Foley, JA et English

al (eds) in new cultural contexts : reflections from Singapore 1988

Bao Zhiming

G"Jntroduction
communities which have a common language, verbal communication is
y a problem. We may come from different regIons with distinct dialectal
tures, or we may belong to different social classes with equally distinct
q,:?c:cellts, but by and large communication is often taken for granted. We
�)-arnU1 up speaking our mother tongue in our daily interaction, and so readily
�::ac:ceIJt it that we pay scant attention to it.
Human contact, and with it language contact, has happened throughout
Language contact is an important external cause of language change.
change is often evolutionary, and its effect can be seen more clearly
the passage of time. Modern English is a product of language contact.
��Jf we compare modern English with, say, Shakespearean English, we can see
�,�learly the phonological, morphological and grammatical differences that
�#me has wrought upon the language. Although the history of English is
�iPunctuated by periods of rapid change due to language contact brought about
�py foreign invasion, the continuity of the language among the general
��population has always been maintained (jespersen, 1923).
;'�L The situation changes radically when groups of people with no common
;�1anguage come together. In Chapter 1, we discussed the linguistic
·;.�consequences of colonization and massive human dislocation, namely, the
;:� mergence of pidgins and creoles and other new varieties of English. In
. ,places such as South and Southeast Asia and West Africa, where England
,imposed colonial rule without establishing a sizeable English settlement, a
small, English-speaking elite emerged, but the general population continued
to speak the local languages. In such multilingual societies, English enjoyed,
and still enjoys, a prestigious status as a language of administration, education,
commerce, law, science and technology. Quite often, as in Singapore, the
local population is composed of different'ethnic groups who speak different
languages, and English functions as a neutral lingua franca for interethnic
communication. Due to influences from local languages, the English language
in such communities has evolved into new varieties which are quite different
from British English, in pronunciation, vocabulary and grammar. Singapore
42 English in New Cultural Contexts

English, Indian English, Nigerian English and so on, are varieties of English
which are transplanted in totally new cultural, social and linguistic
environments. It is not surprising that they develop new pronunciations,
new words, new structures and new meanings that uniquely serve their own
communities.
Pidgins and creoles emerge in a somewhat different sociopolitical
environment. Pidgins arise to fill a communication need in situations where
no common language exists. One such situation is trade. Indeed, the popular
belief is that the word 'pidgin' is derived from the way Cantonese-speaking
Chinese traders pronounce the word 'business'. According to the Oxford
English Dictionary, Pidgin English, which arises to fill the communication
needs of the traders, consists of English words which are 'corrupted in
pronunciation, and arranged according to Chinese idiom.'
Trade is not the only situation which necessitates the emergence of
pidgins. Colonial rule and the slave trade created ample opportunities for
the development of such new forms of language . According to commonsense
understanding, a pidgin is a new language which takes its words from one
language, called the superstrate or lexifier language, and the bulk of its
grammatical structure from another, which is called the substrate language.
This characterization is simplistic, and will be elaborated upon in due course.
The way pidgins are formed is called pidginizatian. Collectively, the superstrate
languages are known as the superstratum, and the substrate languages, the
substratum.
In terms of social and political status, lexifier or superstrate languages are
dominant over substrate languages. Not surprisingly, English, through active
English participation in colonial expansion and the slave trade, lexifies most
pidgins that we know today. The substratum, however, is quite diverse.
Caribbean pidgins have an African substratum, reflecting the geographical
origin of the slave population. The English-based pidgins spoken in Hawaii
have a strong Asian substratum, due to the fact that many labourers were
•

brought from China, Japan and the Philippines to the plantations on the
Hawaiian islands. But these people did not speak the same language, so
substrate influence on the pidgin is varied and difficult to isolate with
certainty.
By definition, pidgins do not have native speakers. When a pidgin acquires
native speakers, as is the case when children of pidgin speakers grow up
speaking it, it becomes a creole. The process whereby a pidgin becomes a
creole, with accompanying increase'in grammatical complexity and lexical
sophistication, is known as crealization. Pidgins and creoles are related in
the way they are acquired, and creoles must have a pidgin as a predecessor.
For this reason the two terms are often used together. Like their pidgin
predecessors, creoles show superstrate and substrate influence, although this
 TheoTi� l of Language Genesis 43

is often masked as their grammar develops and grows in complexity. Pidgin

and creole grammars are never the sum of a superstrate vocabulary and a
substrate syntax.
The exact nature of substrate and superstrate influence on the emergence
of such new languages in general, and pidgins and creoles in particular, is a
complicated matter that defies simple characterization. Students of new
varieties of English, and of pidgins and creoles, never fail to be amazed by
the linguistic diversity and innovation exhibited in such languages. Without
doubt these Englishes are full-fledged systems of communication, with their
own vocabularies and grammars. Despite the efforts of several generations
of researchers, the question of how they arise and acqu ire linguistic structures
remains as controversial as ever. The controversy comes under the rubric of
language genesis , which focuses mainly on pidgins and creoles. The
fundamental question is this: How do the structures of pidgins and creoles
arise through language contact? Since we are only concerned with new
varieties of English, we will restrict our attention to pidgins and creoles
which have English as their lexifier language. The question before us is
therefore narrower. We are concerned with the exact nature of the influence
from the superstrate English and various substrate languages on the structure
of English-based pidgins and creoles and, by extension, on the structure of
other types of New Englishes whose genesis depends heavily on language
contact, such as Colloquial Singapore English, or CSE.
Many theories have been proposed to expla in the processes of
; pidginization and creolization. Two dominant theories are still being debated
.

l to this day. They are the substratist theory, whose advocates claim that the
L structure of pidgins and creoles derives from the substratum, and the
t universalist theory, whose proponents argue that pidgins and creoles reflect
�. �niversal properties of human language. Neither theory is watertight, and
�perhaps some form of synthesis is needed to account for the diverse range
iof features of New Englishes. Before we examine the two approaches, let us
{fa miliarize ourselves with a few important notions associated with the study
·:of pidgins and creoles.

te., 2ife Cycles and the Creole Continuum

i It IS customary to classify pidgins and creoles into four types: jargon, stable
� pidgin, expanded pidgin and creole (Romaine, 1 988) . Typically, jargons
�.mnsist of utterances which are one- or two-words long; in addition, they
;':!ll'e limited in function and exhibit a great deal of variabi!ity among individual
"speakers, Stable pidgins gain in grammatical complexity and stability among
(;speakers, and their vocabulary increases to about two hundred words.
::Expanded pidgins have sophisticated grammatical construction, a large and
'. 44 English In New Cultural Contexts

stable vocabulary and varied social and political functions. A pidgin can
become a creole at any stage, as long as it has acquired native speakers.
Whatever its origin, in structure and function, a creole is much more complex
than its pidgin predecessor. T he development from a jargon to a creole is
often called a life cycle, a notion due to the well-known creolist Robert Hall
(Hall, 1962). Not all creoles have undergone the same life cycle. At least
three types of life cycle have been established among creole languages
(Miihlhausler, 1986; Miihlhausler uses 'stabilized' instead of 'stable');

Type 1 Type 2 Type 3

Jargon Jargon Jargon

I I
stable pidgin stable pidgin

I
expanded pidgin

I
creole creole creole

We can consider jargon, stable pidgin, expanded pidgin and creole as different
phases of a creole's development. As the three life cycles indicate, a jargon
may be creolized aoruptly, skipping the middle two phases; or it may develop
gradually, passing through all the phases before it is creolized. Scholars have
identified Hawaiian Creole English as a Type 1 creole, Torres Straits Creole
English as a Type 2 creole and Tok Pisin as a Type 3 creole.
It is often the case that a creole, its lexifier language and its substrate
language(s} are spoken at the same time, so that the society is effectively
multilingual. In English-based creole societies, Standard English continues
to play an important role in commerce, education and other domains of life,
and its influence on creoles cannot be overestimated. Linguists describe this
mixture of creoles and standard languages as a creole continuum, or post­
creole continuum. T he idea is that people speak creoles with different degrees
of sophistication in their referential meanings and grammatical soundness.
At one end, the speech pattern resembles that of the standard language, i.e.
the superstrate language; this pattern is labelled acrolect. At the other end,
the basilecI, the speech pattern is much more limited in vocabulary and
grammatical complexity. In between we have mesolects, which are varieties
of creoles with varying degrees of linguistic sophistication. T here are other
ways of characterizing the creole continuum. Romaine (1988), for example,
identifies the creole as the basilect, the standard, superstrate language as the
acrolect, and other varieties as mesolects. It must be pointed out that there
are no sharp divisions along the continuum which define the boundaries of
 Theories of Language Gene515 45

