Errors in pronunciation of consonants by learners

of English as a foreign language whose first

languages are Indonesian, Gayo and Acehnese
I N P R i n MATHFW

^'^" paper reports on a study investigating the errors in pronunciation of

consonants made by three groups of EFL learners, native speakers of
Indonesian, Gayo and Acehnese, and possible explar\ations of these errors.
There is a limited amount of previous research on pronunciation errors
made by Indonesian EFL /earners, and none for Gayo and Acehnese
learners. In this study consonantal phoneme pronunciation error data was
collected using four different types of tasks, and analyzed using broad IPA
phonetic transcription.
The findings indicate that for these languages, segmental errors are largely
limited to final stops [p] [b] [t] [d] [k] [g], final sibilants [s] [z] [f] [3],
initial and final affricates [tj] [d^] and interdentals [9] [6]. The most
common error realizations were elision ([o]), devoicing (e.g. [b] realized as
[p]) and lack of release (["]). The error realization data did not completely
match that of previous research for Indonesian learners. In explaining the
data, both transfer and developmental and other factors and their
interaction appear to be at work, as well as factors such as task-type and
level of spontaneity. The study points to many areas of further research
including factors influencing interlanguage phonology such as knowledge of
more than one language, and more complete error analyses using narrow
phonetic transcription.

1. Introduction
The on-going emphasis on communication in ESL/EFL teaching has put a spotlight on the issue of pronunciation. Whilst
native-like pronunciation of a second language may not be possible for the majority of learners, it is essential for language
teachers to try to help EFL learners produce an adequate range of differentiable English phonemes, in order to avoid
misunderstandings and irritation (Dutt 1990:21; Harmer 2001:183-4; NSW AMES 2003:28).
The two main regional languages spoken in the province of Aceh, Indonesia, are Acehnese and Gayo. Acehnese is
spoken mostly in the coastal areas of Aceh, by Acehnese and also by other groups who have moved into the area, such as
Jamee, or who have been dominated by the Acehnese, such as Simeleue. Gayo is spoken by the Gayo of the inland
mountains of Aceh. These are unwritten languages used in daily life alongside the relatively new, national language,
Indonesian, which grew out of trade Malay which had been used in the archipelago for centuries. Indonesian Malay was
chosen as the national language for the yet-to-emerge Indonesia by nationalists in 1928 as part of the Youth Pledge (Balai
Pustaka, p.xii). Indonesian is now the first language of a considerable percentage of Indonesians. All three languages
belong to the Western branch of Austronesian. While teaching English as a Foreign Language (EFL) in Aceh, the author
found that EFL teachers have hunches about English segmental pronunciation errors being made, but that there has been
little systematic research on such errors. Although there have been contrastive and error analyses of various aspects of
syntax and morphology, the acquisition of English phonetics and phonology (sounds and sound patterns) has not been
thoroughly studied in relation to speakers of the many local languages of Indonesia. An additional lacking is that of
pronunciation teaching materials aimed at the Indonesian learner of English, based on error or contrastive analysis. The
aim of this study was to research pronunciation errors as a necessary first step towards generating suitable teaching
materials aimed at the learners of English from these three first language groups.
Celce-Murcia, Brinton and Goodwin (1996), in summarizing the state of the debate regarding the priority of focussing on
the prosodic or the phonemic level, suggest that a pronunciation curriculum ought to address both the segmental and

VOLUME THREE, NUMBER TWO

29

Errors in pronunciation by learners of English as a foreign language whose first lan^Mges are Indonesian, Gayo and Acehnese

suprasegmental features that need to be mastered by a specific group of learners (1996:10). Both levels need to be
researched, but students' errors in pronunciation of consonants first came to the attention of the author as a major
impediment to comprehension, and therefore this problem was chosen as the starting-point for research.
2. Background
The initial step in overcoming pronunciation difficulties is the determination of the errors experienced. The relevance of
a formal study into pronunciation has been highlighted by AGPS (1983), as this small publication seems to be limited to a
prediction of errors based on contrastive analysis of English and Indonesian. Likewise the problems Yong (2001) finds that
Indonesian speakers have in pronouncing English (2001:281). Teaching experience and informal observation confirmed
some areas of difficulty mentioned in these publications, but has also revealed some different ones.
Some work has been done to compare the English and Indonesian phonological systems, notably in Indonesian language
textbooks such as the one written by Macdonald and Soenjono (1967), and in resources for EFL/ESL teachers such as
Learner English (Swan and Smith 2001). However there is a marked absence of research looking at the phonological
system of English alongside those of the regional languages of Indonesia. The author predicted that pronunciation errors
for these groups would be different because the three languages have some differences in their phonological systems. The
relevant phonological information has been summarizedfiromBaihaqi (1977) and Soravia (1984) for Gayo, Durie (1985)
and Wamad Abdullah (1973) for Acehnese, and Macdonald and Soenjono (1967) for Indonesian in the Appendices.
There are different terminologies for and ways of classifying errors, but a simple and useful distinction is that between
transfer and developmental errors. Developmental factors here refer to processes found in first language acquisition, such
as final consonant devoicing (e.g. [b] replaced by [p]), overgeneralization (where one target language sound is substituted
for another), and approximation (Odlin 1989:123; Tarone 1978:19-20). Errors not accounted for by contrastive analysis
can sometimes be explained by looking at the errors made by children acquiring their first language and the processes
they use.
Transfer of linguistic items, including word order, certain vocabulary and sounds, from first language to second language is
another process involved in second language acquisition. Selinker, quoted in Ellis (1994) reports that transfer may also be
"responsible for fossilization" (Ellis 1994:309), meaning that certain L2 (second language) sounds are consistently
replaced by a phonetically close LI (first language) sound; the correct L2 sound is not incorporated into the
interlanguage. Transfer is an important source of errors in second language, but not the only one, although it is more
obvious at the level of the sound system than at other levels of language. For example some phonological errors that were
assumed to be a result of LI interference in later research in child language acquisition and second language acquisition
were shown to be developmental errors. Dulay and Burt's (1974) study of Spanish children acquiring English morphemes
is often quoted in the literature (Ellis 1994:60-1, 308-9,339). They found that a large proportion of errors in acquiring
English could be explained as developmental errors. It has also been found that the acquisition sequence of English
morphemes of groups of children and of adults from different first language backgrounds are very similar, thus showing
the role of developmental processes. However first and second language acquisition sequences differ, with adult second
language acquisition showing transfer errors (Dulay, Burt and Krashen 1982:202, 229). "Transfer and development are
not independent processes but are rather interrelated in complex ways" (Pennington 1992:10).
In addition general processes such as elision, communicative strategies such as simplification, and universal or crosslinguistic characteristics, account for other errors (Driesher and Anderson-Hsieh 1990). Interest in universal constraints
on pronunciation has been renewed in the last few decades in the literature on language universals focussing on language
acquisition (Ellis 1994:428). Some sounds in a target language may be inherently difficult. These are found in relatively
few of the world's languages and/or are hard to articulate. Johansson, quoted in Tarone (1978:19), found a "tendency [in
substituting for target language sounds] to move from the extreme higher and lower positions in the articulation area
toward the middle height, the tongue's rest position". Odlin (1989) noted some correlation between the rarity of a sound,
such as [6] or [tj] found in English and its difficulty. This cross-linguistic evidence can be used to explain pronunciation
errors (Odlin 1989:120).

