
Africans to Europeans was never as high in the Spanish Caribbean as on islands such as Jamaica,
Saint-Domingue and Barbados, where blacks at times outnumbered whites by more than ten to one. In
Cuba and Puerto Rico, the African population never came to represent more than 50% of the total,
even at the height of the sugar plantation boom, and was usually much lower. In the Spanish Caribbean,
a larger proportion of slaves worked away from large plantations, in closer contact with native speakers
of Spanish. Through the system of coartacin many slaves eventually purchased their freedom, and the
free black and colored population became significant. All observers coincide in remarking that blacks
born in the Spanish colonies spoke Spanish with native fluency. Beginning in the final decades of the
18th century, and continuing through the middle of the 19th century, the rapid expansion of the sugar
industry in Cuba created a plantation environment which, if it had lasted long enough, might have
provided a fertile ground for creole formation. However, due to the high mortality of the slaves, the
proportion of recently-arrived bozales speaking little or no Spanish was always quite high, and after
only 2-3 generations the end of the slave trade brought an end to new bozal arrivals. The remaining
African-born blacks gradually acquired Spanish, many moving to urban areas and reinforcing contact
with natively spoken Spanish (cf. Aimes 1907; Castellanos and Castellanos 1988, 1990; Klein 1967;
Knight 1970).
Lpez Morales (1980) finds similar objections to the theory of a former stable Afro-Cuban
creole. He notes that Pichardo (1836) and Bachiller y Morales (1883) both stress that blacks born in
Cuba spoke `normal' Cuban Spanish, and he interprets the remarks which these two authors made
about bozal speech, as well as their imitations, as evidence of a rough pidgin, not a stable creole. As
for a Portuguese component to Cuban bozal Spanish, Lpez Morales observes that by the time of the
sugar plantation boom in the late 18th century the Portuguese supremacy in the Atlantic slave trade had
been replaced by Dutch, French and British dealers. At the same time, he notes a number of lexical
items (drawn from Ortiz 1916: 238-9), many of which are common to West African Pidgin English,
which he claims formed the true `lingua franca' which bozal slaves used with one another while acquiring
Spanish on Cuban plantations. Such words include chapi-chapi < chapear `to chop weeds,'
luku-luku < look `look, see,' ami-ami < nyam `to eat,' tifi-tifi < thief `to steal,' etc. Lpez
Morales agrees with Otheguy (1973)'s reluctance to accept Cuban bozal texts as examples of a prior
Afro-Portuguese pidgin or creole, but he disputes Otheguy's view that the ungrammaticality and
`un-Spanish' nature of many bozal constructions necessarily point to a prior creole of any sort. Few
bozal texts evidence these forms exclusively; nearly all appear counterpoised with unremarkable
Spanish morphosyntactic constructions attributed to the same speakers. For Lpez Morales, this
 xi

represents not a creole but rather `un polimorfismo,' indicador aqu de deficiencias particulares de estos
hablantes en su proceso de castellanizacin ...'

7. Preliminary counterexplanations to claimed creoloid features

With the exception of the preverbal particle ta and genderless third person pronouns, most
features of bozal Spanish which have at one time or another been cited as evidence of a pidgin or creole
Portuguese basis cannot be accepted as probative of an earlier Afro-Hispanic creole language. Among
the more dubious traits are:

(1) Non-inverted questions of the type ?qu t quieres? `what do you want?' (Otheguy 1973).
These constructions are common throughout the Caribbean, and may have been reinforced by Canary
Island immigration. Afro-Iberian creoles exhibit non-inverted questions, but so do non-Africanized
dialects of Spanish, e.g. in the Canary Islands and Galicia.

(2) Categorical use of redundant subject pronouns (Granda 1968, 1971). All Afro-Romance
creoles use obligatory subject pronouns, due to lack of verbal inflection. The same occurs in vestigial
Spanish lacking a creole basis (Lipski 1985), and in many cases of Spanish as a second language.
Given that subject pronouns or clitics are obligatory in nearly all West African languages known to have
come into contact with Spanish, preference for overt pronouns in bozal Spanish would be predicted
without the intermediate stage of a creole.

(3) `Personalized' infinitives with lexical subjects of the type para t hacer eso `for you to do
that' (Alvarez Nazario 1959: 46; Megenney 1984). Such constructions are found not only in
Afro-Iberian creoles, but also in Canary Island and Andalusian Spanish, in Galician and Portuguese, and
throughout Latin America. It is likely that this construction has arisen spontaneously in more than one
area, since it results from the reduction of a marked conjugated form to the maximally unmarked
infinitive; the same process occurs in Spanish child language (Gili Gaya 1960: 29; 1972).

(4) Loss of common prepositions, particularly a and de (e.g. by Alvarez Nazario 1959, Granda
1971, Otheguy 1973, Perl 1982). The same feature is found in nearly all foreign-influenced and
vestigial varieties of Spanish. In contemporary syntactic analyses, de and a may not be underlying
prepositions but rather superficial case-markers, subject to variable deletion during imperfect learning or
linguistic erosion.

(5) Occasional elimination of the copula (Alvarez Nazario 1959, 1974; Granda 1971, Perl
1982). This often occurs in vestigial speech, and given that a large cross-section of West African
languages employ `verbalized adjectives' instead of a combination of VERB + PREDICATE ADJECTIVE,
loss of a copula might be an African areal characteristic rather than a post-creole carryover. In the
Caribbean bozal corpus, there is only a tiny number of cases where the copula has been deleted. Much
more frequent is the use of a default copula, such as son or occasionally ta.

(6) Loss of articles (Alvarez Nazario 1959, 1974; Granda 1971, Perl 1982). This is also
found in vestigial and foreign-influenced Spanish, and in view of the generalized absence of articles in
West African languages, could also be an areal characteristic.
 xii

(7) Postposed demonstratives of the type piera ese [= la piedra esa/esa piedra] `that rock'
(Otheguy 1973). Postposed demonstratives are found in many non-creole dialects of Spanish, as well
as in several creoles. Moreover, Latin American bozal texts yield only a couple of examples, all from
the works of Lydia Cabrera. It would seem that this construction was never common in Afro-Hispanic
language, so no particular value should be accorded to this specific example.

(8) Lack of syntactic complementizers such as que (Granda 1971). Syntactic simplification
through reduction of subordinate structures characterizes all reduced forms of Spanish, and is found in
Afro-Hispanic, Amerindian-Hispanic and Anglo-Hispanic foreigner talk.

(9) Use of the subject pronoun vos, in bozal texts from the Caribbean, where this pronoun is
not normally found. The pronoun (a)bo is found in all Afro-Lusitanian creoles, as well as in Papiamento
and Palenquero. In Caribbean bozal examples, vos is found only in a single text, a 19th century Cuban
villancico from Camagey (C-10):

... Francisco mi pariente

disi que ya vos pari,
como yo quera aguait
lo que v haba paro,
aque me tenei, Se.

The presence of vos in this lone example does not implicate a former creole in Cuba, for vestigial vos,
together with diphthongized verb forms, is amply documented for 19th century Cuba in precisely this
region (Lpez Morales 1965, Pichardo 1985: 12).

(10) Use of the portmanteau preposition/connector na, found in a few Afro-Caribbean texts,
and also found in many Portuguese-based creoles (deriving from the contraction of en + a). A rare
Cuban example is (C-35): atr quitrn pa yeg prisa, prisa, na panadera `behind us the carriage is
arriving quickly, to the bakery,' cuando yo me pi de na caballo `when I dismounted from the horse.'
Brau (1894: 138; PR-4) observed that in 19th century Puerto Rico, `cimarrones bozales' used
expressions such as na- cosina, ne- pueblo, na- casa, etc. for en la cocina `in the kitchen,' en el
pueblo `in the town,' en la casa `in the house.' This form is very limited in Afro-Caribbean texts, and is
not attested for bozal language of other areas.

(11) The use of tener `to have' instead of haber `for there to be' as the existential verb
(Megenney 1984, 1985a; Granda 1968). Most Afro-Iberian creoles, as well as African-influenced
vernacular Brazilian Portuguese, have replaced existential haber by tener/ter. While use of tener in
bozal speech may indeed come from an earlier proto- creole, this is not a necessary conclusion, since
use of existential tener is also found in vestigial Spanish of many regions, and even in some Spanish
dialects with no demonstrable Afro- creole connection (cf. Lipski 1985). Moreover, use of tener with
existential force is quite rare in bozal language; one example (C-25) is: en botica tien de t `in the
medicine chest there is everything.'

(12) The frequent Caribbean preposing of ms in negative expressions (ms nada `no more,'
ms nunca `never again') instead of the more usual phrase- final position has been claimed as the result
of earlier Portuguese- based creole language (Megenney 1985a). A Portuguese connection is quite
 xiii

likely, but the presence of this construction in Caribbean Spanish is more likely due to the heavy Canary
Island influence, in which such constructions (apparently due to earlier Galician- Portuguese maritime
contacts) are common.

8. Undifferentiated 3rd person pronouns in Caribbean bozal Spanish

We turn now to one of the more clearly creoloid features of Latin American bozal Spanish, the
use of third person pronouns undifferentiated for gender. In Afro-Lusitanian creoles, including
Palenquero and Papiamento, third person pronouns are reduced to a single variant each for singular and
plural. Typically the singular variant has the general form e(le), while there is more variation in the plural,
including pronouns of non-Romance origin (e.g. Papiamento nan, Palenquero an). Some
Afro-Caribbean texts show third person pronouns similar to those found in Afro-Iberian creoles. The
general form is elle or nelle; these words are attested for 19th century Cuba, and sporadically Puerto
Rico. Nelle was occasionally used as third person plural. Typical examples include:

Elle estaba en un mortorio. El borbana manda prendeslo. Dentra Tond, elle solito con su
esp, coge dos (C-68)
?Y nelle lo muchachito va pend su Paa de nut? (C-53)
Eso mimo quiere yo, nelle lo mimo, vamo pa la engresia (C-15)
si yo lo ten uno nio como nelle, yo va mur de cuntentamienta (C-34)
yo mir que nelle tiene sangre, ese simbregenza mimo se pi detr la quitrn y arranc corr.
(C-35)
Hora, dipu que nelle coge yebba la gloria, vamo salud Ocha ... Mueco con pritu de mueto
muchacho, que nelle meta dientro (C-20)
Po que junt con la mala compaa y nelle lo pervierte o lo sonsac. (C-65)
Yo tiene la pecho premo pur nelle. Yo ta namor, yo va v si nelle qui s mug ma pur
langresia ... yo pu cas cunelle ... vereme nelle ... nelle toca violn y pone casaca ... t ta
mirando que nelle va llor tova ... (C-41)
nelle que lo s intriuo ... nelle mimo que lo s ... poque nelle ta en la tea ... porque nelle lo
gat ... y nelle mimo disiba que yo s como la miba ... y dimpu que diga nelle que yo
s bruto ignorante ... (C-40)
not quie jabla cun nelle ... y disi que va a tumb mi buj, vereme nelle ... (C-50)
no lo tengo cunelle la may cuusimienta ... nelle lo pue laig uno lintenaso ... dis nelle que la
gente lo gutara uno cumera ne luenga mo ... Cun la geve dielle yo mec sei lichonsito ... no
son cosa que nelle ten, sino que jileo d porelle diese funsin ... (C-33)
nelle tiene un vap ... nelle viene, yo le da ... Neye se ama muj ... neye va acab con
pacfico insurrecto ... (C-26)
toto neye ta carg ... cuando neye mira yo ... neye ta mor de risa ... (C-29)
cuando nei ta ven, ya yo no tiene que da vueta ... sing caballo pa neye ve jodienda la
Tajonera ... (C-21)
varn quit neye ... (C-22)
Neye lo que tiene s un bariga con su yijo lentro (C-54)
yo te ba da un medall pa que tu luse con eye (C-10)
luego nelle va vin a com la buen caliente (C-48)
mucha grasia, sumes, pero nella son honr {PR-7)
Nelle son mala cabesa {PR-7)
 xiv

Nella memo ta casando (C-48)

Yo no quisi di con elle (PR-6)

Alvarez Nazario (1974: 185-97), in the only analysis of this form, feels that semantic replacement of a
preposition plus an article (as in na) has occurred. There is, however, no plausible source in the case of
(n)elle. The [y] represented by ll is presumably derived from ella, ellas and ellos; neither Portuguese
le nor similar forms in Papiamento, Palenquero, So Tomense, etc., provide a source for the [y]. In a
few texts (e.g. C-48 and P-7), the feminine variant nella is found alongside nelle. Elle/nelle may thus
be a spontaneous Afro-Hispanic development which arose in the 19th century Caribbean. However,
the fact remains that this pronoun is not attested in the ample bozal Spanish corpus from 16th-17th
century Spain, from bozal attestions from 19th century Argentina, Uruguay, nor from Mexico and Peru
in the 17th and 18th centuries. The existence of this element only in the 19th century Caribbean is thus
food for thought.

