Prosody, Phonology and Phonetics

Robert Fuchs Editor

Speech Rhythm
in Learner and
Second Language
Varieties
of English
Robert Fuchs
Editor

Speech Rhythm in Learner

and Second Language
Varieties of English
Editor
Robert Fuchs
Department of English
Universität Hamburg
Hamburg, Germany

ISSN 2197-8700 ISSN 2197-8719 (electronic)

Prosody, Phonology and Phonetics
ISBN 978-981-19-8939-1 ISBN 978-981-19-8940-7 (eBook)
https://doi.org/10.1007/978-981-19-8940-7

© Springer Nature Singapore Pte Ltd. 2023

This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part of
the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation,
broadcasting, reproduction on microfilms or in any other physical way, and transmission or information
storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology
now known or hereafter developed.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication
does not imply, even in the absence of a specific statement, that such names are exempt from the relevant
protective laws and regulations and therefore free for general use.
The publisher, the authors, and the editors are safe to assume that the advice and information in this book
are believed to be true and accurate at the date of publication. Neither the publisher nor the authors or
the editors give a warranty, expressed or implied, with respect to the material contained herein or for any
errors or omissions that may have been made. The publisher remains neutral with regard to jurisdictional
claims in published maps and institutional affiliations.

This Springer imprint is published by the registered company Springer Nature Singapore Pte Ltd.
The registered company address is: 152 Beach Road, #21-01/04 Gateway East, Singapore 189721,
Singapore
Rhythmic Patterns of Malaysian English
Speakers

Stefanie Pillai , Anussyia Muthiah, and Wan Ahmad Wan Aslynn

Abstract Previous research on Malaysian English (MalE) has indicated that there
are differences in the way that different ethnic groups produce some segments in
English possibly due to different first languages. However, thus far, there has not
been any published study on the rhythmic patterns of different ethnic groups in
Malaysia. The present study examines the rhythmic properties of speakers from
three ethnic groups: Malay, Chinese and Indians. Since studies have shown that
speaking contexts can affect rhythm, this study also investigates the extent to which
different speaking styles (read and spontaneous speech) affect rhythm in MalE. The
data comprised audio recordings of 12 female speakers from three different ethnic
groups in Malaysia: Malays, Chinese and Indians. The speakers who were between 40
and 45 years old were all fluent speakers of English based on their educational and
professional backgrounds. The speakers were recorded in two speaking contexts.
In the first one, they read a passage, and in the second context, they talked about
themselves and their families. Two metrics were used to examine rhythm in both
these speaking contexts: A normalised Pairwise Variability Index (nPVI) and VarcoV
(the standard deviation of vocalic intervals divided by their means). The results were
compared across the three ethnic groups. Based on the two metrics, there were
no significant differences among the three groups. There were also no significant
differences between the two speaking contexts for all three groups. The findings
suggest that there may be a common rhythmic pattern in MalE that cuts across ethnic
groups.

S. Pillai (B)
Faculty of Languages & Linguistics, Universiti Malaya, 50603 Kuala Lumpur, Malaysia
e-mail: stefanie@um.edu.my
A. Muthiah
Centre for Foundation Studies, Universiti Tunku Abdul Rahman, 43000 Kajang, Selangor,
Malaysia
e-mail: anussyia@utar.edu.my
W. A. Wan Aslynn
Department of Audiology and Speech-Language Pathology, Kulliyyah of Allied Health Sciences,
International Islamic University Malaysia, 25200 Kuantan, Pahang, Malaysia
e-mail: wanaslynn@iium.edu.my

© Springer Nature Singapore Pte Ltd. 2023 79

R. Fuchs (ed.), Speech Rhythm in Learner and Second Language Varieties of English,
Prosody, Phonology and Phonetics, https://doi.org/10.1007/978-981-19-8940-7_4
80 S. Pillai et al.

Keywords Ethnic groups · Chinese · First language · Indians · Malay · Malaysian

English · Pairwise Variability Index (nPVI) · Rhythm · Speaking contexts · VarcoV

1 Introduction

Malaysia is a multilingual country with over 90 languages, and most Malaysians are at
the very least bilingual, speaking, for example, a regional Malay dialect (e.g. Penang
Malay) and/or their heritage languages (e.g. Hokkien). For many Malaysian Chinese,
Mandarin is becoming the first language (Sim, 2012). For some urban Malaysia
Indians, English has become their first language mainly as a result of English-
medium education, which began during the British administration of the country and
continued until the early 1970s (David, 2005; Schiffman, 1995). Because English is
taught in schools, many Malaysians can speak English albeit at varying degrees of
proficiency. This is despite them having learnt English from primary school. Many
Malaysians speak the colloquial form of English. This colloquial variety of English,
sometimes referred to as Manglish, tends to be seen as a deviant or ‘bad’ form of
English. There is, in fact, a tendency to use the term Malaysian English (MalE)
to refer solely to the colloquial variety of MalE rather than as an umbrella term
comprising the various sub-varieties of spoken and written forms of English used in
Malaysia (Pillai & Ong, 2018).