30

MONASH UNIVERSITY LINGUISTICS PAPERS 2005

Errors in prorxunciation by learners of English as a foreign bnguage whose first languages are lt\dorKsian, Gayo and Acehnese

Tarone (1978:18) reminds us that different tasks provide different interlanguage data, making task type another factor in
accounting for pronunciation errors. Corder (1973:268-9) says subjects need to be able to talk and choose their words
freely when being tested for errors, and therefore data are needed from spontaneous samples. Kenworthy (1987:21)
suggests "a spontaneous speech sample, with a bit of reading aloud, mainly to provide the teacher with information of
potential spelling interference problems". Different interlanguage phonology data for each task is to be expected. For
example, reading a list of words, reading a piece of continuous prose and repetition will all produce different results: the
nature of the task influences production (Tarone 1978:18). Labov (1972) isolated various contextual styles including a
careful speech style found in interviews, a reading style and a yet more formal style when reading words in isolation.
These styles cover a range of formal speech tasks. Yet another style is casual style which is difficult to record, though it
can be elicited for in interviews by the use of'danger-of-death' and similar questions. Labov (1972) found that questions
asking for the story of when the interviewee felt in danger, or felt he or she could have died, led to a story narrated rapidly
with great vivacity and emotional involvement. The structure of the language produced and pronunciation was different
from replies to other types of questions (Labov 1972:79-86, 92-4). 'Richer' data may be obtained by using different tasks,
and therefore this research project consisted of four tasks.
3. The research project
3.1 Sample population
Twenty-four volunteer participants were recruited from among the author's students and their friends at universities in
Banda Aceh. They comprised four women and four men from each of the three first language groups (Indonesian, Gayo
and Acehnese). All the participants had studied English as a Foreign Language for six years in high school and for
between one and four years in university either as their major area of study or as a compulsory subject. The eight
participants in each of the three groups were matched for age and major field of study.
3.2 Tasks
The study comprised four tasks selected from those commonly described in the literature. An audio-discrimination task
based on minimal pairs, a word-repetition task and a reading passage task were carried out in a language laboratory,
whilst the fourth task was an interview recorded using audio-cassette player. The interview task was the least
linguistically constrained situation and the repetition task the most constrained.
In pilot tests, errors with consonants in medial position were found in some cases to be the same as errors in initial
position, in other cases the same as errors in final position. Accordingly only initial and final positions in a word were
tested for each of the consonants, except for [3] which only occurs medially and finally.
The aural discrimination task of 158 items based on minimal pairs was recorded by the researcher. The phonemes tested
for were based on errors that appeared in pilot tests. In each item three words, two of which were the same, were heard by
the participants. To test discrimination of each minimal pair, either in word-initial or word-final position, there were three
items, using different words. Two or three 'mis-hearings' out of the three items were deemed to constitute an error. For
example, the first item was:
1 a 'aids' b 'age' c 'age'
In the language laboratory the participants were requested to circle the letter a, b or c showing whether they thought the
first, second or third word was the different one. The answer sheets were collected and incorrect aural discrimination
choices noted, for both phoneme and position. A wrong choice reveals a failure to perceive a contrast between two
phonemes, but does not show which of the phonemes is perceived correctly and which one is undifferentiated. It is
possible that both the phonemes are perceived as yet another phoneme, for example as in the above example, as 'aid' or
even 'ate'. Such information was beyond the scope of this particular test. The aural discrimination test is included in
Appendix B.

VOLUME THREE, NUMBER TWO

31

Errors in pronunciation by learners of English as a foreign language whose first languages are Indonesian, Gayo and Acehnese

In the pronunciation repetition task the consonantal phoneme being tested for was presented in three different words.
The 78 task items were randomized and read by the researcher onto a master audio-cassette. In the language laboratory,
the participants heard a word and were requested to repeat it in the following silence. These repetitions were recorded.
For example, the first four items were:
1 'gross' (final [s])
2 'card' (final [d])
3 'bag' (final [g])
4 'chip' (initial [tj])
The words chosen contained the minimal pair substitutions that had appeared in the pilot tests. Three items were
provided to test each phoneme in word-initial or word-final position. Another native speaker of English was enlisted as
evaluator for this part of the research, as it is too easy to anticipate the errors one is testing for. An error was said to have
been made when either two or three productions of each phoneme was not recognized by the evaluator as the phoneme
being tested for. For each item a list of randomized words including the correct answer were provided for the evaluator to
compare with the participants' responses. For every item the evaluator either circled the word closest to each response, or
noted down where the response was something different. The responses to this second task were also played back by the
author, in particular to determine the sounds being substituted for the above-mentioned different answers. Appendix C is
the pronunciation repetition test.
For the third task, each participant recorded a reading passage of 113 words onto an audio-cassette. The errors revealed
in informal observations and pilot tests were written into this task. The first sentence was:
'Which would you choose - a mug of coffee or a cup of tea?'
The audio-cassettes were played back and each individual error noted down. This task provided information on spelling
interference, which is the influence of spelling on pronunciation, and also on the influence of environment, or which
vowels, consonants or pauses precede and follow the phoneme being tested for. The reading passage test can be found in
Appendix D.
The fourth task was audio-recorded interviews, the main purpose of which was to collect more spontaneous data than
was possiblefi"omthe other, more restricted, tasks. For example the last question, following Labov, was:
16.

