The grammaticalization of lexical items
in pidgin/creole genesis
ANTHONY J. NARO
Universidade Federal do Rio de Janeiro
There are many possible scenarios for the beginnings of pidgin/creole
genesis, but they all have in common a need to communicate in a situation in
which there is no common language among the people involved. The situation
can be momentaneous or sporadic, such as contact between Norwegian and
Russian fishermen in the North Sea in the last century, or it can be drawn out and
continuous, such as contact in Portugal between the Portuguese and African
slaves brought directly from the African coast by the first exploratory expeditions
beginning in the mid 15th century.
In most of these situations one of the groups involved is socially and
economically dominant. In the initial period of European overseas expansion, as
well as during the occupation of the Americas and the colonial era, for example,
the Europeans were the masters and overlords, imposing their will on everyone
else. When we consider the linguistic state of affairs that took hold overseas, we
see that the Europeans did manage to impose their own vocabulary, or set of
lexical items. In fact, the lexicons of most modern day pidgin/creole languages,
from the Caribbean to the Indian Ocean, are composed of more than 90% of
items directly traceable to European languages. For this reason, the language
spoken in Haiti, for example, is called Creole French and is sometimes classified
as a French dialect, simplified French, bastardized French, or with other similar
It is a well known fact that the syntax and semantics of the pidgin and
creole languages do not bear the same degree of similarity to the European
languages that is so apparent in the case of the lexicon. But the lexicon too, of
necessity, underwent drastic changes. For one thing, in the first stages of contact,
when neither side knew anything of the other's language, the lexicon was vastly
reduced in terms of the number of items used. Even as contact progressed, the
range of social contexts in which the contact language was used were usually
quite limited. For example, although China Pidgin English existed for many
decades, its use was limited mainly to trade, and its total vocabulary probably
never exceeded seven to eight hundred items.
In most areas, the language that resulted from contact was appropriated by
those from the socially inferior side. This was necessary because, for diverse
reasons, these people had no language in common among themselves. In the
Americas, for example, slaves were brought from vast areas of the African
38 The grammaticalization of lexical items in pidgin/creole genesis
continent, and there is evidence that a deliberate effort was made to mix people of
different languages precisely in order to hinder their ability to communicate
among themselves and thus lessen chances of rebellion. In other cases, such as in
China and New Guinea, the economic activity created by Europeans attracted
people from such far-reaching areas that the new arrivals had no common
language that they could use to communicate among themselves. Under these
conditions, the contact systems based on European lexicons were taken over by
the other side and came to be used in all of the contexts typical of daily life. Their
environment of use was no longer limited to whatever happened to interest the
dominant side in the contact of superior with inferior.
When the social situations in which the contact language was used
underwent expansion, the lexicon had to expand as well. However, the dominant
Europeans were too far socially removed to be the source of more lexical items,
and, as we have seen, those from the inferior side had no common language from
which to bring in new lexical items. The result of this is that the lexical items that
existed had to be stretched to meet the communicative needs of the population.
This occurred in three ways: (1) the creation of standardized phrases to take the
place of non-existant lexical items, (2) the attribution of new meanings to the
relatively small number of lexical items in use, (3) the attribution of new
syntactic and morphological roles to these items.
The first techniques, the introduction of new fixed expressions, resulted in
phrases that seem picturesque from the European point of view: grass belong
head for 'hair', or ass belong world or big fellow for 'God' in New Guinea, for
example. The second method, enlarging of semantic fields, resulted in certain
lexical items gaining far wider semantic ranges than in the European languages.
A typical case is the preposition na, sometimes thought to be derived from
standard Portuguese na 'in the (fem.)', used in many pidgin/creole systems of
diverse lexical bases with a extensive range of meanings including 'to', 'from', 'in',
'into', 'out of', etc. Another example is found in New Guinea pidgin two-fellow
moon he-die for 'two days passed', in which the lexical item die (as well as moon)
has acquired an additional sense not present in the standard.
It is to the third method of lexical expansion, change of morphosyntactic
role, that we will turn our attention in this paper. Under this category one usually
has in mind examples such as change of the preposition behind to an adverb
meaning 'later' in New Guinea pidgin. An example is:
Behind by-'n-by he come-back again
for 'later it will come back again'. Another case of the same type is the use of a
noun as a verb in
Boatswain gammon me
Anthony J. Naro 39
for 'the boatswain lied to me' (literally 'boatswain game me').
