The relationship between the perception and production of the english contrast pair ʧand ʤ among southern vietnamese speakers of english

ix LIST OF TABLES AND ABBREVIATIONS TABLES Table 2.1 Differences between the Vietnamese and English syllables Table 2.2 Distinctive features specifications of the Vietnamese consonants

VIETNAM NATIONAL UNIVERSITY – HOCHIMINH CITY

UNIVERSITY OF SOCIAL SCIENCES & HUMANITIES

FACULTY OF ENGLISH LINGUISTICS AND LITERATURE

- -

THE RELATIONSHIP BETWEEN THE PERCEPTION AND PRODUCTION OF THE ENGLISH CONTRAST PAIR /t~/ AND /d2/ AMONG SOUTHERN VIETNAMESE SPEAKERS OF ENGLISH

Submitted to the Faculty of English Linguistics & Literature

in partial fulfillment of the Master‟s degree in TESOL

By PHAM THI THUY TRANG

Supervised by

LE HOANG DUNG, PHD

HO CHI MINH CITY, MAY 2011

i

STATEMENT OF ORIGINALITY

I certify that this thesis entitled “THE RELATIONSHIP BETWEEN THE PERCEPTION AND PRODUCTION OF THE ENGLISH CONTRAST PAIR /t~/ AND /d2/ AMONG SOUTHERN VIETNAMESE SPEAKERS OF ENGLISH” is my own work.

This thesis has not been submitted for the award of any degree or diploma in any other institution

Ho Chi Minh City, May 2011

Phạm Thị Thùy Trang

I would also like to thank my teacher, Nguyen Thi Kieu Thu, PHD at the University of Social Sciences and Humanities, HCMC for her ongoing encouragement and support during my master course

Gratitude is also offered to my friend, Nguyen Dang Khoa, graduate student at Australia National University for sending me valuable and precious materials without which I could not have written the thesis

Thanks are also due to the Board of Directors at the Pedagogy University Language Center, Thanh Nien Language Center and Vietnam-Australia Language School, who gave me conditions to conduct my surveys and experiments for the thesis

I owe many thanks to my friends and my colleagues for their support, encouragement and assistance in data collection

Lastly, but certainly not least, I would like to thank my dear husband, Duong Anh Tuan, for his loving care, support and encouragement that has helped me so much in completing this thesis

iii

ABSTRACT

Although the recently-gained insights into the nature of second language acquisition have proved quite beneficial to learners of English, these breakthroughs have not borne many implications to Vietnamese adult learners, especially in the field

of pronunciation training This can partially be attributed to the lesser significance adult learners embrace as compared to children or young learners Additionally, as agreed by many researchers, phonology acquisition is considered as the least universal aspect of language and mostly affected by the interaction of L1 and L2 Thus, the present pronunciation training approach does not prove sufficient for Vietnamese adult learners In this context, this thesis attempted to explore the effects of refining perceptual abilities on improving the pronunciation of English non-native contrast pairs among Vietnamese adults

In the study, it was found that L2 perception plays an important role in language production and acts as a mediator between language input and output The result of the study suggests that, although high perceptual ability does not always guarantee highly intelligible sound production, it is a prerequisite for pronunciation improvement, at least to adult learners, whose L1 skills and knowledge has deeply inculcated in their L2 reflection While the parallel correlation of speech perception and production may not be guaranteed, there exists a strong relationship between these two processes