distinct varieties. To some extent, the continuum reflects the command of

the creole language of various segments of the population. For obvious reasons
the educated elite usually speak the acrolectal varieties, while the less
. ·educated tend to speak mesolectal or basilectal varieties. Creole continua
. make it difficult to describe the grammars of creole languages.
Associated with the notion of the creole continuum is decreolization. This
is the situation when the grammar of the creole becomes indistinguishable
from that of its superstrate language. In other words, the basilectal and the
mesolectal speakers start adopting the acrolectal speech patterns and
'. grammatical structure. This is found, in Bickerton's (1980 ) words, 'wherever
a creole language is in direct contact with its associated superstrate language.'
'. Although it is difficult to determine cases of decreolization, Romaine ( 1 988)
: mentions a few creoles which have been decreolized. In North America,
, Black English· started its life as a creole, and is now in the last stages of
. decreolization. What this means is that the grammar of Black English now
; resembles the grammar of Standard American English in more respects than
:it did previously. Other creoles that are in various stages of decreolization
;. include those spoken in Jamaica, Guyana and Papua New Guinea - places
f where English, the prestigious superstrate language, continues to play an
�. important role: a precondition for decreolization to occur.
r The idea of a creole continuum is not without controversy. Some scholars
� .reject the notion, and argue in favour of diglossia ('two languages' ), with
�complementary communicative functions for the two languages. The term
1: 'diglossia' can be used broadly to refer to two unrelated languages spoken in
f the same community, or narrowly to refer to two functionally distinct varieties
� of English (Ferguson, 1959). Typically, the 'high' variety, the socially more
,
r prestigious one, is used in formal situations such as education, media and
� government; the 'low' variety, which is the less prestigious of the two, is
� used in informal situations. In Singapore, English is one of the official
p anguages, and is the language of government, education, law and commerce,
2 due to the active participation of multinational companies in Singapore's
�:economy. Through universal education, Singaporeans' linguistic repertoire
{includes Standard English (the 'high' variety), CSE, which some consider
: to be creole-like (Platt, 1 975: a cre% id), and local languages - mainly
t Mandarin, several southern Chinese dialects, Malay and Tamil. As far as
�English is concerned, CSE is not a speech continuum, but the 'low' variety
" which is in diglossic co-existence with the 'high' variety, i.e. Standard English.
: The two varieties perform complementary communicative functions in
, Singapore.
The continuum-diglossia controversy is not going to go away easily. The
. core of the matter lies in the inherent difficulty of delimiting the boundary
of a variety of language, whether it is a pidgin, creole, or 'normal' language.
· �': C(;jhe Substratum
In the preceding sections we described the distinction between pidgins and
creoles, and some concepts associated with them. In this section and the
next, we will examine the theoretical controversy surrounding the origin of
pidgins and creoles. Among the theories which have been proposed by
creolists (linguists who study pidgins and creoles) , we will single out the
substratist approach and the universalist approach. These two approaches
are chosen becadse they have the widest acceptance, and the debate they
have stimulated continues to this day. We will examine the substratist
approach first.
As we mentioned earlier, pidgins are languages which arise in communities
where there is no common language. They have a limited vocabulary from
the superstrate, i.e. lexifier, language, and a simplified grammatical structure.
It is generally accepted that the vocabulary of a pidgin comes from the
superstrate language , which is the dominant language in the newly
constituted communities where pidgins arise. The origin of the grammatical
structure of a pidgin is less certain. The core idea of the substratist approach
is that the grammatical structure of a pidgin is, to a large extent, determined
by the grammatical structure of the substrate language or languages. The
substratist approach was first proposed to account for the origin of pidgins.
Later the idea was extended to the genesis of creoles as well. Here we examine
the substratist ideas in light of the structures of pidgins and creoles, without
making a conscious effort to keep them apart.
The evidence of a substratum in a pidgin or creole is not hard to come
by. Researchers have demonstrated parallels in the structure of West African
languages and the Atlantic English-based creoles, which originated in
communities with a large African slave population (Holm, 1 988; Romaine,
1 988; Levebvre and Lumsden, 1989; Arends et aI., 1 995; among many
others). Structural similarities exist even among creoles which are lexified
by different languages. On the strucl'ural affinities of English-based creoles
spoken in the Americas and West Africa, Hancock ( 1 97 1 ) says:

[Tlhe grammatical structure of any one of these English-derived

creoles is far closer to any one of the other creoles...whose words
are not English-derived - such as the French- or Portuguese-derived
creoles - than it is to English.

I n the same spirit, Taylor ( 1 971; 1 977) examines some common grammatical
features of French-based creoles, English-based creoles, Dutch-based creoles
and Iberian-based creoles, and attributes these grammatical features to the
creoles having the same or closely related substrate languages, principally
 Theories of Language Genesis

�.. the African language of Yoruba. We will not di scuss Taylor's grammatical
features, since some of them require specialized knowledge of linguistics.
We will, however, discuss a few fam i liar grrtmmatical features which
substantiate substratist claims.

j. Bislama Pronouns

I Jeff Siegel discusses the pronouns of Bislama, an English-based creole

language spoken in Vanuatu. The pronouns are shown in the following chart
(Siegel, 1996: 255).

Singular Dual Trial Plural

1st person inclusive yumitu(fala) yumitrifala yuml
1st person exclusive ml mitufala mitrifala mifala
2nd person yu yutufala yutrifala yufala
3rd person hem/em tufala trifala olgeta

There is a four-way distinction in person in Bislama: singular, when the

pronoun refers to one individual only; dual, to two individuals; trial, to three
individuals; and plural, to more than three individuals. The distinction
between 'inclusive' and 'exclusive' applies to the non-singular first person
pronoun. When the pronoun refers to the speaker and the listener, it is
inclusive; if it refers only to the speaker, and someone else, it is exclusive.
One noticeable feature of the Bislama pronoun system is the lack of gender
marking: the same pronoun, hem (or em), is used for feminine, masculine
and neuter subjects (she, he, it). It is obvious that the words come from
English, which is the lexifier language. We can easily segment the dual first
person pronoun into 'you-me-two', and the trial into 'you-me-three-fellow'.
But the four-way distinction of number, and the exclusi ve-inclusive
distinction of person, do not come from English. Where do they come from?
The answer to the question can be found in the substratum of Bislama.
The native tongues spoken in Vanuatu all belong to the Austronesian family
of languages. One member of this family is Tangoan, spoken on an island
not far from Vanuatu, which has the following pronoun system:

Singular Dual Trial Plural

1st person inclusive enrarua enratolu enra
- -

1st person exclusive enau kamamrua kamamrua kamamrua

2nd person egko kamimrua bmimtolu kamim
3rd person enta enrarua enratolu enra/enira

The actual pronouns are different: the first person exclusive dual is mitufala
in B islama, but kamamrua in Tangoan; similarly, the second person singular
English in New Cultural Contexts

is yu in Bislama, but egko in Tangoan, and so on. In fact the Bislama pronouns
are derived from English, for example, mitufala is derived from me-twa-fellow,
which explains why it is the first person exclusive dual pronoun; in contrast,
the first person inclusive dual pronoun, yumitufala, is derived from you-me­
two-fellow . The structure of the Bislama pronoun system, however, is exactly
the same as that of Tangoan, in the sense that both systems maintain the
singular-dual-trial-plural distinction and the first person inclusive and
exclusive distinction. Siegel (1996: 256) writes:

T hus it is clear that the forms of Bislama's pronoun system are

derived from its lexifier language, English, but the grammatical
distinctions are from its Oceanic substrate languages.

The Bislama pronoun system is a strong case of substrate influence.

Serial Verb Construction

Another area of substrate influence in the genesis of creoles is the serial

verb construction, in which two or more verbs are used in series in the same
sentence. T his construction is found in Sino-Tibetan languages (Chinese
and Tibetan are the two major languages of this family) and languages in
West Africa, but not in English, or any other European languages. It is,
interestingly, a general feature of creole languages. Arends et al. (1995)
produce the following data from various languages:

(1) a. A tei goni suti di pmgo [Saramaccan]

he take gun shoot the pig
He shot the pig with a gun.

b. Ode adare no twa� nkromata [Akan]

he-take machete the cut-PAST branch
He cut the branch with a machete.