2 Malaysian English

Perhaps, because of this perception, most descriptions of spoken MalE focus on the
colloquial and learner varieties of MalE rather than on the acrolectal variety. There is
still a dearth of studies on the variety used by Malaysians who are dominant or first
language (L1) speakers of English. In general, there still appears to be an assumption
that the speech of fluent speakers does not or should not differ much from standard
spoken British English.
In terms of research on spoken MalE, studies on the pronunciation of MalE have
mainly focused on the segmental and auditory descriptions of its vowels and conso-
nants (e.g. Baskaran, 2004; Phoon & Maclagan, 2009; Pillai, Mohd. Don, Knowles &
Tang, 2010; Tan & Low, 2010). In contrast, the prosodic features of MalE are still
relatively under-researched. This includes the rhythm of MalE. Whilst, previous
studies have shown that there are differences in the way that Malaysians with different
language backgrounds produce English sounds (e.g. Phoon et al., 2013), whether this
extends to rhythmic patterns remains understudied. Traditionally, MalE is described
as being a syllable-timed language (Baskaran, 2008; Tan & Low, 2014; Tongue,
1974). Context-wise, Baskaran (2008) suggests that informal spoken MalE is likely
to be more syllable-timed than the variety used in more formal contexts (e.g. on
television news). However, this has not been fully explored.
Rhythmic Patterns of Malaysian English Speakers 81

Thus far, published studies have mainly focussed on the rhythmic patterns of
Malay speakers of MalE, and not on the other ethnic groups in Malaysia. The present
study attempts to fill this gap by examining the rhythmic patterns of MalE among the
three main ethnic groups in Malaysia: Malay, Chinese and Indian. Ethnic categories
are state-defined, with Malays (along with indigenous groups) making up around
69% of the population. Malaysian Chinese and Indians comprise 23% and 7% of the
total population of about 32.4 million people (www.dosm.gov.my). The majority of
Malaysian Chinese and Indians came to Malaysia in the early nineteenth century.
The Chinese sub-groups include Hokkien, Cantonese, Hakka and Teochew, while
the majority of Malaysian Indians are Tamils from South India.
Because studies have shown that particular speaking contexts can affect rhythm,
this study also investigates rhythmic patterns in two speaking contexts: Read text
and spontaneous speech. In the following sections, we explore previous research on
rhythm with a focus on some of the metrics used to measure rhythm, on the main
languages spoken by the ethnic groups (e.g. Malay, Mandarin, Cantonese, Tamil and
English), as well as, studies on the rhythm of second language (L2) speakers. We
then present the research methods and findings of this study. This will be followed
by a discussion of the findings. The paper concludes by summarising the findings
and recommends further studies on the rhythm of MalE.

3 Measurements of Rhythm

Ramus et al. (1999) focused on the durational variability of consonants and vowels.
They measured the percentage of total vocalic intervals over the entire duration
of an utterance (%V) and the standard deviation of vocalic (/\V) and consonantal
(/\C) intervals in an utterance. According to Ramus et al. (1999), the values for
both /\C and %V are acceptable indicators to describe the rhythmic patterns of
any given language as stress-timed languages were found to have high /\C and low
%V, and syllable-timed one the inverse. These metrics, however, did not account for
differences in speaking rates. Further, as Arvaniti (2009), points out, Grabe and Low
(2002) showed that languages can be categorised differently when different metrics
are used. Grabe and Low (2002) proposed a Pairwise Variability Index (PVI) which is
obtained by calculating the difference in duration between successive intervals. Using
the PVI, they found higher nPVI-V scores for languages like English and German,
which are considered as being stress-timed. Lower scores were found for typically
syllable-timed languages like French and Spanish. The issue was with languages that
were placed between stress- and syllable-timed languages, such as Malay and Tamil.
Based on this finding, Grabe and Low (2002) suggested that there may be no outright
distinction between the two categories of rhythm.
The categorisation of languages as stress- or syllable-timed is also affected by the
speaking context used to elicit data. Arvaniti (2009), for instance, argues that metrics
like the nPVI, are affected by whether the data comprises read (e.g. Low, 1994, 1998;
82 S. Pillai et al.

