You are on page 1of 28

<TARGET "pen" DOCINFO

AUTHOR "Long Peng and Jane Setter"

TITLE "The emergence of systematicity in the English pronunciations of two Cantonese-speaking adults in Hong Kong"

SUBJECT "EWW, Volume 21–1"

KEYWORDS ""

SIZE HEIGHT "220"

WIDTH "150"

VOFFSET "4"

BOOKMARK "The emergence of systematicity in the English pronunciations of two Cantonese-speaking adults in Hong Kong">

English World-Wide 21:1 81–108 (2000) © John Benjamins Publishing Co.

THE EMERGENCE OF SYSTEMATICITY IN THE ENGLISH


PRONUNCIATIONS OF TWO CANTONESE-SPEAKING ADULTS
IN HONG KONG1

Long Peng Jane Setter


State University of New York, Oswego Hong Kong Polytechnic University

This paper describes and analyses the phenomenon of consonant cluster simplifica-
tion in the English of two native Cantonese speakers in Hong Kong. We show that
this process is systematic in that it targets the alveolar plosives and removes them
when they are members of a coda consonant cluster in spite of the fact that the
details of the simplification may vary from subject to subject. We compare this
process to a seemingly similar cluster simplification in native varieties of English
and show that they differ in two key respects. Our study provides evidence of a
systematic morphophonemic alternation in the English of L1 Cantonese speakers,
confirming the observation in a number of sociolinguistic studies that this process
is a linguistic feature of the English of L1 Cantonese speakers.

1. Introduction

One main type of phonological phenomena that has attracted a lot of


attention in studies of first language (L1) sound systems is phonological

1) We would like to thank Patrick Nam Ka Wing and Jenny Cheng Chieh Yan for agreeing
to be our subjects in this study, Ivan Yuen Wing-chiu, our research associate, for his assistance
in this project, and Jean Ann for verifying some of the key transcriptions of the data. We would
also like to thank Edgar Schneider and the conference participants at 1997 Annual Research
Forum of Hong Kong Linguistics Society and First Asia-Pacific Conference on Speech,
Language and Hearing for their constructive comments on the content of this paper. This
research is funded by Hong Kong Polytechnic University CRG Grant No. 351/553.
<LINK "pen-r5">
"pen-r3">
"pen-r16">
"pen-r4">

82 LONG PENG AND JANE SETTER

alternation, or morphophonemic alternation in the terminology of classical


phonemics. This phenomenon is so named because morphemes that are
identical in meaning alternate from one pronunciation to another, depending
on the contexts they appear in. An example of alternation is seen in the
phonetic variations of verb stems such as damn and condemn in L1 English.
These stems have two distinct pronunciations in (1a), one with the stem-final
[n] and one without it. Another example of alternation is exemplified in (1b)
by the plural suffix which may be pronounced as [s], [z] or [Iz], depending
on the final segments of the stems.
(1) a. [dæm] damn [dæm-I]] damning
[k6ndem] condemn [k6ndem-I]] condemning
[dæmn-eIw6n] damnation
[k#nd6mn-eIw6n] condemnation
b. [lIp-s] lips [læb-z] labs [b%s-Iz] buses
[k%f-s] cuffs [k8˜v-z] curves [reIz-Iz] raises
Phonological alternations such as those in (1) are common in many L1 sound
systems and are often predictable from the surrounding phonological and/or
morphological environments. Part of the goal of phonological studies of L1
sound systems is to identify alternations of this sort and to describe the
conditions governing the alternations. Much of modern phonological theori-
sing derives its empirical basis from the sound patterns manifested in
phonological alternation.
Despite its importance in L1 sound systems, phonological alternation
has not attracted much attention as an object of inquiry in the studies of
second language (L2) sound systems. There was some earlier interest in
phonological alternation in L2 following Corder’s (1971) and Selinker’s
(1972) work on interlanguage. They proposed that L2 be viewed as a distinct
system — that is, L2 speakers’ version of the target language — and its
structural properties be studied independently of the target language. This
proposal encouraged some studies of the internal structures of interlanguage
(see e.g. Eckman 1981a; 1981b; 1991; Edge 1991; etc.). But more recent
works on L2 sound systems have focused their attention on other issues. This
study attempts to refocus attention on this important phonological phenome-
non, important because it is central to the understanding of the structural
properties of L2 and its relationships to those of L1 so that L1 and L2 can
<LINK "pen-r5">
"pen-r4">

SYSTEMATICITY IN HONG KONG ENGLISH 83

be compared at the systemic rather than individual lexical item level.2


There is some evidence of phonological alternation in L2 sound
systems. Eckman (1981a; 1981b), for instance, reports evidence of morpho-
phonemic alternations in the English interlanguages spoken by native
speakers of Cantonese, Japanese, Mandarin and Spanish. His analysis of
morphologically related forms such as [pIk] pig and [pI:i] piggy reveals a
number of morphophonemic alternations: Terminal Devoicing for Cantonese
speakers, Schwa Paragoge for Japanese and Mandarin speakers, Terminal
Devoicing and Postvocalic Spirantization for Spanish speakers. Edge’s
(1991) follow-up study of Japanese and Cantonese speakers affirms this
point even though she questions whether these rules are the unique properties
of interlanguage.3 These studies demonstrate that morphophonemic alterna-
tions as a systematic interlanguage phenomenon are worthy of study.
We continue this line of research here, examining the treatment of
English consonant clusters by L1 Cantonese speakers. Our interest in this
topic stems partially from a curiosity about the structural properties of the
English interlanguage of Cantonese speakers. While there has been some

2) Mohanan (1992) addresses explicitly the methodological issues involved in the study of
the non-native varieties of a language. He argues convincingly against using a standard native
variety of a language as the norm and describing the non-native variety as a list of deviations
from the norm. This approach advocates the comparison of the non-native variety with its
native norm at the level of individual lexical items. As the list of deviations is often not
representative of the non-native variety as a whole, it reveals little about the systematic
structural properties of the non-native variety and their similarities to and differences from the
native variety at the systemic level. Instead, he advocates the study of the non-native variety of
a language as an independent system so that comparisons can be made with the native variety
to see how the two systems — not their individual items — may differ or relate to each other.
3) In a replication study, Edge (1991) questions Schwa Paragoge of Japanese speakers and
Terminal Devoicing of Cantonese subjects. In the case of Schwa Paragoge, she points out that
native speakers of English do the same thing, suggesting that it may not be a unique feature of
the English interlanguage system of Japanese speakers. She questions Terminal Devoicing as
a morphophonemic alternation in the English interlanguage of Cantonese speakers. Voiced
consonants in L1 English are often devoiced in medial and initial positions, raising the question
of whether a voicing contrast exists at all in the interlanguage. While her study directly
questions the existence of an alternation in the interlanguage of Cantonese subjects, her central
conclusion is not so much whether these are morphophonemic alternations but whether they are
part of the interlanguage systems. As a matter of fact, while she questions Schwa Paragoge, she
suggests on the basis of her evidence the possibility of Terminal Devoicing as an alternation in
the English speech of her Japanese subjects.
<LINK "pen-r2">
"pen-r10">
"pen-r11">
"pen-r14">
"pen-r4">