Tell me about something very frightening that's happer^d to you.

Each participant had an interview with the author, which was recorded. The audio-cassettes were reviewed, the
interviews in entirety were transcribed and correct and incorrect realizations of the tested phonemes were noted. In this
task an error was regarded as having occurred when the realization of a consonantal phoneme produced by a participant
was not recognized by the author as the phoneme that was supposed to have been produced. The complete interview
schedule is included as Appendix E.
3.3 Interpretation issues
Some limitations in this study need to be mentioned. The three first language group samples were not as large as one
would want, thus reducing the possibility of generalizing the results to the populations from which they came. By using a
language laboratory and audio-recording equipment, information about realizations of the tested phonemes has been
limited to the aural. Video-recording would have given useful visual information about how and what phonemes were
realized.
Further research would need to include more groups, at least, Indonesian first-language speakers, Acehnese first-language
with Indonesian second-language, Gayo first-language with Indonesian second-language, Indonesian first-language with
Acehnese second-language and Indonesian first-language with Gayo second-language. This is because the Acehnese and

32

MONASH UNIVERSITY LINGUISTICS PAPERS 2005

Errors in pronunciation by learners of English as a foreign language whose first languages are Indonesian, Gayo and Acehnese

Gayo participants' data may include interference from Indonesian which is their second language. Many of the
Indonesian first language speakers had either Acehnese or Gayo as their second language, but this has only been
glancingly taken into account in attributing causes to the errors found. In addition it is likely that Qur'anic Arabic, which
is studied by all school-age children in Aceh, has some influence, but this cannot be demonstrated or tested in this
sample.
The master recordings for tasks one and two were made by the researcher, a native speaker of Australian English, and
therefore were biased against some of the participants who said they were more used to an American or local English
speaking model. This was one reason for studying consonant rather than vowel errors. A thorough attempt was not made
to account for 'extensive individual variation' (Hecht & Mulford 1982:325'326) found within the different tasks for each
participant. The study did not look at medial sounds, or at consonant clusters, and vowel environment was not built into
the tasks.
4. Findings and discussion
4.1 By phonological type
The tables on the following pages summarize the most common error realizations of each English phoneme displayed by
each first language group. Each table shows the errors made by 3 or more out of the 8 participants in each group. A blank
cell means errors were made by 2 or fewer participants. As no significant number of errors were found in voiced stops in
initial position they are not included in the following table.
Table I: Informants' realizations of voiced stops
Data source' /
Phonemic feature

Phoneme

Indonesian
,.
aural rep. read interv.

Final voiced stops

b
d

p
t

k\g^

Gayo
aural rep. read interv.

Acehnese
aural rep. read interv.

dAo\t

o
k^

t
k

o\d^ 0
k

From Table 1 we can see that many errors were in fact similar across the three first language groups. Final voiced stops
were mostly devoiced, but lack of release ([^]) and elision ([o]) were also common. In looking at the reasons for these
errors it was found that final stop devoicing (e.g. [p], [t], [k] substituted for [b], [d], [g] can be a developmental process
found in first language acquisition (Hecht and Mulford 1982:324). In terms of first language interference, voiced stops
are not found word final in these three languages, except [b] in Gayo, which was, however, sometimes realised as [v] by
the Gayo group. A simplification or reduction strategy, of course, would also account for the elision. Elision of final stops
is found in child acquisition of English (Cruttenden 1994:144). No significant errors were found in voiceless stops in
initial position so they are absent from Table 2. In English the major distinction between final voiced and unvoiced stops
is found in the length of the preceding vowel and therefore research on vowel errors may throw further light onto this
particular error.
Table 2: Informants' realizations of voiceless stops
Data source/
Phonemic feature

Phoneme

Final voiceless
stops

p
t
k

VOLUME THREE, NUMBER TWO

Indonesian
aural rep. read interv.
p
f
k^

Gayo
aural rep. read interv.
t

t^
kA?

P
0

k^

t'
k'

Acehnese
aural rep. read interv.
p
t^
k^

t^
\C

33

Errors m pronunciation iry leanxers of English as a foreign language whose first languages are Indonesian, Gayo and Acehnese

Final voiceless stops were often unreleased on the reading and interview tasks, even in connected speech, where in
English they would usually be joined onto the following vowel. Twenty participants mispronounced final [k], mostly as
[k^], on the reading task, and nine did the same on the interview task. Such lack of release of final voiceless stops is a
common feature of these Austronesian languages. For example p, t, k are found word-final in Indonesian, but not
released as in [hirup''] [bubut^] [titik'']. We need to remember that in English lack of release of final stops is not
uncommon before a pause (Cruttenden 1994:145). However the consistent unreleased final voiceless stops of Indonesian
English, especially in connected speech, leads to much misinterpretation by native speakers of English.
Table 3; Informants' realizations of sibilants
Data source /
Phonemic feature
- Sibilants

Phoneme

initial
Medial
final

Indonesian
aural rep. read interv.

Gayo
aural rep. read interv.

Acehnese
aural rep. read interv.

d3
z
o
s

3
z

z\I
o
s

Some of the errors in realizations of the sibilants as shown in Table 3 can be explained by over-generalization, [z] and [J]
being often generalized to the more familiar [s]. [J] is available in Acehnese, but is only used for borrowed words in
Indonesian and Gayo (see Appendix A), [z] is only found in borrowed words in these three languages and was mostly
either elided or replaced by [s]. [3] is not found at all in the three first languages under study here, and was mostly
replaced by their borrowed [z].
Table 4: Informants' realizations of affricates
Data source /
Phonemic feature
- Affricates

Phoneme

initial
final

tj
tj
d3

Indonesian
aural rep. read interv.