Low et al., 2000) or spontaneous speech (Deterding, 2001; Nokes & Hay, 2012; Torg-
ersen & Szakay, 2012). Other metrics were developed in an attempt to address the
problem of categorising rhythm. Among them are the normalised standard devia-
tion of consonantal or vocalic interval durations divided by the mean consonantal
(VarcoC) or vocalic durations (VarcoV) respectively which was proposed by Dellwo
(2006). VarcoC and VarcoV aim to control for varying speech rates across speakers.
White and Mattys (2007) report that VarcoV is a good predictor of the difference
between stress- and syllable-timed languages. Knight (2011) also found that vowel-
based metrics were more reliable compared to those that are based on consonants in
a study that employed seven metrics (nPVI-V, rPVI-C, /\V, /\C, VarcoV, VarcoC and
%V on the same material). Fuchs (2016), on the other hand, went further and included
other acoustic features (e.g. duration, loudness and sonority) in a multidimensional
model which included both production and perception data. The significance of this
model is that it provides insights into the various acoustic correlates of rhythm beyond
vocalic and consonantal intervals that appear to be more prominent in a language.
For example, Fuchs (2016) found that loudness and duration are used together in
British English (BrE) as cues to prominence, whereas Indian English (IndE) used
one over the other. For example, “increases in duration are less often accompanied
by increases in loudness” (Fuchs, 2016: 209). This multidimensional model presents
a way forward for future research on rhythm especially for comparisons among
varieties of a language and among different languages.

4 Rhythm in Malay

The rhythmic pattern of Malay is usually described as being syllable-timed (Maris,

1980; Mohd. Onn, 1980; Teoh, 1994). However, Grabe and Low (2002) could not
position Malay in either category of timing. This finding, which was based on PVI
measurements taken from the speech sample of a single speaker of Malay, is contrary
to other observations. For example, these findings differ from those of Wan Aslynn
(2012) who found that the rhythm of Malay was on the syllable-timed end of the
continuum based on recordings of twenty Malay speakers reading a list of ten Malay
sentences. Deterding (2011) also found that Malay is more syllable-timed compared
to, for instance, BrE. He does, nevertheless, express concerns about the use of PVI
to come to this conclusion due to issues of syllabification, determining vowel onset
and offset and speaking rate.

5 Rhythm in Chinese

Studies on rhythm in Chinese languages have mainly focused on Mandarin and

Cantonese. Mandarin (Lin & Wang, 2007; Low & Grabe, 2002) and Cantonese
(Mok & Dellwo, 2008) have been categorised as syllable-timed. For instance, Mok
Rhythmic Patterns of Malaysian English Speakers 83

and Dellwo (2008) compared recordings of six native speakers each of Hong Kong
Cantonese and Beijing Mandarin reading and re-telling (semi-spontaneous speech
context) the North Wind and the Sun (NWS) story. They found that all the rhythmic
measures they used indicated that both Cantonese and Mandarin were syllable-timed
with the former being more syllable-timed. The %V values, for example, were higher
for both these Chinese languages compared to English, German, Italian and French.
The %V values were also lower for the semi-spontaneous speech context compared
to read speech for both Cantonese and Mandarin.

6 Rhythm in Tamil

Unlike descriptions of Cantonese and Mandarin, previous accounts of rhythm in

Tamil have been inconclusive. Keane (2006) points out that this is indicative of the
challenge of categorising a language as either stress- or syllable-timed. Keane (2006:
309) used a variety of metrics to examine the rhythm of colloquial and formal Tamil.
Keane (2006) found that the standard deviation (SD) and PVI of consonantal inter-
vals were significantly higher for formal Tamil. However, there was no significant
difference for the nPVI-V. Keane (2006) suggests that these findings may be due to
cross-word consonant clusters and the slower speech rate in formal Tamil. Similar to
Malay, Low and Grabe (2002) found that Tamil had a vocalic nPVI value that placed
it among languages that were neither clearly syllable- or stress-timed.

7 Rhythm in Varieties of English

The issues raised in relation to the discrepant findings on rhythm by different metrics
can be seen in attempts to classify different varieties of English. ‘New’ varieties
of English, including MalE, tend to be classified as syllable-timed compared to
‘native’ varieties (typically the ‘inner circle’ varieties) (Mesthrie & Bhatt, 2008).
Fuchs (2016), for example, found educated IndE to be more syllable-timed compared
to BrE. InE had lower nPVI-V and VarcoV values for both read and spontaneous
speech, and there were significant differences in the values between the two vari-
eties of English for both speaking contexts. Among the characteristic that led to the
perception of IndE being more syllable-timed was that there was less variation in
vocalic durations compared to BrE. Similarly, in Singapore English (SgE), succes-
sive vowels were found to not differ much in terms of duration (Low et al., 2000).
SgE was found to have smaller PVI values compared to BrE (Low et al., 2000).
Deterding (2001) and Tan and Low (2014) also found that SgE was syllable-timed.
However, Tan and Low (2014) found that MalE, which is perceived to be similar to
SgE, was even more syllable-timed based on their PVI values. There was less vowel
reduction in MalE compared to SgE in both read and spontaneous speech. Thus,
there were differences in terms of how syllable-timed these two groups of Malay
84 S. Pillai et al.