84 LONG PENG AND JANE SETTER

interest in the sound system of this interlanguage in recent years, hardly any
study, apart from the devoicing of final consonants in Eckman (1981a),
provides systematic and detailed descriptions of its alternation patterns. Our
interest also stems from an earlier investigation of this topic in Eckman
(1987) as well as a number of works on the social and political status of
English in Hong Kong (Luke and Richards 1982; Platt 1982; Bolton and
Kwok 1990; Li 1999). All these studies compare the pronunciations of some
lexical items in English of L1 Cantonese speakers and their corresponding
forms in L1 English. On the basis of such comparisons, they point to the
simplification of consonant clusters as one of the linguistic features of the
English of L1 Cantonese speakers. As the claim of cluster simplification
relies on the comparison of interlanguage forms with those of the target
language in these studies, it is not clear whether the simplification of
consonant clusters is a systematic feature of the English interlanguage of L1
Cantonese speakers. The goal of this study is to determine whether the
process of simplification is systematic and, if so, to describe the conditions
that trigger this process.
Our study shows that consonant cluster simplification is systematic in
the L2 English of the two subjects analysed in this paper. It targets the
alveolar plosives [t] and [d] and removes them when they are members of a
coda consonant cluster. Our study also reveals that the two subjects may
differ substantively in their alternation patterns. In spite of this, the funda-
mental condition underlying the alternation is identical in that it is triggered
by the presence of a consonant cluster. Our study provides evidence of a
systematic morphophonemic alternation in the English interlanguage of L1
Cantonese speakers. It confirms the observation in Luke and Richards
(1982), Platt (1982), Eckman (1987), and Bolton and Kwok (1990), suggest-
ing that this process is a linguistic feature of the English of at least some
Cantonese speakers in Hong Kong.
This paper is organised as follows. In section 2, we address the main
methodological issues. section 3 provides a detailed description of the
patterns of alternation as evidenced in the English speech of each of the two
subjects. In section 4, we discuss the status of this alternation as an inter-
language phenomenon. Finally, section 5 explores some implications of this
study.
SYSTEMATICITY IN HONG KONG ENGLISH 85

2. Methodology

2.1. Subjects

The data documented here were obtained from two adult native speakers of
Cantonese in Hong Kong. In (2), we present some relevant personal data of
the two subjects for information.
(2) Subject Sex Age Native Time of Formal Time in an English-
Language English Training speaking country

Subject 1 Female 20 Cantonese 15 years None

Subject 2 Male 23 Cantonese 15 years 1 month in Britain

At the time of research, the two subjects were in their second year of study
in a BA degree in Contemporary English Language at the Hong Kong
Polytechnic University. Prior to their university enrolment, they attended
English-medium schools for a total of 13 years (6 of primary school and 7
of secondary school). As in most English-medium schools,4 subjects other
than English were offered in Cantonese even though they are supposed to be
taught in English. During these 13 years, the two subjects averaged 7 to 8
hours a week of English instruction. For the two years they spent at the
Hong Kong Polytechnic University, they had roughly 6 hours a week of
formal English instruction and 9 hours of English linguistics, all of which
were conducted in English. Apart from one month Subject 2 spent in Britain,
neither subject spent any time in an English-speaking country. In general, the
subjects’ exposure to English is fairly representative of the experience of
Hong Kong students enrolled in English-medium schools. What distinguishes
these two students from others is the two years they spent in a BA English
programme. They can conduct conversations on daily topics in English
without much difficulty and their proficiency in English may be character-
ised as upper intermediate.

4) In Hong Kong, public-funded schools are classified into two types: (a) English-medium
and (b) Chinese-medium. English-medium schools are required to use English as the medium
of instruction. In reality, textbooks are in English, but the language of classroom is often
Cantonese.
86 LONG PENG AND JANE SETTER

2.2. Data collection

The data reported here are based on a list of 576 word pairs, totalling 1152
words. Each pair consists of two morphologically related forms: one morpho-
logically derived word with one of the selected suffixes and one without this
suffix. In determining the suffixes for inclusion in this study, we consider
one key criterion: whether a suffix begins with a vowel or consonant. We
include both vowel-initial and consonant-initial suffixes in the study so that
the different patterns of simplification under suffixation can be studied. We
follow three criteria in selecting the words for inclusion in the word list.
First, as we are interested in the behaviour of stem-final consonants in
different morphological contexts, the words are carefully selected so that
they represent the full spectrum of different consonant endings that English
allows both in terms of types of consonants and in terms of cluster types.
Second, as there are restrictions on whether a stem can take the selected
suffix, the list includes only those that can take the corresponding suffix.
Lastly, we include in the list words of various grammatical classes so as to
ensure that the sample is representative.
The subjects were recorded individually and completed the recording in
two sessions. At the beginning of each session, they were asked to identify
those words in this list which they were not familiar with to prevent them
from guessing the pronunciations of words on the basis of spellings. They
were instructed to read each word twice; their pronunciations were recorded
on a Sony Minidisc Recorder (MZ-R3). For most of the words in the list, the
two readings were identical. Occasionally, there were discrepancies between
the two readings of a word. In such cases, we ascertained the pronunciation
by placing the relevant word in sentential contexts. The subjects were not
informed of the goal of research or the aspect of their speech under investi-
gation.
Once the recordings were completed, the pronunciations of the two
subjects were transcribed independently by three people: the two authors and
a research assistant. The transcriptions were compared to determine whether
they were consistent with respect to the feature under investigation. The
comparison reveals that the transcriptions were consistent in more than 90%
of the cases. Where there were differences, the recordings were checked
again to determine the correct transcription.
<LINK "pen-r1">
"pen-r4">

SYSTEMATICITY IN HONG KONG ENGLISH 87

We are well aware that reading a word list may not represent the most
authentic form of communication. But we settle on a word list for three
reasons. First, word lists allow us to elicit a large number of morphologically
related forms so that the conclusion drawn here is based on a substantial
amount of empirical data. Second, word lists have the advantage of minimis-
ing the influence of performance errors such as stutters, slips of the tongue,
etc., which do not accurately reflect the linguistic competence of a speaker.
Lastly, but most importantly, reading a word list represents arguably the
most careful type of speech as the speaker is more able to focus on the
individual sounds of a word in reading a word list than in natural conversa-
tions. According to Bayley (1996: 105), this is a style of speech in which L2
speakers are least likely to drop alveolar plosives. If it turns out that L2
speakers delete alveolar plosives in a style in which they are least expected
to do so, the data so uncovered provides the most convincing type of
evidence for the existence of a phonological pattern.
Although the data reported here are based entirely on the reading of
word lists, the pronunciations of some words were double-checked in
sentential contexts and in natural, spontaneous speech. These pronunciations
are consistent with those collected by means of the word list insofar as the
pattern reported here is concerned. In addition, the data collected here are
consistent with those of Eckman (1987), whose data are obtained through
word lists, sentence reading and natural conversations. We doubt seriously
that the patterns documented here are a reflection of the way in which the
data were collected. If any discrepancy may arise from the chosen speech
style, our expectation would be that there would be more — not fewer —
cases of alveolar plosive deletion as the speech becomes less formal and
more casual as shown in Bayley (1996).