Gayo
aural rep. read interv.

k
d3

t
tJ

k
ts

t\o\tI

Acehnese
aural rep. read interv.

t
tJ

tJ

d3
tJ

t
t

tJ

As articulatorily difficult and unfamiliar phonemes, the affricates mostly revealed the expected reduction to a stop
phoneme. Contrastive analysis, as presented in the AGPS (1983), of the phonemes of these three languages and English
would suggest that the affricates will be error-prone, as they are not found in Gayo, Acehnese or Indonesian. The nearest
sounds phonetically are the voiced and voiceless alveopalatal stops (see appendix). Developmental processes would also
explain the substitution of universally common stops for the affricates, as children produce stops before they produce
affricates (Odlin 1989:123). However no explanation has yet been found for the substitution of the voiced and voiceless
affricate for each other.
The interdentals were generally substituted by alveolar stops or sibilants. Interdental sounds are relatively rare in the
world's languages; they are not found in these Austronesian languages, except Acehnese which has a "lamino alveodental fricative with a wide channel area" which sounds to an English speaker like a sound between an [s] and a [G]
(Durie 1985:12-13). Some of the Indonesian participants have Acehnese as their second language. I think this is the
explanation for substitution of interdentals for sibilants. It is therefore not surprising that these participants had
difficulties with them. The strategy of overgeneralization would explain why the high-frequency and articulatorily close
alveolar stops and sibilant fricatives were substituted for the interdentals. There were in fact many more different
realizations of [0] and [6] compared to the realizations of the stops. This data seems to point to a large variety of

34

MONASH UNIVERSITY LINGUISTICS PAPERS 2005

Errors in pronunciation by learners of English as a foreign language whose first languages are Iruionesian, Gayo and Acehnese

substitutions being made where the phoneme is unavailable in the first language, compared to more consistent
substitutions where the phoneme is more familiar.
Table 5: Informants' realizaticms of interdentals
Data source/
Phonemic feature
- Interdentals

Phoneme

initial

9
d

final

e
d

Indonesian
aural rep. read interv.

Gayo
aural rep. read interv.
t
d
t

Acehnese
aural rep. read interv.

t
d
t\d\0

4.2 Errors by task type

Spelling interference is seen in some of the differences between the error data from the reading passage task and the
other three tasks. Spelling interference comes from at least two places: various English sounds are spelled with the same
letters; a first language sound is spelled with a letter that stands for a different sound in English. Thirteen out of the
twenty-four participants substituted [k] for the initial [tj] in realizing 'chocolate'. Digraphs are used much more widely in
English than in Indonesian and seemed especially difficult for the participants to pronounce. The role of < h > in <ch>
and <sh> and < t h > in indicating a different phoneme may not have been apparent to the participants. For example in
'seethe' the interdental was realized as [t] by some participants. 'Squash' was realized as [skwDs], which shows that the
'sh' spelling for [f] has probably not registered. The author has noticed people in shops in Aceh saying [skwos] for
orange squash. In the other tasks [s] was also often substituted for [J], and would be explained as a developmental factor,
where the universal [s] is substituted for other sibilants. Another obvious case of spelling interference was the common
production of some other sibilant in the place of [3] in 'leisure', where in other words's' stands for [z], [s] or [J]. The [z]
in 'zany' was occasionally replaced with [d3]. [z] in Acehnese is often pronounced [(J] (alveopalatal fricatized stop)
which is phonetically close to [d^]. 'Grease', as a surname, was pronounced with a final [s] or [z]; both pronunciations
were accepted. Final [z] in words denoting plural were mostly elided. Spelling would clearly point to the production of an
alveolar sibilant. However the three Austronesian languages in question do not use a grammatical marker for plurals, and
neither do they have word final consonant clusters. Both a reduction strategy and developmental factors would account
for simplification of the plural form to the singular.
The interview provided information on environmental, lexical and grammatical factors which may have influenced the
pronunciation of certain phonemes. Some participants who elided final [d] in fact consistently failed to mark the past
tense endings on regular verbs. There may be a number of factors at work here: absence of final [d] in first language,
absence of tense markers in first language, delayed acquisition of past tense morpheme as well as developmental factors.
In fact, tense suffixes often occur in consonant clusters, which did not form part of this study. Other participants only
elided or did not release final [d] in very common words, such as 'did' and 'and'. They were perhaps more careful with
less familiar words. One participant did not release the final [d] of words he repeated after the interviewer. Yet another
participant did release final [d] if the following word began with [b]. The words 'English' and 'language' were overrepresented in the interview data and account for almost all of the final [f] and [ds] errors in the interview columns in
Table I. 'English' was sometimes realized with a final [s] which is similar to its Indonesianized form'Inggeris'. The final
sound of 'language' produced a great variety of realizations and simplification strategies and spelling interference both
seem to account for most of them. The most common, as shown in the table, was devoicing. Initial [6] realized as [d] was
self-corrected by some participants. No difference was noticed in pronunciation errors between answers to Labov's
'danger of death' question and the other interview questions.
The aural discrimination task showed the least incidence of error. It seems that although participants heard the
differences between certain sounds they could not necessarily produce such a difference. Both the aural discrimination

VOLUME THREE, NUMBER TWO

35

Errors in pronunciation by learners of English as a foreign language whose first languages are Indonesian, Gayo and Acehnese