Based on the findings, some suggestions were made with the hope to integrate the new teaching approach into the textbooks The suggestions are on how teachers deal with pronunciation teaching in the classroom and how syllabus designers can allocate the materials in Vietnam situation for the learners to get the most from the current textbooks

iv

TABLE OF CONTENT

STATEMENT OF ORIGINALITY i

ACKNOWLEDGEMENT ii

ABSTRACT iii

TABLE OF CONTENT iv

LIST OF TABLES AND ABBREVIATIONS ix

TABLES ix

ABBREVIATIONS ix

LIST OF FIGURES AND CHARTS x

Chapter 1 INTRODUCTION 1

1.1 RATIONALE 1

1.2 STATEMENT OF THE PROBLEM 2

1.3 BACKGROUND TO THE STUDY 5

1.3.1 The practice of pronunciation teaching to adult learners in HCMC 5

1.3.2 Propositions on the relationship of speech perception on speech production 6

1.4 PURPOSES OF THE STUDY 8

1.5 RESEARCH QUESTIONS 8

1.6 SIGNIFICANCE OF THE STUDY 9

1.7 DEFINITIONS OF TERMS 9

1.7.1 Phonological contrast 9

1.7.2 Phoneme, phone, consonant and vowel 9

1.7.3 Phonemic perception 10

1.8 ORGANIZATION OF THE THESIS 10

Chapter 2 LITERATURE REVIEW 11

v

2.1 THE RELATIONSHIP BETWEEN SPEECH PERCEPTION AND SPEECH PRODUCTION

11

2.1.1 Propositions concerning perception-production relationship 11

2.1.2 Empirical evidence of perception-production relationship 13

2.1.2.1 Evidence from Spanish speakers of English 14

2.1.2.2 Evidence from adult learners of French 16

2.1.2.3 Evidence from bilinguals of Asian languages 18

2.1.3 Counter-evidence of perception-production relationship 22

2.1.3.1 Counter-evidence from Swedish and Finnish learners of English 22

2.1.3.2 Counter-evidence from Japanese speakers 23

2.1.4 Discussion 24

2.2 THE NATURE OF L2 SPEECH PERCEPTION 25

2.2.1 Speech Learning Model 25

2.2.2 Perceptual Assimilation Model 27

2.2.3 Feature Competition Model 28

2.2.4 Summary 30

2.3 COMPARISON OF ENGLISH AND VIETNAMESE SOUND SYSTEM 31

2.3.1 Differences in syllable composition 31

2.3.2 Differences in distinctive features 32

2.3.3 The contrast pair /t~/ and /d2/ 33

2.4 HIGH-VARIABILITY METHOD – THE ELICITATION PROCEDURES FOR ENHANCING PHONETIC PERCEPTION 34

2.5 PRINCIPAL FACTORS INFLUENCING THE PROCESS OF SPEECH ACQUISITION 37

2.5.1 Native L1 transference 37

2.5.2 Exposure to the second language 39

2.5.3 Motivation 39

vi

2.6 SUMMARY 40

Chapter 3 METHODOLOGY 42

3.1 RESEARCH QUESTIONS 42

3.2 RESEARCH DESIGN 42

3.2.1 Research participants 42

CHART 3.1 The process of subject selection 44

3.2.2 The speakers 45

3.2.3 Research procedure 45

3.2.3.1 The stimuli appraisal 47

3.2.3.2 Experimental teaching 47

3.2.3.3 The collection of perception data 48

3.2.3.4 The collection of production data 51

3.2.3.5 The treatment of perception and production data 52

3.2.3.6 The assessment of production data 52

3.2.3.7 The treatment of the assessment data 54

Chapter 4 DATA ANALYSIS AND FINDINGS 56

4.1 DATA ANALYSIS 56

4.1.1 The analysis of perceptual training outcomes 56

4.1.2 The analysis of native listeners‟ preference judgement of the subjects‟ production 60

4.1.3 The analysis of native listeners‟ identification of minimal pairs 63

4.1.4 The relationship between the perception and production of /t~/ and /d2/ 64

4.2 Findings 70

4.3 Summary 72

CHAPTER 5 RECOMMENDATIONS AND CONCLUSION 73

5.1 RECOMMENDATIONS FOR TEACHERS OF ENGLISH 73

vii