(2) a. A waka go na wowoyo [Sranan]

he walk go PREP market
He went to the market.

b. Oguang koo ahabang mu [Akan]

he-flee-PAST go-PAST bush m

He fled into the bush.

To which we can add the following from CSE [(3a) is taken from Ho, 1992]:
 Theories of Language Genesis 49 ;-

(3) a. I follow her go to market [CSE]

I went to the market with her.

b. wo gen ta qu shichang [Chinese]

I follow her go market
I went to the market with her.

c. I walk go market [CSE]

I walked to the market.

d. wo zou qu shichang [Chinese]

I walk go market
I walked to the market.

Saramaccan and Sranan are two creole languages spoken in Surinam, for
which Akan is one of the African substrate languages, and Chinese is a
major language spoken in Singapore. The English lexical source is quite
dear in the data. The serial verb construction of Saramaccan, Sranan and
eSE, which does not exist in English, has parallels in the substrate languages.
In ( 1 ), the first verb, tei in Saramaccam and ode in Akan, expresses an
'instrument' meaning; in (2), the second verb, go in Sranan and koo in Akan,
t:xpresses the 'destination' meaning. In English, these meanings are typically
expressed with the help of prepositions, as indicated in the glosses. The
CSE sentences are almost a word-for-word translation of the Chinese
equivalent. The similarity in structure is striking. Although English is the
source of the words, the grammatical structure comes out of the substrate
languages, which happen to have the serial verb construction. As far as
serialization of verbs is concerned, CSE, Saramaccan, Sranan, and other
Atlantic creoles with African substrata, have more in common with each
other than with English, their common lexifier language.

Already in Singapore English

The system of tense and aspect is another area of substrate influence. Here,
we will consider the function of already in CSE. Many researchers have
worked on the use of already , and shown that its functions are related to the
Chinese Ie (Platt and Weber, 1980; Kwan-Terry, 1989; Ho, 1 992; Bao, 1 995).
A note on the term 'Chinese' is in order. For all practical purposes we can
consider 'Chinese' as a cover term for a group of mutually unintelligible
languages. These languages are identified by the geographic locations in
which they are spoken, and are often called Chinese dialects. Since the
majority of the Chinese population in Singapore trace their origin to southern
Fujian, the most common Chinese dialect is Hokkien (Fujian in pinyin).
1;.'. so English in New Cultural Contexts

Other major Chinese dialects which are represented in Singapore are

Teochew, Cantonese, Hakka and Hainanese. Although their phonologies
differ consider ably the dialects nevertheless share a core vocabulary and
,

the basic syntactic structure. For this reason, we will use the term 'Chinese',
and transcribe the data in pinyin in accordance with Mandarin pronunciation,
whenever explicit reference to a particular dialect is not necessary. In
Mandarin, Ie is etymologically the same as liau in Hokkien.
Already in CSE marks two aspects, the perfective and the inchoative.
Aspects provide a perspective on the actions being reported, and the
perfective and the inchoative have opposite aspectual meanings: the former
means the completion of an action or state, and the latter means the
beginning of an action or state. In English, the perfective aspect is typicalli,

expressed with the past tense - Mei Mei cried reports a completed event;
and the inchoative aspect is expressed with the help of words like start or
begin - Mei Mel started to cry reports an event that has just been initiatedi
In Hokkien, however, both aspectual meanings are expressed by liau: the
expression chat bun liau - 'eat rice liau' - could have the perfective meanin �
of having eaten the meal, or the inchoative meaning of starting to eat the
meal. Context of use will disambiguate the expression. j
Now consider the CSE data below (Kwan-Terry, 1989): i
(4) a. I cannot go inside already.

b. T he tongue red already. i

T he tongue has turned red. 1
,1
j
Both sentences in (4) have the inchoative reading: in (4a), the speake
I
�
does not have the permission to go inside now, with the implication that he
or she had the permission earlier - in other words, the speaker 'has started ]
OJ
to be in the new state. Similarly, in (4b), the tongue is now in the state
redness, which it was not previously. T he sentences below are perfectiv '
(Kwan-Terry, 1989): '
1
(5) a. I eat the cake already.
I have already eaten the cake.
1
I
j
J
,
b. Close all the doors already.
1
I have already closed all the doors.

From the sentences in (4) and (5), we can make the observation that alread
1�
is inchoative when it is used with state predicates (such as red) or modal
verbs (such as can ) , and perfective when it is used with event predicat �
(eat, close) . T his is evident in the following sentences (Kwan-Terry, 1989)1
 Theories of Language Genesis

[,(6) a, Alice fell down in the hole already. (perfec tive)

b. The red car go already. (perfective)

i,:
; , c. Now my school is close already, (inchoative)

d. I don't want it already. (inchoa ti ve)

:, :
I

I'Since already occurs in the same syntactic position for both the perfective
I:and inchoative interpretations, we would expect it to be ambiguous in some
I'sentences. This is the case for the sentence below:
f
�(7) I eat durian already.
" 1. I have already eaten (the) durian. (perfective)
I;
�� , ' 11. I now eat durians. (inchoati ve)
1,
��'\'

� rrhe sentence is ambiguous between the perfective reading (7i) and the
b, inchoative reading (7ii). Again, context of use provides cues to disambiguate
� the sentence.
How does already acquire the two uses? To be sure, the two aspects are
r
Fi

'semantically compatible with the lexical meaning of already in English. But

� ,English expresses the perfective aspect with have and the verb's past participial
�;form (l have eaten durians) and uses words like start, begin or now to express
r the inchoative aspect. Already does not perform any aspectual function
r;:directiy. Therefore the source of CSE already cannot be found in English.
It We will show below that the aspectual function of already comes from the
W ,
substrate language of Chmese.
{, The Mandarin Ie can occur in two pOSitions - after the verb, or at the
�end of the sentence. These are exemplified below:

a. Zhangsan chi-Ie pll1ggll0.

Zhangsan eat-LE apple
Zhangsan has eaten/ate the apple.

b. Zhangsan chi pingguo Ie.

Zhangsan eat apple LE.
i. Zhangsan starts to eat the apple. (event)
ii. Zhangsan now eats apples. (state)

,. When Ie is post-verbal, it expresses the perfective aspect, as shown in (Sa).

; When it is used at the end of a sentence, as in (Sb), it expresses the inchoative
,

.' aspect: as a state, chi pingguo Ie means that Zhangsan now eats apples, implying
f'"
,�' 52 English in New Cultural Contexts

that he did not eat apples before; as an event, the sentence means that
Zhangsan starts to eat the apple. Note that the gloss of pingguo differs as
well: in the event interpretation (8b-i), pingguo/apples refers to a specific
apple or specific apples; in the state interpretation (8b-ii), it does not.
For state predicates such as know, believe and red, le expresses the beginning
of the state, regardless of where it occurs in the sentence:

(9) a. Zhangsan zhidao zhe-jian shi leo

Zhangsan know this matter LE.
Zhangsan knows the matter.

b. Zhangsan zhidao le zhe-jian shi.

Zhangsan know LE this matter.
Zhangsan knows the matter.

c. shitou hong Ie. [ef. (4b)]

tongue red LE.
The tongue is red/has turned red.

This is exactly the situation we saw with already.

From the above discussion it is clear that the aspectual function of already
in CSE is derived from the Chinese leo We summarize the aspectual functions
of the Mandarin Ie, and CSE already in the chart below:

With Event Predicate With State Predicate

Ie perfective (verb-final) inchoative (verb-final)
inchoative (sentence-final) inchoative (sentence-final)
already perfective inchoative
inchoative
•

Again, we see that the word already is derived from English, the superstrate
language, but the aspectual functions are derived from Chinese, the major
substrate language of CSE.

Substrate Influence in Phonology

Substrate influence in the pronunciation of a pidgin or creole hardly needs

emphasizing. When we learn a foreign language, our pronunciation will most
likely be influenced by the speech habits we have acquired in our native
tongue. Foreign accents are not something new in our experience with
language. In the formation of a creole, the phonological features of the
substrate languages will show up here and there, and the features of the
 Theories of Language Genesis

superstrate, particularly those which are not found in the substrate, will be
dropped, or otherwise modified to facilitate pronunciation. Here we will
consider how the distinction between long and short vowels in English fares
in pidgins and creoles.