speakers were. Both group of speakers were undergraduate students in their home
countries, and were similar in age, but what distinguished them was their medium of
instruction in schools, i.e. English in Singapore and Malay in Malaysia. Tan and Low
(2014: 211) suggest that “the difference in the educational and social environments of
Malaysia and Singapore” could account for the different rhythmic patterns in MalE
and SgE found in their study. Based on their findings, it was felt that the PVI was
better at distinguishing differences in rhythmic patterns between these two varieties
of English. The PVI was also able to distinguish between Maori English being more
syllable-timed than Pakeha English in New Zealand English (Szakay, 2006).
In other varieties of English, the results are not always conclusive. Mok and Dellwo
(2008) found contradictory results for Cantonese English and Mandarin English using
different metrics to measure rhythm. With VarcoC and %V, the values suggested that
these two varieties of English were more syllable-timed. Other metrics, such as
/\C and nPVI-V, indicated that they were more stress-timed. These discrepancies
were attributed to a lower speaking rate said to be common among L2 speakers and
the lengthening of syllables (Mok & Dellwo, 2008). In general, it does appear that
Cantonese speakers of English are likely to sound more syllable-timed because of
the lack of vowel reduction and differentiation between the durations of syllables
(Setter, 2006).
Other studies have indicated the possible influence of one’s L1 on L2. Sarmah
et al. (2009) found that L2 speakers of American English whose L1 was Thai had
rhythmic values (based on PVI-V and %V values) that were very different from those
of L1 English speakers. Gut (2003) also found that speakers’ L1 seemed to influence
the production of German as an L2. However, Gut (2012) cautions that different
speakers may yield different results when it comes to distinguishing native and non-
native varieties of the same language because speakers who are acquiring an L2 may
or may not transfer the rhythmic patterns of their L1 on their L2. In fact, Li and Post
(2014) point out that research in L2 prosody fails to give consistent evidence for
rhythmic differences in L2 speech. This may in part be due to the metrics used.

8 Methods

The following sections describe the method used in this study to examine the rhythmic
patterns among three ethnic groups in two speaking contexts.

8.1 Speakers

Twelve female Malaysian speakers from three different ethnic groups were selected
for this study: Four Malays, four Chinese and four Indians. These ethnic categories
are based on state-defined ones. The reason for selecting these ethnic groups is that
they are the three major ethnic groups in Malaysia.
Rhythmic Patterns of Malaysian English Speakers 85

The average age of the speakers was 43 years. Purposeful sampling was used to
select the speakers based on the following criteria:
. Born and grew up in Malaysia.
. Primary and secondary education in national schools in Malaysia.
. Fluent speakers of English based on their tertiary qualifications (Bachelor’s degree
and higher in English language teaching or English Literature, and Linguistics),
job experience (English language lecturers at a university) and frequency of using
English in different contexts.
Based on these criteria, a total of 14 out of 30 potential speakers who were
identified agreed to be recorded. Almost all of the potential speakers were females,
and the 14 speakers who consented to participate in the study were all females. For
this study, we only selected those who identified their ethnic group as Malay, Indian
and Chinese, and hence, two of the 14 speakers were removed. The L1 of all four
Malay speakers was Malay. Cantonese was the L1 of two of the Chinese speakers,
while Hokkien was the L1 of the remaining two Chinese speakers. Language shift
to English among Malaysian Indians was mentioned in the Introduction section and
thus, not surprisingly, three of the Indian speakers said that English was their L1. Only
one of the speakers said that Tamil was her L1. English was the home and dominant
language for the speakers who declared English as their L1 (see Pillai, 2006). These
speakers grew up speaking English rather than their South Indian heritage languages
and are not fluent in these languages (Tamil for two of them and Malayalam for the
other). They are the first generation in their families to use English as an L1. For all
the other speakers, especially the non-Malay speakers, English was used extensively
at home with their children. All twelve speakers of this study are similar in terms
of educational backgrounds and professions as all of them are English language
lecturers with more than 10 years of experience of teaching English at the time of
the recordings.

8.2 Data

Since differences have been reported in the prosodic features of read speech and
conversational speech (e.g. Howell & Kadi-Hani, 1991), this study examined rhythm
in two speaking contexts: read text and spontaneous speech. For the read speech
context, the subjects were recorded reading the NWS passage while the spontaneous
speech consisted of a short interview with the speakers about themselves and their
families.
86 S. Pillai et al.