Here we will examine a change of category of a more radical type, where
a lexical item looses its semantic content and becomes a functional item within
the grammar. This type of change is called 'grammaticalization'. A typical case of
this sort can be found in the history of the future with gonna (going to) in
standard colloquial English. Originally a construction such as
I am going to read my paper
had a purposive meaning involving actual physical dislocation in space along the
lines of 'I am traveling (from Rio de Janeiro to Granada) in order to read my
paper'. From this meaning the future developed: if I intend to do something, that
thing hasn't happened yet, i.e., it is in the future. The modern English going to
construction now has both meanings, and they can sometimes be difficult to
distinguish. The phonologically reduced auxiliary verb form gonna, however, can
have only the grammaticalized future meaning, in which the lexical content of
physical motion associated with the verb to go is no longer present. In fact, it is
now possible to use both going to and gonna with clauses in which the semantic
element of purpose or intent is virtually impossible and the future is the only
reading possible. An example is:
I am going to (gonna) hear the chime when the alarm goes off.
The common Romance future (Portuguese: cantarei; Spanish cantaré,
French je chanterai, etc.) also developed through a process of grammaticalization
from the construction cantare habeo, with loss of the lexical content of
possession in habere and the later extension to intransitive verbs. In this case the
phonological reduction to the former lexical item, the verb habere, has proceeded
further to a clitic stage, not yet attained in the case of colloquial English gonna.
One of the most intriguing aspects of pidgin/creole grammatical structure
is the verb and its associated categories of number and tense/aspect. In general,
lexical items such as nouns, verbs, adjectives, and so on tend to have only one
invariant phonological form in pidgin/creole languages. In the case of verbs the
most important consequence of invariance of form is the fact that the tense/aspect
and number elements of meaning must be expressed by means of independent
lexical items rather than through bound inflections. The verbal root itself does not
express any particular tense or aspect. When used in a clause, it is up to the
listener to interpret this aspect of meaning in accord with the context. In a
Portuguese-based pidgin, for example, sol nascer might mean 'the sun comes up',
'the sun came up', 'the sun will come up', 'the sun is coming up', 'the sum was
coming up', 'the sun will be coming up', and so on, according to the discourse
context in which it occurs. In case the context is not clear in this regard, the
40 The grammaticalization of lexical items in pidgin/creole genesis
speaker might add an adverb such as agora 'now', ontem 'yesterday', amanhã
'tomorrow', an adverbial phrase such as daqui a pouco 'in a little while', or an
adverbial particle such as já 'already' or logo 'right away'. This would clarify the
In the 1970's, New Guinea Tok Pisin was undergoing a period of
expansion and stabilization as a language used in the full range of social
situations of daily life. In its earlier stages, the future in Tok Pisin was indicated,
optionally, by the use of any lexical element with a future meaning such as the
adverbs close-to 'soon' or behind 'later', or the more general and vague adverbial
expression by-'n-by 'in a while', usually in sentence initial position. Sankoff &
Laberge (1982) studied the expansion process and found that by, a shortened
form of by-'n-by, was beginning to take on properties typical of a grammatical
future marker. The characteristics they list as differentiating the new
grammaticalized future marker by from the previous lexical adverbial expression
by-'n-by are: (1) reduction in phonological material, from the full bisyllabic form
to only one syllable, frequently with a reduced shwa-like vowel; (2) reduction in
stress, from the earlier primary word stress level to that typical of unstressed
syllables in full lexical categories; (3) redundant use of by even when the future
meaning is conveyed by lexical items such as adverbs; (4) tendency toward
placement of by immediately before the verb, rather than in sentence initial
position, especially with full noun phrase subjects. In this way the full lexical
expression by-'n-by is converted to the reduced grammatical item by. In other
words, grammar arises from the lexicon.