Tense vs. Lax Vowels

Hall (1966) describes the loss of contrast between 'tense' vowels and 'lax'
vowels in pidgins and creoles. Tense vowels are pronounced with a noticeable
degree of tension in the tongue muscles, whereas l"x vowels are pronounced
with relaxed tongue muscles. English maintains the distinction between /if
beat and /II bit; between lei bait and lei bet; between lui fluke and lui book;
and between 101 coat and hi caught. In English, the tense vowels are typically
lengthened or diphthongized; so Ii, e, u , 01 are commonly transcribed as Ii:,
eI, U:, �u/. Some linguists transcribe the diphthongs leI, �ul as lei, oul, bringing
out the common effect of the tongue's muscular tensing on the vowels. In
Neo-Melanesian, a creole spoken in Papua New Guinea, the tense-lax
distinction is lost, as shown in the following chart:

Words English Neo-Melanesian Sound change

leg kg leg >e
I

pussy pusi PUSl u>u

forgive fJrglv pOglp J >0; I> i

The tense-lax vowel pairs Ie , el, Ii, II, lu, ul and 10, JI become four vowels
transcribed as Ie, i, u , 01 respectively. Whatever the phonetic realizations of
the phonemes Ie, i, u, 01, the fact remains that the distinction between
tenseness and laxness is no longer contrastive in Nco-Melanesian. In other
words , due to substrate influence, the phonemic inventory no longer contains
the tense-lax pairs of vowels. The loss of the tense-lax distinction is quite
common across New Englishes.
Exactly the same sound changes happen in CSE. According to Tay (1982)
and Hung (1 996), CSE has five simple vowels, as shown below:

1 U

e J
a

plus the schwa I�/. Among the vowels, there is no length contrast, nor is
there tense-lax contrast. So, the follOWing words are homophonous in CSE:
beat/bit [bit], bet/bat [bcd, pool/pull [pul], caught/cot [btJ and lark/luck [IokJ.
In S tandard English, they are distinct.
l'"""-'" J"T cngulf! In f'Jew l,uttural Contexts

The vowel system of eSE shows clear influence from the speech habits
of its substratum, which includes the Austronesian language of Malay. Malay
has the same number of vowels as eSE (Othman, 1 990):

u
e 0

plus the schwa /d/. The vowel qualities differ slightly, notably for the mid
vowels. In the articulation of Malay Ie, 0/, the tongue position is higher
than that for English /£, J/. Nevertheless, the eSE vowel system is remarkably
similar to that of Malay.
The eSE and Neo-Melanesian vowel systems are representative of the
vowel systems of many New Englishes, including pidgins and creoles. Tok
Pis in, a pidgin English spoken in Papua New Guinea, and Sranan, a creole
English spoken in Surinam, have the same vowel contrasts, which are close
to those of eSE:

Tok Pisin Sranan

u u
e o e o
Q Q

(Todd, 1 984) ( Adamson and Smith, 1995)

In addition to the five oral vowels, Sranan has five nasalized vowels, whereas
Tok Pisin does not have nasalized vowels at all. It would seem mysterious
why totally unrelated languages have exactly the same vowel contrasts. There
are two plausible explanations for this. First, the five-vowel contrasts we see
in eSE, Neo-Melanesian, Tok Pisin and Sranan are a direct result of
simplification, i.e . the loss of the long-short and tense-lax distinctions in
the vowels of S tandard British or Am�rican English. Second, the substrate
influence is similar, even though the New Englishes do not have the same
substrate languages. For eSE, Neo-Melanesian and Tok Pisin, the main
substrate languages are members of the Austronesian family; Sranan, on the
other hand, has an African substratum. According to Holm ( 1 988), African
languages, such as Bantu, have the following vowel pattern:

Front Back
H igh u
Mid e 0

Mid £ J
Low a
 Theories of Language Genesis

Two mid vowel pairs, /e, c/ and /0, 'J/, are phonemic only in a few languages.
For many African languages, then, there are only five vowel phonemes: two
high vowels, two mid vowels and on e low vowel. Such a vowel pattern is
exactly that found in Sranan and other New Englishes with an Austronesian
substratum. It is possible that the simplified vowel system of New Englishes
results from simplification and substrate influence.

Pre-nasalized Vowels

There are two types of stop consonant which are found in some New
Englishes, but not in Standard English: the co-articulated stops and pre­
nasalized stops. Co-articulated stops are made at two places of articulation
simultaneously. Two common stops of this type are /kp, gb/, which are
produced at the velar region and the lips. These stops are found in many
African languages and New Englishes with an African substratum, such as
Krio and Nigerian Pidgin English. Typically, they occur in loan words from
the substrate language<;
Pre-nasalized stops, symbolized as /mb, nd, Qg/, are different from sequences
of nasals and stops with the same place of articulation, such as mb in bombard.
They are single phonemes, not sequences of distinct phonemes. Pre-nasalized
stops are produced by lowering the soft palate to let airstream enter the
nasal cavity before releasing the stop. T hey are found in Tok Pisin: tambu
'taboo' and em i ndai 'he fainted' (Todd, 1984). It is not clear from Todd's
description whether these stops are phonemic. Most likely, they are the
phonetic realizations of voiced stops (lb, d, g/). T hey do occur as phonemes
in some languages. In Saramaccan, a New English spoken in Surinam, pre­
nasalized nasals have been analysed as distinct phonemes (Holm, 1988) .
Incidentally, they are phonemes in African languages, which are the
substratum of Saramaccan.

Phonotactically Conditioned Sound Change

The combination of phonemes into syllables and words is subject to language­

specific conditions known as phonotactic conditions or constraints. T he
phonotactic conditions determine the possible phonological shapes of words
in a given language. Chinese, for example, does not allow consonant clusters
at all, while English allows up to three consonants in word-initial position,
e.g. spring. In Chinese, only nasals and voiceless plosives (in some dialects )
can appear at the end of a word, but in English, we have a wider range, from
nasals (pen), to plosives (stop), to fricatives (leaf). An English word can
even end in more than one consonant (six). Thus, Chinese and English
differ considerably in their phonotactic conditions.
In CSE, and many other New Englishes, we can observe two types of
III" 56 English in New Cultural Contexts

sound effect due to the phonotactics of the language. These constraints

concern consonant clusters and the phonetic realization of stops in word­
final position. We look at the cluster simplification first. To be sure, CSE
allows roughly the same range of consonant clusters in word-initial position:
C (pay), CC (pray) and CCC (spray). So there is no Simplification there.
However, in coda position, the final stop is deleted (for ease of comparison
we use the same vowel and consonant symbols):

Standard English CSE

list [lIst] [lIs]
past [pa:st] [pa:s]
send [send] [sen]
friend [frend] [fren]
mask [ma:sk] [ma:s]

The final consonant is not deleted when a vowel follows it:

listing [hsuJ)]
sending [sendIJ)]
masking [ma:skiJ)]

So the best analysis is to assume that the phonemic representation of words

like list, send and mask in CSE is the same as the Standard English input,
and the word-final stops are deleted to simplify the coda c lusters.
When a word-final stop is preceded by a vowel, it is unreleased, and the
vowel comes to an abrupt stop. Since voiced stops devoice word-finally, the
fol lowing RP minimal pairs are homophonous (the symbol � means·
unreleased; only relevant phonetic details are shown):

tap ta[p�] debt delt"] back

tab ta[p�] dead dele] bag
•

The devoicing of Ib, d, gl occurs as well.

The phenomena exemplified above have reflexes in the local languages. i

In both Malay and Chinese, consonants do not form clusters in word-final.

positions, and only voiceless stops appear in that position, and are realized ;
as unreleased. So here are two cases of substrate influence.
However, the substrate influence is rather weak, since other types of
cluster, such as [ks] in six, are attested in CSE, and onset clusters are the
same as those in Standard English. The weakness of substrate influence in:
CSE is due to the strong influence of Standard English. In the development
of the phonology of CSE, Standard English has won the tug of war over.
local languages.
 Theories of LangJ'd�€ lJ enes!s

Phonetic Realization

Substrate influence can be seen in phonetic realization of phonemes as well.