9. Verbal constructions with ta

The most indisputably creole element found in some Caribbean bozal texts, which has formed
the centerpiece for theories which claim a previous Afro-Hispanic creole, is the use of ta, in
combination with a verbal stem derived from the infinitive lacking final /r/:

? o que t no ta quer a m? (PR-5)

P
Siempre ta regal dinero a m (PR-5)
Horita ta ben pa c (C-71)
Ro seco ta corre mamba (C-56)
Como que yo ta cuch la gente que habla tanto ... yo ta mir gente mucho (C-29)
Primero ta llor na m. (C-66)
yo ta yor poque Calota ya ta mor. (C-70)
S, pe, yo ta rob un gana jab. (C-21)
Como no va ta contenta si ut vovi cuando yo ta acod mi yari yari que et en Guin.
(C-21)
Cuando ojo ta llor a narice toca su parte. (C-25)
Amo ta ped leche. (C-18)
Ya branco ta debarat cosa (C-22)
pavo real ta buc palo (C-20)
Ta ju, ta puj m, si (C-67)
ta dic: tu buca la cosa bueno (C-26)

This construction, unlikely to have arisen spontaneously from an unstructured Spanish pidgin, is identical
to verb phrases in Iberian-based creoles throughout the world, including Palenquero and Papiamentu in
Latin America, Guinea-Bissau and Cape Verde Kriol in Africa, Portuguese-based creoles in India, Sri
Lanka, Macau and Malaysia, and in Philippine Creole Spanish (Chabacano). Many investigators,
including Otheguy (1973), Granda (1968), Megenney 1984, 1985a), Perl (1982), etc. have regarded
the presence of ta in Afro-Caribbean bozal Spanish as virtually conclusive proof that an Afro-Hispanic
creole, similar to Palenquero and Papiamento, was once spoken throughout the Caribbean, and perhaps
even in South America. The existence of ta + Vinf in these creoles is a strong bit of evidence in favor of
a common origin or at least mutually shared influences (but cf. Lipski 1987a, forthcoming b). In other
 xv

respects, the verbal systems of these creoles share fewer similarities. In the languages just mentioned,
ta is variously used for present/imperfective and durative aspect. For the past/perfective, ya/ja is the
most common variant, but other forms are also used; Papiamento, for example uses a, as does the
Ternateo dialect of Philippine Creole Spanish. In Palenquero, ba is used as an imperfective marker,
although its syntactic properties are different from those of ta. Papiamento and Cape Verde crioulo
make some use of Spanish/Portuguese imperfect verb forms. No Latin American bozal text shows
consistent use of any past/perfective particle; in particular, ya appears in the same positions as in
Spanish.
There is even more variation among creoles to represent future/irrealis: Papiamento has lo
(apparently from Ptg. logo `later'), Palenquero has tan, Philippine Creole Spanish has di or ay, and so
forth. Latin American bozal texts, on the other hand, use no particle to signal futurity; either the simple
present or a periphrastic Spanish future with va are used. The latter element is never used consistently
enough to be considered an innovative particle. This casts considerable doubt on the prior existence of
a uniform Afro-Hispanic creole, since only one component of the usual three-particle creole verb system
is found in attested bozal language. A re-evalution of the role played by ta in bozal Spanish is called
for.
It is striking that among the scores of Afro-Hispanic texts, from Spain and all of Latin America
and spanning nearly 400 years, the combination ta + Vinf is found ONLY (1) in a very small number of
texts, (2) in the 19th century, (3) in Cuba and Puerto Rico. Even in the 19th century Afro-Caribbean
corpus, constructions based on ta alternate with the archetypical bozal pattern of partially or incorrectly
conjugated verb forms. No use of ta occurs in the large Afro-Hispanic corpus from Argentina,
Uruguay or Peru, nor in scattered texts from other regions. The Dominican Republic, which shares the
island with speakers of Haitian creole (which makes ample use of pre-verbal particles, including ta <
perfective te + future/irrealis (v)a as conditional), some of whose elements have penetrated into hybrid
border dialects, shows no evidence of ever having had ta + Vinf or other preverbal particles, despite the
existence of other possibly postcreole phenomena (Gonzlez and Benavides 1982, Megenney 1990).
In a few cases it might be possible to argue that spontaneous developments took place, e.g.
where ta is clearly derived from esta(r) acting as either a locative verb or in combination with an
adjective: Yo no pue TA quieto ya (C-34) `I can't be still now'; Nnge TA bueno ... (C-20) `The
nange [tree] is good.' In other instances, phonetic erosion of what was once a gerund, ending in
-ando or -(i)endo is also a possibility: Que to mi cuepo me ET tembl (C-26) `My whole body is
trembling'; pavo real TA bucn palo (C-20) `the peacock is looking for a tree'; yo EST CORT un
caas `I am cutting sugar cane' (C-31). In other cases, however, the verbs in question are habitual or
durative, contexts where Spanish would not use any combination involving estar. This residue is among
the few sure indicators of an infusion of creole elements into bozal language.
An overview of Puerto Rican bozal texts reveals only a small number of cases of the particle ta
combined with an invariant verb stem (Alvarez Nazario 1974: 193- 4), as compared with other
examples in which no non-Hispanic verb forms appear. By far the greatest number of cases of ta occur
in texts from Cuba, beginning towards the middle of the 19th century, carrying through to the early 20th
century. Even in Cuba, numerous bozal texts from the same time period lack any use of ta, employing
instead the prototypical bozal verbal system in which verbs were reduced to the third person singular or
to a bare infinitive (or in the case of the copula ser, to son).

10. Cabrera's `transcriptions' and Afro-Spanish hybrids

Given the importance of Lydia Cabrera's writings, in particular anthropological- religious studies
 xvi

such as El monte, to Afro-Caribbean creole theories,4 it is useful to consider how accurate and
representative such texts might be as specimens of bozal speech in Cuba and elsewhere in the
Caribbean. Cabrera's works are extensive and captivating, and the long segments written in African
languages and in bozal and vernacular Cuban Spanish sound so convincing that they seem to be the
transcription of actual recordings. For Granda, Perl and Otheguy, Cabrera's Afro-Cuban
anthropological studies are not only the most accessible source of bozal speech, but also the most
trustworthy, given the author's clearly positive attitudes towards Afro-Cuban culture and her avowed
intention of describing it accurately. Missing in these writings is the mocking humor which characterized
the habla de negros in the Cuban `teatro bufo,'5 and in other writings from Spain and Latin America
(although in such texts as Francisco y Francisca [C-21] and Refranes de negros viejos [C-25] some
glimpses of this tendency emerge). At the same time, many of her writings employ literary techniques of
magical realism, complete with possible exaggeration of linguistic usage. For Ortiz (1940: 9):

estos cuentos [Cabrera's Cuentos negros: JML] vienen a las prensas por una
colaboracin, la del folklore negro con su traductora blanca. Porque tambin el texto
castellano es en realidad una traduccin, y en rigor sea dicho, una segunda traduccin.
Del lenguaje africano ... en que las fbulas se imaginaron, stas fueron vertidas en Cuba
al idioma amestizado y dialectal de los negros criollos. Quiz la anciana morena que se
las narr a Lydia ya las recibi de sus antepasados en lenguaje acriollado. Y de esta
habla tuvo la coleccionista que pasarlas a una forma legible en castellano ... la autora ha
hecho tarea difcil pero leal, y por tanto, muy meritoria, conservando a los cuentos su
fuerte carcter extico de fondo y de forma ...

For Ortiz, then, the bozal language of Cabrera's writings is (at least) twice-removed from what might
have once been an Afro-Cuban creole (cf. also Martnez Gordo 1982: 52). This position contrasts
radically with, e.g., Granda's view that Cabrera's literary representation of bozal Spanish is as faithful as
a tape recording or verbatim transcription. This is not to say that Cabrera used bozal Spanish to
represent conversations held mostly or entirely in African languages, although many older Afro-Cubans
did use African languages amongst themselves. In personal conversation, the late Lydia Cabrera
described to the present writer the manner in which older bozales spoke throughout Cuba (not just in
Havana, as Granda seems to imply), but both from her writings and her personal recollections it is clear
that no bozal speaker used creoloid forms exclusively. Most combined creoloid, pidgin, and standard
Spanish forms, not necessarily because the bozal language was `decreolizing,' but because the creoloid
elements often arose spontaneously as African-born Cubans evolved in their acquisition of Spanish.
It is not insignificant that in Cabrera's writings, as well as in nearly all Cuban literary works in
which any Afro-Hispanic language with creoloid tendencies appears (cf. the appendix for a
representative list of items consulted for the present study), it is always older, bozal Africans, never
native-born Cuban blacks, who exhibit creoloid traits. This fact led Lpez Morales (1980: 109) to
question `si los hijos de estos hombres ya no son congos, ya manejan un espaol cubano estndar,
desconociendo en muchas ocasiones la lengua africana de sus padres, ?qu tipo de transmisin es sta?'
Although there are some creoloid forms in bozal texts, nothing in the texts themselves or in known facts
of Cuban history would suggest that these African-born slaves had learned such items from a FORMER
generation, either in Africa or in Cuba. If there were a stable Afro-Hispanic creole from which bozales
learning Spanish could draw elements, there should be convincing evidence that NATIVE-born black
Cubans spoke this creole, even more consistently (i.e. to the exclusion of both standard Spanish and
pidgin items) than did bozales. Such evidence is entirely lacking.
 xvii

There is an even more problematic aspect of texts like El monte, which adds another degree of
indeterminacy to the reconstruction of Afro-Caribbean bozal Spanish. This involves the fact that, not
only are the most creoloid features proferred by the oldest, African-born blacks, but most are an
integral part of Afro-Cuban religious rituals, which implies fossilization and transmission in automatistic
form to initiates. Thus Lpez Morales (1980: 108):

En estos textos donde los informantes negros hablan de sus religiones, supersticiones,
magias y folklore, hay ejemplos de naturaleza morfosintctica y lxica ... que han sido
tomados con valor de muestra de la pervivencia de una lengua criolla. Sin embargo,
slo se trata de ejemplos de estadios lingsticos individuales, aunque por fuerza
coincidentes en hablantes de la misma lengua materna, que denuncian una adquisicin
imperfecta del espaol. Todos ellos aparecen en boca de bozales, ninguno en labios
criollos.

Most of the speakers implicated in the use of creoloid items in bozal Spanish were native speakers of
an African language, from which they routinely drew lexical elements and morphosyntactic structures
when speaking to religious initiates, other Africans with knowledge of the same or similar African
languages, or sympathetic listeners such as Lydia Cabrera. Among the linguistic strategies employed by
such bozal speakers was the free creation of hybrid structures, employing a Spanish (or pidgin Spanish)
morphosyntactic frame with an African lexical core. Thus in Lachataer's ?Oh, mo Yemay! (C-47)
we find the Yoruba/Spanish chant:

A la mofil
Chang t mol ...

From Cabrera's writings come:

pa nkam [enter the religious ceremony], coge huevo ese, pasa cara, pasa cuerpo, limpia bien
y cuando te limpio to ut pu cog Mpeg' (C-17)
?Por qu t coge owo Elgbara? Si mimo dic t ta ol y te va agarr pinado su
papalote ... (C-20)
Ese otro yo me lo va yun y a Migu no pasa n ... (C-20)
Olofi ya oku, Olofi ta mirando, ya ik. (C-20)
Olofi no est aro. Olof est ddara. (C-20)
Cmo va s mano branco, si ta af, ta prieto yo ... (C-22)
Vamo sir (C-22)
Ahora ese ewe, to ese palo ta sacramentao ... (C-22)
?Ay mi maro! ?ok mi, ok mi! (C-18)
Maana yo ik (C-22)

From the Afro-Cuban song `Elegua quiere tamb' by Celia Cruz (Castellanos 1983: 57) comes the
fragment:

No hay Orisa como Elegua pa la il, porque siempre et of ... Ochn ta weye weye ... `There
is no god like Elegua for the [ceremonial] house, because [he] is always watching ...'
 xviii

These examples show the incorporation of Yoruba elements into Spanish sentences, sometimes
slightly modified, and in other cases without modification. Thus owo `money,' yun (Yoruba jeun) `to
eat,' ol `thief,' of `to watch' (< fe `(s)he watches'), aro `sick,' oko `husband,' af `dark,' ewe
`leaves,' ? ir `to play,' etc. However, the texts also reflect Yoruba morphosyntax, including the at
times tenouous or non-existent morphological difference between nouns, verbs and adjectives, and the
attempt to create hybrid verbal constructions using Spanish elements to replace Yoruba particles, in
conjunction with a Yoruba stem. In several instances, the results closely resemble verbal structures in
Afro-Iberian creoles such as Papiamento, Palenquero or Cape Verdian Crioulo, but in which the case
for an independent development can be made quite strongly.
Consider, for example, the Hispano-Yoruba hybrid transcribed by Cabrera: `Olofi ya oku,
Olofi ta mirando, ya ik.' In this combination, ya seems to be acting as a preverbal particle, much as in
Afro-Iberian creoles. In reality, however, this sentence embodies a subtle form of code-switching. In
Yoruba, the root for `die' is k. This element can also be used as an adjectival verb, meaning `be
dead.' Thus k can mean either `(s)he dies/died' or `(s)he is dead.' When modifying a noun, the
adjectival form is kk. The corresponding noun is ik `death.' In the pidginized Yoruba used by the
Abaku/Lucum Afro-Cuban religious cults, much of this grammatical information has been lost, so that,
e.g., ik can be used as a verb/adjective: maana yo ik `tomorrow I will die/be dead' (cf. also
Cabrera 1970c: 160). When combined with ya, as in Olofi ya oku ... ya ik `Olofi is already
dead/already died,' there seems to be a prototypical creole formation in which ya operates as a
preverbal [+anterior] particle.6 In many Iberian-based creoles, in Africa, Asia and the Caribbean, the
[+anterior] particle is derived from Sp. ya/Ptg. ja, almost always combined with an Ibero-Romance
verbal stem. In such cases, ya has evolved from a free adverbial to a preverbal clitic particle, always as
part of the verb phrase. In the Spanish/Yoruba hybrid constructions transcribed by Lydia Cabrera, the
fact of code-switching precludes analyzing ya as a clitic or verbal particle; rather, it is still behaving as an
adverbial adjunct.
The case of mimo dic t ta ol `he himself says that you are a thief/are robbing' appears to
embody ta as a preverbal particle. In Yoruba, ol is a noun meaning `thief,' while the verb `rob' is jal.
Once more, in the pidginized Yoruba used in El monte, this distinction is usually lost, either because the
speakers in question were not truly native speakers of Yoruba (whether or not they were born in
Africa), or because code-switching between two languages with radically different morphosyntactic
patterns resulted in canonical full forms from Yoruba being used when a Yoruba item was inserted in a
Spanish sentence.7 The hybrid combination t ta ol is thus ambiguous, meaning both `you are (a)
thief,' and `you steal/are stealing.' In the first interpretation, ta is used incorrectly as a copula with a
predicate nominal, a usage also found in vestigial or semi-fluent usage in other regions. In the second
meaning, ta is more clearly derived from the progressive auxiliary est, routinely pronounced as ta in
vernacular Cuban Spanish. The resulting configuration, however, is superficially identical to
quintessential Iberian-based creole constructions derived from ta plus the Romance infinitive. The
combination vamo sir `let's play,' involves a Yoruba verb used identically to a Spanish infinitive; the
superficial form of the Yoruba verb resembles a Spanish infinitive. Many of the other Yoruba items
transcribed by Cabrera have the same canonical shape as Ibero-Romance infinitives: af, mol, ol,
etc. In the hybrid constructions characteristic of Afro-Cuban speech these items are not behaving as
verbal infinitives, whether or not derived from a Yoruba verbal stem. Similarly, in the hybrid
constructions, ta < est is not behaving as an auxiliary for a progressive construction, much less as a
preverbal particle signalling tense, mood, or aspect, as occurs in legitimate Afro-Romance creoles. It is
a simple copula, introducing a patently foreign insertion.8 Despite the analysis of Yoruba-Spanish
hybrid constructions in a fashion different from that of TMA PARTICLE + VERBAL STEM found in
 xix

