See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/29751792

Phonetic correlates of unintelligibility in Vietnamese-accented English

Article
Source: OAI

CITATIONS READS

5 202

1 author:

Una Cunningham
Stockholm University
48 PUBLICATIONS 540 CITATIONS

SEE PROFILE

All content following this page was uploaded by Una Cunningham on 20 June 2015.

The user has requested enhancement of the downloaded file.

 Proceedings, FONETIK 2009, Dept. of Linguistics, Stockholm University

Phonetic correlates of unintelligibility in Vietnamese-

accented English
Una Cunningham
School of Arts and Media, Dalarna University, Falun

Abstract There are a number of possible ways in

which intelligibility can be measured. Listeners
Vietnamese speakers of English are often able can be asked to transcribe what they hear or to
to communicate much more efficiently in writ- choose from a number of alternatives. Stimuli
ing than in speaking. Many have quite high vary from spontaneous speech through texts to
proficiency levels, with full command of ad- sentences to wordlists. Sentences with varying
vanced vocabulary and complex syntax, yet degrees of semantic meaning are often used
they have great difficulty making themselves (Kirkpatrick, Deterding et al. 2008) to control
understood when speaking English to both na- for the effect of contextual information on intel-
tive and non-native speakers. This paper ex- ligibility. Few intelligibility studies appear to be
plores the phonetic events associated with concerned with the stimulus material. The ques-
breakdowns in intelligibility, and looks at com- tion of what makes an utterance unintelligible is
pensatory mechanisms which are used. not addressed in these studies. The current pa-
per is an effort to come some way to examining
Intelligibility this issue.
The scientific study of intelligibility has passed
through a number of phases. Two strands that Learning English in Vietnam
have shifted in their relative prominence are the The pronunciation of English presents severe
matter of to whom non-native speakers are to challenges to Vietnamese-speaking learners.
be intelligible. In one strand the emphasis is on
Not only is the sound system of Vietnamese
the intelligibility of non-native speakers to na- very different from that of English, but there are
tive English-speaking listeners (Flege, Munro et also extremely limited opportunities for hearing
al. 1995; Munro and Derwing 1995; Tajima, and speaking English in Vietnam. In addition,
Port et al. 1997). This was the context in which there are limited resources available to teachers
English was taught and learned – the majority of English in Vietnam so teachers are likely to
of these studies have been carried out in what pass on their own English pronunciation to their
are known as inner circle countries, which, in students.
turn, reflects the anglocentricism which has University students of English are intro-
characterised much of linguistics. The other duced to native-speaker models of English pro-
strand focuses on the position of English as a nunciation, notably Southern educated British,
language of international communication, but they do not often have the opportunity to
where intelligibility is a two-way affair between speak with non-Vietnamese speakers of Eng-
a native or non-native English-speaking speaker lish. Most studies of Vietnamese accents in
and a native or non-native English-speaking English have been based in countries where
listener (Irvine 1977; Flege, MacKay et al. English is a community language, such as the
1999; Kirkpatrick, Deterding et al. 2008; Rooy U.S. (Tang 2007) or Australia (Nguyen 1970;
2009).
In-gram and Nguyen 2007). This study is thus
The current study is a part of a larger study unusual in considering the English pronuncia-
of how native speakers of American English, tion of learners who live in Vietnam. The
Swedish, Vietnamese, Urdu and Ibo are per-
speech material presented here was produced by
ceived by listeners from these and other lan-
members of a group of female students from
guage backgrounds. Vietnamese-accented Hanoi.
speech in English has been informally observed
to be notably unintelligible for native English-
speaking listeners and even for Vietnamese lis- Vietnamese accents of English
teners there is great difficulty in choosing The most striking feature of Vietnamese-
which of four words has been uttered (Cun- accented English is the elision of consonants, in
ningham 2009). particular in the syllable coda. This can obvi-
 Proceedings, FONETIK 2009, Dept. of Linguistics, Stockholm University