In this connection, we will discuss three cases: aspiration, palatalization and
devoicing. As is well-known, the voiceless stops in Standard English are
realized as aspirated, except before /s/; compare pan [phan], tan [than], can
[khan] with span [span], stan [stan], scan [skan] . In New Englishes, however,
the voiceless stops are typically realized as unaspirated regardless of what
precedes them. CSE is typical in this respect; there is no difference in the
phonetic realization of the voiceless stops in the two groups of words
mentioned above. One may argue that this is due to substrate influence,
since in Malay, an Austronesian language, voiceless stops are unaspirated.
But there is a wrinkle in this argument. The majority of the population in
Singapore is Chinese, speaking dialects of southern China which contain
voiceless, aspirated stops. Whatever the explanation, it is a strong tendency
for voiceless stops to be unaspirated in New Englishes .
Palatalization is a common phonological process which affects consonants,
.. particularly alveolars and velars. In casual speech in English, /s/ becomes [J]
(miss you) and /t/ becomes /tI/ (meet you). Since /I, tI/ are palato-alveolars,
the change to [J, tI] and other palatal sounds, is called palatalization. The
'
� English example is a typical case of palatalization: a sound becomes palato-

2 alveolar or palatal when it precedes the high vowel /i:/ or the semi-vowel
� /i/. There is articulatory reason for this change: to pronounce /i:, j/, the body
�. of the tongue has to be raised towards the palate, which will likely influence
� the pronunciation of the immediately preceding alveolars and velars.
" Palatalization is a very common sound change among the world's languages.
:. Among New Englishes, palatalization is responsible for the occurrence
� of a number of 'new' sounds, the most frequent of which is the palatal nasal
f' /Jl/. In CSE, this occurs before the high vowel /i/ and the semi-vowel /i/: oni
.lOJli] 'only', nonya [noJlja] 'nonya' . Almost certainly in CSE fJJ] is not a
:I:hstinct phoneme, but an allophone of In/. Its phonemic status in other
'New Englishes, such as Cameroon Pidgin English analysed in Todd ( 1984),
'is not clearly established. In Surinamese Creole Englishes, the velar stops
jk, g/ are palatalized before front vowels: gei or djei 'to resemble'. Here, the
,-palatal [di] is not phonemic, but an allophone of /gl (Holm, 1988).
�', Devoicing is a common process in New Englishes. Voiced consonants,
:�specially the stops /b, d, g/ and the affricate /d3/, become voiceless at the
end of a word. In Cameroon Pidgin English and Tok Pisin, Ib, d, g/ are
;aevoiced in rub, bad, big. In CSE, they are not only devoiced, but also
:imreleased. In the case of CSE, substrate influence can be established. In
Malay and the southern dialects of Chinese, the word-final stops are all
L
� VOiceless and unreleased.
�.
pIIin:)(j t!ngllS/I In New Luttural Lomexrs

Word Structure

Hall ( 1944) describes the grammar of Chinese Pidgin English in some detail.
According to the author, this pidgin was in use since the eighteenth century
in the treaty ports in central and southern China, 'as a medium of intercourse
between Chinese and foreigners.' Apparently it was never creolized' since
nobody acquired it as a native language. Two features in Hall's data show
clear influence from Chinese. In Chinese Pidgin English, the word side is
used as a suffix to indicate a location. Thus we find the following expressions:
Shanghai-side 'at Shanghai', office-side 'at the office', and you house-side 'at
your home'. Corresponding to these expressions, we have, in Chinese,
Shanghai na-bian (Shanghai that-side), bangongshi na-bian (office that-side),
and ni jia na-bian (you house that-side). In other words, the source of the
locative use of side is na-bian 'that side' or zhe-bian 'this side'. The sentence
My go club-side is a direct translation from Chinese: wo qu julebu na-bian.
CSE exhibits the same kind of substrate influence as well. One often
hears expressions like Clementi that side, Changi Village that side, where that
side has the same locative function as -side in Chinese Pidgin English. Both
are translations of the Chinese na-bian.
Number in English is expressed through inflection. So we have one book/
two books. Chinese is not an inflectional language, and makes use of classifiers
such as ben in counting (yi-ben shu 'one book /liang-ben shu 'two books').
'

Not surprisingly, Chinese Pidgin English has forms like two coolie, three
rickshaw. Interestingly, the word piece can used as a classifier in the above
two expressions: two piece coolie, three piece rickshaw. The substrate influence
is evident in the lack of number inflection and in the use of the classifier
piece.

Some Problems with the Substratist Approach

In some ways the substratist approach is intuitively satisfying. Our common

sense tells us that our native speech hilbits stay with us for a long time, and
will not change easily in new linguistic environments. T here are, however,
problems with the substratist approach in the actual analysis of linguistic
facts. Here we will mention three.
Hall (1962) argues that, unlike 'normal' languages which are transmitted
from generation to generation, a pidgin has a life cycle. It begins its life as
a 'minimum language' to fulfil the need for communication in communities
where no commor, language is spoken. Therefore, its early, 'jargon' stage is
simple in both vocabulary and grammar, and its functions are vastly reduced.
At the phase of 'expanded pidgin', it has gained in vocabulary and
grammatical complexity as well as communicative functions, particularly
when it acquires native speakers, i.e. is creolized. If the superstrate language
 Theories of LanglldgL' Genes,

continues to be spoken in the community and commands the position 0

prestige, the speakers will have the sociopolitical, cultural and commercia
incentive to adopt the speech standards of the superstrate language. Thl
creole then experiences decreolization, until it becomes indistinguishabll
from the superstrate language, whereupon it has reached the end of its life
If pidginization reduces the vocabulary size and grammatical complexity 0
the superstrate and substrate languages, creolization is just the opposite: th(
vocabulary, grammatical structure and functions are expanded, sometime!
vastly so. When we examine the features of various pidgins, we find that th{
substrate influence is the strongest in the jargon stage, and weakens towardE
the later stages. As the pidgin grows in grammatical complexity, the source
of influence becomes difficult to isolate. Undoubtedly expanded pidgim
contain grammatical constructions attributable to substrate influence,
especially those which are inherited from early stages; substratists cannot
account for the full complexity of creole structure.
Another problem with the substratist approach is the difficulty in
identifying the substrate languages themselves. In some situations, it is
possible to identify the source of substrate influence, as is the case with Tok
Pisin, Bislama and CSE. In the Atlantic creoles developed out of the slave
trade, slaves who were brought to the New World spoke entirely different
languages. So the pidgin which arose in such a situation was not only used
for communication between slaves and their owners, but for communication
among slaves as well. This was the situation for Surinam creoles (Arends et
al., 1995). The lack of homogeneity in the substratum makes it all but
impossible to attribute a grammatical structure to influence from a specific
substrate language. And the passage of time also obscures the origin of
grammatical features in a creole.
Finally, the substratist theories may fail to give a full account of a
grammatical construction, even though the substrate source can be clearly
identified. This is the situation with c luster simplification in CSE, which
we have seen in an earlier section. While it is true that no more than one
consonant can appear in word-final position in Chinese, the substrate
account does not explain why cluster simplification applies only to word­
final posit ion, leaving word-initial c l u s ters intact. This type of
'underexplanation' can also be found in substrate explanations of other
grammatical constructions.

e&" C(;jhe Universalist Approach

While the substrate theorists are preoccupied with creole constructions which
are traceable to substrate languages, the un iversalist theorists are impressed
by the structural similarities among creoles with diverse substrata. The core
60 English in New Cultural Contexts

process for the substratists is pidginization; for the universalists, Lt LS

creolization. There are many versions of the universalist approach to creole
studies, which are discussed in such works as Todd (1984), Romaine (1988)
and Arends et at. (1995 ) ; here we will introduce one such approach, the
Language Bioprogramme Hypothesis, developed by the noted creolist Derek
Bickerton.
The Language Bioprogramme Hypothesis is heavily influenced by
generative linguistics developed in the 1960s by Noam Chomsky and his
colleagues. S o , b efore we examine the content of the Language
Bioprogramme Hypothesis, let us outline Chomsky's view of language and
language acquisition.