Afro-Iberian creoles, the superficial similarity with the latter phenomena has led to the inclusion of the
hybrid constructions in the `evidence' of a former creole status for Cuban bozal Spanish, as well as to
the purported similarities with Afro-Portuguese pidgins and creoles. Among Cuban bozales, purely
Spanish constructions based on ta + Vinf were used in addition to the hybrid Spanish+Yoruba
constructions, but not necessarily as part of a pan-Caribbean creole pattern. It is more likely that,
having established a productive hybrid system in which Spanish ta could serve as a copula or generic
verb introducing inserted borrowed elements, this same combination was extended to uninflected
Spanish verbs (i.e. infinitives), with similar use. In other words, speakers of Yoruba and other African
languages in which the noun/adjective verb distinction is marked differently from Spanish, or not at all,
would freely combine Spanish verbs, nouns and adjectives with ta to create complex verbs with simple
meanings.
The assertion that ta was used in Afro-Hispanic language to introduce elements regarded as
foreign is difficult to
substantiate, given the
limited nature of the
Cuban bozal corpus
and the fact that
Spanish-speaking
observers would not be
likely to notice the
non-Spanish fashion in
which such
combinations were
being used. One
possible method is to
look for cases where
ta/est is combined
with verbal stems in
which progressive
aspect is entirely
lacking, perhaps even
where a perfective
aspect can be assumed.
Also, one might look
for use of ta/est with
adjectives in a fashion
which could be
re-analyzed as a simple
verb, as well as
combinations involving
ta/est with the
meaning of a simple
verb. Some Caribbean
bozal verbal
constructions based on
ta which lack a
 xx

progressive aspect, or
where the following
element is not based on
the Spanish gerund,
include:

? o que t no ta quer a m? `Why don't you love me?' (PR-5)

P
Siempre ta regal dinero a m `He always gives me money' (PR-5)
Horita ta ben pa c `Soon she (will) come here' (C-71)
Yo ta llor pocque Calota ya ta mor `I am crying because Carlota died' (C-70)
la campana ta pin, pin, ?brgame Dios la tragn que lo tiene la critiano! `The bell is going "pin,
pin. Good God, how these Christians are dressed up!' (C-29)
Yo lleg, quit cachucho, cuch msica bonito ... yo ta mir gente mucho ... `I arrive(d), take
off my hat, listen to the pretty music ... I looked at all the people' (C-29)
moso ta mirando neye, neye ta mor de risa, `The boys are looking at them [the girls], [the girls]
are dying of laughter (C-29)
Lamo ta rega m, yo siempre con soramienta no cuch ... `The master always scolds me,
annoyed, I never listen' (C-29)
Ta tent dol en su cabeza. Ti fiebre. `He has a headache and fever.' (C-21)
S, pe, yo ta rob un gana jab, y dipu yo rob una yegua. `Yes, Father, I stole a hen, and
then I stole a mare' (C-21)
S, que progrosa son buey, poque cuando nei ta ven ya yo no tiene que da vueta y vueta y
vueta lo trapiche `Oxen are very dangerous, because when they come, I don't have to turn the
sugar mill' (C-21)
Ay Di, cuando mi maro ta juy, si mdico dic yo ta embaraz `Oh, God, when my husband
left, the doctor told me I was pregnant' (C-22)
Miamo Juan ta mur `Master Juan (has) died' (C-11)
soldao ta cabrn, mamb piliando con machete afil y cort cabeza. `The soldiers are bastards,
Mambs fighting with sharp machetes, cutting off heads' (C-1)

In all these examples, ta cannot be adequately analyzed as indicating progressive aspect. This
provides indirect evidence that ta was being used neither as a preverbal particle nor as an auxiliary verb,
but rather as part of a hybrid form in which ta introduced foreign elements. In fact, there is evidence
that in the more rudimentary forms of Caribbean bozal Spanish, it was the bare uninflected infinitive
(lacking final /r/) that was used as the default verb. In slightly more fluent forms of bozal Spanish, the
infinitive might be supplemented with `auxiliary' verbs such as (es)t and va, while in the most fluent
varieties of bozal Spanish, the 3rd person singular verb form normally acted as the default (although son
was more commonly used as the default copula), at times in alternation with fully conjugated Spanish
verbs. Examples of the bare infinitive in bozal speech include:

La vieja Asuncin nunca jabl (C-63)

yo tambin me calent ... y cuando cuch campana, yo me va pa la Tamisa (C-29)
No, si, yo no mat ninguno, yo sent atr quitrn pa yeg prisa, prisa, na panadera (C-35)
yo tindora, ya yo jabl mimo hoy don Ciriaco ... (C-15)
Na dot, n com lo chicharr caliente, beb de l'agua fra, y to la noche pas de lo catre a lo
tib ... (C-21)
 xxi

ya yo no s si lon gato mat la juta o si la juta mat lon gato ... (C-20)
Yo llev ya mucho tiempo comiendo con mano, y quer dame guto com con tened y
cuchillo lo mimo que gente rica, porque viejo no quer mor sin met pinchacito tened dentro
carne sabroso (C-32).
Can pa no seb pa n. Can pa tira tiro paf y se cay ... (C-72)
yo lo dis po bien suya (C-34)
En la guerra yo peli ... (C-36)
Bueno, sumes, siende como dis la nia ... (GC-44)
Yo sab que o Raf son guardiero tu buj ... (C-34)
Uno biyete que yo cumpr la loter yo me sac (C-34)
Yo llev ventid muelto, aqu va clito ... (C-38)
Torcuato ten que hablale ... Torcuato cog guerrillero, Torcuato cambi viejo po bueyes ...
(C-52)
si mur, mij sera (PR-7)

Examples of the use of son as the default copula include:

Cura que no son de ac, ta caram como chiva (C-29)

t son dueo e tierra (C-14)
?qu son ese? (C-6)
Tu boca son la capuyo de la susena fragante (C-10)
Ya yo son libre, yo ta cas (C-34)
porque t son mis antojos (C-15)
hoy son da grande (C-15)
?qu cosa mimo son ese? (C-21)
T no s congo bruto musulungo, t s congo luanda (C-21)
?T dic que yo son tu coraz y no retuece pecuezo pollo? (C-21)
maana son da corobata (C-24)
?T son chimosa? (C-45)
Pap son ma diablo que yo (C-26)
Calabela to nosotro son de all, to nosotro son familia africana ... (C-22)
Calamba muj so run (C-48)
t son bruto (C-20)
t no son valiente (C-13)
?Y quine son not? (C-50)
Mosotro no son cas por le iglese (C-68)
Yo que lo son mu agradiso a la fav que me lo jase ... (C-44)
t son negro que no pu ta caggao, tu mare no son rico ... (Castellanos 1983)
si t son cubano, di gabanso, y no te pasa na. (C-37)
t son criollito sinvegenza... (C-26)
Ust son muchacho hoy, pero ust son hombre maana (C-17)
mi padre son coturero, mi madre son coturera (C-38)
ud. son lo mimo que la madrecita pa nosotros (C-63)
lamo son amo brabo ... nella son honr (PR-7)

Many of the examples given above also illustrate the use of the 3rd person singular verb as the default.
 xxii

This is most often noticeable when the subject is yo. When the subject is t, normal Cuban loss of final
/s/ cannot be ruled out. The Caribbean bozal corpus contains very few examples of 3rd person singular
verbs combined with plural subjects (e.g. nosotros or ustedes), but nothing suggests that verbs behaved
any differently in these cases. More examples include:

yo empea mi ropa (C-56)

Yo alebanta sojo (C-29)
A m no bebe aguariente, mi ama (C-51)
Yo sabe lav, planch, jas dulse y cosin (C-9)
Aqu ta yo. (C-53)
yo mi ama Frasico Mandinga (C-57)
yo bota lan garaf (C-10)
ya yo ta contento (C-15)
yo no quiere mi compae con pata candel (C-21)
yo llama Mayombe sacur (C-24)
Yo va con ut, s se. (C-20)
yo no tiene la culpa. (C-67)
Aguaita que yo mata. (C-59)
Yo va a na San Sava a compr un poco de tasao (C-69)
Tot pjaro negro, el amo, yo llama Eugenio (C-49)
S, Taita, yo entiende (C-11)
yo est cort un caas (C-31)
yo va sal pa lamprasa (C-64)
Yo va pregunt a too la gente si conoce a mi yijo Eulogio (C-65)
Yo tiene la pecho premo pur nelle (C-41)
yo va donde est mi capat pa entregalo ese botella (C-50)
Yo no me ama Jos, me ama Cirilo (C-68)
Yo va pa la gallera (C-36)
si yo sabe, yo dise a su mes (C-44)
Yo tiene crupa na ma de to lo que ta pasando (C-34)
yo va s ut lo criollo cuento de mi tiela (C-26)
yo tiene mala cabeza ... (C-55)
Yo va mat mi materi (C-61)
yo le da yeba pa quit anbenenao (C-63)
yo no puere aguant m (PR-7)

Examples such as the ones above characterize the vast majority of the Caribbean bozal corpus, in
comparison with `creoloid' combinations involving ta, and show that Afro-Hispanic pidgin in the
Caribbean did show evidence of a move toward consistency, but not in the direction of TMA PARTICLE
+ VERB combinations. Rather, bozal Spanish adopted minimally inflected verb forms, at first based on
the infinitive and then gravitating toward the most unmarked conjugated form (3rd person singular).

11. Caribbean bozal Spanish in contact with Papiamento

In partial summary, Latin American bozal Spanish exhibits only one recurring trait which
converges with other Afro-Iberian creoles, namely verbal constructions based on ta. Undifferentiated
 xxiii

third person pronouns might also fit in this category, although the pronouns in question do not find as
close a match among known creoles. These two traits, however, do not characterize all or even most
Latin American bozal texts. Rather, their distribution is strikingly limited in both time and space. All
known examples occur in 19th century Cuba (together with 2-3 collateral examples from Puerto Rico).
Earlier Caribbean bozal texts show no signs of the constructions in question, nor do 18th and 19th
century Afro-Hispanic documents from the Dominican Republic, Mexico, Venezuela, Peru, Ecuador,
Argentina and Uruguay, not to mention the extensive Peninsular Spanish bozal corpus of 16-18th
century texts. Reconstruction of the vestigial Afro-Hispanic language spoken by the negros congos of
Panama (Lipski 1990) also fails to reveal evidence of earlier verbal constructions involving ta, nor of
undifferentiated third person pronouns. This very limited distribution undermines sweeping claims of a
previous stable Afro-Hispanic creole spoken widely throughout the Caribbean and elsewhere. At the
same time, the creoloid traits discussed above must be accounted for in the cases where they appear.
This dilemma may be at least partially resolved by taking a closer look at the recruitment of the labor
force in the 19th century Caribbean, and by offering a closer examination of many lesser-known
examples of Caribbean bozal Spanish.
In the final decades of the 18th century, the sugar plantation boom caused an unprecedented
demand for laborers in Cuba, Brazil, and to a lesser extent Venezuela, Trinidad, and Puerto Rico. The
African slave trade had slowed to a trickle, and was rapidly reinitiated without concern for separating
slaves speaking the same languages. Slaves were carried en masse from West Africa to Latin America,
resulting in the existence of Yoruba- KiKongo- and Ewe-speaking groups well into the 20th century.
By the first few decades of the 19th century, anti-slavery movements in Europe were strong, and slaving
ships en route to the Americas were intercepted and confiscated. A burgeoning contraband slave trade
ensued, and the Dutch slave station at Curaao was instrumental in making up the difference between
the slaves coming from Africa and the total needs of the Spanish colonies. For much of the colonial
period, the Dutch had maintained an asiento on Curaao, from which slaves were reshipped to
Spanish, French and English possessions in the Caribbean. The asiento was revoked in 1713, but
clandestine traffic from Curaao and St. Eustatius continued past this point, transshipping African slaves
throughout the Caribbean.
The importance of Curaao in the history of the African slave trade to Cuba and Puerto Rico is
well documented. For nearly two centuries, the Dutch slave depot at Curaao supplied the authorized
and clandestine slave traffic to Cuba and Puerto Rico, together with other sources of African labor.
There is no indication that Curaao played a predominant role during the period prior to the 19th
century, although the flow of slaves was constant. In the first decades of the 19th century, European
(particularly British) attempts to stop the flow of African slaves to the Americas became significant, and
the transshipment of slaves from one Caribbean island to another rose in importance, creating a
Caribbean-wide shell game which was difficult to interdict in its entirety. Western Puerto Rico and
eastern Cuba received numerous slaves from Curaao, as well as from the Danish colony in the Virgin
Islands, and from St. Barthelmy, Martinique and Guadeloupe (Morales Carrin 1978: 39).
Alvarez Nazario (1970) traces the arrival of slave and free blacks from Curaao in Puerto Rico,
during the course of the 18th and early 19th century. This study was the first to appreciate the
significance of a text originally published by Pasarell (1951: 124), which purports to represent the use
of Papiamento in early 19th century Puerto Rico.9 For Alvarez Nazario (1970: 4), the text `ofrece
pruebas de primera mano que establecen el arraigo definitivo y claro en nuestro suelo por entonces de
sectores poblacionales usuarios del papiamento, con races que se remontan posiblemente en el tiempo
a los siglos XVII y XVIII ... cuando este instrumento expresivo va definiendo y consolidando
histricamente sus caracteres de lengua criolla del Caribe ...' According to him, the language in question
 xxiv