ously be related to the phonotactic constraints Vowel quality

operational in Vietnamese, and it is clearly a The semantic load of the distinction between
problem when speaking English which places a the KIT and FLEECE vowels is significant.
heavy semantic load on the coda in verb forms This opposition seems to be observed in most
and other suffixes. Consonant clusters are gen- varieties of English, and it is one that has been
erally simplified in Vietnamese-accent English identified as essential for learners of English to
to a degree that is not compatible with intelligi- master (Jenkins 2002). Nonetheless, this dis-
bility. Even single final consonants are often tinction is not very frequent in the languages of
absent or substituted for by another consonant the world. Consequently, like any kind of new
which is permitted in the coda in Vietnamese. distinction, a degree of effort and practice is
Other difficulties in the intelligibility of required before learners with many first lan-
Vietnamese-accented English are centred in guages, including Vietnamese, can reliably per-
vowel quality. English has a typologically rela- ceive and produce this distinction.
tively rich vowel inventory, and this creates
problems for learners with many L1s, including
Vietnamese. The distinction between the vow-
els of KIT and FLEECE to use the word class
terminology developed by John Wells (Wells
1982) or ship and sheep to allude to the popular
pronunciation teaching textbook (Baker 2006)
is particularly problematic.
Other problematic vowel contrasts are that
between NURSE and THOUGHT (e.g. work vs
walk) and between TRAP AND DRESS (e.g.
bag vs beg). The failure to perceive or produce
these vowel distinctions is a major hinder to the
intelligibility of Vietnamese-accented English. Figure 1. F1 vs F2 in Bark for S15 for the words
Vowel length is not linguistically significant bead, beat, bid, bit.
in Vietnamese and the failure to notice or pro- Fig.1 shows the relationship between F1 and
duce pre-fortis clipping is another source of un- F2 in Bark for the vowels in the words beat,
intelligibility. Another interesting effect that is bead, bit and bid for S15, a speaker of Viet-
attributable to transfer from Vietnamese is the namese (a 3rd year undergraduate English major
use in English of the rising sac tone on sylla- student at a university in Hanoi). This speaker
bles that have a voiceless stop in the coda. This does not make a clear spectral distinction be-
can result in a pitch prominence that may be tween the vowels. As spectral quality is the
interpreted as stress by listeners. most salient cue to this distinction for native
Vietnamese words are said to be generally speakers of English (Ladefoged 2006; Crutten-
monosyllabic, and are certainly written as den 2008), the failure to make a distinction is
monosyllables with a space between each sylla- obviously a hinder to intelligibility.
ble. This impression is augmented (or possibly
explained) by the apparent paucity of connected Vowel duration
speech phenomena in Vietnamese and conse-
quently in Vietnamese-accented English. Enhanced pre-fortis clipping is used in many
varieties of English as a primary cue to post-
vocalic voicelessness (Ladefoged 2006; Crut-
Analysis tenden 2008). It has been well documented that
A number of features of Vietnamese-accented phonologically voiced (lenis) post-vocalic con-
English will be analysed here. They are a) the sonants are often devoiced by native speakers
vowel quality distinction between the words of many varieties of English (e.g. Cruttenden
sheep and ship, b) the vowel duration distinc- 2008). This means that in word sets such as
tion between seat and seed, and c) the causes of bead, beat, bid, bit native speakers will signal
global unintelligibility in semantically mean- postvocalic voicing primarily by shortening the
ingful sentences taken from an earlier study prevocalic vowel in beat and bit. In addition,
(Munro and Derwing 1995). native speakers will have a secondary dur-
ational cue to the bead, beat vs bid, bit vowel
distinction where the former vowel is system-
 Proceedings, FONETIK 2009, Dept. of Linguistics, Stockholm University

atically longer than the latter (Cruttenden fusion (as illustrated in Figures 1 and 2 above.
2008). The word often is liable to be pronounced with
a substitution of /p/ for the /f/ at the end of the
first syllable, as voiceless stops are permissible
in coda position in Vietnamese while fricatives
are not.
Let us then see what happens when speaker
V1, a 23-year old male graduate student of Eng-
lish from Hanoi, reads this sentence. In fact he
manages the beginning of the utterance well,
with an appropriate (native-like) elision of the
/d/ in friends. Things start to go wrong after that
with the word sheep. Figure 3 shows a spectro-
gram of this word using Praat (Boersma and
Weenink 2009). As can be seen, the final con-
Figure 2. Average vowel and stop duration in ms sonant comes out as an ungrooved [s].
for S15 for 5 instances of the words bead, beat, bid,
bit.