Chomsky and Universal Grammar

The theory of language that is associated with Noam Chomsky is mentalistic

in philosophical orientation. Language is a mental faculty specific to the
human species, and the j ob of linguists is to specify its structure in rigorous
and formal terms. Chomsky expresses this idea in many of his writings (e.g.
1986). According to him, a child is born with the basic mechanism of ,
acquiring language, which he calls the initial state. As a child acquires
language, his mind 'matures' from the initial state to the final state, which
is the fully formed adult grammar of the language. Between the two states
there are various intermediate states, when the child's developing mental
grammar is not exactly the same as the target grammar. Schematically we
can represent the stages of child language acquisition as follows:

where So is the initial state, S 0 are the intermediate states, and S is the
I, <1<n n

final state . Although the states in the above schema are theoretical
idealizations, we can imagine the one-word stage, the two-word stage, and
so on, to be the products of the grammar in various states of the child's
linguistic development. I t must be emphasized that S represents a
hypothetical mental state, which allows us to conceptualize the process of
child language acquisition. The initial, innate state is also called Language
Acquisition Device, or LAD. At the present time LAD has no b iological
reality, and remains a controversial concept of generative linguistics.
A related, and equally controversial, concept proposed by Chomsky is
Universal Grammar. Universal Grammar is our genetically pre-programmed
gift of the mind which will eventually 'grow' into the grammar of our
languages. Since languages differ enormously, how does Universal Grammar
account for the differences that exist among languages ? In Chomsky's
conception, Universal Grammar consists of two parts - principles and
 Theories of Language Genesi

parameters. Principles are fixed, ava i lable to all languages. Parameters an

l ike switches, which can be switched on or off. T he differences amon!
languages, big or small; result from specific settings of the parameters. Tht
enterprise of generative linguistics is to d iscover the principles and enumerat(
the parameters that together determine the range of possible grammars 0
natural language. T he relat ionship between Universal Grammar anc
grammars of particular languages is schematizcd below:

Universal Grammar

G I
G n

Universal Grammar is the content of the initial state, So; it is a more explicitly
articulated theoretical concept. Chomsky ( 1 986: 1 46) summarizes the
concept of Universal Grammar as follows:

[We} may think of UG [Universa l Grammar] as an intricately

structured system, but one that is on ly partially "wired up." T he
system is associated with a finite set of switches, each of which has
a finite number of positions (perhaps two ) . Experience is required
to set the switches. When they are set, the system functions.

Switches are parameters, with either the 'on ' or 'off' positions. The child's
linguistic experience is obtained from the en vi ronment created by his or
her caretakers, and this helps him or her decide whether to set a certain
parameter in the 'on' or 'off' position .
To illustrate how the parameters of Universal Grammar work, we will
discuss one common parameter: the so-cal led pro-drop parameter, which
determines whether sentences of a language can have m issing subj ects.
English sentences typically require a subject; where there is no plausible
subject, a dummy is inserted to satisfy this requirement ( Muysken and
Veenstra, 1 995 ) :

( I O) a. He eats a'. * eats

b. I t seems that . . . b'. * seems that . ..

In ( l Ob ) , it is the dummy subj ect, and '. . .' stands for any sentence. Spanish,

* Indicates unacceptable forms.

62 English i n New Cultural Contexts

by contrast, allows null subjects. The two ungrammatical sentences in (lOa')

and ( l Ob') are grammatical in Spanish:

(lO) c. Come 'eats'

d. Parece que ' " , seems th at .. '
,

The grammaticality judgement is due to the setting of the pro-drop parameter:

English is [-pro-drop], Spanish is (+pro-drop]. I t turns out that other
grammatical contrasts can be attributed to the same parameter. Consider
the following two contrasts:

(10) e. Spanish: H a venido Juan.

English: *has come John.
J ohn has come.

f. Spanish: Ha sido devorada la oveja por el lobo.

English: *has been devoured the sheep by the wolf.
The sheep has been devoured by the wolf.

In ( lOe), the subject Juan/John appears after the verb; in ( l 0f) , which is
passive, the original object (la oveja/the sheep) appears from the passivized
verb. The Spanish examples are grammatical, whereas their English
counterparts are not. The three syntactic features are all associated with the
parameter [pro-drop]. We summarize the effect of this parameter below:

( +pro-drop] [-pro-drop]
M issing subject yes yes
Post-verbal subject yes no
Post-verbal object in passive yes no
•

What appear to be unrelated grammatical features of English and Spanish

are united by a single parameter. The parametric study of language attempts
to derive grammatical features from a small set of parameters which are
universally available in the biologically endowed Universal Grammar.
Although we still do not have the full set of such parameters, the approach
nevertheless has its conceptual attractiveness.

Bickerton and the Language Bioprogramme Hypothesis

Bickerton is dissatisfied with the substrate account of creole genesis. The

* Indicates unacce ptable forms.

 Theories of Languu!:t' Genesis

. . basic observation is this: in a slave community, a pidgin arises to fill the

basic communicative needs. It has a small vocabulary and a small repertoire
"
. of grammatical constructions, and performs a lim ited range of communicative
functions. With little or no direct contact with either the superstrate or the
\' substrate languages, children growing up in the community acquire the pidgin

f as their first language - in other words, the first-generation p idgin is rapidly

, nativized to become the mother tongue of the second generation. According
t to Bickerton ( 1981), this is the case with Hawaii Creole English, which
evolved from a jargon to a creole in one generation, bypassing the stabilized
pidgin and expanded pidgin stages. As has been observed by many students
of creole languages, the vocabulary, structure a nd function of the creole are
much r icher than those of its pidgin predecessor. T he question is, where
does the additional structure in the creole come from ? T he extreme substratist
would argue for a substrate source. T his position is difficult to maintain in
the case of Hawaii Creole English. Due to Hawa ii's geographical isolation,
children of slaves and labourers who had been brought to the plantations
had limited access to other languages, whether English or the native languages
of their parents. Nevertheless, they managed to acquire a creole which has
a larger vocabulary and a richer repertoire of grammatical constructions. In
relevant respects, creolization is language acquisition, except that a pidgin
is being acquired as a native language of a speech community. Here the
situation is quite severe: children with pidgin-speaking parents are not
exposed to the same range ofgrammatical constructions as children ofparents
who speak 'normal' languages. T he substratist approach fails to give an
adequate account of the grammatical aspect of creolization. It is in this
connection that B ickerton ( 19 8 1 ; 1984) prop oses his Language
Bioprogramme Hypothesis as a theory of creole genesis. In the Introduction
of his influential book, Roots of Language, Bickerton ( 198 1 : xii) says of his
task:

I shall examine the relationship between the variety of Creole

English spoken in Hawa ii and the pidgin which immediately
preceded it, and I shall show how several elements of that creole
could not have been derived from its antecedent pidgin, or from
any of the other languages that were in contact at the time of creole
formation, and that therefore these elements must have been, in
some sense, " invented.". . .I shall discuss some ... of the features which
are shared by a wide range of creole languages and show some striking
resemblance between the "inventions" of Hawaii and "inventions"
of other regions which must have emerged qu ite independently; . ..

Inventions, or innovative grammatical features of creoles, particularly those

� 64 English in New Cultural Contexts

which are shared across pidgins and creoles, are the tough nut for the
substratists to crack.
Chomsky's influence in the formulation of the Language Bioprogramme
Hypothesis is obvious. In Bickerton's words, 'the infrastructure of language
is specified at least as narrowly as Chomsky has claimed' (Bickerton, 1984) .
If Chomsky's claim of a biologically endowed LAD is right, then child
language acquisition is a process of parameter setting, which can be
accomplished only through exposure to linguistic input. If a child hears pro­
drop sentences, she 'sets' the pro-drop parameter as [+pro-drop]; if the language
being acquired is not pro-drop, the child will be exposed to data where
subjects are necessary for the well-formedness of sentences. Hence, the
parameter will be set [-pro-drop]. The same process is repeated for other
hypothesized parameters in Universal Grammar. When acquiring a 'normal'
language such as English and Spanish, a child will have positive evidence
for setting the parameters. Since a pidgin is simplified in structure, a child
may have positive evidence for some parameters, but not for all of them.
Bickerton hypothesizes that each Universal Grammar parameter has a default
setting, which is activated in the absence of positive evidence in the linguistic
environment. A child is born with Universal Grammar as part of her
biological endowment. When she is exposed to English data, the parameters
are set accordingly, and the child's Universal Grammar fleshes out to be the
grammar of English; when she is exposed to other languages, the same process
takes place, to create grammars of those languages. In a pidgin, she is exposed
to limited data, and lacks positive evidence for some parameters. The
grammar of the emerging creole will then contain structures from the pidgin,
and new structures which result from the default settings of some of the
parameters of Universal Grammar. This is, in a nutshell, the essence of
Bickerton's Language Bioprogramme Hypothesis.
It is difficult to find relevant empirical facts to support the Language
Bioprogramme Hypothesis, because of the lack of the right social conditions
for creolization. Bickerton thinks that Hawaii, because of its relative isolation
in the late nineteenth century and early twentieth century, provides just
the right social environment for the study of creolization that bears on the
Language Bioprogramme Hypothesis. In Bickerton ( 1981; 1984) , Hawaii
Pidgin English and H awaii Creole Engl ish are compared, and five
grammatical and semantic features are isolated. These features exist in Hawaii
Creole English but not in the pidgin which preceded it - a case of
grammatical innovation brought about by creolization. The five features
are as follows (Bickerton, 198 1: 1 7 ) :
a. movement rules
b. articles
c. verbal auxiliaries
 l , lell' te� UJ Lllnguage UeneSlS