represents the vestiges of Papiamento transplanted to Puerto Rico several generations prior to the
attestation in question, and partially remodeled through contact with evolving bozal and criollo Spanish
of Puerto Rico. The most significant aspect of this discovery, amply recognized by Alvarez Nazario, is
the fact that the language of these genti di Cors was familiar enough to observers in early 19th century
Puerto Rico as to require no special introduction or translation. Alonso (1975: 57), in the classic work
El jbaro, also referred to the presence of `criollos de Curazao' in 19th century Puerto Rico, evidently
an unremarkable phenomenon in his day.10
Granda (1973) extended the scope of the inquiry, tracing the transshipment of Africans from
Curaao throughout the Caribbean, and documenting the survival of folkloric texts in Venezuela in a
language which is much more transparently Papiamento. In the case of Cuba, Granda points to the
observations of the Dutch traveller Bosch (1836: 226), also mentioned by Hesseling (1933: 265-6),
who encountered Papiamento speakers in Cienfuegos, as well as in the Virgin Islands. According to
Granda's interpretation, Bosch's previous knowledge of Papiamento as spoken in Curaao would
assure that he was not mistaking a local Afro-Cuban creole or pidgin for legitimate Papiamento.
Hesseling himself did not rule out the possibility that Bosch was confusing the Dutch-based creole
Negerhollands with Papiamento, although according scant probability to such a hypothesis.11 Given
Cubans' negative attitudes towards the speech of bozales, it is unlikely that many observers had either
the experience or the inclination to differentiate the Afro-Hispanic pidgin spoken by African bozales
sent directly to Cuba and the already well-established Afro-Iberian creole in use on Curaao.
St. Thomas in the Virgin Islands was often a way-station for shipments originating in Curaao,
and according to Bosch (1836) also harbored a population of Papiamento speakers during the 19th
century. This provides a dual vector by which elements of Papiamento could enter Cuba and Puerto
Rico during the early decades of the 19th century. It is not insignificant that the `Papiamento' text
documented in Puerto Rico comes from Mayagez, a port on the western end of the island where slave
traffic with St. Thomas and Curaao was most intense.
The Virgin Islands connection suggests an even more subtle way in which Papiamento elements
could have trickled into bozal language of Cuba and Puerto Rico. In the early 19th century, when the
(Danish-controlled) Virgin Islands played a key role in the clandestine slave trade to Cuba and Puerto
Rico, the Dutch-based creole Negerhollands was still the principal language of the black population.
Hesseling (1933) has analyzed many features of 19th century Negerhollands as bearing the earlier
influence of Papiamento, stemming from a time when the Dutch also controlled parts of the Virgin
Islands, and transfers of Africans from Curaao to St. Thomas and St. Croix were frequent. One of the
traits attributed to Papiamento is the extension of bo < na bobo to include meanings far removed from
the original spatial value of `over'; this extension exactly corresponds to Papiamento riba as compared
to Spanish/Portuguese arriba. It will be shown below that riba/arriba occurs in bozal Caribbean texts
with an expanded range of values comparable to those of Papiamento. Although there is no evidence of
Negerhollands ever being spoken in stable populations in either Puerto Rico or Cuba, it is nearly
inevitable that some Negerhollands speakers were transferred to these Spanish colonies. In the process
of acquiring Spanish, parallel structures from Negerhollands would likely serve as the basis for
translation; combinations which had previously been calqued from Papiamento could thus enter bozal
Spanish from a language twice-removed. The Virgin Islands-Negerhollands connection is worth
considering as a potentially, as yet unproved, collateral source of creole-like elements in Caribbean
bozal Spanish.

12. Other Papiamento elements in Caribbean bozal Spanish

 xxv

Closer scrutiny of bozal texts, principally from Cuba, reveals the existence of several other
elements which point to a direct Papiamento influence; in many instances, these forms are found
precisely in the text which exhibit the benchmark creole features just surveyed. These same items are
conspicuously absent in bozal Spanish texts from other regions. Among the more salient features which
coincide partially or totally with Papiamento are:

(1) The Papiamento word for `child, son, daughter' is yiu, with an intrusive initial /y/ not clearly
derivable from Spanish
hijo. The form yijo
appears in Cuban bozal
texts from the 19th
century, but is unknown
in bozal texts from
other regions:

Mi yijo, gayina negro son mucho, y toto pone gebo blanco (C-53)
no ta sufr mi yijo (C-63)
Yija de mi pecho son (C-15)
Si maana yijo fri, ?quin llora su madrina? (C-24)
ay, yijo, yo no tiene carabela aqu. (C-22)
Yo ta compa to yijo. (C-22)
si, yijo, es mo el quimbomb (C-20)
mi yijo Eulogio, naco y criao en el Guatao (C-65)
?T no ve uno yegua para que anda con la yijo suyo como quien la tiene orgullo porque saca
lotera? (C-60)
Neye lo que tiene s un bariga con su yijo lentro. (C-54)
tu son mi yijo, arrea, vamo ... Yo no tiene muj, no tiene yijo ... (C-26)

Considering that intervocalic /x/ is weakly pronounced in Caribbean Spanish, the spelling yijo probably
represented a word similar or identical to Papiamento yiu. It is likely that the frequent combination mi
hijo, pronounced with intrusive hiatus-breaking and hypocorrect /y/, is at the root of the transformation
of hijo to yiu/yijo. However, while intervocalic /y/ is quite weak in Papiamento, it is generally strong in
Caribbean Spanish, so that independent parallel development of yijo in Cuba is rendered unlikely.

(2) In Papiamento, the word for `today' is awe, which is not easily derivable from either
Spanish hoy or Portuguese hoje.12 Forms similar or identical to Papiamento awe, and with identical
meaning, appear in several Cuban bozal texts. Examples include:

Poquitico fat pa que seora mur ago (C-35)

Age memo, amito (C-66)
ahuoy lo va a jas Pancha ... Ma ahuoy, let mi corasn ... ahuoy bariga yo saca ... Ahuoy
vamo ta mosotro como pec dentro lagua ... (C-33)
ahuoy cun ese cumeria (C-34)
ag da tamb to mundo baila (C-26)
ahuoy mimo po la maanito (C-34)
 xxvi

(3) In Papiamento, awor is the word for `now.'13 Among existing or attested Spanish dialects,
aguora/ahuora appears only in
19th century Cuba, in bozal
texts or as the representation of
illiterate rural speakers in areas
with a strong Afro-Hispanic
presence. Examples of
ahuora/aguora in Cuban bozal
texts include:

? qu yo dic ahuora, eh? ... ahuora s mi pecho est girviendo como agua que pela engallina
Y
(C-15)
y ahuora que no lo ve ... donde ahuora yo s otra v congo y trabajaore la muelle ... dende
ahuora yo s Jos mimo ... Ahuora a trabaj (C-40)
con toa esa bemba se larga ahuora mimo de aqu ... vamo a ve si ahuora ot me entiende ...
ahuora s verd que no pue m ... hasta ahuora yo no ten guto pa conoc a not
(C-50)
Prusumpueto que ahuora narie lo habra diotro cosa ma que de la Jpera (C-34)
... que bravu diese lo pega ahuora como uno pache de brea en la sojo de uno buticaria (C-34)
... la Cula ta gualando, aguola en la cafet (C-48)
Camina, pcaro, que aguora t lo va pag (C-35)
aguora yo jabla ot (Guayabo, rumores del Mayabeque, 1881; cited in Ortiz 1924: 12).

Although Ortiz and Dihigo suggest that ag and ahuora both derive from ahora/agora `now,' several
texts, including the writings of Creto Gang (C-34) and El quitrn (C-35) maintain the two separate,
with ag meaning `today' and ahuora in the sense of `now,' just as with Papiamento awe and awor.

(4) In Papiamento, riba (< Sp., Port. arriba) is a preposition meaning `on, upon': Kiko tin
riba mesa? `What is on the table?' Arriba is not used this way in Spanish, although the combination
arriba de occasionally appears with the meaning `above, over.' Among Cuban bozal texts riba/arriba
appears on several occasions as a preposition, with the meaning `on, up':

ya par rriba tngue ... ya par rriba jagey ... ya par rriba nange ... yo sube arriba palo
(C-20)
pone can riba alifante ... ut sienta riba pelo y va arastrao como en coche ... ech vara
arriba ngombe yo brinca volante ... mi pecho ta roncando, parece toro galano que et nriba la
loma ... (C-26)
Ese trepa riba palo (C-21)
Arriba entoto me juran ganga (C-24)
arriba negro, culpa siempre ta guind (C-25)
Sava ariba loma. (C-69)
no lo guanta que moca lo para riba su yo (C-33)
Yo s, moca fueron, mira como ta arriba la mesa. (C-37)
lo tiniba lumbaniyo ma pa riba la fundiyo (C-34)
Y Rupeto, qu negro sient riba la vagaso (C-34)
lo sal cribindo ariba Lan Faro Sindutri en luenga mandinga? (C-34)
 xxvii

o haba ariba de la pap muchu diabro pintao (C-34)

(5) In Papiamento, the word for `say, tell' is bisa (< Sp., Port. avisar). Among non-African
Spanish dialects, (a)visar is most commonly used in the sense of `advise, warn,' but rarely with the
meaning `say tell.' Dihigo (1946: 199) gives the form bisar, whose sole example is the Afro-Cuban
text from Santa Cruz (1908: 132), reproduced below. Although Dihigo does not explicitly list this form
as pertaining to the Afro-Cuban lexicon, the choice of examples is instructive. Among Cuban bozal
texts, however, the latter meaning is frequently associated with (a)visar:

Nia Paulita am yo, bisa negra pa ni ... Bisa Andr que ta geno ... (C-66)
Rob, visa mi seora sen que yo ta nel ro (C-21)
Don Jos, yo va ahora mimo a la cabildo pa avis too carabela pa que viene tiempla juna
tambor ... (C-15)
Gente desconfi, peligro siempre ta avis Ay Di, cuando mi maro ta juy ... (C-25)
Amo tuyo quiquirib un da, t avis m. ?T entiende? (C-11)
Madre Oc avis pa que jable con vo. (C-1)
Cuando ley Mech contenda con ley ingl, n ta sentao en su trono y vis que baco ingl ta la
baha ... no avisa no, poque hata dipu de mueta yo cuido mi muj... si hay malo, avisa pa el t
avisa pa l. (C-26)

(6) In Papiamento, the first person singular subject pronoun is (a)mi, with the longer form being
emphatic or contrastive. Since the 16th century, no form of bozal Spanish has used (a)m as subject
pronoun (cf. Lipski 1991), but this form reappears in a few 19th century Cuban bozal texts. Also
found in the latter documents is the use of disjunctive m as object pronoun, instead of the usual clitic
me. This pattern is not found in any monolingual variety of Spanish, but is parallel with Papiamento
usage. Found occasionally is the use of pa m `mine,' identical in structure to Haitian pa-m (<
pa-mwe), and possible a calque introduced by Haitian speakers who were especially frequent in eastern
Cuba. Examples of these uses of m in Cuban bozal texts include:

A m no bebe aguariente, mi ama (C-51)

Ay, siora, nigua no deja camin a m. (C-35)
Ah, amito, perdona m ... M no sabe, amito ... mi no sabe n (C-66)
colazn pa m ta brincando dentro la pecho como la cuebro (C-15)
No se, vegenza no e pa m, e pa amo Tom. (C-16)
s, i, contramayor manda m ... Ta ju, ta puj m, si (C-67)
El amo mata m ... (C-49)
T dise m, yo calla ... Eyo lleban m una casa. (C-11)
Ay, Flancico, t perdona m ... (C-37)
Nl cr va agarr m ... (C-26)
la sojo suyo, que lumbra como cocuyo, ta mirando yo nam. A m no ... Lamo ta rega m ...
(C-29)
Ecucha Encarna, m no guta eso ... ella dise, m ba casa ma Cecilia ...(C-12)

In the examples just given, use of m is combined with bozal language in which some agreement, verb
conjugation, syntactic subordination, etc. can be found. In other words, the level of approximation to
Spanish is above that of an incipient pidgin such as might be spoken by a recently-arrived African with
 xxviii

no prior knowledge of Spanish.14

(7) Although the undifferentiated third person pronoun (n)eye does not have a direct
counterpart in Papiamento, the variant ne, also appearing in Cuban bozal texts, corresponds to a variant
of Papiamento e (usually found after que, in comparative constructions), e.g. ta un Dios tin y no tin
otro mas que N `there is only one God and no other' (Hesseling 1933). Examples from the bozal
corpus include:

? u nim son ese que ne parese maj? (C-4)