2.3. Analytic procedures

The currently popular methodology of second language research employs two


basic practices. One practice of second language research mixes the data
from different subjects. The second practice mixes the data collected through
different styles of speech such as reading of word lists, reading of passages,
natural conversations, etc. The fundamental motivation of these practices is
to ensure that the data so gathered provide a reliable and representative
<LINK "pen-r9">

88 LONG PENG AND JANE SETTER

measure of the overall language proficiency of a selected group of L2


speakers. These practices may be appropriate for a large-scale quantitative
study of speech, whose goal is to identify the shared linguistic characteristics
of L2 speakers as a group. But they are not appropriate for the present
investigation, the goal of which is to describe the phonological conditions
that trigger the alternation of alveolar plosives in the speech of each subject.
The problem with mixing data from different subjects and styles of speech
is that these practices may obscure the patterns in the data and lead to false
claims of unpredictability or variability.
Consider the first practice of mixing data from different subjects. It is
well known that L1 speakers may differ in speech and their grammars may
vary from speaker to speaker. Such inter-individual differences are well
known especially in sociolinguistic studies of speech (see e.g. Labov 1989
for a summary of L1 differences with respect to the deletion of [t] and [d]
in American English). The term “idiolect” or “dialect of self” recognises
linguistic differences that may exist among different speakers of the same
language or dialect. When it comes to second language speech, the linguistic
variations that may exist among different speakers are much more acute due
to individual differences in language proficiency, the amount of formal
instruction, language exposure, personal motivation, language aptitude, etc.
In light of these differences, the combination of data from different subjects
may distort the underlying regularity and lead one to conclude that a
particular phenomenon is unpredictable when it is not. Consider the data of
the two Cantonese subjects in this study as an illustration. As we will show
later, Subject 1 deletes alveolar plosives in certain environments, one of the
environments being tri-consonantal clusters in (3a). Subject 2 retains alveolar
plosives in the same environment, as in (3b).
(3) a. [%sept] accept [%seps] accepts
[%k#nt] account [%k#ns] accounts
b. [6sept] accept [6septs] accepts
[%k#nt] account [%k#nts] accounts
Mixing (3a) and (3b) may lead to the conclusion that the phenomenon is not
predictable, as [t] is sometimes deleted and sometimes retained in the same
environment. But as we will show in the following sections, alveolar plosive
deletion is systematic, predictable in the speech of each subject, even though
their patterns of deletion are not completely identical.
<LINK "pen-r12">
"pen-r6">
"pen-r1">

SYSTEMATICITY IN HONG KONG ENGLISH 89

The second practice of mixing data gathered from different styles of


speech poses the same problem. Different phonological rules may operate in
different styles of speech. This is well known in L1 studies. An example is
a rule of deletion that targets alveolar plosives in a tri-consonantal cluster in
L1 English. According to Gimson (1994: 215), this rule operates only in
rapid, colloquial speech. In careful speech, retention — not deletion — is the
norm. If the data collected from rapid, colloquial speech in (4a) are com-
bined with those of careful speech in (4b), they are likely to lead to the
conclusion that deletion is either unpredictable or optional in that it does not
apply obligatorily.
(4) a. [d%s( )mæn] dustman [k6~s( ):"˜d] coastguard
b. [d%s(t)mæn] dustman [k6~s(t):"˜d] coastguard
As different rules may operate in different styles of speech in L1, it is a
possibility that there may be different rules governing different styles of
second language speech. Many second language studies have identified the
differences arising from different styles of L2 speech (see, for instance,
Bayley 1996; Major 1994). For this reason, non-judicious mix of the data of
different styles of speech may obscure the underlying regularity just as the
mix of data of different subjects.
Our goal here is to describe the phonological conditions that trigger the
alternation involving alveolar plosives. To achieve this goal, we believe that
it is important to (a) describe the conditions governing the alternation of one
subject at a time and (b) describe the conditions governing the alternation in
one style of speech at a time. Only after the patterns of individual subjects
and those of different styles of speech are identified does it make sense to
compare one subject with another and one style with another to determine
the area of similarity and extent of variation between the two subjects and
different styles. For this reason, we analyse the data of each subject based
on the reading of word lists.

3. Analysis

In this section, we describe a pattern of alternation involving alveolar


plosives. In the English of the two Cantonese subjects, the alveolar plosives
[t] and [d] which are present in one morphological paradigm may disappear
90 LONG PENG AND JANE SETTER

in another morphological paradigm. Our analysis reveals that the variations


are systematic within the English interlanguage of each speaker and may be
accounted for by a deletion rule, triggered by the presence of a consonant
cluster in the syllable coda. The data documented here provide evidence that
a process of consonant cluster reduction operates within the English inter-
languages of the two Cantonese subjects.

3.1. The consonants and their distribution

Let us consider briefly the inventory of consonants and their distribution


before discussing the patterns involving alveolar plosives. We can identify a
total of twenty-five surface consonants in the English of our Cantonese
subjects. Twenty-three may appear in the syllable onset, as illustrated in (5).
(5) Onset consonants in the English of the two Cantonese subjects
labial dental alveolar velar glottal

stops p, b t, d k, :
fricatives f, v q, ð s, z, w, Š h
affricates tw, dŠ
nasals m n
liquids l, r
glides w j

Of the twenty-three segments, the two underlined interdental fricatives — [q]


and [ð] — are rare, replaced frequently by [f] and [d], respectively, in the
syllable onset. The number of consonants that appear in the syllable coda is
more restricted; only twenty out of twenty-five consonants may function as
the coda:
(6) Coda consonants in the English of the two Cantonese subjects
labial dental alveolar velar glottal

stops p, b t, d k, : ‘
fricatives f, v s, z, w, Š
affricates tw, dŠ
nasals m n ]
liquids
glides w j
SYSTEMATICITY IN HONG KONG ENGLISH 91

In the English of our subjects, [q], [ð], [l], [h] and [r] are not found in the
syllable coda. [q] and [ð] are replaced by [f]; the lateral [l] is vocalised as
[~]. The voiced obstruents are not as common as their voiceless counterparts,
due to the fact that they are often devoiced in the syllable coda. The voice-
less velar [k] is rare in the coda, as it is often replaced by a glottal stop.
Consonant clusters are attested more often in syllable onsets than in
syllable codas. A maximum of three consonants may appear in the syllable
onset, though such complex onsets are rare, limited in both number and type.
Generally, up to two consonants may fill the coda position. Complex codas
are even more restricted in number and type. By far the most common type
of complex codas consists of a nasal followed by an obstruent. Complex
codas with two obstruents such as [kt], [pt], [ft], [ks] are attested less
frequently because of the deletion and glottalisation of certain plosives. Coda
clusters made up of a lateral followed by another consonant such as [lt], [lp],
[lk], [lm], etc. are not attested due to the vocalisation of [l] in the coda position.