8.3 Procedure

All recordings were carried out in a quiet room. The Kay Elemetrics Computerized
Speech Lab (CSL) Model 4500 was used to record the speakers at a sampling rate
of 44,100 Hz using a high-quality dynamic microphone placed a few inches from
the mouth of the speakers. Written consent was obtained from all speakers prior
to the recordings. For the read text, speakers were provided with the text to read
through once so that they were familiar with it before being recorded. As for the
spontaneous speech context, questions were posed to the speaker to elicit responses
about themselves and their families.

8.4 Segmentation of Speech

The NSW passage consisted of five sentences, 113 words and 142 syllables. All five
sentences were analysed from the read text of each speaker. These recordings, which
had been transcribed orthographically, were further examined and annotated using
Praat version 5.3.82 (Boersma & Weenink, 2014). The recordings for each speaker
were segmented in text grids using Praat into the following elements: Text, vocalic
and consonantal units. As shown in Fig. 1, tier 1 contains the text and tiers 2 and 3
show the segmented consonantal and vocalic elements respectively. The location of
boundaries for vocalic and consonantal units were identified and labelled based on
the wideband spectrograms (White & Mattys, 2007). The duration was first measured
from left-to-right for vocalic and intervocalic intervals. Vowels in sequences, such
as in fricative-vowel and vowel-nasal, were identified following the criteria used by
Grabe and Low (2002) where applicable. In fricative-vowel sequences, such as in the
word ‘sun’, the vowel was measured when the noise pattern of the voiceless fricative
/s/ ended. Vowel-nasal sequences (e.g. disputing) were segmented by observing the
formant movement of both the nasal and vowels. Similar to Deterding (2001) and
Tan and Low (2014), phrase final syllables were excluded for both read and informal
spontaneous speech to avoid the effect of phrase-final syllable lengthening on the PVI
measurements. Further, pauses between intonational phrases were excluded from the
analysis.
The 12 interviews consisted of 33 utterances each. These utterances were deter-
mined from the time speakers started speaking until the point that they paused or
were silent. The utterances produced by the speakers could be a word, a short phrase,
or a sentence. The whole interview was analysed since these interviews were short
with an average of 1 min and 49 s per interview. The recording for each speaker
was segmented similar to the read speech. Other considerations included taking into
account contracted forms, for example, the contraction of the words they are to
they’re (Deterding, 2001). Further, in spontaneous speech, it is common for speakers
to hesitate (e.g. silent or filled pauses), but hesitations were not measured. Repetitions
of lexical items in a stretch of speech were also excluded from the analysis of the
current study. The first correct lexical item produced by the speaker was taken into
Rhythmic Patterns of Malaysian English Speakers 87

Fig. 1 Screenshot of annotations in Praat for read speech

consideration. Other forms of interruptions, such as laughter and also silent pauses
of three hundred milliseconds and above, were also excluded from the analysis.

8.5 Measurements

The vocalic intervals and syllable durations were measured based on the spectrograms
and the auditory segmentation. In this study, two metrics, nPVI-V and VarcoV, were
used to measure vocalic variability between the two speaking styles produced by the
speakers from the three ethnic groups. As previously explained, the PVI is used to
measure the differences in duration of successive intervals. The nPVI is the mean of
the differences between successive intervals divided by the sum of the same intervals
where nPVI-V is the normalised Pairwise Variability Index for vocalic intervals.
VarcoV is calculated as the standard deviation of vocalic interval duration divided
by mean vocalic interval duration and then multiplied by 100. These two metrics
were selected as they have been found to be “robust to variation in speech rate and
relatively robust to variation in sentences, speakers and transcribers” (Fuchs, 2016:
56).

9 Findings

Table 1 presents the overall nPVI-V and VarcoV values for the Malay, Chinese and
Indian speakers in the two speaking contexts. A one-way ANOVA indicated that
there were no significant differences between the mean nPVI-V values among the
88 S. Pillai et al.

three ethnic groups for read [F(2, 9) = 0.64, p = 0.55] and spontaneous speech
[F(2, 9) = 0.38, p = 0.69]. As shown in Table 1, the average nPVI-V values were
slightly higher in read speech for the Chinese speakers. The Indian speakers had
lower average nPVI-V values in spontaneous speech. Consistent with these findings,
no significant differences were found between the nPVI-V for read (M = 56.46, SD
= 2.87) and spontaneous speech (M = 54.48, SD = 5.36): t(11) = 1.22, p = 0.25
based on a two-tailed correlated samples t-test.
A one-way ANOVA indicated that there were no significant differences between
the mean VarcoV values among the three ethnic groups for read (F(2, 9) = 2.08,
p = 0.18) and spontaneous speech [F(2, 9) = 0.59, p = 0.57]. As can be seen in
Table 2, the difference between the average VarcoV values for both speech contexts
among Chinese speakers was 7.93 with a higher VarcoV value in spontaneous speech.
Higher VarcoV values were also found for the other two groups with a difference
of 1.07 for the Indian group and 8.76 for the Malay one. No significant differences
were found in terms of vocalic variability for read (M = 62.82, SD = 5.96) and
spontaneous speech (M = 68.73, SD = 10.04): t(11) = 1.95, p = 0.08 based on a
two-tailed correlated samples t-test. Figures 2 and 3 show the cross comparisons of
VarcoV and nPVI-V for read and spontaneous speech respectively. The two metrics,
nPVI-V and VarcoV, were not significantly correlated in both read (r = 0.01, p =
0.97) and spontaneous speech (r = 0.56, p = 0.06).