Q
ne te mea, ne te caga, ne te tumba, ne te mete rabo la culo, y se va ... Na dot, n com lo
chicharr caliente ... N tiene un gequito aqu, un gequito all (C-21)
Ne mur jay tiempo ... ne contr lo ro la suete lo rey ekoi, y varn quit neye ... (C-22)
Yo va cur n cun su mec. (C-67)
Cuando ley Mech contenda con ley ingl, n ta sentao en su trono ... N mir po teojo ... si
n no tiene serrucho, y manque negro jabla mucho mucha yuca hay que ray ... ?hijo de quien
n? (C-26)

(8) One area in which similarities between 19th century Caribbean bozal Spanish and
Papiamento may indicate more than a fortuitous resemblance involves vowel harmony. The data are not
as extensive and clear as the previously-mentioned phenomena, but the parallels are worth exploring.
Papiamento exhibits numerous cases in which etymological final vowels from Spanish or Portuguese
were replaced by a copy of the stressed vowel, in a fashion which transcends simple raising of
unstressed /e/ and /o/ to [i] and [u] (common in Papiamento but also found in Portuguese and in some
regional dialects of Spanish). Typical cases include aa < ao, biaha < viaje, caya < calle, siboyo <
cebolla, staa < estao, paa < pao, shinishi < ceniza, dede < dedo, etc. (cf. Birmingham 1970:
25-6). Pretonic vowels were sometimes also drawn into the harmonizing process: rospond <
responder, sosod < suceder. There is no indication of the chronology of this vowel harmony in the
history of Papiamento, whether it occurred only during the formative stage (when paragogic and
epenthetic vowels were also occasionally added), or whether the tendency persisted as an active
process in later stages of the language.
A comparative study of bozal Spanish from earlier time periods (16th-18th century), in Spain
and Latin America, gives evidence of both vocalic epenthesis and occasional vowel harmony, typified
by the frequent occurrence of sioro/siolo < seor and Dioso < Dios. In the 19th century Caribbean,
however, vocalic epenthesis rarely occurred; widespread elimination of syllable-final consonants created
the appropriate syllabic structure (usually consisting of a single consonant+vowel combination) without
recurring to additional vowels. In the 19th century bozal corpus from Cuba and Puerto Rico, however,
there are several interesting instances of final vowel harmony which cannot be explained by unstressed
vowel raising. This is in addition to the dozens of examples where final /o/ is raised to [u] and pretonic
and postonic /e/ is raised to [i], creating configurations very similar to those of Papiamento. Among the
instances of possible vowel harmony gleaned from the Caribbean bozal corpus are:

bngama < vlgame (C-35); beye < bella (C-10); botelle < botella (C-50); buene
< buena (C-21); bunite < bonito (C-68); caserite < cas(er)ita (C-68); clava <
esclavo (C-68); cosito < cosita (C-34); cuele < cuela (C-50); cusitu < cosita
(C-15); diaquelle < de aquella (C-34); dielle < de ella (C-34); dinele < dinero
(C-57); dinera < dinero (C-68); entierre < entierro (C-68); Francisque <
 xxix

Francisco (C-68); grese < iglesia (C-29); jierre < hierro (C-68); linde < linda
(C-68); meme < mismo (C-68); tguiri < tigre (PR-2).

There are not enough examples to base a strong claim of Papiamento carryover, but the examples just
cited all appear in texts in which other Papiamento-like elements occur. When added to the other
features just enumerated, they strengthen the notion that direct contact with Papiamento played a role in
shaping 19th century Cuban and Puerto Rican bozal speech.

(9) One final trait of many Cuban bozal texts also reminiscent of Papiamento is also worth
mentioning, as it may help illustrate the type of influence which a well-developed creole might influence
on the pidignized Spanish of newly-arrived African slaves. Papiamento is known for its creation of
nouns from verbal stems through use of the suffix -mento, e.g. papiamento `language, speaking' <
papi `to speak,' etc. (Lenz 1927: 292-3; 1928). Similar forms are of course found in both Spanish
and Portuguese, and indeed one or both of these languages provided the model for Papiamento itself.
In the latter language, however, greater flexibility is found in creations with -mento, including many
combinations not found in Spanish or Portuguese. Among bozal Spanish texts from 19th century Cuba
a number of nouns in -miento are found which do not correspond to etymological Spanish items, and
which are not to be found in bozal texts from other regions or time periods. These include:

alegramiento (C-40); avisamienta (C-34); bautizamiento (C-39); butisamienta

(C-34); cagamiento (C-34); calenturiamento (C-34); chiflamienta (C-26; C-33;
C-34); cumpaamiente (C-34); cuntentamiento/cuntentamienta (C-34);
didicamienta (C-34); digutamiento (C-34); dormiento (C-39); gopiamiento
(C-34); gisiamiento (C-34); imbrujulamiento (C-34); pidimienta (C-34);
presiamienta/presiamiento/prisiamiento (C-33; C-34); pruvechamienta (C-34);
relambimiento/rilambimienta (C-33; C-50); simprisionamiento (C-34);
sincuntansiamiento (C-34); sofocamiento (C-34); soramienta/soramiento (C-34;
C-36); titulamiento (C-34); trivimiento (C-33)

13. A model for bozal-Papiamento linguistic contacts

The trustworthiness of the bozal texts analyzed above varies widely, although despite the wide
range of motives which apparently prompted the works to be written, and the attitudes towards
Africans displayed in the works, there is surprising homogeneity of linguistic details. Among the most
accurate observations were those of the late Lydia Cabrera, who respected Afro-Cuban culture and
whose abilities as a transcriber of other linguistic manifestations of Afro-Cubans enabled her to
penetrate many subcodes used during the last century. Significantly, every one of the possible
Papiamento traits enumerated above appear in Cabrera's extensive writings; hers are in fact the only
texts in which all traits converge. Fernando Ortiz was another sympathetic and meticulous observer;
although he did not offer extensive transcriptions of Afro-Hispanic language, the descriptions he did
offer coincide with independently-documented Afro-Hispanic linguistic manifestations, and can be
assigned a high degree of credibility. The aristocratic Emilio Bacard Moreau was also an accurate and
objective observer of bozal speech. Even such obviously burlesque work as the writings of `Creto
Gang' reveal an almost complete coincidence with more objectively described Afro-Cuban language.
A comparison of the texts in which Papiamento-like elements occur reveals a tendency for the clustering
of two or more such features in each work. This provides a partial verification of the hypothesis that
 xxx

direct transfer of Papiamento elements may have contributed to the development of Caribbean bozal
Spanish.
The influence of Papiamento on Afro-Caribbean Spanish could never have been intense or
dominant; otherwise the bozal corpus would reveal `purer' instances of Papiamento instead of a
scattering of isolated forms. Even the putatively Papiamento text from Puerto Rico shows few
unmistakably Papiamento forms. The raising of final unstressed /o/ to [u], although typical of most
Papiamento dialects, is not unknown in other Spanish dialects, and is found in bozal Spanish from 17th
century Spain and in Latin American bozal language from the 17th-19th centuries. The same holds for
the pronunciation of de as di, and of gente as genti (indeed, the Papiamento word is hende).
Papiamento has tur `all' corresponding to Spanish todo. In other bozal texts, flapping of intervocalic
/d/ has given turo or turu. The word tutur in the `Papiamento' text from Mayagez may be a hybrid of
Papiamento and Afro-Hispanic forms. Julands does correspond to the Papiamento pronunciation (cf.
Spanish holands), although archaic aspiration of Spanish h is also common in bozal language. The
text is very Hispanized, and was evidently produced by second- or third-generation speakers of
Papiamento, or by a mixed group of Spanish- and Papiamento-speaking singers.
By the time of the bozal Cuban and Puerto Rican texts which contain the Papiamento-like
forms, it can be assumed that few if any monolingual Papiamento speakers were to be found in these
areas. The Papiamento forms surveyed above seeped into Caribbean bozal Spanish over as much as a
century, presumably spurred by the linguistically privileged position which Papiamento-speaking slaves
or free laborers would have in comparison with African-born bozales. The latter would speak only the
barest rudiments of Spanish, and would be at the mercy of interpreters or commands presented in the
simplest terms. Regardless of the possible existence of a Portuguese-based pidgin among earlier
generations of African slaves, bozales brought to Cuba and Puerto Rico in the late 18th and early 19th
centuries were usually whisked out of Africa by legal or illegal means. Few were held in Portuguese
slaving stations such as So Tom or Cape Verde where they would have acquired a Portuguese-based
pidgin. In addition, by the 18th century the Portuguese primacy in the African slave trade to Spanish
America had been superceded by British and Dutch entrepreneurs, in whose slaving empires there is no
evidence of pidgin Portuguese. An African captive arriving in Cuba or Puerto Rico would then have to
rely more heavily on intercommunication with speakers of the same African languages, an increasingly
frequent phenomenon in the latter decades of the slave trade. A speaker of Papiamento, on the other
hand, would experience fewer difficulties in understanding Spanish; some may have already been in
contact with (Venezuelan) Spanish on Curaao. This would give Curaao natives an advantage in the
slave communities which could well be translated into positions of relative authority and influence within
slave groups in Cuba and Puerto Rico. Bozal Africans acquiring survival-level Spanish would most
naturally turn to fellow laborers whose abilities in some variety of `Spanish ' were superior to their own;
indeed in many instances this might provide the only contact with Spanish on a regular basis.
Since many Papiamento forms are so similar to Spanish as to attract little notice, it is impossible
to determine the precise Papiamento contribution to Caribbean bozal Spanish. For example, although
Papiamento has no nominal/adjectival inflection for gender, the variable appearance of noun-adjective
agreement in bozal texts does not rule out Papiamento influence. Bozal speakers covered a wide range
of fluency in Spanish, and while those in closest contact with Spanish might well acquire a partial
concordance system, those whose contact with Spanish was minimal would develop the sort of
haphazard and precarious agreement found among currently-existent Spanish pidgins and vestigial
varieties.15 If a bozal Spanish speaker were to encounter Papiamento nouns and adjectives in which
agreement was totally missing, this would not be noticed, since the `basilectal' bozal pidgin speaker
would not have an accurate standard of comparison. Thus a single speaker might learn some Spanish
 xxxi

words from native or near-native speakers, with correct or nearly correct Spanish agreement, and other
words from Papiamento speakers in which all agreement was absent. Only if a bozal speakers became
totally fluent speakers in Papiamento would all inflection be eliminated, and such individuals were
obviously very few in number. Since all descriptions and imitations of bozal speech were written by
native speakers of Spanish with no apparent knowledge of Papiamento, the subtle differences between
the variable concordance of bozal Spanish and the uniformly agreementless Papiamento might escape
notice.
Papiamento elements would potentially be borrowed by bozal Spanish speakers in direct
proportion to the structural similarity with (vernacular varieties of) Spanish. Thus, for example, verbal
combinations based on ta + Vinf sound similar to those of vernacular Spanish in which est is reduced to
ta (Lipski forthcoming, a). In modern Papiamento, ta is sometimes followed by the Spanish gerund
(thus converging with the Spanish progressive forms) instead of the usual uninflected stem derived from
an earlier infinitive; if such alternation is not just the result of recent partial decreolization, but has always
characterized Papiamento, then bozal speakers would identify ta + INF and ta + GERUND constructions,
using both as free variants. Both forms, as well as several intermediate varieties, are found among the
texts of the Cuban bozal corpus.
In a similar vein, awor/ahuora vs. ahora and awe/age vs. hoy have identical function in
Spanish and in Papiamento/bozal speech, and a bozal speaker would not likely notice identify the forms
as belong to two separate languages. Putative Papiamento forms identified in bozal Caribbean texts are
convergent with equivalent Spanish forms, in both syntax and general phonological shape. Papiamento
features which are not shared by Spanish are not attested in the bozal corpus. For example, the
marking of nominal plurality by postposing of the third person plural pronoun (e.g. homber `[the] man'
vs. hombernan `[the] men') is a process totally at odds with Spanish pluralizing strategies, and is not
found in even the most creoloid bozal Spanish texts. Future/irrealis in Papiamento is marked by lo,
which is placed before subject pronouns (e.g. lo mi bai `I will go'), and between a subject full NP and
the verb (e.g. Maria lo bai `Mary will go'). Although Papiamento lo is plausibly derived from Spanish
luego/Portuguese lgo, this identification is lost in modern Papiamento, and a bozal speaker would not
likely notice this subtle and syntactically variable particle, which does not converge behaviorally with any
single Spanish element. In bozal Spanish, future, when signalled at all, is based on the Spanish
periphrastic construction with ir, usually via the form va. Papiamento signals past/perfective with
preverbal a, which structurally occupies the same position as Spanish auxiliar verbs. However in
spoken Papiamento, a frequently fuses with subject pronouns (e.g. mi a vini > ma vini `I came'), so
that its existence as a separate element is often obscured and would not likely be noticed by a bozal
speaker. Bozal texts show no consistent signalling of past/perfective. Usually just an unconjugated verb
is used, and occasionally a Spanish perfective or preterite form appears.