3.2. Subject 1

Our discussion of Subject 1 begins with forms that take the vowel-initial
suffix: -ing. This discussion is followed by consideration of forms that take
consonant-initial suffixes: (a) -s; (b) -ful and (c) -ment. We show that the
alveolar plosives [t] and [d] do not undergo elision in a form taking the
vowel-initial suffix and that elision is seen only in forms that take the three
consonant-initial suffixes. This distinction supports the conclusion that the
deletion of alveolar plosives is a consonant cluster reduction process.
Consider first the -ing suffix data. In what follows, we present the
phonetic forms of words without the -ing suffix in the left column and the
phonetic forms of words with the -ing suffix in the right column. The data
clearly show that the stem-final alveolar plosives [t] and [d] are retained,
regardless of whether a consonant or a vowel immediately precedes the
alveolar plosives.
(7) English -ing suffix
[æ‘t] act [æktI]] acting
[li˜d] lead [li˜dI]] leading
[%sept] accept [%septI]] accepting
[%k#nt] account [%k#ntI]] accounting
[hend] hand [hændI]] handing
92 LONG PENG AND JANE SETTER

[maInt] mind [maIndI]] minding


[Invest] invest [InvestI]] investing
[6dŠ%st] adjust [6dŠ%stI]] adjusting
The data in (7) are important in comparison with those presented in (8)
through (10), which show that the alveolar plosives may be elided when a
consonant-initial suffix is attached. We include in (7) some of the same
words as those that appear in (8) through (10) so that a comparison can be
made between (7) and (8)–(10) about the different behaviours of alveolar
plosives when different suffixes are attached.
Let us turn now to the data with consonant-initial suffixes, which are
grouped in terms of: (a) words ending in a single [t] or [d]; (b) words ending in
[Ct] or [Cd] (C stands for a consonant other than [s] here); and (c) words ending
in [st] clusters. These groupings reflect the fact that the patterns of deletion vary
along these lines. In (8), we present a broad phonetic transcription of words with
or without the third person singular or plural suffix, which is realised as [s] or
[6s]. The data show that the alveolar plosives [t] and [d] in all three contexts are
deleted uniformly under suffixation even though they are present in forms
without the suffix and with the -ing suffix in (7).
(8) English plural or 3rd person singular suffix
a. In words ending in [t] or [d], [t] or [d] is deleted when -s is
attached
[b6~t] boat [b6~s] boats
[æ‘t] act [æs] acts
[pr6m6~t] promote [pr6m6~s] promotes
[li˜d] lead [li˜s] leads
[ri˜d] read [ri˜s] reads
[sk6~d] scold [sk6~s] scold
b. In words ending in [Ct] or [Cd] clusters other than [st], [t] or [d]
is deleted when -s is attached
[%sept] accept [%seps] accepts
[%k#nt] account [%k#ns] accounts
[frend] friend [frens] friends
[6b#nd] abound [6b#ns] abounds
c. In words ending in [st] clusters, [t] is deleted when -s is attached
[Invest] invest [Invess] invests
[6dŠ%st] adjust [6dŠ%ss] adjusts
[dentIst] dentist [dentIss] dentists
[lI]:wIst] linguist [lI]:wIss] linguists
SYSTEMATICITY IN HONG KONG ENGLISH 93

Note that the output of elision is a geminate [ss] in (8c), which is not
permitted in the native varieties of English. But in reading the word list
prepared by us, our Cantonese subjects have no problem in prolonging the
production of [s],5 which we transcribe as [ss]. It is also worth noting that
Subject 1 (and Subject 2) systematically treats words with a single final [s]
such as [b%s] differently from words with a final [st] cluster. Words that end
with a sibilant take the [6s] variant of the plural or third person singular
suffix: e.g. [b%s6s]. Words with a word-final [st] cluster take the [s] variant
of the suffix; there is no evidence of a schwa present in (8c).
Given the data in (8), one may be tempted to conclude that [t] or [d] is
deleted in the environment of another consonant. This analysis does not
properly characterise the deletion process in Subject 1 once we consider the
data involving other consonant-initial suffixes. In (9) where the data
involving the -ful suffix are presented, we see that [t] or [d] is removed only
if it is preceded immediately by another consonant. [t] or [d] is retained in
suffixed forms if it is immediately preceded by a vowel. Some alveolar
Ÿ
plosives are partially devoiced, which we transcribe as [d].
(9) English -ful suffix
a. In words ending in [t] or [d], [t] or [d] remains when -ful is
attached
[fraIt] fright [fraItff~] frightful
[fru‘t] fruit [fru‘tff~] fruitful
Ÿ
[da~t] doubt Ÿ
[da~tff~] doubtful
[f6:et] forget [f6:etff~] forgetful
[rI:"‘d]Ÿ regard [rI:"dff~] regardful
Ÿ
[ni˜d] need Ÿ
[ni˜dff~] needful
b. In words ending in [Ct] or [Cd] clusters other than [st], [t] or [d]
is deleted when -ful is attached
[hend] hand [hænff~] handful
[maInt] mind [maInff~] mindful
c. In words ending in [st] clusters, [t] is deleted when -ful is attached
[tr%st] trust [tr%sff~] trustful
[teIst] taste [teIsff~] tasteful
[rest] rest [resff~] restful

5) Occasionally, we do see the pronunciation in which the alveolar plosive and the third
person singular are both omitted. So invests has the pronunciation of [Inves] rather than
[Invess].
94 LONG PENG AND JANE SETTER

The same pattern of deletion is found in forms taking the -ment suffix, as
shown in (10).
(10) English -ment suffix
a. [t] or [d] remains or turns into [‘] when -ment is attached
[steIt] state [steI‘m6nt] statement
[k#mIt] commit [k#mItm6nt] commitment
[6l#t] allot [6l#tm6nt] allotment
[6sft] assort [6sftm6nt] assortment
[In.æt] enact [In.ætm6nt] enactment
[6mend] amend [6medm6nt] amendment
b. In words ending in [Ct] clusters other than [st], [t] is deleted when
-ment is attached
[6pfInt] appoint [6pfInm6nt] appointment
[dIs6pfInt] disappoint [dIs6pfInm6nt] disappointment
[k6m"nt] command [k6manm6nt] commandment
c. In words ending in [st] clusters, [t] is deleted when -ment is
attached
[e‘dŠ%st] adjust [e‘dŠ%sm6nt] adjustment
[6rest] arrest [6resm6nt] arrestment
[Invest] invest [envesm6nt] investment
The data in (9a) and (10a) suggest that the mere presence of another conso-
nant following the stem-final [t] or [d] such as [f] of -ful or [m] of -ment is
not sufficient to trigger deletion of alveolar plosives. Where alveolar
plosives are deleted, they are preceded by another consonant.
To summarise, the stem-final alveolar plosives are retained in some
morphological contexts such as when they appear without the suffixes or
with the vowel-initial -ing suffix. Alveolar plosives may be deleted when
they occur with suffixes that begin with a consonant. Specifically, when the
plural or third person singular suffix is attached, [t] or [d] is deleted regard-
less of whether it is preceded by a vowel or consonant. But when -ful and
-ment are attached, alveolar plosives undergo elision when another consonant
precedes them. These patterns are summarised in the Table in (11):
SYSTEMATICITY IN HONG KONG ENGLISH 95

(11) Subject 1
-ing -s -ful -ment

t,d → t,d /V__ing t,d → 0 /V__s t,d → t,d /V__ful t,d → t,d,‘/V__ment
t,d → t,d /C__ing t,d → 0 /C__s t,d → 0 /C__ful t → 0 /C__ment
t → t /s__ing t → 0 /s__s t → 0 /s__ful t → 0 /s__ment