and the repetition tasks limited the error data to phonemes found in English, as the task items were English words. Task
three and in particular task four have error data which is influenced by the relative frequency of sounds in English. In
these two tasks first language interference was more noticeable as the research design allowed for transcription including
non-English phones. Therefore lack of release of final stops and use of the glottal stop showed up in the datafi^omthese
latter two tests, but would not have been picked up on the repetition task, or would have been analyzed as elision.
4.3 Errors by language group
If we examine the errors made by each of the first language groups separately, the preponderance of errors in realizations
of consonant phonemes in final position for the Indonesian group is not surprising. Of the phonemes in this study only
[pl. [t], [k], [f] and [s] can appear word-final in Indonesian (see Appendix A). It is suggested in AGPS (1983) that
errors occur with voiced stops because "no voiced stops occur finally in BI [Bahasa Indonesia/Indonesian language]"
(1983:6). Indonesian stops in final position are unreleased and voiceless (AGPS 1983:6) and this also shows up in the
data in this study. It may also be noted that [3], [9], [6], [tj] and [d^] are not found in Indonesian at all, and there were
significant errors found in at least one task for each of these sounds. In Indonesian [f] is a borrowed sound but now
interchangeable with [p], likewise the borrowed [v] is frequently interchanged with [f] and [p]. This information
explains some of the less common errors made in producing the labial sounds (not shown in any of the tables). Borrowed
[J] is interchangeable with [s] in Indonesian, and this explains the error where [s] is substituted for [J].
Gayo has [f] only as a borrowed sound and no [v], yet [f] and [v] were often substituted for [p] and [b] on the repetition
task. Hypercorrection may be one explanation for this. In addition it may be that fricatized allophones of the phonemes
[p] and [b] cause no problem as there are no labial fricatives in Gayo, and this has been carried over into English
(Jakobson 1941:51-52). [z] and [J] are available in Gayo only in words borrowed from Qur'anic Arabic, and showed
errors on the reading task in particular.
Like Indonesian, Gayo has no [9], [6], [3],' [tj] and [d3], and the data show errors with these sounds. With regard to
dropping final phonemes, Gayo has no final [d] and [g], and this may be the explanation as to why [d] was deleted or
devoiced and [g] sometimes reduced to an unreleased [k]. In Gayo final [k] can be realized as glottal stop, and in a
similar fashion to Indonesian final voiceless stops are not released, which would seem to account for the error data of
these sounds.
Unlike Indonesian and Gayo, Acehnese does have [J] in its phonemic inventory, hence this was the most common
substitution for [3], and on the reading passage task also for initial [tj]. The Acehnese participants more often realized
[9] as [s] than the other two groups. This has probably occurred because some dialects of Acehnese have a sound close
to [9] as an allophone of [s]. Acehnese has no [6] or affricates and there were significant errors in each of these sounds.
Acehnese has a [z] found initially and medially in words borrowed from Arabic. In the error data final [z] was a plural
marker and elided in the reading passage, but sometimes devoiced on the repetition task. One explanation is that
devoicing of final sibilants is also a developmental process (Hecht and Mulford 1982:322). Realization of the final stops
and [dsl can be accounted for in a similar manner to the error data of the Indonesian first language speakers.
4.4 Limitations
It is worth mentioning that some errors which would have been anticipated on the basis of the contrastive analysis
hypothesis did not turn up in the present study, and unanticipated errors did turn up, which highlights a possible
weakness in the contrastive analysis hypothesis. For example, Gayo speakers often realized [p] as [f] despite having [p] in
their phonemic system but only a borrowed [f]. In explaining error data the fact that most of the participants are fluent in
two of the three languages needs to be remembered, because transfer can come from any languages spoken, including
Qur'anic Arabic in the case of Aceh, or the interaction of these languages.
English teachers in Indonesia in general and Aceh in particular comment that final [s] and [z] are often not produced in
suffixes, in consonant clusters or on their own. Acehnese phonology does not allow for [s] in final position. In addition,

36

MONASH UNIVERSITY LINGUISTICS PAPERS 2005

Errors in pronunciation by learners of English as a foreign language whose first languages are Indonesian, Gayo and Acehrxese

all three of these first languages have neither a method of marking plural forms on nouns nor of marking person on verbs.
My colleagues, native-speakers of Indonesian, Gayo and Acehnese, who have studied contrastive analysis, tell me that
this is why I notice a lack of final [s] and [z] in their speech. In pilot tests using only non-English department students
the non-production of final [s] and [z] appeared as a notable characteristic of Acehnese English in particular. However
non-production of final [s] did not show up in this data analysis. Possible explanations for this absence are: the test tasks
didn't contain enough or a wide enough range of items testing for realizations of final [s]; the participants have been
taught to pronounce final [s]; all the participants have Indonesian as either first or second language and Indonesian does
have a final [s].
5. Conclusion
This study revealed patterns of errors in the realizations of certain English phonemes. It comes as no surprise that the
errors are found in features and environments that are not found in these three first languages (see Appendix A). When
voiced final stops ([b], [d], [g]) were realized erroneously, they were mostly devoiced. This can be accounted for as both
a transfer and a developmental process. Voiceless final stops ([p], [t], [k]) were mostly unreleased, which is the case in
the phonology of the three first languages involved. For the sibilants, final [z] elision predominated, [3] was substituted
by the alveolar sibilant [z], and [J] was largely replaced by the alveolar [s]. Interdentals were realized in a large variety of
ways. Affricates were mostly reduced to the nearest stop. In seeking explanations of these results, transfer and
developmental factors, spelling interference, general processes, communicative strategies, and the interaction of these
factors all appear to be involved.
The error data shows significant similarity across the three first language groups. This is most probably because the
languages are fairly closely related, and all the participants had Indonesian as their second if not first language. It also
points to the influence of universal developmental processes, which affect second language pronunciation as they do first
language acquisition. What is also interesting is that the differences in error data across the three first language groups
can probably be explained as due to transfer processes.
As evident in the five tables, each of the four tasks produced different error data because tasks and levels of spontaneity
in response affect pronunciation differently. Hence the reading task showed some errors from spelling interference.
Overall the least number of errors realized found on the aural task.
6. Recommendations
A teaching program for the pronunciation of consonant phonemes in English for these three first language groups could
order the items to be included based on the frequency of errors, information on which is found in this error analysis. The
number of participants who realized each target phoneme erroneously could be used as an indication of difficulty, the
more common errors being addressed first. (Complete individual error data is not included here, but is available from the
author). Allocation of pronunciation teaching time needs to take into account the frequency of errors. But in addition
the functional load and frequency of such consonants in English, as described by Catford (1987:88-9) and their
articulatory difficulty should also be considered, so that time would not be wasted on rarely-occurring sounds. In fact the
error data from the interview task was influenced by the frequency of the consonants in free speech; only sounds that are
both common in English speech and realized erroneously a third of the time showed up. Lowes (MA TESOL 1990),
seeking a foundation on which to develop teaching materials, suggests an EFL pronunciation teaching order for Castillian
Spanish speakers based on the frequency, functional load, difficulty, error gravity and other factors of the phonemes of
English. Lowes has taken into account both the role of first language and the relative importance of the different
phonemes of English (1990:53-5). Similar work could be done for Indonesian learners of English.
The findings from the different tasks show that even when sounds can be differentiated by ear they are not necessarily
differentiated in speaking, and vice-versa. Teaching pronunciation should include aural and oral practice, expecting
increased perception of differences between the phonemes of English to go hand in hand with production of adequately
different realizations of them. The repetition test could provide data on decontextualized words. There are sounds that