Table 1 Average nPVI-V and standard deviation values

Speakers Read speech Spontaneous speech
nPVI-V VarcoV nPVI-V VarcoV
Chinese 1 53.56 77.70 50.39 71.80
Chinese 2 56.00 62.12 57.09 65.80
Chinese 3 55.62 59.22 56.82 67.05
Chinese 4 51.25 62.51 56.83 88.62
Average 54.11 (2.19) 65.39 (8.34) 55.28 (3.27) 73.32 (10.53)
Indian 1 57.68 64.99 43.88 49.01
Indian 2 59.22 66.63 59.88 80.24
Indian 3 60.46 66.23 54.40 66.31
Indian 4 61.30 61.19 58.03 67.74
Average 59.67 (1.57) 64.76 (2.48) 54.05 (7.15) 65.82 (12.84)
Malay 1 56.38 58.53 50.87 68.53
Malay 2 54.28 61.97 48.41 70.07
Malay 3 53.84 53.34 52.59 56.73
Malay 4 57.92 59.36 64.61 72.92
Average 55.61 (1.90) 58.30 (3.61) 54.12 (7.20) 67.06 (7.13)
Overall average 56.46 (54.96) 62.82 (5.96) 54.48 (5.60) 68.73 (10.04)
Note Standard deviations (SD) are in parenthesis
Rhythmic Patterns of Malaysian English Speakers 89

Table 2 nPVI-V and VarcoV values in some varieties of English

Variety nPVI-V VarcoV
Read Spontaneous speech Read Spontaneous speech
BrE (Fuchs, 2016) 61.3 58.3 53.2 51.7
IndE (Fuchs, 2016) 55.6 52.4 46.3 45.7
MalE (present study) 56.46 54.48 62.82 68.73
MalE (Tan & Low, 2014) 41.21 − 37.49 −
SgE (Tan & Low, 2014) 47.3 − 44.15 −

Fig. 2 VarcoV and nPVI-V for read speech for all speakers

Fig. 3 VarcoV and nPVI-V for spontaneous speech for all speakers
90 S. Pillai et al.

10 Discussion

The analysis of nPVI-V and VarcoV values indicate no significant differences among
the three ethnic groups for both read text and spontaneous speech which suggests that
the three groups had similar rhythmic patterns. There was also no significant differ-
ence between the two speaking contexts suggesting that both had similar rhythmic
patterns. There was a tendency for the nPVI-V and VarcoV values to be slightly
lower (but not significantly) in the spontaneous speaking context indicating less
vocalic variability in this speech context compared to the read text.
As previously mentioned, there is a tendency for ‘new’ varieties of English to be
categorised as syllable-timed compared to a variety like BrE suggesting that there is
less variability in these varieties. However, the values for both speaking contexts were
higher than those reported in previous studies. Tan and Low (2014), for example,
reported much lower PVI and VarcoV values for Malaysian and Singapore English
(see Table 2). The PVI-V value for IndE for both read and spontaneous speech (Fuchs,
2016), on the other hand, is closer to the ones in this study. However, this was not
the case for VarcoV. The PVI-V values for IndE reported by Fuchs (2016) are lower
that what he reports for BrE. While it is acknowledged that Fuch’s results are not
based on exactly the same type of data as the one in this study (see Table 2), they can
provide an overview of where MalE is positioned in the stress- and syllable-timed
continuum. The figures in Table 2 suggest that MalE has a similar rhythmic quality
to IndE based on nPVI-V values, indicating that like IndE, MalE has less vocalic
variation than BrE, and is therefore, more syllable-timed.
The difference between the current study and other studies on MalE are the
age group of the speakers, educational backgrounds, dominant home language and
professions. These factors may have affected the rhythmic patterns found in both
speech styles in this study. Firstly, the speakers of Tan and Low (2014) were students
in their twenties whereas the speakers of the current study are English lecturers with
an average age of 43. Secondly, the dominant home language for each speaker is
different as some use English as their home dominant language. Thirdly, they are
from different educational and professional backgrounds. The results of the present
study suggest that despite different declared L1s, including English, the Malaysian
speakers in this study exhibited similar rhythmic patterns.