The deletion of alveolar plosives takes place not merely because of the
presence of a following consonant: alveolar plosives are elided under
suffixation of -ful and -ment only when they are adjacent to another conso-
nant. These properties of elision can be interpreted as a reflection of the role
of syllable structure. Put simply, the elision of alveolar plosives is syllable-
driven in the English of Subject 1. The alveolar plosives are deleted only if
they are members of a coda consonant cluster. In the case of the stems with
the singular or plural suffix, stem-final alveolar plosives are part of the
syllable coda as the suffix consonant [s] can be syllabified only into the
same syllable coda as the stem-final alveolar plosives, as shown in (12a). In
contrast, two patterns emerge from the -ful or -ment suffixation. As the
suffix-initial consonants of -ful and -ment function as the onset, an alveolar
plosive may form a cluster only if it is preceded by another consonant (12b);
otherwise, it will be the only coda consonant in the syllable (12c). The
different surface effects arise from the distinct syllable configurations that
alveolar plosives find themselves in:
(12)
a. s b. s s c. s s s

b 6 ~ t s h ae n d f f ~ 6 l # t m e n t

∅ ∅ n/a

Now that the condition triggering the deletion is clear, we are in a position
to state the rule of deletion. In (13), we specify a comma between C and
____ to express the fact that [t]/[d] may be the first or second member of a
coda cluster, as illustrated by (12a) and (12b). The specification of the right
bracket restricts the rule to coda consonants in a syllable. The reference to
the suffix conveys the fact that the deletion takes place only under suffix-
ation as coda consonant clusters consisting of [t] or [d] are permitted in
unsuffixed forms such as [Invest] invest.
96 LONG PENG AND JANE SETTER

(13) t/d → ∅ ⁄ C, ____ ]s suffix


So far, we have considered stems that end with alveolar plosives. What about
other stem-final consonants? From the data we have collected, we have
evidence that these consonants do not participate in this deletion process in
the same way in the speech of Subject 1. Under suffixation, these conso-
nants may be devoiced if they are voiced obstruents and they may become
glottal stops if they are either bilabial or velar plosives. But in contrast with
alveolar plosives, these stem-final consonants are not deleted. Some stems
taking the -s and -ment suffixes are shown in (14).
(14) [æpsf˜p] absorb [æpsf˜ps] absorbs
[b"˜f] bath [b"˜fs] baths
[tj~tfrIl] tutorial [tj~tfrIls] tutorials
[qI]k] think [qI]‘s] thinks
[6ses] assess [6sesm6nt] assessment
[p%nIw] punish [p%nIwm6nt] punishment
[e]geItw] engage [e]geIdŠm6nt] engagement
[rIfaIn] refine [rIfaInm6nt] refinement
[Inh%ns] enhance [Inh%nsm6nt] enhancement
[6reIntw] arrange [6reIndŠm6nt] arrangement
Since consonants other than alveolar plosives remain intact even as members
of a coda cluster (e.g. [Inh%nsm6nt] enhancement), we restrict the application
of the rule in (13) to alveolar plosives.

3.3. Subject 2

One characteristic that distinguishes Subject 2 from Subject 1 is that Subject


2 retains [t] in more environments than Subject 1, with elision restricted to
a subset of the environments seen in the English of Subject 1. In spite of this
difference, the main condition triggering deletion is identical in that the deletion
is triggered by a presence of a consonant cluster in the syllable coda.
In a stem with a final alveolar plosive, this plosive remains intact when
a vowel-initial suffix is attached to this stem. In (15), we include only those
words that end either with a [nt] or [st] cluster as they are the only words
that alternate when consonant-initial suffixes are attached in (16) through (18).
SYSTEMATICITY IN HONG KONG ENGLISH 97

(15) English -ing suffix


[Invest] invest [InvestI]] investing
[6dŠ%st] adjust [6dŠ%stI]] adjusting
[rest] rest [restI]] resting
[teIst] taste [teIstI]] tasting
[tr%st] trust [tr%stI]] trusting
[%rest] arrest [%restI]] arresting
[%pfInt] appoint [%pfIntI]] appointing
[dIs6pfInt] disappoint [dIs6pfIntI]] disappointing
In contrast, alveolar plosives may undergo deletion when a consonant-initial
suffix is attached to a stem. Consider the affixation of the plural or third
person singular suffix which is also realised as [s] or [6s] in Subject 2. In a
stem in which [t] is the only consonant in the final position, this alveolar
plosive remains when -s is attached. A stem-final [d] is devoiced when -s is
attached. This is exemplified in (16a). In (16b), where [t] or [d] is a member
of a consonant cluster with [p] and [n], this alveolar plosive is still retained
when the -s suffix is attached. Deletion is seen only where the alveolar
plosive [t] is preceded by the fricative consonant [s] as in (16c).
(16) English plural or 3rd person singular suffix6
a. In words ending in [t] or [d], [t] remains and [d] turns into [t] when -s
is attached
[b6~t] boat [b6~ts] boats
[pr6m6~t] promote [pr6m6~ts] promotes
[ækt] act [æts] acts
[li˜d] lead [li˜ts] leads
[ri˜d] read [ri˜ts] reads
[sk6~d] scold [sk6~ts] scolds

6) There are some data that appear to suggest that the bilabial and velar plosives [p] and [k] may be
elided under suffixation: e.g. [ækt] act vs. [æts] acts in (16a) and [da~pt] doubt vs. [da~tf#~] doubtful
in (17a). Careful examination of all the collected data reveals two aspects that distinguish the elision of
these plosives from that of alveolar plosives considered in this paper. First, the elision of these plosives
does not appear to be systematic as these plosives are not always deleted under similar conditions: cf.
[6sept] accept vs. [6septs] accepts in (16b) and [qI]k] think and [qI]ks] thinks in (23). Second, our
corpus includes little data from either of the two subjects which show the presence of consonant clusters
with [p] or [k]. This is due to the fact that these two consonants are often glottalised. Consequently,
there is not enough data to demonstrate the elision of these plosives in spite of the fact that our
sample word list contains many examples of words with consonant clusters that include [p] or [k]
as part of the cluster. For these two reasons, our discussion does not include the elision of [p] or [k].
98 LONG PENG AND JANE SETTER

b. In words ending in [Ct] or [Cd] clusters other than [st], [t] remains and
[d] turns into [t] when -s is attached
[6sept] accept [6septs] accepts
[%k#nt] account [%k#nts] accounts
[frend] friend [frents] friends
[6b#nd] abound [6b#nts] abounds
c. In words ending in [st] clusters, [t] is deleted when -s is attached
[Invest] invest [Invess] invests
[6dŠ%st] adjust [6dŠ%ss] adjusts
[dentIst] dentist [dentIss] dentist
[lI]gwIst] linguist [lI]gwIss] linguists
The same pattern of deletion is found under the suffixation of -ful. The dark
lateral of the suffix is vocalised, giving rise to the pronunciation of -ful as
[−f#~].
(17) English -ful suffix
a. In words ending in [t] or [d], [t] or [d] remains or turns into [‘] when
-ful is attached
[frut] fruit [fru‘f#~] fruitful
[da~pt] doubt [da~tf#~] doubtful
[f#get] forget [f#getf#~] forgetful
[dIsrIspet] disrespect [dIsrIspetf#~] disrespectful
[rig"t] regard [rig"tf#~] regardful
[nit] need [nidf#~] needful
b. In words ending in [Ct] clusters other than [st], [t] remains when -ful is
attached
[Ivent] event [Iwentf#~] eventful
[hen] hand [hentf#~] handful
[maIn‘] mind [maIntf#~] mindful
c. In words ending in [st] clusters, [t] is deleted when -ful is attached
[rest] rest [resf#~] restful
[teIst] taste [teIsf#~] tasteful
[tr%st] trust [tr%sf#~] trustful
The pattern of deletion is slightly different in the case of stems taking the
-ment suffix. The alveolar plosive is dropped as long as another consonant
precedes it regardless of whether it is [s] or not. This is exemplified in (18b)
and (18c).
SYSTEMATICITY IN HONG KONG ENGLISH 99