VOLUME THREE, NUMBER TWO

37

Errors in pronunciation by learners of English as a foreign language whose first languages are Indonesian, Gayo arul Acehnese

are adequately realized in the contextualized speech of the reading and interview tests but not clearly differentiated from
other phonemes when said alone. Teachers need to be aware that learners will probably have adequate pronunciation for
functioning in conversation before their English is error-free for tasks such as reading out lists. This is because a feature of
conversation is redundancy such as repetitions, time markers in addition to past tense, and numbers in addition to
plurals, that provide clues to meaning even when pronunciation is poor. A needs analysis for the learners would show
their purpose in learning English and how accurately they need to speak it. Teachers could use the error data from the
task that is most relevant to their students to plan pronunciation teaching.
Information on the Indonesian, Gayo and Acehnese sounds systems is to help the teacher further discover what the
English learners are or are not doing and help them widen their repertoire of sounds to include the ones needed to be
intelligible in English.
7. Further research
This research is a first step towards describing and accounting for the pronunciation errors of consonant phonemes of a
few first language communities found in Indonesia. Errors in realizing English vowels, intonation and stress are all
important issues which need further research to fill out the Indonesian EFL picture. In terms of depth, studies of EFL of
Indonesians of various first languages using narrow phonetic transcription would shed more light on pronunciation errors,
and provide information that could be used for teaching strategies. For example Acehnese has an allophone of [s]
phonetically between [s] and [6]. Narrow rather than broad phonetic transcription would show if it is being used by
Acehnese speakers for either English [s] and/or [0]. In addition in the author's experience this sound can usefully be used
as a starting-point to help teach English [6].
In order to improve EFL pronunciation teaching in Indonesia, error analyses such as the one carried out in this present
study need to be followed up by research into a suggested teaching order for pronunciation.
Interlanguage phonology in multilingual communities is complex. Error analysis needs to consider not only the various
factors such as transfer, development and use of strategies but also the interaction of the various languages that the
learners bring to the task of learning the target language. Research mentioned earlier by Pennington (1992) and also
Hecht and Mulford (1982) into attempting to isolate these factors and analyse their effects and interaction has not
proved conclusive yet. For example elision is variously claimed as a development factor, transfer, or a communicative
strategy of simplification, amongst others. Dreasher and Anderson-Hsieh (1990) remind us that this large field of
research therefore needs much more attention. It is hoped that this study could also be a starting point for research into
the neglected area of the phonology of the languages of Indonesia.
Notes
1. The data sources refer to the four sources of spoken text provided by informants. Aural' refers to the audio
discrimination tasks based on 158 items in minimal pairs. 'Rep.' refers to the pronunciation repetition task consisting of
consonantal phonemes in 78 task items. 'Read' refers to recording passage. 'Interv.' refers to the audio-recorded interview
conducted with each informant, consisting of a range of questions including a Labovian narrative question based on
informants' experience of a frightening event.

38

MONASH UNIVERSITY LINGUISTICS PAPERS 2005

Errors in pronunciation by learners of English as a foreign language whose first languages are Indonesian, Gayo and Acehnese

Appendix AConsonant phoneme inventory

bilabial labio-dental
Stop

English
Indonesian

P b
P b

t
t

d
d

Acehnese

P b
P b

[z]
[z]
[z]

Gayo
Fricative English
Indonesian
Acehnese
Gayo
Affricate

English

Nasal

English
Indonesian
Acehnese
Gayo

Lateral

[f]
[f]
[fl

[V]

e a

Indonesian
Acehnese
Gayo

m
m
m
m rr1

n
n
n
n n

glottal

k g

t 4
I 4
I 4
J 3

k g

k g
k g

h
h
h
h

[X]
[X]
[X]

m
d3

0
Ji
Ji
Ji

n
J}

1
1
I
1
w

w
w
[w]

- no
final
w, j ,
- no final b, d,
g, J, (|, ji,
-nofinal b, d, k, g, J, (|, ji, w,
- no final d,
g, J, 4> J^> w, j ,

J
J
J
j

English
Indonesian
Acehnese
Gayo

English
Indonesian
Acehnese
Gayo

pos [-alveolar palatal velar

tj

English
Indonesian
Acehnese
Gayo

Continuant English

TriU/flap

dental alveolar

r
r
r
J,
h
[v], [x],
[z], [J]
[x], 1, r, , [z], j"
[x],
[z], [f], weak nasals

Key
n = weak nasal (not as strongly nasal as an ordinary one)
J = laminal and fiicatized t (IPA = laminal)
[ ] = borrowedfiromArabic or Dutch
Allophones
Indonesian

mostly found in borrowed sounds

/f/: [f] interchangeable with [p]
/v/: [v] interchangeable with [f] and [p]
/z/: [z] interchangeable with [4] and [d]

VOLUME THREE, NUMBER TWO

39

Errors in pronunciation by learners of English as a foreign language whose first languages are Indonesian, Gayo and Acehnese

1^1'- Lf] interchangeable with [s]

/w/: [w] interchangeable with [w] and [P]
Gayo

/k/: [k], syllable final [?] no complete allophonic information

Acehnese

All stops, M and /r/ have aspirated variants

The nasals have a set of incomplete variants in some dialects
/h/ has a voiced variant
In some dialects: Ixl: [r] [r] [K] and /s/: [s] [] [9]
(Baihaqi 1977; Durie 1985; Macdonald and Soenjono 1967; Soravia 1984; Wamad Abdullah 1973)

Appendix B^Test 1
Recorded by researcher on a master audio-cassette
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36