In Papiamento, as well in many Portuguese-based creoles (principally in
Asia), the future or, in more technical terms, the irrealis mood, is indicated by
means of the grammatical particle lo, derived from the lexical adverb logo 'right
away'. Like by in New Guinea, lo is unstressed and is used redundantly even
when the semantic content of futureness is already established by lexical items
present in the clause or extra linguistic context. Although we do not have a fully
documented history of Portuguese pidgin, it is possible to establish earlier usage
by examining the imitation of the pidgin found in certain 16th century plays. One
particularly interesting example, found in the anonymous play Auto dos enanos,
is a scene in which two Portuguese peasants are trying to communicate with a
Castilian noblewoman. No Africans or other non-Europeans are present in this
particular scene, although there were many Africans in Portugal at the time. The
Spanish lady is consistent in always using standard Castilian, but the Portuguese
peasants are not very good at understanding her. At one point, she complains,
saying to the peasants "You don't understand", to which one of them replies:
Se vós falar como nós
If you speak like us,
logo nós entender vós
We will understand you
logo ess' ora.
And reply immediately.
Anthony J. Naro 41
Notice that all of the verbs in the Portuguese peasant's speech are in the
infinitive and the pronouns are all fully stressed non-clitic forms, in typical
pidgin fashion. There are two occurrences of logo in the peasant's speech. The
first is used as a general future marker, in a way quite parallel to by-'n-by in New
Guinea pidgin. The second, with the original meaning of logo 'right away', is
reinforced with the expression ess' ora 'right now' because the meaning of logo
had already been weakened to the point where it no longer contained the lexical
semantic element of immediateness. In the course of time, logo was
phonologically reduced to just lo and became completely detached from the
lexicon and integrated into the morpho-syntax, in a way parallel to the
development of by from by-'n-by.
We see then, that the beginning of the pidginization process consists of a
reduction and simplification on the linguistic levels of the lexicon and the
grammar, parallel to the narrowing of the social contexts in which the pidgin is
used to the interaction of lower and upper social strata. When the social situations
of language use expand upon the appropriation of the language by a community
of speakers for use in the full range of situations typical of daily life, the lexicon
and grammar must expand as well, in order to adequately fulfill the new
functionality. One of the ways in which this occurs is the transformation of full
lexical items to morpho-syntactic items, a process known as grammaticalization.
In effect, the lexicon suffers a minor loss in order to expand the grammar. The
lexicon's loss is only temporary, however, as it too undergoes expansion through
the processes outlined above. In the long run, a new full language, capable of
fulfilling all of its social roles, is created from the limited resources of the pidgin.
HOPPER, Paul J. & CLOSS TRAUGOTT, Elizabeth (1993) Grammaticalization.
Cambridge: Cambridge University Press.
HOLM, John (1988) Pidgins and creoles. Volume I: Theory and Structures. Cambridge:
Cambridge University Press.
NARO, Anthony J. (1978a) "Pidginization, creolization, and natural change". Revista
Brasileira de Lingüística 5:123-138.
NARO, Anthony J. (1978b) "A study on the origins of pidginization". Language 54: 314-
NARO, Anthony J. (1998) "O uso da concordância verbal no português SUBSTANDARD do
Brasil: atualidade e origens", in Substandard e Mudança no Português do Brasil, ed.
by Sybille Große & Klaus Zimmermann. Frankfurt am Main: TFM, 1998.
NARO, Anthony J. & VOTRE, Sebastião J. (1998a) "Mecanismos funcionais do uso
lingüístico". DELTA (Documentação de Estudos em Lingüística Teórica e Aplicada).
42 The grammaticalization of lexical items in pidgin/creole genesis
NARO, Anthony J. & VOTRE, Sebastião J. (1998b) "Discourse motivations for linguistic
regularities: verb/subject order in spoken Brazilian Portuguese". Probus: International
Journal of Latin and Romance Linguistics (to appear).
NARO, Anthony J. & VOTRE, Sebastião J. (1992) "Mecanismos funcionais do uso da
língua: função e forma". DELTA (Documentação de Estudos em Lingüística Teórica e
NARO, Anthony J. & PEREIRA SCHERRE, Maria Marta (1993) "Sobre as origens do
português popular do Brasil". DELTA (Documentação de Estudos em Lingüística
Teórica e Aplicada). 9:437-45.
NARO, Anthony J. & PEREIRA SCHERRE, Maria Marta (1997) "Variable Concord in
Portuguese: the situation in Brazil and Portugal". In Current issues in pidgin and
creole linguistics , ed. by John McWhorter. Amsterdam: Benjamins (to appear).
SANKOFF, G. & S. LABERGE (1980) "On the acquisition of native speakers by a
language". The Social Life of Language 195-210, ed. by. Gillian Sankoff.
Philadelphia: University of Pennsylvania Press.