(18) English -ment suffix

a. In words ending in [t] or [d], [t] or [d] remains or turns into [‘] when
-ment is attached
[steIt] state [steI‘m6nt] statement
[k#mIt] commit [k#mI‘m6nt] commitment
[6l#t] allot [6l#tm6nt] allotment
[Inet] enact [Ine‘m6nt] enactment
b. In words ending in [Ct] clusters other than [st] clusters, [t] is deleted
when -ment is attached
[k#ntent] content [k#ntenm6nt] contentment
[%pfInt] appoint [%pfInm6nt] appointment
[dIs6pfInt] disappoint [dIs6pfInm6nt] disappointment
c. In words ending in [st] clusters, [t] is deleted when -ment is attached
[6dŠ%st] adjust [6dŠ%sm6nt] adjustment
[%rest] arrest [%resm6n] arrestment
[Invest] invest [Invesm6nt] investment
The various patterns of deletion for Subject 2 are summarised in (19).
(19) Subject 2
-ing -s -ful -ment

t,d → t,d /V__ing t,d → t /V__s t,d → t,d,‘/V__ful t,d → t,d,‘/V__ment


t,d → t,d /C__ing t,d → t /C__s t,d → t /C__ful t → 0 /C__ment
t → t /s__ing t → 0 /s__s t → 0 /s__ful t → 0 /s__ment

One consistent aspect of deletion in (19) is that the alveolar plosive is


removed if it is preceded by [s] and followed by a consonant-initial suffix.
We can state the deletion affecting these alveolar plosives as follows.
(20) t/d → ∅ ⁄ s ____ ]s suffix
What distinguishes (20) from the deletion rule of Subject 1 in (13) is the
specification of [s] as part of the conditioning environment. Apart from this,
the presence of a preceding [s] and a consonant-initial suffix ensures that [s]
and [t] are both members of the coda. The alveolar plosive deletion of
Subject 2 is thus triggered by a condition similar to that seen in Subject 1.
In (20), we specify “suffix”, not specifically “a consonant-initial
suffix” as the conditioning environment of the deletion rule. There is no
need to refer to a particular type because of the specification of the syllable
boundary. When a stem with a final [st] cluster takes a vowel initial suffix,
100 LONG PENG AND JANE SETTER

this final [t] is syllabified as an onset, rendering the deletion rule inapplica-
ble as in (21a). Hence, it is not necessary to refer to a consonant-initial
suffix to prevent deletion from applying to forms taking the vowel-initial
suffix. In (21b), the attachment of a consonant-initial suffix prevents the
syllabification of [t] as the onset; hence, an alveolar plosive is subject to
elision in such a configuration.
(21) a. s s s b. s s

h # n 6 s t I r e s t f # ~

n/a ∅

Apart from the deletion in (20), we also see the deletion of the alveolar
plosive when it is preceded by [n] and followed by the -ment suffix. We
state this deletion as (22).
(22) t/d → ∅ ⁄ n ___ ]s -ment
The pattern of deletion expressed by the rule in (22) reflects the internal
variability of interlanguage. Why the stem-final [t] of a [nt] cluster is elided
under the suffixation of -ment, but not under the suffixation of -s or -ful, is
not clear to us. As our central goal here is to document the systematic
properties of alveolar plosive deletion, we will leave the precise formulation
of the deletion rules as in (20) and (22).
We have so far considered only words with alveolar plosives in the
stem-final position. Just like Subject 1, Subject 2 systematically treats other
stem-final consonants differently from alveolar plosives. Some data are
presented in (23) as a comparison:
(23) [b"˜f] bath [b"˜fs] baths
[qI]k] think [qI]ks] thinks
[wfsp] wasp [wfsps] wasps
[ju˜s] use [ju˜sf#~] useful
[IkwIp] equip [IkwIpm6nt] equipment
[6men] amend [6menm6nt] amendment
[æv"˜ns] advance [æv"˜nsm6n] advancement
[6reIntw] arrange [6reIndŠm6nt] arrangement
<LINK "pen-r4">

SYSTEMATICITY IN HONG KONG ENGLISH 101

3.4. The scope of the alveolar plosive deletion

We have just provided a description of the alveolar plosive deletion in the


English of two Cantonese speakers. One natural question arising from this
discussion is to what extent this deletion is a characteristic of English of
speakers whose first language is Cantonese. While we cannot provide a
definitive answer to this question yet (as it requires detailed analyses of the
English of many more speakers than conducted here), we have some evi-
dence that this deletion is found in the English of other speakers whose
native language is Cantonese, even though details of deletion may vary from
speaker to speaker.
Eckman (1987) reports a study of consonant cluster reduction in English
spoken by six non-native speakers. Two of the subjects are native Cantonese
speakers aged 24 and 25 respectively. One speaker has spent two months in
the US at the time of research and is listed as having the English proficiency
of a beginner. The other speaker has been in the US for 15 months and is
classified as having an intermediate English proficiency. Though Eckman’s
study does not directly compare morphologically related forms, there are
some data that suggest that similar patterns of alveolar plosive deletion and
retention are found in the English of these two subjects. One subject — the
intermediate speaker — has a deletion that resembles that of Subject 1 in our
study, exemplified in (24a). The other subject — the beginner — shows the
retention of [t] that is similar to what we see in Subject 2 of our study. This
is shown in (24b).
(24) English of two Cantonese speakers in Eckman (1987)
a. [æt] act [æks] acts
[s7t] sect [s7ks] sects
[6dapt] adopt [6daps/6dats] adopts
b. [at] opt [ats] opts
[æt] act [æts] acts
[ækt] act [ækts] acts
[apt] opt [apts] opts
We do not, unfortunately, have evidence of the deletion of alveolar plosives
in Subject 2 as Eckman’s study contains only a limited amount of data with
morphologically related forms. But the pattern of the [t] retention in (24b) is
identical to that of Subject 2 of our study in (16). Recall that Subject 2 in
<LINK "pen-r8">