11 Conclusion

In sum, it can be concluded that all three ethnic groups exhibited similar rhythmic
patterns to each other, suggesting a common pattern for MalE. There was also no
discernible difference in the rhythmic patterns of the two speaking contexts: Speakers
seemed to have similar patterns when reading a text and speaking spontaneously.
Despite the higher PVI and VarcoV values than a previous study, MalE can still be
classified as being more syllable-timed.
Rhythmic Patterns of Malaysian English Speakers 91

However, the two metrics used in this study were not able to capture the prominent
features in the acoustic signal which were regulated across the data in order to describe
the rhythmic property of MalE. In line with Arvaniti (2009), a more robust speech
feature which is able to withstand the influence of those independent variables might
be more accurate in characterising the rhythm of any languages. This may mean that
the researchers might need to go beyond segmental duration measurements. Wan
Aslynn (2012) and Fuchs (2014) suggest that features like syllable duration, inter-
intensity minima duration and single vocalic duration are worth looking into in order
to underpin the regularity across speakers and sentences.
In addition, the results of this study need to be treated carefully due to the size of
sample and the speaking contexts used to elicit the data. The status of English for the
speakers (i.e. fluent bi-/multilinguals and dominant speakers of English) may also
have influenced the findings. Thus, future studies on rhythm in Malaysian English
should include a bigger and more diverse sample. For example, the sample should
look at whether there are differences between fluent speakers and English language
learners or speakers for whom English is an L1.

References

Arvaniti, A. (2009). Rhythm, timing and the timing of rhythm. Phonetica, 66, 46–63.
Baskaran, L. M. (2004). Malaysian English. Morphology and syntax. In B. Kortmann, K. Burridge,
R. Mestrie, E. W. Schneider, & C. Upton (Eds.), A handbook of varieties of English (Vol. 2,
pp. 1073–1085). Mouton de Gruyter.
Baskaran, L. M. (2008). Malaysian English – Phonology. In R. Mesthrie (Ed.), Varieties of English:
Africa, South and Southeast Asia (pp. 278–291). Mouton de Gruyter.
Boersma, P., & Weenink, D. (2014). Praat: Doing phonetics by computer (version 5.3.82). Retrieved
from http://fon.hum.uva.nl/praat/.
David, M. K. (2005). Reasons for language shift in Peninsular Malaysia. Journal of Modern
Languages, 15(1), 1–11.
Dellwo, V. (2006). Rhythm and speech rate: A variation coefficient for deltaC. In P. Karnowski, & I.
Szigeti (Eds.), Language and language processing: Proceedings of the 38th Linguistic Colloquium
(pp. 231–241). Piliscsaba 2003. Frankfurt: Peter Lang.
Deterding, D. (2001). The measurement of rhythm: A comparison of Singapore and British English.
Journal of Phonetics, 29(2), 217–230.
Deterding, D. (2011). Measurements of the rhythm of Malay. Proceedings of the 17th International
Congress of Phonetic Sciences, Hong Kong, 17–21 August 2011, 576–579.
Fuchs, R. (2014). Integrating variability in loudness and duration in a multidimensional model of
speech rhythm. Evidence from Indian English and British English. In Proceedings of speech
prosody 7. Dublin (pp. 290–294). Retrieved from https://www.isca-speech.org/archive_v0/Spe
echProsody_2014/pdfs/47.pdf
Fuchs, R. (2016). Speech rhythm in varieties of English—Evidence from educated Indian English
and British English. Springer.
Grabe, E., & Low, E. L. (2002). Durational variability in speech and the rhythm class hypothesis.
In C. Gussenhoven, & N. Warner (Eds.), Laboratory phonology (Vol. 7, pp. 515–546). Mouton
de Gruyter.
Gut, U. (2003). Prosody in second language speech production: The role of the native language.
Fremdsprachen Lehren Und Lernen, 32, 133–152.
92 S. Pillai et al.