5.1.1 Teacher should apply a balanced focus on both speech perception and motor skill in

pronunciation training 73

5.1.2 Teachers should follow a systematic procedure in teaching pronunciation 74

5.2 Recommendations for syllabus designers 76

5.2.1 The syllabus should provide more time allotment for the pronunciation training 76

5.2.2 Syllabus designers should take into consideration specific influence of L1 on L2 acquisition 77

5.2.3 Syllabus designers should recognize the importance of aural input in oral production 78

5.3 Limitation 78

5.4 Recommendation for future research 79

CONCLUSION 81

REFERENCES 84

APPENDICES 93

APPENDIX 1 SURVEY 93

APPENDIX 2 WORD LIST FOR RECRUITMENT TEST 97

APPENDIX 3 LIST OF UNFAMILIAR WORDS FOR SUBJECTS 98

APPENDIX 4 PERCEPTUAL PRETEST (and POST-TEST) (SPEAKER 4) 99

APPENDIX 5 GENERALIZATION TEST 1 (SPEAKER 6) 100

APEENDIX 6 GENERALIZATION TEST 2 (SPEAKER 1) 101

APPENDIX 7 TRAINING SESSIONS 102

A TRAINING SESSION 1 (SPEAKER 1) 102

B TRAINING SESSION 2 (SPEAKER 2) 103

C TRAINING SESSION 3 (SPEAKER 3) 104

D TRAINING SESSION 4 (SPEAKER 4) 105

E TRAINING SESSION 5 (SPEAKER 5) 106

F TRAINING SESSION 6 (SPEAKER 1) 107

viii

G TRAINING SESSION 7 (SPEAKER 4) 108

H TRAINING SESSION 8 (SPEAKER 3) 109

APPENDIX 8 PRODUCTION TEST (Pre-test and Post-test) 110

APPENDIX 9 RESULT OF PERCEPTUAL TRAINING 111

A Pre-training perceptual test result 111

B Post-training perceptual test result – Group A (experimental group) 112

C Post-training perceptual test result – Group B (control group) 113

APPENDIX 10 RESULT OF NATIVE LISTENERS‟IDENTIFICATION OF MINIMAL PAIRS 114

A Native listeners‟ identification test result - Group A (experimental group) 114

B Native listeners‟ identification test result - Group B (control group) 115

ix

LIST OF TABLES AND ABBREVIATIONS

TABLES

Table 2.1 Differences between the Vietnamese and English syllables

Table 2.2 Distinctive features specifications of the Vietnamese consonants

Table 2.3 Distinctive feature specifications of the English consonants

Table 4.1 Perceptual test result at pretest phase

Table 4.2 Perceptual test result at posttest phase

Table 4.3 The two-way ANOVA test result of perceptual test scores

Table 4.4 The T-test result of perceptual test scores

Table 4.5 Native listeners‟ identification test result

Table 4.6 The T-test result of native listeners‟ identification of minimal pairs

Table 4.7 Individual Vietnamese trainee perception and production accuracy scores at pre-test and at post-test

ABBREVIATIONS

1 L1 First language

2 L2 Second language

3 HCMC Ho Chi Minh City

4 ESL English as the Second Language

5 EFL English as the Foreign Language

x

LIST OF FIGURES AND CHARTS

FIGURES

Figure 1.1 Hewlett‟s (1990) model of phonetic and phonological processing

Figure 4.1 Percent correct perceptual identification performance for the experimental group and the control group at the pretest, post-test and two tests of generalization Figure 4.2 The native speakers‟ preference judgement of the subject production

Figure 4.3 (a) Vector plot of individual Vietnamese subjects‟ perceptual identification accuracy and production identification accuracy from pretest to post-test (Subjects 1-5)

Figure 4.3 (b) Vector plot of individual Vietnamese subjects‟ perceptual identification accuracy and production identification accuracy from pretest to post-test (Subjects 6-10)