d. for-to complementization
e. relativization and pronoun-copying

To better appreciate Bickerton's argument, we will take a look at three

features, namely, articles, movement rules, and for-to complementization,
in that order. All examples are quoted from Bickerton ( 198 1 ) .

Articles

In English, the referential capabilities of noun phrases are related to articles:

the definite article the and the indefinite article alan. In Standard English,
the specific and non-specific distinction applies only to noun phrases, and
is specified with the articles the and alan respect ively. A/an is used when the
noun phrase does not refer to any particular object, e.g. a man/an apple.
When we talk about specific things, we use the: the man/the apple. In Hawaii
Pidgin English, the articles are used sporadically and unpredictably. There
are two articles, the definite artic le da {the} and the indefinite article wan
(one). The following examples illustrate:

(11) a. hi get da hawaian waif.

He has a Hawaiian wife.

b. oni tek tu slais da bred.

I only take two slices of bread.

Here, the use of the defin ite article da is unpred ictable. In Standard English,
one would use a in the first sentence, and no article at all in the second. In
Hawaii Creole English, by contrast, da is used for noun phrases with spec ific
reference only:

(11) c. eefta da boi, da wan wen jink deet milk, awl da maut soa .
Afterward, the mouth of the boy who had d runk that milk was
all sore.

The indefinite article wan is used for first mention , where the NP reference
is unknown to the listener:

(11) d. hi get wan blaek buk, daet buk no du eni gud.

He has a black book. That book doesn't do any good.

The following noun phrases have no article:

(11) e. dag smat.

The dog is smart.
 f. yang fela dei no du daet.
Young fellows don't do that.

g. mi ai get raesh.
As for me, I get a rash.

The 'bare' noun phrase expresses meanings which call for different markings
in English, as the glosses indicate. The generic use of the noun phrase in
( l Ie) is marked with the in Standard English (dag vs. the dog), the plural use
in ( 1 1 f) with plural -s (yang fela vs. young fellows) , and the indefinite use in
(l Ig) with the indefinite article a. Bickerton terms all three uses non-specific.
In Hawaii Creole English, the unmarked noun phrase groups together three
categories which are kept distinct in Standard English. So, in referential
possibilities, the noun phrase in Hawaii Creole English differs not only from
its immediate predecessor, Hawaii Pidgin English , but also from the
superstrate language, English. One might argue that the noun phrase is
influenced by the substrate languages. But the substrate languages, among
them Japanese, do not have articles at all. Bickerton concludes that the
non-specific uses of unmarked noun phrases are an invention as Hawaii
Pidgin English is turned into Hawaii Creole English.

Movement Rules

Word order in Hawaii Pidgin English and Hawaii Creole English is another
invention which Bickerton thinks supports the Language Bioprogramme
Hypothesis. The word order of a sentence is expressed in terms of the relative
order of the subj ect (S) , the verb (V) and the object (0) . A typical English
sentence has the order SVO (John loves Mary) , and a typical Japanese
sentence has the order SOv. A language can be characterized in terms of
the word order of its typical sentences. So, English and Chinese are SVO
languages; Japanese is an SOy language, and so on. The word order of Hawaii
Pidgin English and Hawaii Creole EnglisQ. is summarized in Table 1 (adapted
from Bickerton, 1 981).

Word Order Hawaii Pidgin English Hawaii Creole English

SVO yes yes
SOy yes no
VS yes yes
VOS no yes
OSV no yes
OVS no yes

Table 1. Word Order in Hawaii Pidgin English and Hawaii Creole English
 Theories of Langua!;<' Genesis 6

both the pidgin and creole, the SVO order is the common and unmarked
. ice for all speakers. In Hawaii Pidgin English, word order preference
,,"c signals ethnicity; Japanese labourers prefer SOY order and Filipino
' ''''''' rn

rers prefer VS order. These orders are characteristic of the native

Ud!;C:� of the speakers, betraying substrate influence in Hawaii Pidgin

">,1,,, "".1. In Hawa ii Creole English, however, word order is uniform for all

, and the SOY order is notably missing. All other orders are derived
the basic order SVO through movement rules which move either the
(0) or the predicate ( V + 0) to the beginning of the sentence. VOS
derived by moving the predicate to the front; OSV by moving the object;
by moving the predicate first, and then the object. VS occurs in both
"",�, waii Pidgin English and Creole English. But in Hawaii Pidgin English it
.. __

is a result of the verb-first word order, possibly from Philippine languages.

Hawaii Creole English, it is the result of rules: 0 is deleted after VO is
1I'�'rn (HI Pr1 to the front to yield VOS. As for SOY, which does not exist in Hawaii

English despite the fact that it exists in its predecessor, the movement
rules will not derive it from the basic SVO: there is no fronting of the object
predicate.
How does the word order of Hawaii Creole English arise? Japanese is a
language, and its word order is SOY. Yet SOY does not exist in
ii Creole English. Substrate influence may exist in Hawaii Pidgin
but not in Hawaii Creole English which followed it. Bickerton ( 1 98 1 :
) argues that substratists are unable to explain the distribution of word
: order shown in Table 1 . He concludes:

[I]t can hardly be accidental if that particular distribution turns out

to be exactly what is generated if one assumes basic SVO order
(which is virtually mandatory when you have no other means of
marking the two major cases) plus a rule which moves either of the
two major constituents, NP and Vp, to sentence-initial position.
We may therefore claim that the rules which move NPs and VPs
cannot have been acquired inductively by the original [Hawaii
Creole English] speakers, but must, in some sense of the term, have
been "invented" by them ab ova.

For-to Comp1ementization
In English, embedded infinitive clauses may be introduced by for, as the
following sentences show:

(12) a. Mary bought this for you to read.

b. Mary prefers (for) Bill to go.
." 68 English in New Cultural Contexts

c. Mary prefers ( *for) to go.

Whether for is obligatory or not depends on the matrix verb. In ( 1 2a) , for
is obligatory ; i t introduces the infin itive clause you to read. In ( l 2b ) , for is
optional, the infin itive clause Bill to go can appear without it. In infinitive
clauses without an overt subject, as in ( 1 2c ) , for cannot be used at all. When ·
a sentence or clause is used to complete the meaning of a predicate phrase,
it is called a sentential complement, and the grammatical phenomenon is .
known as sentential complementization. Since the infinitive clauses in ( 1 2a) .
and ( 1 2b ) are introduced with for, Bickerton calls the phenomenon for-to
complementization.
In Hawaii Pidgin English, sentence-embedding of any kind is rare. Hawaii J

Creole English has sentential complements, marked by fa (from English for) :

and go. In Hawa ii Pidgin English, fa is used as a preposition, and go is used '
as a main verb, as a marker of imperatives, and 'as a preverbal modifier of ,

extremely indeterminate meaning and wildly fluctuating d is tribut ion' :

( Bickerton, 1 98 1 : 3 1 ). They are never used as complementizers, for the : ,

simple reason that Hawaii P idgin English has no sentential complements. !

Go and fa as complementizers are illustrated below:

( 13) go as complementizer

a. de i wen go ap dea erl i in da mawning go plaen.

T hey went up there early in the morning T O plant.

b. so a i go daun kiapu go push.