Gut, U. (2012). Rhythm in L2 speech. In D. Gibbon (Ed.), Speech and language technology (pp. 83–
94). Cambridge University Press.
Howell, P., & Kadi-Hani, K. (1991). Comparison of prosodic properties between read and
spontaneous speech material. Speech Communication, 10(2), 161–169.
Keane, E. L. (2006). Rhythmic characteristics of colloquial and formal Tamil. Language and Speech,
49(3), 299–332.
Knight, R. A. (2011). Assessing the temporal reliability of rhythm metrics. Journal of the
International Phonetics Association, 41(3), 271–281.
Li, A., & Post, B. (2014). L2 Acquisition of prosodic properties of speech rhythm: Evidence from
L1 Mandarin and German learners of English. Studies in Second Language Acquisition, 36(2),
223–255.
Lin, H., & Wang, Q. (2007). Mandarin rhythm: An acoustic study. Chinese Linguistics and
Computing, 17(3), 127–140.
Low, E. L. (1994). Intonation patterns in Singapore English. Unpublished master’s thesis.
Department of Linguistics, University of Cambridge, Cambridge, United Kingdom.
Low, E. L. (1998). Prosodic prominence in Singapore English. Unpublished Ph.D. thesis. University
of Cambridge, Cambridge, United Kingdom.
Low, E. L., Grabe, E., & Nolan, F. (2000). Quantitative characterizations of speech rhythm:
‘Syllable-timing’ in Singapore English. Language and Speech, 43(4), 377–401.
Maris, Y. (1980). The Malay sound system. Kuala Lumpur: Fajar Bakti.
Mesthrie, R., & Bhatt, R. M. (2008). World Englishes: The study of new linguistic varieties. Key
topics in sociolinguistics. Cambridge University Press.
Mohd. Onn, F. (1980). Aspects of Malay phonology and morphology: A generative approach.
Universiti Kebangsaan Malaysia.
Mok, P., & Dellwo, V. (2008). Comparing native and non-native speech rhythm using acoustic
rhythmic measures: Cantonese, Beijing Mandarin and English. Speech Prosody 2008, Camp-
inas/Brazil, 6–9 May 2008, 423–426.
Nokes, J., & Hay, J. (2012). Acoustic correlates of rhythm in New Zealand English: A diachronic
study. Language Variation and Change, 24(1), 1–31.
Phoon, H. S., & Maclagan, M. (2009) The phonology of Malaysian English: A preliminary study.
In L. J. Zhang, R. Rubdy, & A. Lubna (Eds.), Englishes and literatures-in-English in a globalised
world, Proceedings of the 13th International Conference on English in Southeast Asia (pp. 46–60).
Singapore: National Institute of Education, Nanyang Technological University.
Phoon, H. S., Abdullah, A. C., & Maclagan, M. (2013). The consonant realizations of
Malay-, Chinese- and Indian-Influenced Malaysian English. Australian Journal of Linguistics,
33(1), 3–30.
Pillai, S. (2006). Malaysian English as a first language. In M. K. David (Ed.), Language choices and
discourse of Malaysian families: Case studies of families in Kuala Lumpur, Malaysia (pp. 61–75).
Petaling Jaya: SIRD.
Pillai, S., & Mohd. Don, Z., Knowles, G., & Tang, J. (2010). Malaysian English: An instrumental
analysis of vowel contrasts. World Englishes, 29(2), 159–172.
Pillai, S., & Ong, L. T. (2018). English(es) in Malaysia. Asian Englishes, 20(2), 147–157.
Ramus, R., Nespor, M., & Mehler, J. (1999). Correlates of linguistic rhythm in the speech signal.
Cognition, 73(3), 265–292.
Sarmah, P., Gogoi, D. V., & Wiltshire, C. (2009). Thai English: Rhythm and vowels. English
World-Wide, 30(2), 196–217.
Schiffman, H. (1995). Language shift in the Tamil communities of Malaysia and Singapore.
Southwest Journal of Linguistics, 14(1–2), 151–165.
Setter, J. (2006). Speech rhythm in World Englishes: The case of Hong Kong. TESOL Quarterly,
40(4), 763–782.
Sim, T. W. (2012). Why are the native languages of the Chinese Malaysians in decline? Journal of
Taiwanese Vernacular, 4(1), 63–95.
Rhythmic Patterns of Malaysian English Speakers 93

Szakay, A. (2006). Rhythm and pitch as markers of raciality in New Zealand English. In P. Warren, &
C. Watson (Eds.), Proceedings of the 11th Australasian International Conference on Speech
Science & Technology (pp. 421–426). Auckland: University of Auckland.
Tan, R. S. K., & Low, E. L. (2010). How different are the monophthongs of Malay speakers of
Malaysian and Singapore English? English World-Wide, 31(2), 162–189.
Tan, R. S. K., & Low, E. L. (2014). Rhythmic patterning in Malaysian and Singapore English.
Language and Speech, 57(2), 196–214.
Teoh, B. S. (1994) The sound system of Malay revisited. Kuala Lumpur: Dewan Bahasa Pustaka.
Tongue, R. K. (1974). The English of Singapore and Malaysia. Eastern Universities Press.
Torgersen, E., & Szakay, A. (2012). An investigation of speech rhythm in London English. Lingua,
122(7), 822–840.
Wan Aslynn, W. A. (2012). Instrumental phonetic study of the rhythm of Malay. Unpublished Ph.D
thesis. Newcastle University, United Kingdom.
White, L., & Mattys, S. L. (2007). Calibrating rhythm: First language and second language studies.
Journal of Phonetics, 35(4), 501–522.