CHARTS

Chart 3.1 The process of subject selection

Chart 3.2 The research procedure

1

Chapter 1 INTRODUCTION

This thesis reports the result of an experiment conducted to investigate the relationship between the perception and production of the English contrast pair /t~/ and /d2/ among Southern adult Vietnamese speakers of English, with an aim to develop new techniques for the modification of the structure of learners‟ phonetic system, thus facilitate communication outcomes This section takes into consideration the rationale for the study, statement of the problem as well as presents the background for the study The introduction also points out purposes of the study and introduces research questions and hypotheses

1.1 RATIONALE

In recent years, the goal of language teaching has been pivoted around endowing learners with communicative competence Hymes (1985) proposes that communicative competence is exhibited through the ability to handle effectively real, natural conversations and necessarily encompasses linguistic and pragmatic aspects Apart from grammatical, lexical and discourse mastery, phonological competence remains an indispensible part at the linguistics side The generation of acceptable pronunciation has been acknowledged as a crucial integration of any instances of successful communication In fact, Garrigues (1999) essentially points out that the foundation of effective spoken communication is good pronunciation If a converser utters well- and clearly-enunciated tokens, it should be much easier for their audience interlocutors to grasp their meaning with relative ease Otherwise, mispronunciation stands high probability of leading to misunderstanding and, in quite a few cases, leading to utter communication breakdown or termination In this sense, pronunciation has recently attracted more and more attention of language researchers and practitioners, and has claimed more and more substantial share in EFL curriculum It is recognized as a fundamental skill which students should acquire,

“phonological loop” (Newton, 2009) In essence, the phonological loop is the practice

of the brain repeating the words or the phrases over and over again, either consciously

or subconsciously, to retain it in the working memory or to help it transfer to the term memory As far as this concerned, a learner without a consistent and stable pronunciation of a word cannot allot it to the long-term memory because it cannot be held in the phonological loop Conclusively, the phonological patterns, apart from the grammatical structure representations, exert a substantial effect on the size of the working memory among learners of a second language It is thus important to attribute substantial attention to pronunciation in the EFL classroom to help learners quickly establish stable pronunciation and internalize the patterns and rules that work within the language

long-In consideration of the roles of intelligible pronunciation, it is proved amply justified for ESL/ EFL learners to systematically learn to produce consistent acceptable sounds from the very beginning of L2 learning Also, teachers should attach significance to the practice of pronunciation teaching and learning in EFL classroom Suitable remedial strategies must be planned and pondered should any arisen issue intervene in this acquisition process

1.2 STATEMENT OF THE PROBLEM

Communicative teaching approach and student-centred philosophy have been introduced into Vietnam for the last ten years and with this trend, pronunciation training has gained a larger portion in communication-oriented syllabi However, the influence of this movement has not extended as far as expected It is observed that the

3

introduction of English phonological systems in the current widely-used English textbooks may, to some extent, benefit young learners, yet it seems insufficient to make noticeable modifications to adults‟ existent sound patterns

On addressing the complications of training adult Vietnamese learners to produce comprehensible English, a lot of language practitioners and teachers have aired their deep concern about the situation With many foreigners‟ comment that Vietnamese and Thai speakers of English are among the most incomprehensible speakers in Asia, Ha C Tam (2002) admits the failure of the present communicative teaching oriented syllabus in improving adult Vietnamese learners‟ pronunciation In her observation, although many of them can achieve high proficiency in other language areas as reading and writing, just a few can make themselves easily understood in direct interactions with foreigners She also attributes this failure to several phonological factors, among which mal-production of segmental components,

or discrete sounds, is the most basic aspect causing communication breakdown, especially with learners of low levels Also, some foreign ESL teachers in HCMC (Guan, 2010; Douglas, 2009; Dalana, 2007) reflected on the experience of their struggle to rectify Vietnamese adults‟ articulation and maintained that the improvement has not been quite significant over the last ten years