So I went down to Kiapu TO push (clear land with a bulldozer).

c. ai gata go haia wan kapinta go fiks da fom .

I had TO hire a carpenter 10 fix the form.

( 14) fa as comp lementizer

a. aen dei figa, get sambadi fo push demo
And they figured there'd be someone TO encourage them.

b. mo beta a bin go hanalulu fo bai maiself.

It would have been better if I 'd gone to Honolulu TO buy it
myself.

* Indicates unacceptable forms.

 Theories of Language Genesis o�

c. hau yu ekspek a gai fo mek pau hiz haus?

How do you expect a guy TO finish his house ?

; po and fa are used

in places where English would use to. Syntactically, the
C two words have the same grammatical function of introducing embedded
:
. sentences; semantically, go is used to describe actions which have taken
" place, whereas fa is used to describe actions which are hypothetical. The
'!ealized-untealized distinction is clearly indicated in the English glosses.
t ompared with its immediate predecessor, Hawaii Creole English exhibits
: three new grammatical a n d sema n t i c fe a t u re s . Firs t , sentential
;' complementization has been invented; second, the preposition fa and the
��mperative marker go have acquired the new function of complementizer;
{third, the realized-untealized distinction has been invented in sentential
complementization. Bickerton argues that these new features could not have
,been inherited from Hawaii Pidgin English, since the latter does not have
:shese features; nor from any of the contact languages. They were, therefore,
St nvented ab ova.
-J
- Summary
,

;�Inventions result from the intricate interplay of the universal principles and
��arameters of the language bioprogramme. In generative linguistics, the
.;notion of constituency plays an important role. All linguistic processes,
:
�mong them movement, apply to constituents, not random strings of words.
:hThe fact that the Hawaii Creole English movement rules move only
,:constituents (objects 0 and predicates V O) to the beginning of the sentence
"

·]s a result of universal principles. Exactly which principles and parameters

rare involved, Bickerton does not elaborate. The insight remains that the
:�ovative aspects of creole grammar cannot simply be attributed to substrate
� apguages, as staunch supporters of the substrate approach would argue. A
<plausible explanation of creolization lies in a thorough appreciation of
,1'inguistic universals. The innovative grammatical features conform to
,. � lversal constraints which the human language bioprogramme imposes on
' e structure of language .

�Some Problems with the Language Bioprogramme Hypothesis

.;me Language Bioprogramme Hypothesis has generated much excitement
'among students of pidgins and creoles, and it has been the subject of scholarly
{debate ever since Bickerton first proposed it. While it is attractive as a theory
jpf creole genesis, i t is neverthe less problematic in its conceptual
,underpinnings and empirical explanation. Conceptually, it is based on the
,View of language as a biologically determined mental faculty, but so far we
1<> have no biological evidence for this . B ickerton's language bioprogramme,
just like Chomsky's LAD, remains a mentalistic construct. A theory of creole
genesis based on such a construct is open to criticism from people with an
empiricist philosophical persuasion. In addition, even if we accept the
bioprogramme, we still do not know what constitutes a linguistic universal.
And the features which Bickerton attributes to the bioprogramme can b " ,
interpreted in other ways. The specific vs. non-specific distinction in Hawaii
Creole English is said to have arisen from the bioprogramme. But one can
say, with equal plausibility, that the English definite article the is retaine&
in Hawaii Creole English, and the indefinite article is replaced by wan. It'
does not bear on the bioprogramme at all, whether or not we accept the
Language Bioprogramme Hypothesis as an adequate theory of creole genesis.'
Empirically, it is difficult, if at all possible, to test the validity of the.
Language Bioprogramme Hypothesis in real-life situations. Although there
are many pidgins in the world which have undergone creolization, not all
of them can be used as evidence either to support or refute the Language,
Bioprogramme Hypothesis. Bickerton ( 1 98 1 ) stipulates two conditions. A
creole which is relevant for the Language Bioprogramme Hypothesis must"
be a language which:

1. arose out of a prior pidgin which had not existed for more than
a generation;

2. arose in a population where not more than 20 percent were native

speakers of the dominant language and where the remaining 80
percent was composed of diverse language groups.

The first criterion excludes cases like Tok Pisin, among others. The one- l
generation requirement minimizes the effect of prolonged contact with both !
the superstrate and substrate languages, which complicates the study of the :
language bioprogramme. The second criterion imposes on the creole 1
community a sociological condition 'which is difficult to obtain, even in j
such an isolated place as the Hawaii islands at the tum of the century. In : ,
order to verify the claims of the Language Bioprogramme Hypothesis, one '
needs to examine creoles spoken in communities with little contact with
the outside world, so that innovative grammatical features could be attributed
to the language bioprogramme, rather than contact languages. Under such
stringent requirements, most creoles do not qualify as relevant evidence.
According to Bickerton, Hawaii Creole English is just about the only ideal
creole for observing the language bioprogramme. Other researchers, among
them Janson ( 1 984) and Romaine ( 1 988), disagree, arguing that the features
of Hawaii Creole English which Bickerton discusses can in fact be interpreted
 Theories of Language GenesIs

ch a way that they do not bear on the language bioprogramme at all.

"ation is a case of language change. T he deve lopment of Hawaii Creole
Ish f rom Hawaii Pidgin Eng l ish is the refore a na logous to the
VeU)pnflern of Modern English from Middle English, and of Middle English
Old English. One need not invoke the hypothetical mental construct
language bioprogramme.
The mentalist foundation of the Language bioprogramme hypothesis
much debate of a philosophical na ture. A universalist theory of
,creOle genesis without the mentalist view of language is capable of accounting
the mechanisms of pidginization and creolization. Language universals
, re general tendencies found in the world 's languages, and new languages
more likely to acquire features which would conform to the universals,
against them . For example, the bilabial nasal /m/ ls found ·in all languages,
the inter-dental fricatives Ie, 0/ in only a few, among them , English.
ot surprisingly, alL New Englishes have the nasa l /m/, and no New English
the fricatives Ie, 6/. By assuming that language is a biologically endowed
tnental organ we do not necessarily provide a better account of the emergence
language.
Despite its conceptual and e mpi rical p roblems, t he La nguage
loprogramme hypothesis is nevertheless an interesting theory which opens
a host of q uestions for future research in p idgins and creoles. The debate
that has ensued has advanced our understanding of the structure and function
language in general, and pidgins and creoles in particular. T he research
has inspired continues.

�onclusion
In this chapter we examined two influential theories of pidgin and creole
. languages, namely the substratist approach and the universalist approach.
In their extremes the claims made by the two approaches are diagonally
opposed. T he former looks to the linguistic substra tum for explanation,
whereas the latter dismisses it as insignificant . From the perspective of the
pidgin and creole data, the substratist approach g ives a good account of
pldginization, where substrate influence is the strongest ; but it fails in
. accounting for creolization, particularly those constructions of creoles which
come neither from pidgin predecessors nor contact languages. T he
universalist approach gives a good account of creoliza.tion, attributing it to
the language bioprogramme. Creolization does not create structures that
would violate universal constraints on possible human languages imposed
by the bioprogramme. In their extreme forms the two theories are certainly
not tenable. T he real nature of pidginization and creolization perhaps lies
somewhere between the two extreme positions. One must acknowledge the
English in New Cultural Contexts

role of the substrate languages in the emergence of pidgins and subsequently

of creoles, and at the same time recognize the fact that pidgins and creoles,
like 'normal' l anguages, do not violate universal principles of language.
The two approaches need not be mutually exclusive, as Bickerton assumes.
Mufwene ( 1 986) says, 'most of the features of pidgins and creoles that the
substrate hypothesis has been claimed to explain are not really accounted
for unless some universal principles are accepted to apply at some stage of
the formation of these languages.' Pidgins and creoles are human languages;
as such they conform to universal principles which define human language
in general. The grammatical features selected from substrate languages in
the processes of pidginization and subsequent creolization are restructured
in strict accordance with the universals of language, whether or not they
constitute the bioprogramme.
It might be instructive to end this chapter with a quote from B ickerton
( 1 986 ) , which outlines the possibility of incorporating the l inguistic
substratum into a universalist theory.

A universalist account by no means rules out substratum influences

in principle; on the contrary, a universalist account ... must assume
that syntax includes a number of open options, where choices might
indeed be determined by substrataI influence .