14. Summary

The preceding discussion is not intended as a total rejection of a possible creole origin of bozal
Spanish in the Caribbean, but includes a call for caution in not underestimating the complexity of the
ethnolinguistic environment in which Afro-Hispanic language evolved in the 19th century Caribbean.
Written attestations of Afro-Hispanic language are varied and at times misleading, including accurate
reproductions of bozal speech and humorous exaggerations. Basing claims of a previous creole basis
on a small and non-representative sample of texts yields conclusions which suggest that a stable creole
was once spoken in Cuba and Puerto Rico. The evidence surveyed above, however, points to a much
different conclusion. Caribbean bozal Spanish always represented a pidginized version of regional
 xxxii

vernacular Spanish, with the latter varieties already characterized by considerable phonetic and
morphological reduction. Superimposed on this common bozal language was an infusion of true
Afro-Hispanic creole, in the form of Papiamento structures used by laborers imported from Curaao.
The combination of historical demographics and textual attestations renders the possibility of fortuitous
similarities extremely unlikely. The Papiamento-like features found in the Cuban and Puerto Rican bozal
corpus are not found in other Spanish-speaking regions, even in Afro-Hispanic language, and are not
found in all Caribbean bozal texts. These features generally cluster in the bozal texts in which any one
of them occurs. The time and place also coincide: creoloid Papiamento-like features are found only in
19th century Cuban and Puerto Rican texts, precisely at the time when Papiamento speakers were
known to have been living in these two countries. The Papiamento contribution may on occasion have
been supplemented by calques from French Creole and even Negerhollands, given the documented
presence of speakers of these languages in 19th century Cuba and Puerto Rico.
Bozal speakers in contact with Papiamento may have used items from the latter language to
extend their own pidginized Spanish, and in some cases a restructuring of bozal in the direction of a
more Papiamento-like creole may have occurred. In general, however, the creole influence on
Caribbean Spanish appears to have been discrete, contributing words and occasional verbal
combinations but never completely recasting the authentically local bozal dialects.
Also contributing to the notion that Caribbean bozal Spanish was a stable creole with
Afro-Lusitanian roots were hybrid African-Spanish combinations, often involving a Spanish auxiliary
verb such as (es)t or va plus an unassimilated African element. The resulting configurations often bear
a superficial similarity with Afro-Iberian creole PARTICLE+VERB constructions, but are better analyzed
as code-switches or unassimilated borrowings. The same analysis, it has been suggested, can also be
extended to bozal verbal combinations involving only Spanish elements, in which the `verb' is employed
as an unanalyzed morpheme, combined with the rudiments of an auxiliary system, or used alone. The
presence of `particle + verb' constructions in prominent bozal texts such as those of Lydia Cabrera has
created the impression of a much more sophisticated and creole-like verbal system in Caribbean bozal
Spanish than is warranted.
To further refine the hypotheses presented here, the search must be extended, to bring to light
more bozal texts from Cuba and Puerto Rico, and also from Venezuela, where transplanted nuclei of
Papiamento speakers are known to have arisen. Much remains to be done before the reconstruction of
Afro-Hispanic language in the Caribbean can be considered complete; the preceding remarks are
offered as a contribution to the ongoing enterprise.
 xxxiii

Notes
1
Megenney (1985b ) gives one example of a 17th century Afro-Mexican text in which
Portuguese elements predominate. However, later bozal texts from Mexico contain only Spanish items.

2
The word nan, in nan gato, is identical to the Papiamento plural marker and third person
plural pronoun; Alvarez Nazario (1959, 1974: 167, 185-87) considers the use of nan in some
Afro-Caribbean bozal texts as related to a possible common African substrate underlying both bozal
Spanish and Papiamento. Wagner (1949: 158-9) comes to a similar conclusion with respect to the
item lan used in Cuban bozal texts. In Papiamento, however, pluralization is marked by placing nan
AFTER the noun to be pluralized; preposed nan has no function in this language. Moreover, in the
example cited by Pichardo, it is likely that what is written as nan was actually a prenasalized stop on the
following word: como ngato `like [a] cat.' This process is well-documented for bozal language (Lipski
1992), as in the formation of generic comparatives by como + NOUN, where other Spanish varieties
would prefer an indefinite article.
3
The `article' lan, like nan, has been considered by Alvarez Nazario to reveal an African
origin, possibly related to KiKongo or Kimbundu elements in the Gulf of Guinea Afro-Portuguese
creoles. Alvarez Nazario (1974: 167, 185-197) postulates that the original form was nan, and that the
change nan > lan took place through the influence of the definite article la. In Puerto Rican bozal
Spanish, both lan and nan are found, but in Cuban texts, lan (with occasional variant lon) occurs
almost exclusively. If the occurrence of lan/nan in Cuba and Puerto Rico stems from a common
extraterritorial source, then the existence of both forms in Puerto Rico and the predominance of the
former in Cuba would suggest an evolution lan > nan, initiated and only partially completed in Puerto
Rico. The opposite development would be suggested only if it could be demonstrated that lan/nan was
attested in Cuba significantly before appearing in Puerto Rico, having undergone the putative evolution
nan > lan before the latter form was transferred to Cuba, via an as yet unattested route of linguistic
transplantation. However, a comparative search of Afro-Hispanic texts from Spain and Latin America
(Lipski forthcoming a) shows that lan occurs from the early 17th century on, both in Spain and in
Spanish America, including Puerto Rico. The almost total restriction of nan to 19th century Puerto
Rican texts thus suggests a route of evolution opposite to that suggested by Alvarez Nazario, namely lan
> nan, if in fact the two items are related etymologically. Regardless of the relationship of nan and lan,
the analysis proposed in Lipski (1987b, forthcoming a) views lan as merely an article followed by a
prenasalized obstruent at the beginning of the next word. No Papiamento or other Afro-Iberian creole
element is required in this equation.
4
Thus for example Granda (1972: 12) discloses that `estaba a punto de abandonar mi
bsqueda ... cuando el extraordinario libro El monte, de la gran investigadora Lydia Cabrera ... me
proporcion, al fin, los datos que ya desista de conseguir. Cantos de carcter religioso y, sobre todo,
trozos de conversaciones mantenidas por ancianos negros habaneros con la autora son transcritos por
Lydia Cabrera con toda fidelidad y, sin lugar a dudas, ofrecen caractersticas lingsticas criollas, an
ms puras que los textos puertorriqueos ... en Cuba, todava el habla criolla ms autntica y menos
desnaturalizada ... era usada corrientemente, no hace an veinte aos, por las generaciones negras
ancianas en la propia capital y de modo totalmente espontaneo y normal.'
 xxxiv
5
Perl (1989b) explicitly defends the use of a few presumably authentic texts such as El monte
and the exclusion of a larger literary corpus, including the Cuban `teatro bufo' and many novels, stories
and poems, given the tendency to exaggerate and stereotype in literature. Although this caution is no
doubt warranted, not all observers are so pessimistic about the linguistic value of Afro-Cuban literary
texts. Thus for example Garca et. al. (1981) and Garca Gonzlez et al. (1984), echoing the sentiments
of Leal (1975: 18), find considerable linguistic accuracy in the `teatro bufo,' as regards bozal speech
and in the other typical stereotypes portrayed in this comic drama.
6
The form oku is probably a combination of the {3sg.} subject clitic , the stem k, and the
rhetorical final -o, common in spoken Yoruba.
7
This same process is observable, e.g. in Ese otro yo me lo va yun `I will eat that other
one.' In Yoruba jeun is an intransitive verb meaning `eat.' As such, it can never be combined with an
object clitic or NP, as in the example cited by Cabrera. The transitive verb `eat' is je; however in the
Spanish-Yoruba hybrid, the longer intransitive form has been adopted. Alvarez Nazario (1974: 216)
observes that `negros criollos' (i.e. those born in Spanish America) are often unaware of the internal
structure of borrowings from Yoruba and other African languages, and that `los descendientes de los
antiguos esclavos emplean en expresiones mixtas el espaol y el anag [= Yoruba: JML]: "me voy
pa(ra) el inl" ...'
8
The pattern is similar to the constructions based on hacer + ENGLISH WORD, at times found in
the Spanish of Mexican-Americans from California: hacer fix `to fix,' hacer type `to type,' etc. (cf.
Reyes 1976).
9
The text in question comes from a pamphlet describing festive activities realized in the south of
Puerto Rico in 1830, celebrating the birth of the heiress to the throne of Fernando VII, who would
become Isabel II. Among the songs and dances described in the pamphlet is the following song,
attributed to the `mulatos holandeses que residan en el Sur' (Pasarell 1951: 124):

Tutur genti de Cors

Celebrado Rey Fernandu
Mambatindu ha tambur
Di contentu boy bailandu.
Comparsa di Julands
Celebra a Reyna Cristina
Nan Princesa naroyna
Novo astro boy mir.

The language of the text, while clearly written in a type of `jerga' (the term used by Pasarell), is not
Papiamento, although bearing a number of resemblances to the latter language, as will be seen below.
However, the attribution of this text to natives of Curaao, and the references to Curaao and its history
in the song itself, suggest that some form of Papiamento was once to be found among the `mulatos
holandeses' residing in Puerto Rico.
10
A more contemporary example, from a poem published in 1947 (Rodrguez de Nolla (1947:
63; PR-8) is:
 xxxv

Ao nuevo dand, ao tabin

This latter example is reminiscent of Papiamento ta bini; dand finds no ready explanation. Puerto
Rican bozal texts provide no examples of undifferentiated third person programs or other creoloid
features.
11
There are other indications, however, that Papiamento was not widely known in 19th century
Cuba, at least by that name. The attestations just mentioned come from the eastern end of the island,
which was sparsely populated during most of the 19th century, and where cultural contact was very
limited with Havana, the center for cultural diffusion and writing. Most educated Cubans of the time
period were familiar primarily with life in urban areas, particularly Havana, and were not aware of the
speech of peons on remote sugar plantations. Thus the Havana resident Bachiller y Morales (1883:
102- 3) noted that `en mi dilatada vida, ni o hablar del papiamento, ni hubiera conocido su existencia a
no haber salido de Cuba' [in all my long life, I never heard of Papiamento, nor would I have learned of
its existence if I had not left Cuba]. Since Bachiller y Morales' (1812-1889) period of observation
would have begun just shortly after the visit of Bosch, this indicates that caution must be used in claiming
widespread use of Papiamento in 19th century Cuba. A mitigating circumstance is that the term
Papiamento was rarely used by outsiders to describe the speech of Curaao; terms such as espaol
araado or espaol degenerado were more common.
12
Birmingham (1970: 21) proposes that Pap. awe developed from Spanish hoy `... involving
... the breaking of a diphthong ... the Spanish diphthong [oj] is broken into two separate syllables [o]
and [i], and further, that the [o] has opened to [a] and the [i] to [e]. The semiconsonant [w] is then
produced to facilitate pronunciation. This process is not at all unlike the one that is observed in certain
varieties of American English, particularly in the South, in which the word boy is pronounced ['bowI].'
This may be a possible route of evolution, but unlike in the Southern dialects just mentioned, where
breaking is generalized and not confined to specific lexical items, Papiamento awe is unique in the
breaking of a Spanish diphthong. Ortiz (1924: 11) gives the form ag, with the meaning `ahora.'
Dihigo (1928) refers to this item as a `forma adverbial que lleva en s la caracterstica del habla vulgar en
boca de la clase de color,' proposing a step-by-step evolution from ahora. In nearly all the exmples in
which the form occurs, however, it is possible to substitute hoy, especially since at the vernacular level
in many Spanish-speaking regions (particularly in Central America), hoy is used in the general sense of
`now.' Ortiz (1924: 11) glosses ahuoy as `hoy,' suggesting the influence of Congo (i.e. KiKongo)
guau or oguau `now.' In Palenquero, the word for `today' is also ag, a fact which reinforces
theories which link these two creoles to an earlier common source. Schwegler (1989: 17), rejecting
earlier claims of an African etimology for ag, correctly recognizes the word as of Ibero-Romance
origin, and cites the variant ge in Asturian-Leonese, as well as Ortiz' mention of Cuban Spanish. In
Cuba, however, this word was exclusively found in Afro-Hispanic speech, usually among bozales but
sometimes extending to native-born Afro-Cubans. In the case of Caribbean bozal Spanish, the
influence of archaic dialectal forms from the Iberian Peninsula is quite unlikely, while direct transfer from
Papiamento is a more plausible explanation.
13
Birmingham (1970: 29) suggests an alternative variant *aguora for the Portuguese/Old
Spanish agora. Maduro (1960: 11) cites the Murcia variant agua [awa] `now' in connection with
Papiamento awor, although without explicitly claiming the former word as an etymon. Ortiz (1924: 12)
 xxxvi

suggests KiKongo guau/oguau `now' as a possible contributing factor. The variant aguola (exhibiting
intervocalic shift of /r/ > [l], typical of Afro-Hispanic pidgin) is also attested among Afro-Cubans (Ortiz
1924: 12; Dihigo 1928).
14
At least one example of the latter type of speaker also appears in the bozal corpus, in the
story `Los chinos' by Alfonso Hernndez Cat (C-46). In these short fragments, use of m as subject
pronoun is also found:

m no importar guardias. m tener un machete y mater todos de noche, igual que en matadero.
m saber bien.