As a teacher of English in EFL context for quite a few years, the writer of this thesis has acceptably come to the realization that the majority of Vietnamese learners, especially those commencing to experience L2 exposure after puberty, often bewilder native speakers by mispronouncing English sounds Among common pronunciation mistakes, it often remains the case that these learners, for one reason or another, either utter one sole phonetic representation for two English contrastive sounds or distort the sounds by replacing them with L1 alternatives This usually occurs when the two L2 sounds are not contrastive in L1 or when an L2 sound shares some prominent features

4

with an L1 equivalent The production of misleading phone representatives, as a result, hinders the learners from maintaining smooth interactions and this communication gap cannot simply be compensated by their competence in other linguistic fields This breakdown has also taken its toll when it comes to the issue of vocabulary learning So, their poor vocabulary repertoire can partly be blamed on the failure to retrieve a consistent and stable phonetic image in the mind By all means, something should be done immediately to ameliorate this situation

The poor pronunciation among Vietnamese adult learners can be partially traced back to two factors Firstly, it can be attributed to the largely incompatible phonological systems in Vietnamese and English languages, which embrace different phonetic characteristics and prominent features In fact, it is noticed that adult learners are the most vulnerable to non-native phonetic contrasts in L2 learning, which usually

leads to mispronunciation (e.g Tahta et al., 1981; Flege and Hillenbrand, 1987) For example, Japanese learners have troubles with the contrast pair /r/ - /l/ (Logan et al.,

1991) while Chinese speakers often cannot distinguish between /v/ and /w/ Considering Vietnamese speakers, contrastive works of Nguyen D Liem (1989), Cheung (1986) and So & Dodd (1992) reveal that the following non-native phonetic contrasts will cause great trouble and confusion for learners: /t~/ and /d2/, /8/ and /5/, /n/ and /7/, /z/ and /2/, /d/ and /5/ Usually, they will produce one sole phonetic representation for the two members of each pair Secondly, it has been argued that adult learners, with their specific characteristics and learning styles, should have more speech input in order for them to produce acceptable output Thus, special treatment should be offered to these adult learners, especially in their dealing with certain non-native, or novel, contrasts

Although there have been extensive research and studies on the pronunciation remedial strategies for adult learners on dealing with novel L2 sounds in Japan,

5

China, Korea and other Asian countries, there seem to be few investigations into the issue in the context of Vietnam Since Vietnamese language also embraces distinctive features that set itself apart from English and other Asian languages, an insight into a certain approach to rectify Vietnamese adults‟ production of certain non-native sounds proves to be worth consideration

1.3 BACKGROUND TO THE STUDY

1.3.1 The practice of pronunciation teaching to adult learners in HCMC

In HCMC, there exist a large number of language centers which offer a lot of general English courses for adult learners like Thanh Nien Language center, SIFONE, Language center of Pedagogy University, Vietnam-Australia language School, ELITE and VUS

Except for some minor changes, the teaching syllabi for these courses usually conform to the available syllabi presented in certain textbooks – New English File, New Cutting Edge, International Express, New Lifelines Although all of these textbooks follow modern communicative teaching philosophy and ethics, an insight into their pronunciation syllabi reveals these two main issues Firstly, since the books were designed for a wide range of target markets, the treated segmental aspects are not closely related to the problem a certain learner with a certain L1 background may cope with In fact, all the phones in the English sound system are tackled with equal amount of time and focus throughout the whole textbook series Thus, if teachers stick

to the existent syllabi, the challenging L2 sounds will not be highlighted and sufficiently dealt with Secondly, since these textbooks are meant to address a wide range of target learners, adults seem not to get a lot of benefits from the presented pronunciation teaching approach Therefore, the textbooks and the syllabi should be suitably adapted to match with specific learners with specific pronunciation problems