102 LONG PENG AND JANE SETTER

our study drops [t] only if it is preceded by [s] under the -s suffixation. He
retains [t] if it is preceded by a vowel or preceded by a consonant other than
[s]. For example, his pronunciations of acts and accepts are [æts] and
[6septs]. This pattern of [t] retention is identical to that in (24b).
In addition, three students under our supervision have looked into the
deletion of consonants in the English of Cantonese speakers in Hong Kong.
They have examined a total of nine speakers of various education back-
grounds ranging from secondary students to university students. There is
clear evidence that similar patterns of alveolar plosive deletion are present in
all nine subjects, though the details of the deletion differ from subject to
subject (as is evident from the two subjects in this study). Some of the data
in Hau (1997: 7) are presented here for comparison:
(25) [6fæt] affect [6fæs] affects
[IndIkeIt] indicate [IndIkeIs] indicates
[6sæpt] accept [6sæps] accepts
[r6~st] roast [r6~ss] roasts
Apart from the above, we have conducted some casual experiments checking
to what extent deletion of alveolar plosives is “psychologically real” with
respect to the two subjects in our study. We have tried to teach our subjects
the standard pronunciations of some of the words with alveolar plosives and
elicit their pronunciations of these words a few days later. We find that our
subjects consistently drop the alveolar plosives in spite of the teaching,
suggesting that the alveolar plosive deletion is entrenched firmly in the
phonological systems of these two speakers.
In light of the evidence, we do not believe that the alveolar plosive
deletion documented here is an isolated property of the English of the two
subjects we study here. Evidence appears to suggest that it is symptomatic of
the English of a substantial number of Cantonese speakers, though we do not
know how widespread this deletion might be in English spoken by Canton-
ese speakers, nor do we know the extent and limit of variation from speaker
to speaker. Answers to these questions will have to await detailed studies of
a larger number of speakers of various backgrounds.
To summarise the discussion up to this point, we have described the
patterns of the alveolar plosive deletion in English of two subjects. In spite
of the fact that the details of the deletion may vary from one subject to
another, there are three aspects of alveolar plosive deletion that are shared by
SYSTEMATICITY IN HONG KONG ENGLISH 103

the two subjects. First, the deletion targets alveolar plosives. Second, the
deletion happens under the condition that a consonant cluster is present in
the syllable coda. Third, the alveolar plosive [t] of a [st] cluster is elided
under suffixation in the English pronunciations of both subjects. These
shared properties of deletion suggest that deletion is not random and the
variations associated with it are limited. The two properties of deletion —
invariance and variability — seen in the interlanguage rule of alveolar
plosive deletion are shared by phonological rules in any native language
grammar. Thus, interlanguage rules are fundamentally identical with rules in
the grammars of any first language.

4. The origin and status of deletion

In this section we discuss the origin and status of alveolar plosive deletion.
We show that this deletion cannot have originated from Cantonese because
there is no evidence of cluster reduction due to the lack of consonant clusters
in the native vocabulary of Cantonese and there is nothing unique about
deletion in the resolution of consonant clusters in English loans of Canton-
ese. The comparison with alveolar plosive deletion in the colloquial speech
of the native varieties of English reveals two distinct properties, reducing the
possibility that our subjects may have acquired the rule of the target language.
Taken together, the evidence appears to suggest that the rule of alveolar
plosive deletion is part of the interlanguage grammars of the two speakers.
Cantonese syllables are maximally [CVC] or [CVV]. Consonant clusters
are not permitted in onsets or codas within a Cantonese syllable. Given the
structure of syllables, consonant clusters either in the syllable onset or coda
cannot arise from the mere concatenation of different syllables such as
[CVC.CV] as the two adjacent consonants are syllabified into different
syllables. Logically, the only way consonant clusters may emerge in either
the onset or coda is through the deletion of nuclear vowels such as [CVC.CV] →
[CVCC]. To the best of our knowledge, there is no evidence of such vowel
deletion nor consonant cluster reduction in the native vocabulary of Cantonese.
Due to the British presence in Hong Kong, Cantonese has acquired a
large number of English loan words. Examination of these loan words
suggests that there is nothing unique about deletion in the resolution of
<LINK "pen-r6">
"pen-r17">
"pen-r20">
"pen-r15">
"pen-r19">

104 LONG PENG AND JANE SETTER

consonant clusters. As shown in (26), consonant clusters of English borrow-


ings are routinely simplified by vowel insertion or by consonant deletion or
by a combination of the two strategies. Note that the numerals following the
words under the Cantonese loan column mark the tones that these words
have acquired in Cantonese.
(26) English Loans in Cantonese
[tIps] tips tip1 si6
[sI]k] sink si]1
[s#ft] soft so1 fu4
[k"˜st] cast ka1 si2
There is nothing in these loan words that resembles the rule of alveolar
plosive deletion seen in the English of the two subjects considered in this
paper (for two recent discussions of Cantonese loans, see Silverman 1992
and Yip 1993).
Could the two subjects have acquired a rule of the target language? There
is evidence that a similar rule of alveolar plosive deletion operates in present-day
colloquial speech. According to Gimson (1994: 215, 261), Roach (1991: 127),
and Yallop (1994: 60), consonant clusters may be simplified in present-day
colloquial English, specifically, in rapid, colloquial speech. This process
eliminates alveolar plosives in the middle of a tri-consonant cluster in a word
or a sentence. Some examples provided in Gimson and Cruttenden are:
exactly, mostly, facts, wristwatch, restless, handsome, landscape, friendship,
kindness, next day, last chance, first light, drift by, bend back, tinned meat,
send round, hold tight, etc. Comparison of colloquial English and English
data of our subjects reveals some clear similarities in deletion. In both
colloquial English and English of our subjects, alveolar plosives are being
targeted for elision, and deletion takes place predominantly in a tri-consonan-
tal cluster.
In spite of the similarities, there are two properties that crucially
distinguish the deletion of our subjects from that of the native varieties of
colloquial English. First, the deletion of our subjects is based on word lists,
in other words, what you might call “careful speech”, as the subjects can
monitor their speech much more closely when reading word lists than when
they conduct normal conversations. According to Gimson (1994: 215),
deletion of alveolar plosives is seen only in rapid, colloquial speech of
English; retention — rather than deletion — of alveolar plosives is the norm
<LINK "pen-r7">
"pen-r16">
"pen-r12">
"pen-r18">

SYSTEMATICITY IN HONG KONG ENGLISH 105

in careful speech. Second, in the speech of Subject 1, the deletion is seen in


a bi-consonantal cluster as well as in a tri-consonantal cluster: [b6~s] boats,
[pr6m6~s] promotes, [li˜s] leads, [ri˜s] reads. To the best of our knowledge,
deletion of alveolar plosives does not take place in similar environments in
casual or careful speech of the native varieties of English. Therefore, though
we cannot rule out the influence of colloquial English as it provides part of
the input our subjects receive, the evidence does not appear to support the
claim that our subjects acquire the rule of deletion seen in colloquial English.
We suspect that the choice and specific form of deletion evolve from
the process of acquisition. In learning a second language, learners must try
to approximate the target language as closely as possible while at the same
time reconciling the structural requirements of the target language with those
of their first language. The result is often an intermediate system. As
Selinker (1972) proposes and many subsequent studies have demonstrated,
this system may be distinct both from the learner’s first language and the
target language, something which Selinker refers to as “interlanguage”. We
think that the patterns of deletion reported here are part of the interlanguage,
belonging neither to their first language nor to the target language, even
though both may have contributed to its development.
As a matter of fact, the choice of deletion as opposed to epenthesis in
English of our subjects appears to be consistent with the general preference
of deletion over insertion in resolving coda clusters in second language
acquisition. A number of studies show that L2 learners’ treatment of conso-
nant clusters is dependent on syllable positions; L2 learners are more likely
to employ vowel epenthesis rather than consonant deletion to break up
consonant clusters in the onset, while they tend to resort to deletion rather
than epenthesis when it comes to clusters in the coda (see Major 1987;
Tarone 1980; Hancin-Bhatt and Bhatt 1996; etc.). That our subjects resort to
deletion in resolving coda clusters is consistent with this tendency towards
deletion in syllable codas.