So, as is apparent from Figure 2, speaker S15

manages to produce somewhat shorter vowels
in bid and bit than in beat and bit. This is the
primary cue that this speaker is using to dis-
similate the vowels, although not, unfortunately
the cue expected as most salient by native lis-
teners. But there is no pre-fortis clipping appar-
ent. This important cue to post-vocalic voicing
is not produced by this speaker. In conjunction
with the lack of spectral distinction between the
vowels of bead, bead vs. bid, bit seen in Figure
1, the result is that these four words are per- Figure 3. The word sheep as pronounced by speaker
cieved as indistinguishable by native and not- V1.
native listeners (Cunningham 2009).
Now there is a possible explanation for this. As
Sentence intelligibility mentioned above, final /f/ as in the word if is
often pronounced as [ip]. This pronunciation is
A number of factors work together to confound viewed as characteristic of Vietnamese-
the listener of Vietnamese-accented English in accented English in Vietnam – teacher and thus
connected speech. Not only is it difficult to per- learner awareness of this is high, and the feature
ceive vowel identity and post vocalic voicing as is stigmatised. Thus the avoidance of the final
in the above examples, but there are a number /p/ of sheep may be an instance of hypercorrec-
of other problems. Consider the sentence My tion. It is clearly detrimental to V1’s intelligi-
friend’s sheep is often green. This is taken from bility.
the stimuli set used for the larger study men- Another problematic part of this sentence by
tioned above. The advantage of sentences of V1 is that he elides the /z/ in the word is. There
this type is that the contextual aids to interpret- is silence on either side of this vowel. Again,
ability are minimised while connected speech this leads to intelligibility problems. The final
phenomena are likely to be elicited. There are difficulty for the listener in this utterance is a
here a number of potential pitfalls for the Viet- matter of VOT in the word green. V1 has no
namese speaker of English. The cluster at the voice before the vowel begins as is shown in
end of friend’s, especially in connection with figure 4. The stop is apparently voiceless and
the initial consonant in sheep can be expected the release is followed by a 112ms voiceless
to prove difficult and to be simplified in some aspiration. This leads to the word being misin-
way. The quality and duration of the vowels in terpreted by listeners as cream.
sheep and green can be expected to cause con-
 Proceedings, FONETIK 2009, Dept. of Linguistics, Stockholm University

Acoustical Society of America 97(5): 3125-

3134.
Ingram, J. C. L. and T. T. A. Nguyen (2007).
Vietnamese accented English: Foreign ac-
cent and intelligibility judgement by listen-
ers of different language backgrounds, Uni-
versity of Queensland.
Irvine, D. H. (1977). Intelligibility of English
speech to non-native English speakers. Lan-
guage and Speech 20: 308-316.
Jenkins, J. (2002). A sociolinguistically based,
empirically researched pronunciation sylla-
bus for English as an international language.
Figure 4 the word green as pronounced by speaker Applied Linguistics 23(1): 83-103.
V1. The marking shows aspiration after the release Kirkpatrick, A., D. Deterding, et al. (2008). The
of the initial stop. international intelligibility of Hong Kong
English. World Englishes 27(3-4): 359-377.
Ladefoged, P. (2006). A Course in Phonetics.
Conclusion Boston, Mass, Thomson.
So it can be seen that the intelligibility of these Munro, M. J. and T. M. Derwing (1995). Proc-
Vietnamese speakers of English is a major essing time, accent, and comprehensibility
problem for them and their interlocutors. Not in the perception of native and foreign-
only do they have non-native pronunciation fea- accented speech. Language and Speech 38:
tures that are clear instances of transfer from 289-306.
their L1, Vietnamese, they also have other, Nguyen, D. L. (1970). A contrastive phonologi-
spontaneous, modifications of the target cal analysis of English and Vietnamese.
sounds. This is part of the general variability Canberra, Australian National University.
that characterises non-native pronunciation, but Rooy, S. C. V. (2009). Intelligibility and per-
when the sounds produced are as far from the ceptions of English proficiency. World Eng-
target sounds as they are in the speech of V1, lishes 28(1): 15-34.
communication is an extreme effort. Tajima, K., R. Port, et al. (1997). Effects of
temporal correction on intelligibility of for-
References eign-accented English, Academic Press Ltd.
Tang, G. M. (2007). Cross-linguistic analysis of
Baker, A. (2006). Ship or sheep: an intermedi- Vietnamese and English with implications
ate pronunciation course. Cambridge, Cam- for Vietnamese language acquisition and
bridge University Press. maintenance in the United States. Journal of
Boersma, P. and D. Weenink (2009). Praat: do- Southeast Asian American Education and
ing phonetics by computer. Advancement 2.
Cruttenden, A. (2008). Gimson's Pronunciation Wells, J. C. (1982). Accents of English. Cam-
of English. London, Hodder Arnold bridge, Cambridge University Press
Cunningham, U. (2009). Quality, quantity and
intelligibility of vowels in Vietnamese-
accented English. Issues in Accents of Eng-
lish II: Variability and Norm. E. Waniek-
Klimczak. Newcastle, Cambridge Scholars
Publishing Ltd.
Flege, J. E., I. R. A. MacKay, et al. (1999). Na-
tive Italian speakers' perception and produc-
tion of English vowels. Journal of the
Acoustical Society of America 106(5):
2973-2987.
Flege, J. E., M. J. Munro, et al. (1995). Factors
affecting strength of perceived foreign ac-
cent in a 2nd language. Journal of the

View publication stats