The use of m in this case could indicate an initial contact with Papiamento, an acquaintance with an
Afro-Lusitanian creole, or a carryover from West African Pidgin English. By the 19th century, the
Portuguese contribution to the Cuban slave market was minimal, and there is no evidence of arrivals,
e.g. from So Tom, but rather directly from the African mainland, or via other Caribbean islands such
as Curaao and St. Eustatius. On the other hand, the transfer of West African Pidgin English forms to
19th century Cuban bozal Spanish has been well-documented, e.g. by Ortiz (1916) (cf. also Garca
1973: 350), so yet another source of creoloid structures in Caribbean bozal Spanish might be
postulated.
15
Such a situation appears to have obtained in Cape Verde, where vestiges of Portuguese
agreement are found, whether through more recent decreolization or as long-standing traces of the
original Portuguese system. The same holds for Palenquero, where Spanish-like agreement on some
elements may suggest recent re-borrowings from Spanish or a trait which has carried over from the
earliest stages.
 xxxvii

References

Acosta-Rubio, Ral. 1976. Quiquirib Mandinga (se lo llev el diablo). Miami: Ediciones Universal.
Aimes, Hubert H. S. 1907. A history of slavery in Cuba 1511 to 1868. New York: G. P. Putnam's
Sons.
Alonso, Manuel. 1975. El jbaro. Barcelona: Editorial Vosgos.
Alvarez Nazario, Manuel. 1959. Notas sobre el habla del negro en Puerto Rico durante el siglo XIX.
Revista del Instituto de Cultura Puertorriquea 2.43- 8.
_____. 1970. Un texto literario del papiamento documentado en Puerto Rico en 1830. Revista del
Instituto de Cultura Puertorriquea 47.9-20.
_____. 1972. El papiamento: ojeado a su pasado histrico y visin de su problemtica del presente.
Atenea (Mayagez) 9.9- 20.
_____. 1974. El elemento afronegroide en el espaol de Puerto Rico. San Juan: Instituto de Cultura
Puertorriquea, 2nd ed.
Aretz de Ramn, Isabel and Luis Felipe Ramn y Rivera. 1955. Resumen de un estudio sobre las
expresiones negras en el folklore musical y coreogrfico de Venezuela. Archivos Venezolanos
de Folklore 3(4)65- 73.
Bacard Moreau, Emilio. 1914. Va crucis. Havana: Imp. El Cubano Libre/Barcelona: Imp. Vda. de
Tasso.
_____. 1916-7. Doa Guiomar. Havana: Imp. El Siglo XX de A. Miranda.
_____. 1972. Filigrana. Madrid: Obras completas de E. B. M.m reeditadas por Amalia Bacard
Cape.
Bachiller y Morales, Antonio. 1883. Desfiguracin a que est expuesto el idioma castellano al
contacto y mezcla de razas. Revista de Cuba 14.97- 104.
Ballagas, Emilio. 1946. Mapa de la poesa negra. Buenos Aires: Pleamar.
Barnet, Miguel. 1966. Biografa de un cimarrn. Cuba: Instituto de Etnologa y Folklore.
Bentez del Cristo, Ignacio. 1930. Los novios catedrticos. Archivos del Folklore Cubano
5(2).119- 46.
Berenguer y Sed, Antonio. 1929. Tradiciones villaclareas, tomo I. Habana: Imprenta y Papalera de
Rambla, Bouza y Ca.
Birmingham, John. 1970. The Papiamentu language of Curaao. Ph. D. dissertation, University of
Virginia.
Bosch, Gerardus-Balthasar. 1836. Reizen in West Indi, vol 2. Utrecht: N. van der Monde.
Brau, Salvador. 1894. Puerto Rico y su historia. Valencia: Imprenta de Francisco Vives Mora.
Buxton, Thomas Powell. 1839. The African slave trade. Philadelphia: Merrihew and Thompson,
Printers.
Cabrera, Lydia. 1970a. La sociedad secreta Abaku. Miami: Editorial C. R.
_____. 1970b. Refranes de negros viejos. Miami: Ediciones CR.
_____. 1970c. Anag--vocabulario lucum (el yoruba que se habla en Cuba). Miami: Coleccin del
Chiherek.
_____. 1971. Ayapa: cuentos de jicotea. Miami: Ediciones Universal.
_____. 1975. Anaforuana: ritual y smbolos de la iniciacin en la sociedad secreta Abaku. Madrid:
Ediciones R.
_____. 1976. Francisco y Francisca: chascarrillos de negros viejos. Miami: Editorial C. R.
_____. 1979. Reglas de congo. Miami: Editorial C. R.
_____. 1980. Yemay y Ochn. New York: CR (Eliseo Torres).
 xxxviii

_____. 1983. El monte. 3rd ed. Miami: Editorial C. R.

_____. 1989. Los animales en el folklore y la magia de Cuba. Miami: Ediciones Universal.
Cabrera, Manuel. 1973. Fragment of poem. Iniciacin a la poesa afro-americana, ed. by Oscar
Fernndez de la Vega and Alberto Pamies, 122-131. Miami: Ediciones Universal.
Cadilla de Martnez, Mara. 1953. La poesa popular en Puerto Rico. San Juan: Imprenta Venezuela
[doctoral thesis, Universidad de Madrid, 1933].
Calcagno, Francisco. 1977. Romualdo: uno de tantos. Noveletas cubanas, ed. by Imeldo Alvarez,
277-388. Havana: Editorial de Arte y Literatura.
Castellanos, Isabel. 1983. Elegua quiere tambo: cosmovisin religiosa afrocubana en las canciones
populares. Cali: Universidad del Valle.
Castellanos, Jorge and Isabel Castellanos. 1988. Cultura afrocubana 1: el negro en Cuba,
1492-1844. Miami: Ediciones Universal.
_____. 1990. Cultura afrocubana 2: el negro en Cuba 1845-1959. Miami: Ediciones Universal.
Creto Gang (Bartolom Jos Crespo y Borbn). 1975. Un ajiaco o la boda de Pancha Juta y
Canuto Raspadura. Teatro bufo xiglo XIX, antologa, tomo I, ed. by Rine Leal, 47-93.
Havana: Editorial Arte y Literatura.
Cruz, Mary. 1974. Creto Gang. Havana: Instituto Cubano del Libro `Contemporneos.'
Estrada y Zenea, Ildefonso. 1980. El quitrn. Havana: Editorial Letras Cubanas.
Feijo, Samuel (ed.). 1965. Sabidura guajira. Havana: Editorial Universitaria.
_____. 1979. Cuentos cubanos de humor. Havana: Editorial Letras Cubanas.
_____. 1980. Cuarteta y dcima. Havana: Editorial Letras Cubanas.
_____. 1981. Cuentos populares cubanos de humor. Havana: Editorial Letras Cubanas.
Fernndez Marrero, Jorge. 1987. Ein historiographisches Dokument zur "Lengua Bozal" in Kuba: el
Catecismo de los Negros bozales (1795). Beitrge zur Afrolusitanistik und Kreolistik, ed. by
Matthias Perl, 37-45. Berlin: Akademie der Wisschschaften der DDR, Zentralinstitut fr
Sprachwissenschaft, Linguistische Studien 172.
Friedemann, Nina S. de and Carlos Patio. 1983. Lengua y sociedad en el Palenque de San Basilio.
Bogot: Instituto Caro y Cuervo.
Galn, Natalio. 1983. Cuba y sus sones. Valencia: Pre-Textos.
Garca, Angel Cristbal, Gonzalo Mndez Vzquez, Ricardo Reyes Perera, and Jos F. Garca. 1981.
Apuntes sobre el lxico del teatro bufo en el siglo XIX. Islas 69.169-209.
Garca, Calixto. 1973. El negro en la narrativa cubana. Ph. D. dissertation, City University of New
York.
Garca Gonzlez, Jos, Ricardo Reyes Perera, Luis Alfaro Echevarra, and Susana Carreras Gmez.
1984. Algunos fenmenos morfosintcticos en textos populares cubanos del siglo XIX. 1984.
Islas 77.61-76.
Garca Herrera, Rosala. 1972. Observaciones etnolgicas de dos sectas religiosas afrocubanas en
una comunidad lajera, La Guinea. Islas 43.143-181.
Gelabert, Francisco de Paula. 1875. Cuadros de Costumbres cubanas. Havana: Imprenta de la
Botica de Santo Domingo.
Gili Gaya, Samuel. 1960. Funciones gramaticales en el habla infantil. Ro Piedras: Universidad de
Puerto Rico. Publicaciones pedaggicas, serie II, nmero 24.
_____. 1972. Estudios sobre lenguaje infantil. Barcelona: Bibliograf.
Girn, Socorro. 1991. Vida y obra de Ramn Mndez Quiones. San Juan: Instituto de Cultura
Puertorriquea. 3rd ed.
Gonzlez, Carlisle and Celso Benavides. 1982. ?Existen rasgos criollos en el habla de Saman? El
 xxxix

espaol del Caribe, ed. by Orlando Alba, 105-132. Santiago de los Caballeros: Universidad
Catlica Madre y Maestra.
Goodman, Morris. 1987. The Portuguese element in the American creoles. Pidgin and creole
languages, essays in memory of John E. Reinecke, edy b. Glenn Gilbert, 361-405. Honolulu:
University of Hawaii Press.
Granda, Germn de. 1968. La tipologa `criolla' de dos hablas del rea lingstica hispnica.
Thesaurus 23.193- 205.
_____. 1970. Un temprano testimonio sobre las hablas `criollas' en Africa y Amrica. Thesaurus
25.1-11.
_____. 1971. Algunos datos sobre la pervivencia del "criollo" en Cuba. Boletn de la Real Academia
Espaola 51.481-491.
_____. 1972. Estado actual y perspectivas de la investigacin sobre hablas criollas en
Hispanoamrica. Anuario de Letras 10.5-27.
_____. 1973. Papiamento en Hispanoamrica (siglos XVII- XIX). Thesaurus 28.1- 13.
_____. 1976. Algunos rasgos morfosintcticos de posible origen criollo en el habla de reas
hispanoamericanas de poblacin negra. Anuario de Letras 14.5-22.
Guirao, Ramn. n. n. Cuentos y leyendas negras de Cuba. Havana: Ediciones Mirador.
_____. 1938. Orbita de la poesa afrocubana 1928-1937. Havana: Ucar Garca.
Hernndez Cat, Alfonso. 1946. Los chinos. Cuentos cubanos contemporneos, ed. by Jos Antonio
Portuondo, 27-33. Mexico: Editorial Leyenda.
Hesseling, D. C. 1933. Papiamento en Negerhollands. Tijdschrift voor Nederlandsche Taal- en
Letterkunde 52.265- 88.
Holm, John. 1989. Pidgins and creoles, volume II: reference survey. Cambridge: Cambridge
University Press.
Klein, Herbert. 1967. Slavery in the Americas: a comparative study of Virginia and Cuba. Chicago:
University of Chicago Press.
Knight, Franklin. 1970. Slave society in Cuba during the nineteenth century. Madison: University of
Wisconsin Press.
Lachataere, Rmulo. 1938. ?Oh, mo Yemay! Manzanillo, Cuba: Editorial "El Arte."
Lapesa, Rafael. 1980. Historia de la lengua espaola. Madrid: Gredos, 8th ed.
Laurence, Kemlin. 1974. Is Caribbean Spanish a case of decreolization? Orbis 23.484- 99.
Lavia, Javier (ed.). 1989. Doctrina para negros: explicacin de la doctrina cristiana acomodada a la
capacidad de los negros bozales, de Nicols Duque de Estrada. Barcelona: Sendai.
Leal, Rine. 1975. Teatro bufo, siglo XIX, antologa, tomo I. Havana: Editorial de Arte y Literatura.
Lenz, Rodolfo. 1927. Gramtica del papiamento. III: lexicologa. Anales de la Universidad de Chile
ser. 2, vol. 5.291-327 [reprinted in Lenz 1928].
_____. 1928. El papiamento, la lengua criolla de Curazao. Santiago de Chile: Balcells y Ca.
Lipski, John. 1985. Creole Spanish and vestigial Spanish: evolutionary parallels. Linguistics
23.963-984.
_____. 1986a. Golden Age `black Spanish': existence and coexistence. Afro-Hispanic Review
5(1-2).7-12.
_____. 1986b. Convergence and divergence in bozal Spanish. Journal of Pidgin and Creole
Languages 1.171- 203.
_____. 1987a. The construction ta + infinitive in Caribbean bozal Spanish. Romance Philology
40.431-450.
_____. 1987b. On the origin and use of lan/nan in Caribbean bozal Spanish. Beitraege zur
 xl

romanischen Philologie 24.281-290.

_____. 1990. The speech of the negros congos of Panama. Amsterdam: John Benjamins.
_____. 1991. On the emergence of (a)m as subject in Afro-Iberian pidgins and creoles. Linguistic
studies in medieval Spanish, ed. by Ray Harris-Northall and Thomas Cravens, 39-61.
Madison: Hispanic Seminary of Medieval Studies.
_____. Forthcoming a. Spontaneous nasalization in the development of Afro-Hispanic language.
Journal of Pidgin and Creole Languages, in press.
_____. Forthcoming b. Origin and development of ta in Afro-Hispanic creoles. Atlantic meets Pacific:
proceedings of the first meeting of the Society for Pidgin and Creole Linguistics, ed. by Francis
Byrne and John Holm. Amsterdam: John Benjamins, in press.
Lpez, Jos Florencio [Jacan]. 1879. Nadie sabe para quin trabaja. Matanzas: Imprenta El
Ferro-Carril.
Lpez Morales, Humberto. 1980. Sobre la pretendida existencia y pervivencia del `criollo' cubano.
Anuario de Letras 18.85- 116.
Maduro, Antoine. 1960. Loque a sobra den e macutu di dicho, refran, proverbio, frase i palabranan di
nos lenga i nan nificacin na ulands. Curaao: Ciro J. Cardoze.
Malpica La Barca, Domingo. 1890. En el cafetal. Havana: Tipografa de "Los nios hurfanos."
Martnez Gordo, Isabel. 1982. Lengua "bozal" como lengua criolla: un problema lingstico. Santiago
46.47-53.
Mason, J. Alden and Aurelio Espinosa. 1918. Porto Rican folklore: dcimas, Christmas carols,
nursery rhymes and other songs. Journal of American Folklore 34.143- 208.
Megenney, William. 1984. Traces of Portuguese in three Caribbean creoles: evidence in support of
the monogenetic theory. Hispanic Linguistics 1.177- 89.
_____. 1985a. La influencia criollo- portuguesa en el espaol caribeo. Anuario de Lingstica
Hispnica (Valladolid) 1.157- 80.
_____. 1985b. Rasgos criollos en algunos villancicos negroides de Puebla, Mxico. Anuario de
Letras 23.161-202.
_____. 1990. Africa en Santo Domingo: la herencia lingstica. Santo Domingo: Museo del Hombre
Dominicano.
Mellado y Montaa, Manuel. 1975. La casa de Taita Andrs, semi-parodia de la casa de Campo.
Juguete cmico del gnero bufo en un acto. Teatro bufo siglo XIX, tomo I, ed. by Rine Leal,
263-303. Havana: Editorial Arte y Literatura.
Merlin, Mara de las Mercedes Santa Cruz y Montalvo, Comtesse de de. 1974. Viaje a La Habana.
Havana: Editorial de Arte y Literatura.
Montenegro, Carlos. 1934. El negro Torcuato. Bohemia 26 (Aug. 26, 1934), pp. 46-7, 153-4.
Montes Huidobro, Matas. 1987. Teora y prctica del catedratismo en Los negros catedrticos de
Francisco Fernndez. Miami: Editorial Persona.
Morales Carrin, Arturo. 1978. Auge y decadencia de la trata negrera en Puerto Rico (1820- 1860).
San Juan: Instituto de Cultura Puertorriquea.
Mora Delgado, Martn. 1972. Sofa. Havana: Biblioteca Bsica de Autores Cubanos, Instituto
Cubano del Libro.
_____. 1975. La familia Unzazu. Havana: Editorial Arte y Literatura.
Naro, Anthony. 1978. A study on the origins of pidginization. Language 54.314- 47.
Ortiz, Fernando. 1916. Hampa afro-cubana: los negros esclavos. Havana: Revista Bimestre
Cubana.
_____. 1940. `Prejucio,' prologue to Lydia Cabrera, Cuentos negros de Cuba. Havana: Imp. La
 xli