5. Conclusion

We have described a pattern of phonological alternation involving alveolar


plosives in the English pronunciations of two Cantonese speakers. The
<LINK "pen-r2">
<DEST "pen-r11">
"pen-r4">
"pen-r1">

106 LONG PENG AND JANE SETTER

alternation is motivated by the surface variations seen between forms taking


no suffix or a vowel-initial suffix and related forms with consonant-initial
suffixes. The variations are shown to be systematic and accounted for by
rules of deletion that target alveolar plosives in the syllable coda.
The pattern documented here is important in that it provides evidence
for the existence of consonant cluster reduction in the English of Hong Kong
speakers. The evidence is unique in that it does not take the target language
as the norm of reference. Rather, it relies on the description of the variations
seen in different morphological paradigms in the English interlanguage of
Hong Kong speakers. It offers the first piece of evidence confirming the
systematic character of consonant cluster reduction observed in Luke and
Richards (1982: 59), Eckman (1987), and Bolton and Kwok (1990: 153). The
pattern reported here is also significant in that it demonstrates that morpho-
logically-related forms may alternate predictably in interlanguage, just like
native varieties of human languages where phonological alternations of this
kind are common.
This paper identifies one area of English that is potentially difficult for
L1 Cantonese speakers, an area that classroom instruction can focus on. The
reduction of alveolar plosives in words such as [æs] acts and [li˜s] leads can
lead seriously to problems of comprehension as they are rendered completely
identical to words such as ass and lease. If any classroom hour is devoted to
English pronunciation, it should focus on those aspects of pronunciation that
hinders communication. The data reported here identifies one such area.
To conclude, the study of the internal structures of interlanguage may
reveal the systematic properties of that interlanguage. We hope that our study
will stimulate more interest in the systematic study of the internal structures
of interlanguage so that we can come to a better understanding of the nature
and properties of interlanguage grammars and their relations to first language
grammars.

References

Bayley, Robert. 1996. “Competing constraints on variation in the speech of


adult Chinese learners of English”. In Robert Bayley and Dennis R.
Preston, eds. Second Language Acquisition and Linguistic Variation.
Amsterdam, Philadelphia: Benjamins, 97–120.
<DEST "pen-r12">
"pen-r2">
"pen-r3">
"pen-r4">
"pen-r5">
"pen-r6">
"pen-r7">
"pen-r8">
"pen-r9">
"pen-r10">
"pen-r11">

SYSTEMATICITY IN HONG KONG ENGLISH 107

Bolton, Kingsley and Helen Kwok. 1990. “The dynamics of the Hong Kong
accent: Social identity and sociolinguistic description”. Journal of Asian
Pacific Communication 1:147–72.
Corder, S. Pit. 1971. “Language continua and the interlanguage hypothesis”.
In S. Pit Corder and Eddy Roulet, eds. The Notions of Simplification,
Interlanguages, and Pidgins, and their Relation to Second Language
Pedagogy. Geneva: Librairie Droz, 11–7.
Eckman, Fred R. 1981a. “On predicting phonological difficulty in second
language acquisition”. Studies in Second Language Acquisition 4:18–30.
———. 1981b. “On the naturalness of interlanguage phonological rules”.
Language Learning 31:195–216.
———. 1987. “The reduction of word-final consonant clusters in inter-
language”. In Allan James and Jonathan Leather, eds. Sound Patterns in
Second Language Acquisition. Dordrecht, Holland: Foris Publications,
143–62.
———. 1991. “The structural conformity hypothesis and the acquisition of
consonant clusters in the interlanguage of ESL learners”. Studies in Second
Language Acquisition 13:23–41.
Edge, Beverly. 1991. “The production of word-final voiced obstruents in
English by L1 speakers of Japanese and Cantonese”. Studies in Second
Language Acquisition 13:377–93.
Gimson, A. C. 1994. Gimson’s Pronunciation of English. Revised by Alan
Cruttenden. London: Edward Arnold.
Hancin-Bhatt, Barbara and Rakesh M. Bhatt. 1996. “Optimal L2 syllables”.
Studies in Second Language Acquisition 19:331–78.
Hau, Tak Ming. 1997. “Self English pronunciation profile”. BA thesis, Hong
Kong Polytechnic University.
Labov, William. 1989. “The child as linguistic historian”. Language Varia-
tion and Change 1:85–98.
Li, David C. S. 1999. “The functions and status of English in Hong Kong: A
post-1997 update”. English World-Wide 20:67–110.
Luke, Kang Kwong and Jack C. Richards. 1982. “English in Hong Kong:
Function and status”. English World-Wide 3:47–64.
Major, Roy C. 1987. “A model for interlanguage phonology”. In Georgette
Ioup and Steven H. Weinberger, eds. Interlanguage Phonology: The
<DEST "pen-r20">
"pen-r13">
"pen-r14">
"pen-r15">
"pen-r16">
"pen-r17">
"pen-r18">
"pen-r19">

</TARGET "pen">

108 LONG PENG AND JANE SETTER

Acquisition of Second Language Sound Systems. New York: Newbury


House, 101–24.
———. 1994. “Chronological and stylistic aspects of second language
acquisition of consonant clusters”. Language Learning 44:655–80.
Mohanan, Karuvannur P. 1992. “Describing the phonology of non-native
varieties of a language”. World Englishes 11:111–28.
Platt, John T. 1982. “English in Singapore, Malaysia, and Hong Kong”. In
Richard W. Bailey and Manfred Görlach, eds. English as a World Lan-
guage. Ann Arbor: University of Michigan Press, 384–414.
Roach, Peter. 1991. English Phonetics and Phonology: A Practical Course.
Cambridge: Cambridge University Press.
Selinker, Larry. 1972. “Interlanguage”. International Review of Applied
Linguistics 10:209–31.
Silverman, Daniel. 1992. “Multiple scansions in loanword phonology:
Evidence from Cantonese”. Phonology 9:289–328.
Tarone, Elaine. 1980. “Some influences on the syllable structure of inter-
language”. International Review of Applied Linguistics 18:139–52.
Yallop, Collins. 1994. English Phonology. Australia: Fast Books.
Yip, Moira. 1993. “Cantonese loanword phonology and optimality theory”.
Journal of East Asian Linguistics 2:261–91.

Long Peng Jane Setter


Linguistics Program English Department
State University of New York, Oswego Hong Kong Polytechnic University
Oswego, NY 13126 Hung Hom, Kowloon
USA Hong Kong
e-mail: faspeng@earthlink.net e-mail: egjanes@polyu.edu.hk

You might also like