40

a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a

aids
age
pour
zip
sot
lip
grow
nigh
thigh
suit
measures
grief
march
loath
mass
Ruth
verve
age
roof
thick
thawed
leave
rids
though
kin
Jock
perch
fan
vole
graze
shoe
Asia
life
chin
lop
poo

b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b

age
ate
porch
sip
sot
live
growth
night
thy
soothe
measures
grief
march
low
mass
roof
verb
aid
rue
sick
ford
leaf
ridge
though
chin
jock
purr
pan
vole
grace
shoe
Ada
lie
shin
lop
poo

c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c

age
age
pour
sip
shot
lip
growth
nigh
thy
soothe
meshes
grieve
mark
loath
mash
Ruth
verb
aid
rue
sick
thawed
leaf
rids
dough
kin
chock
purr
pan
foal
grace
chew
Ada
life
shin
lock
pooch

37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72

a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a

robe
goad
bilge
feel
ridge
fie
char
Joyce
writ
chill
bow
chew
sip
bays
V

aitch
rave
tick
vow
bile
vole
ache
zap
car
dog
thin
rill
harp
heap
lib
bats
badge
door
mouth
chew
math

b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b

robe
goat
builds
veal
rich
thigh
tar
joys
rich
chill
bow
Jew
ship
bathe
V

age
rape
tick
bow
pile
bole
aitch
sap
char
dock
fin
rill
half
heap
live
batch
bad
thaw
mouse
too
math

c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c

rove
goat
bilge
veal
rich

fie
tar
joys
writ
kill
both
chew
sip
bathe
bee
aitch
rave
thick
bow
bile
vole
ache
zap
char
dock
fin
chill
harp
heave
live
bats
bad
door
mouth
chew
massed

MONASH UNIVERSITY LINGUISTICS PAPERS 2005

Errors in pronunciation by learners of English as a foreign language whose first languages are Indonesian, Gayo and Acehnese

73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115

a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a

beg
lass
viper
had
leaf
rack
door
seize
pea
hock
leisure
thy
high
goes
push
know
vile
lop
pat
thigh
shock
tight
miff
Joyce
boo
cat
lab
hark
thigh
tab
late
laid
lobe
chore
pain
pate
thin
baa
thin
myth
row
myth
perk

b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b

peg
lash
viper
had
leap
rag
jaw
cease
peak
hop
ledger
die
high
goes
puss
knows
file
lob
pat
tie
shock
tight
myth
choice
boob
catch
lab
harp
die
tap
lay
lathe
low
shore
vain
fate
thin
bag
sin
mist
row
miss
perk

c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c

peg
lass
piper
hat
leap
rack
door
seize
pea
hock
leisure
die
hide
go
puss
knows
vile
lob
fat
tie
sock
tie
myth
choice
boo ,
catch
lap
harp
die
tap
lay
laid
lobe
chore
pain
pate
din
baa
sin
myth
rope
miss
perch

116 a
117 a
118 a
119 a
120 a
121 a
122 a
123 a
124 a
125 a
126 a
127 a
128 a
129 a
130 a
131 a
132 a
133 a
134 a
135 a
136 a
137 a
138 a
139 a
140 a
141 a
142 a
143 a
144 a
145 a
146 a
147 a
148 a
149 a
150 a
151 a
152 a
153 a
154 a
155 a
156 a
157 a
158 a

thigh
see
Sue
writhe
oak
note
Bro
rid
see
dare
hope
loaf
swathe
node
seethe
Jew
rate
lathe
math
fuse'n
pore
vile
rich
lea
each
growth
batch
badge
rogue
babe
toe
aye
pert
gross
catch
jay
writhe
tin
live
tag
hath
size
cup

b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b

sigh
sees
zoo
write
oh
node
brogue
ridge
seek
there
hoe
loath
sways
no
seed
do
wraith
late
mat
fusion
bore
pile
Ritz
leaf
eats
grossed
batch
bat
row
babe
toed
ape
purge
growth
cadge
day
writhe
thin
life
tag
hath
scythe
cup

c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c

thigh
see
zoo
writhe
oak
note
brogue
ridge
seek
dare
hoe
loaf
swathe
node
seethe
do
rate
late
math
fusion
pore
pile
rich
lea
eats
grossed
bat
bat
row
bay
toe
ape
purge
growth
catch
day
ride
thin
live
tack
hat
size
cuff

Appendix C^Test 2
The words in italics were recorded by the researcher on to a master audio-cassette
1
2
3

gross
Car
baa

groat
cart
bag

growth
card
back

VOLUME THREE, NUMBER TWO

grows

grow

bad

41

Errors in pronunciation by learners of English as a foreign language whose first languages are Indonesian, Gayo and Acehnese

4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53

42

Tip
coo
bathe
poor
lease
sop
suit
math
pain
Zack
loaf
lass
zoo
hoe
ship
thin
bain
bad
bit
seed
try
ricks
pain
aitch
ass
chill
sin
H
too
strife
doe
live
cop
say
thou
den
Lee
sick
muss
surf
beep
sap
pledger
cat
day
Bro'
gyp
Lea
can
thigh

ship
kook
bay
port
lees
shop
soothe
mat
fain
Jack
low
lash
sue
hose
zip
tin
vain
badge
fit
seize
trice
rib
vain
A
as
shiU
shin
ate
toot
stripe
though
lie
cob
sake
vow
then
leap
tick
mush
serve
bee
zap
pleasure
fat
jay
broke
chip
lead
pan
tie

gyp

chip

bays
poured
lea

bade

bait

sue
massed
bain
shack
lope

sued
mass
vain
sack
lobe

sues

shoe
sip
fin
pain
batch
pit
seethe
trite
rick
fain
ache
ash
till

sin
fain
bads

bat

seat

see

riff
deign
ate

rip
Thane
age

bain

Jill

A
strive
life
cough
dow
Zen
leaf
thick
sir
beaks

leave

beak

beef

pat
brogue
dip
leak
fan
fie

league
sigh

MONASH UNIVERSITY LINGUISTICS PAPERS 2005

Errors in pronunciation by learners of English as a foreign language whose first languages are Indonesian, Gayo and Acehnese