Vernica.
_____. 1985. Los bailes y el teatro de los negros en el folklore de Cuba. Havana: Editorial Letras
Cubanas.
Otheguy, Ricardo. 1973. The Spanish Caribbean: a creole perspective. New ways of analyzing
variation in English, ed. by Charles-James Bailey, Roger Shuy, 323-339. Washington:
Georgetown University Press.
Pasarell, Emilio. 1951. Orgenes y desarrollo de la aficin teatral en Puerto Rico. Ro Piedras,
Editorial Universitaria, Universidad de Puerto Rico.
Perl, Matthias. 1982. Creole morphosyntax in the Cuban "habla bozal." Studii ? i Cercet? ri
Lingvistice 5.424-433.
_____. 1984. Las estructuras de comunicacin de los esclavos negros en Cuba en el siglo XIX. Islas
77.43- 59.
_____. 1985. El fenmeno de descriollizacin del `habla bozal' y el lenguaje coloquial de la variante
cubana del espaol. Anuario de Lingstica Hispnica (Valladolid) 1.191-201.
_____. 1987. "Habla bozal"--eine spanisch-basierte Kreolsprache? Beitrge zur Afrolusitanistik und
Kreolistik, ed. by Matthias Perl, 1-17. Berlin: Akademie der Wisschschaften der DDR,
Zentralinstitut fr Sprachwissenschaft, Linguistische Studien 172.
_____. 1989a. Algunos resultados de la comparacin de fenmenos morfosintcticos del "habla
bozal," de la "linguagem dos musseques," del "palenquero" y de lenguas criollas de base
portuguesa. Estudios sobre el espaol de Amrica y lingstica Afroamericana, 369-380.
Bogot: Instituto Caro y Cuervo.
_____. 1989b. El "habla bozal" ?una lengua criolla de base espaola? Anuario de Lingstica
Hispnica (Valladolid) 5.205-220.
Pichardo, Esteban. 1836. Diccionario provisional casi- razonado de vozes cubanas. La Habana: Imp.
La Antilla.
_____. 1866. El fatalista. Havana: Imprenta Militar de M. Soler.
Postma, Johannes. 1990. The Dutch in the Atlantic slave trade 1600-1815. Cambridge: Cambridge
University Press.
Ramos, Jos Antonio. 1963. Caniqu, Trinidad, 1830. Habana: Consejo Nacional de Cultura.
Reinecke, John. 1937. Marginal languages: a sociological survey of the creole languages and trade
jargons. Ph. D. dissertation, Yale University.
Reyes, Rogelio. 1976. Language mixing in Chicano bilingual speech. Studies in southwest Spanish,
ed. by J. Donald Bowen and Jacob Ornstein, 183-188. Rowley, Massachusetts: Newbury
House.
Rodrguez, Luis Felipe. 1969. Cinaga. Miami: Mnemosyne Publishing Inc.
_____. 1971. Marcos Antilla, relatos de caaveral. Havana: Instituto Cubano del Libro.
Ruz Garca, Armanda. 1957. Ms all de la nada. Santa Clara: Offset Cancio.
Snchez Maldonado, Benjamn. 1961a. La herencia de Canuto, zarzuela bufa en un acto. Teatro
bufo, siete obras, tomo I, 83-120. Universidad Central de las Villas.
_____. 1961b. Los hijos de Thala o bufos de fin del siglo. Teatro bufo, siete obras, tomo I,
pp. 217-255. Universidad Central de las Villas.
Sandoval, Alonso de. 1956. De instauranda aethiopum salute; el mundo de la esclavitud negra en
Amrica. Bogot: Impresa Nacional de Publicaciones, facsimile edition.
Santa Cruz, Mara de. 1908. Historias campesinas. Havana: Imprenta y Librera de M. Ricoy.
Schwegler, Armin. 1989. Notas etimolgicas palenqueras: "casariambe," "tnganan," "ag,"
"monicongo," "maricongo," y otras voces africanas y pseudo-africanas. Thesaurus 44.1-28.
 xlii

Sosa, Francisco. 1974. Sistema fonolgico del espaol hablado en Cuba: su posicin dentro del
marco de las lenguas "criollas." Ph. D. dissertation, Yale University.
Surez y Romero, Anselmo. 1947. Francisco. 2nd ed. La Habana: Ministerio de Educacin.
Thompson, R. W. 1961. A note on some possible affinities between the creole dialects of the Old
World and those of the New. Creole language studies number II, ed. by Robert B. Le Page,
107-113. London: Macmillan.
Valds Bernal, Sergio. 1978. Las lenguas africanas y el espaol coloquial de Cuba. Santiago
31.81-107.
Valkhoff, Marius. 1966. Studies in Portuguese and creole. Johannesburg: Witwatersrand University
Press.
Van Wijk, H. 1958. Orgenes y evolucin del papiamentu. Neophilologus 42.169- 82.
Villaverde, Cirilo. 1979. Cecilia Valds. Critical edition by Raimundo Lazo. Mxico: Porra.
_____. 1981. Excursin a vuelta abajo. Havana: Editorial Letras Cubanas.
Wagner, Max Leopold. 1949. Lingua e dialetti dell'America spagnola. Florence: Edizioni "Le Lingue
Estere."
Whinnom, Keith. 1965. The origin of the European-based creoles and pidgins. Orbis 14.509-527.
Ziegler, Douglas- Val. 1981. A preliminary study of Afro- Cuban creole. Unpublished MS, San
Diego State University.
 xliii

Appendix 1: List of Cuban bozal texts

Acosta-Rubio, Ral (1976). Quiquirib Mandinga (se lo llev el diablo) {C-1}

Anon.
`Cantar para matar culebras' (Guirao 1938) {C-2}
`Canto congo de cabildo' (Guirao 1938) {C-3}
`Canto de comparsa ta Julia' (Guirao 1938) {C-4}
`Canto funeral' (Guirao 1938) {C-5}
`Cantos de cabildo' (Guirao 1938) {C-6}
`Dilogo' (Guirao 1938) {C-7}
`Papelito "habla" lengua' (De la Iglesia n.d.: 76) {C-8}
`Villancico' (Ballagas 1946: 92) {C-9}
`Yo bota lan garafo' (Guirao 1938) {C-10}

Bacard Moreau, Emilio

Doa Guiomar {C-11}
Filigrana {C-12}
Va Crucis {C-13}

Barnet, Miguel (1966). Biografa de un cimarrn {C-14}

Bentez del Cristo, Ignacio (1930). `Los novios catedrticos' {C-15}

Berenguer y Sed, Antonio (1929). Tradiciones villaclareas {C-16}

Cabrera, Lydia
Anaforuana (Cabrera 1975) {C-17}
Ayapa, cuentos de jicotea (Cabrera 1971) {C-18}
Cuentos negros de Cuba [1936] (Cabrera 1989) {C-19}
El monte (Cabrera 1985) {C-20}
Francisco y Francisca (chascarrillos de negros viejos) (Cabrera 1976) {C-21}
La sociedad secreta Abaku, narrada por viejos adeptos (Cabrera 1970b) {C-22}
Los animales en el folklore y la magia de Cuba (Cabrera 1989) {C-23}
Por qu, cuentos negros de Cuba (Cabrera 1972) {C-24}
Refranes de negros viejos (Cabrera 1970a) {C-25}
Reglas de congo: palo monte mayomb (Cabrera 1979) {C-26}
Yemay y Ochn (Cabrera 1980) {C-27}

Cabrera Paz, Manuel

`Exclamaciones de un negro' (Guirao 1938) {C-28}
Fragments of poem (Cabrera Paz 1973) {C-29}

Calcagno, Francisco (1977). Romualdo: uno de tantos {C-30}

Conseguegra Guzmn, Ismael (Feijo 1979)

 xliv

`Yo est cort un caas' {C-31}

``Yo quer met pinchacito tened dentro carne sabroso' {C-32}

Creto Gang (Bartolom Jos Crespo y Borbn).

`Un ajiaco o la boda de Pancha Juta y Canuto Raspadura' [1847] (Leal 1975) {C-33}
Assorted fragments (Cruz 1974) {C-34}

Estrada y Zenea, Ildefonso (1980). El quitrn [1880] {C-35}

Feijo, Samuel (ed.)

Cuarteta y dcima (Feijo 1980) {C-36}
Cuentos populares cubanos de humor (Feijo 1981) {C-37}
Sabidura guajira (Feijo 1965) {C-38}

Fernndez, Francisco
`El bautizo'(Montes Huidobro 1987) {C-39}
`El negro cheche' (Montes Huidobro 1987) {C-40}
`Los negros catedrticos'(Montes Huidobro 1987) {C-41}
`Minu' (Leal 1982) {C-42}
`Polticos de Guinea' (Leal 1982) {C-43}

Gelabert, Francisco de Paula (1875). Cuadros de Costumbres cubanas {C-44}

Guirao, Ramn (n. d.). Cuentos y leyendas negras de Cuba. {C-45}

Hernndez Cat, Alfonso (1946). `Los chinos' {C-46}

Lachataer, Rmulo (1938). ?Oh, mo Yemay! {C-47}

Lpez, Jos Florencio [Jacan] (1879). Nadie sabe para quin trabaja {C-48}

Malpica La Barca, Domingo (1890). En el cafetal {C-49}

Mellado y Montaa, Manuel (Leal 1982). `La casa de Taita Andrs, semi-parodia de la casa de
Campo. Juguete cmico del gnero bufo en un acto' [1880] {C-50}

Merlin, Mara de las Mercedes Santa Cruz y Montalvo, Comtesse de de (1974). Viaje a La Habana
{C-51}

Montenegro, Carlos (1934). `El negro Torcuato' {C-52}

Mora Delgado, Martn

La familia Unzazu (Mora Delgado 1975) {C-53}
Sofa (Mora Delgado 1972) [1891] {C-54}

Nicols Duque de Estrada. Doctrina para negros, explicacin de la doctrina cristiana acomodada
 xlv

a la capacidad de los negros bozales [1797] (Lavia 1989) {C-55}

Ortiz, Fernando (1985). Los bailes y el teatro de los negros en el folklore de Cuba {C-56}

Pichardo, Esteban
Diccionario provincial casi razonado de vozes y frases cubanas [1836/1865]
(Pichardo 1985) {C-57}
El fatalista (Pichardo 1866) {C-58}

Ramos, Jos Antonio (1963). Caniqu, Trinidad, 1830 {C-59}

Rodrguez, Jos Silvio (Leal 1982). `La esquina de la viajaca.' {C-60}

Rodrguez, Luis Felipe

Cinaga (Rodrguez 1969) {C-61}
Marcos Antilla, relatos de caaveral (Rodrguez 1971) {C-62}

Ruz Garca, Armanda (1957). Ms all de la nada. {C-63}

Snchez Maldonado, Benjamn (1961)

`La herencia de Canuto, zarzuela bufa en un acto' [1896] {C-64}
`Los hijos de Thala o bufos de fin del siglo' [1896] {C-65}

Santa Cruz, Mara de (1908). Historias campesinas {C-66}

Surez y Romero, Anselmo (1969). Francisco, el ingenio o las delicias del campo {C-67}

Villaverde, Cirilo.
Cecilia Valds o la loma del ngel (Villaverde 1977) {C-68}
Excursin a vuelta abajo (Villaverde 1981) {C-69}

Villa, Ignacio (Guirao 1938: 183-6)

`Calota ta mor' {C-70}
`Drumi, Mobila' {C-71}

Zell, Rosa Hilda (1953). `La sombra del caudillo' {C-72}

 xlvi

Appendix 2: List of Puerto Rican bozal texts

Anon.
`Dcima de negros' (Cadilla de Martnez 1953: 111) {PR-1}
`Nanqui toy ma mkinley' [Puerto Rico, 1898] (Mason 1918: 361) {PR-1}
`Yo so un negrito angolo' (Cadilla de Martnez 1953: 308) {PR-3}

Brau, Salvador (1894: 138). Puerto Rico y su historia {PR-4}

Caballero, Ramn. `La juega de gallos o el negro bozal' [Ponce, 1852] (Alvarez Nazario 1974)
{PR-5}

Derkes, Eleuterio. `Tio Fele' [Ponce, 1883] (Alvarez Nazario 1974) {PR-6}

Mndez Quiones, Ramn. Alternate fragments from `?Pobre Sinda!' [1884] (Girn 1991: 399-411)
{PR-7}

Rodrguez de Nolla, Olga (1947: 63-4). `Ao nueve dand, dao nuevo tabin' {PR-8}