6

Also, from her class observation, the writer of this thesis noticed that the pronunciation teaching to adults is quite similar to other target learners Usually, the language input, or language exposure, does not guarantee acceptable production of novel L2 sounds Endowed with only one or two times listening to native speakers before producing the sounds, the adult learners still cope with a lot of troubles and challenges with some contrast pairs like /t~/ and /d2/, which are not contrastive in L1 Furthermore, teachers generally concentrate mainly on training motor aspects of phone production such as lip shape, tongue position and jaw movement rather than provide students with audio discrimination cues between the sounds In other words, focus is usually shifted on the articulation phase of sound production and does not tackle the deeper language breakdown – L2 perception

All in all, an overview of the present teaching practice has revealed some flaws

in pronunciation training to Vietnamese adult learners Firstly, the contemporary prevalent teaching syllabus does not address the specific problems those learners cope

in learning L2 phonological system, or the specific sounds they characteristically have troubles with, for example /t~/ and /d2/ Secondly, adult learners have not received suitable remedial strategies for acquiring these novel L2 sounds In fact, teachers usually dedicate most of their effort on training articulatory motor skills and do not focus much on sound perception – the input side – in their pronunciation training Hence, the philosophy behind this pronunciation teaching practice- the covert belief and assumption that sound production can be improved by mere direct motor training-should be critically reflected and revised

1.3.2 Propositions on the relationship of speech perception on speech production

A lot of linguists and researchers, in their studies of the nature of speech production, draw on the impact of enhanced sound perception on refined sound production Actually, the linkage between speech perception and production has

7

arisen as a long-standing issue in speech science and experimental phonetics, and has intrigued a variety of theories concerning its nature (e.g Liberman and Mattingly, 1989; Stevens and Blumstein, 1981; Diehl and Kluender, 1989) Apart from the theoretical concern of perceptual training in the classroom practice, much of the empirical evidence justifies the corresponding need to study the relationship between perception and production An increasing number of positive findings of the relationship have supported this perceptual-based hypothesis of speech accuracy For example, Schneiderman, Bourdages & Champagne (1988), as studying the effectiveness of perceptual training on speech accuracy, found that learners of French with different L1 backgrounds, including English, Chinese, Tamil, Hindu, Turkish, Spanish and German, achieved an encouraging progress in discriminating and producing French words and phrases after the training Statistical analysis also demonstrated a significant correlation between perception and production post-training In much the same way, it was reported that perceptual training has improved, though inconsistently, the intelligibility of the production output of Japanese learners of English (Bradlow et al., 1997)

Based on this argument and evidence, the process of speech production is closely-linked to speech perception In this way, the amount of exposure to language and the quality of the language stimulants in the speech perception stage greatly influence the quality of speech production Specifically, in training Vietnamese adult learners to acceptably produce the non-native phonetic contrast like /t~/ and /d2/, it may be beneficial if teachers plan some suitable strategies to improve their perception first Since little preliminary research has been done into the linkage between sound perception and sound production in Vietnam context, the proposition should be further tested before it is applied in reality

8

1.4 PURPOSES OF THE STUDY

In consideration of insufficient empirical evidence of the relationship between sound perception and sound production in Vietnam context, the present research was conducted to:

(i) explore how the production of /t~/ and /d2/ among Southern Vietnamese speakers change in correlation with the change in their perception of the two sounds;

(ii) investigate the relationship between sound perception and production with the case of /t~/ and /d2/;

(iii) offer some possible recommendations on the current pronunciation teaching of non-native sounds to Vietnamese adult learners

1.5 RESEARCH QUESTIONS

Based on the purposes of the study, an experiment was designed to investigate the relationship between speech perception and speech production in the case of /t~/ and /d2/ among Vietnamese adult learners Since L1 background also exerts some influence on the production of L2, the subjects taking part in the experiment were selected within homogenous population – adult speakers with Southern Vietnamese accent This investigation was meant to answer the following questions

1 How does the production of /t~/ and /d2/ among adult Southern Vietnamese speakers change in correlation with the change in their perception of these two sounds?

2 Might there be a positive relationship between the perception and production

of /t~/ and /d2/?