54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78

ate
chew
pea
mack
rope
tab
Ada
fee
true
Madge
gray
rope
pan
Rick
ruch
ague
Jew
force
pub
V
cease
van
grays
pine
Ritz

aye
too
fee
Madge
rove
ta
Asia
V
truced
mat
graze
row
fan
ridge
rues
A
due
fort
pucks
pea
sees
pan
gray
kine
writ

aid
shoe
bee
match
row
tat
pea
tTMth

mads
grace
robe
ban
rich
rouge
ache
you
forth
puck
Dee
see
Dan
grate
fine
rids

Jew
V
mat
robe
tap
bee
truce
mad
rove
van
writs
rude
aid
do
for
puff
bee
seat
ban
grace
rid

mats

match
wrote
writ
ruse

forced
pup
thee
seethe
than

, ridge

fee
fan

rich

Appendix D^Test 3: Reading passsage

Which would you choosea mug of coffee or a cup of tea? As children we could only have a glass of hot chocolate or
cold orange squash. I think that wasn't very fair. My brother Jim Howett always tried to cadge two teas for us from the
zany cook, Rob Grease, but he never got them. We would watch our parents taking their leisure with the forbidden drinks
safe from us, listening to music in the lounge. We would peek through the shutters at them with grave faces, and seethe
with rage, biting back tears at the unfairness of life. It was the basis of our nursing extended grudges against them.
Appendix E^Test 4: Interview schedule
1. Where were you bom, and in what year?
2. Tell me something about the
(ethnic group)?
3. What ethnic group does your mother's mother belong to?
4. What ethnic group does your mother's father belong to?
5. What ethnic group does your father's mother belong to?
6. What ethnic group does your father's father belong to?
7. What language did you first learn as a child, from whom?
8. What language do your parents use with you?
9. What was the second language you learnt, where?
10. How did you learn
(second language)?
11. What was the third language you learnt, where?
12. What was the fourth language you learnt, where?
13. What was the fifth language you learnt, where?
14. Which sounds do you have difficulty pronouncing in English?

VOLUME THREE, NUMBER TWO

43

Errors m pronunciation by learners of English as a foreign language whose first languages are Indonesian, Gayo and Acehnese

15. Why are you studying English? or Why did you decide to study in the English department?
16. Tell me about something very frightening that has happened to you.
References
AGPS. 1983. Asian language notessome likely areas of difficulty for Asian learners of English. No. 3. Indonesian/Malawi. 2nd
edition. Canberra: Australian Government Publication Service.
Baihaqi, A.K. et al. 1977. Struktur bahasa Gayo.[The structure of the Gayo language.] Banda Aceh: IAIN Jamia'ah ArRaniry and Universitas Syiah Kuala.
Balai Pustaka. 2002. Kamus Besar Bahasa Indonesia [Comprehensive Dictionary of Indonesian] 3rd edition. Jakarta:
Education Department.
Catford, J. 1987. Phonetics and the teaching of pronunciation: a systematic description of English phonology. In J. Motley
(ed.), Current perspectives on pronunciation: practices anchored in theory, 83-100. Washington DC: TESOL.
Celce-Murcia, M., Brinton, D.M. and J.M. Goodwin 1996. Teaching pronunciation. A reference for teachers of English to
speakers of other languages. Cambridge: Cambridge University Press.
Corder, S. Pit. 1973. Introducing applied linguistics. Harmondsworth: Penguin.
Cruttenden, A. 1994- Gimson's pronunciation of English. London: Edward Arnold.
Dreasher, L. and J. Anderson-Hsieh 1990. Universals in interlanguage phonology: the case of Brazilian ESL learners.
Papers and studies in contrastive linguistics 26:69-92.
Dulay, M., Burt, H. and S. Krashen 1982. Language two. New York: Oxford University Press.'
Durie, M. 1985. A grammar of Acehnese on the basis of a dialect of North Aceh. Dordrecht: Foris.
Dutt, M. 1990. Brid.ng the gulf: some features of Malayali pronunciation of English. Unpublished Masters thesis, Institute of
Education, University of London.
Ellis, R. 1994. The study of second language acquisition. Oxford: Oxford University Press.
Harmer, J. 2001. The practice of English language teaching: London: Longman.
Hecht, B. and R. Mulford 1982. The acquisitio'n of a second language phonology: Interaction of transfer and
developmental factors. Applied Psycholinguistics 3:313-28.
Jakobson, R. 1941- Child language aphasia and phorwb^cal universals. (A.R. Keiler, Trans.).
Kenworthy, J. 1987. Teaching EngUsh pronunciation. London: Longman.
Labov, W. 1972. Sociolinguistic patterns. Oxford: Blackwell.
Lowes, R. 1990. Towards a rationale for the development of materials for the teaching of pronunciation to speakers ofCastillian
Spanish. Unpublished Masters thesis, Institute of Education, University of London.
Macdonald, R. and Soenjono 1967. Indonesian reference grammar. Washington DC: Georgetown University Press.
NSW AMES. 2003. English language handbook for speakers of Iruhnesian. Sydney: AMES.
Odlin, T 1989. Language transfer. Cambridge: Cambridge University Press.
Pennington, M. 1992. A review of recent research in second language phonology with implications for practice. City Polytechnic
of Hong Kong.
Tarone, E. 1978. The phonology of interlanguage. In J. Richards (ed.), Understanding second and foreign language learning,
15-33. Rowley, Mass.: Newbury House.
Soravia, G. 1984. A sketch of the Gayo language. Unpublished manuscript, Gruppo Linguistico Catanese, Catania.
Yong, J. 2001. Malay/Indonesian speakers. In M. Swan and B. Smith (eds), Learner English, 279-95. Cambridge:
Cambridge University Press.
Wamad Abdullah. 1973. A contrastive study ofAchehnese and English cor\sonant phonemes. Singapore: paper for 9"'" 4-month
course.
Ingrid Mathew has lectured and tutored in ESL/EFL in universities in Indonesia and at RMIT and the University of Melbourne.
Her major fields of interest are teacher training, communication skills, translation and interpretation, phonetics and phonology,
pronunciation, dictionary-making and culture studies.

44

MONASH UNIVERSITY LINGUISTICS PAPERS 2005