9

3 What are the methodological implications for the teaching of non-native English sounds to adult Vietnamese learners?

1.6 SIGNIFICANCE OF THE STUDY

If the relationship between the perception and production of the contrast pair /t~/ and /d2/among Southern Vietnamese speakers are to be proved, then this result lends significance to the practice of pronunciation teaching in Vietnam Firstly, it provides a novel insight into the principles behind pronunciation teaching, at least at segmental level Secondly, it serves as a guideline for teachers and syllabus designers

to plan their teaching strategies for Vietnamese adult learners Thirdly, it can serve as

a background for future studies on the issue of improving pronunciation among Vietnamese learners

1.7 DEFINITIONS OF TERMS

1.7.1 Phonological contrast

Gimson (1989), Nunan (1991) and Spencer (1997) share the similar view that a

phonological system (i.e sound system) of a language is built upon the idea of sound

contrasts By selecting one type of sound as opposed to another, one can distinguish

one word from the other In other words, sound contrasts serve to distinguish

meanings in a language For instance, in English, one can distinguish the words

“char” and “jar” by choosing the first consonant either as /t~/ or /d2/ and such

distinction between the sounds /t~-d2 / contributes to a contrast, /t~-d2 / is thus known

as a "contrast-pair”

1.7.2 Phoneme, phone, consonant and vowel

Crystal (1991) defines "phoneme" as the minimal unit in the sound system of a language while a "phone" is the physical realization of a phoneme The former thus

represents an abstract level of analysis while the latter refers to the actual articulatory

10

movement (of lips, tongue and vocal tract, etc.) and acoustic events heard in speech

In general, there are two types of phones - consonants and vowels Phonetically, consonants are those sounds produced by narrowing the vocal tract so that airflow is either completely blocked or restricted and audible friction is heard Vowels, on the other hand, are those sounds produced in the absence of such vocal tract constriction

1.7.3 Phonemic perception

In defining "phonemic perception", many researchers (e.g Ferguson, 1975; Juiesz

& Hirsch, 1972; Shvachkin, 1973) acknowledge that it includes three aspects -

auditory sensations, interpretation and classification of speech sounds into

meaningful contrastive categories which signify the meaning of a word Bearing the

same principle, Barton (1980) concisely defines "phonemic perception" as the

"process of discriminating and classifying speech sounds into minimal units that signify meaning differences (1980, p.97)

1.8 ORGANIZATION OF THE THESIS

The thesis is divided into five chapters The First chapter presents the introduction to the thesis, which focuses on rationale of the research and problem addressing, points out the background to the study, overviews the current practice in L2 phonology teaching and declares the purposes of the study The Second chapter gives a literature review of the previous findings of the perception-production relationship, the high-variability training approach and proposed theories of L2 speech perception Chapter Three describes the research design Chapter Four of the thesis reports on the results and findings of the study while the Fifth chapter elaborates on possible implications and recommendations drawn from the findings

11

Chapter 2 LITERATURE REVIEW

The present study is an attempt to investigate the relationship between the perception and production of /t~/ and /d2/ and explore possible changes to the pronunciation teaching of non-native English sounds to Vietnamese adult learners Thus, the purpose of the literature review is to find out the nature of L2 sound acquisition and how an L2 novel sound is integrated into the existent L1 phonological system, which helps to explain why the present teaching approach is not sufficient to rectify adult learners‟ pronunciation habit of non-native sounds The literature review

is also meant to revise the principles and empirical evidence regarding the relationship between speech perception and production as the background for the present study

2.1 THE RELATIONSHIP BETWEEN SPEECH PERCEPTION AND SPEECH PRODUCTION

This section takes into consideration the established argument, both theoretically and empirically, in regards to the correlation between L2 perception and production The section focuses on the heterogeneous stances proposed by different theorists as well as on related experimental evidence provided in the works of some renowned researchers

2.1.1 Propositions concerning perception-production relationship

The association between speech perception and speech production has ever been

a concern in the field of speech science, with many researchers long since divided over the core nature of this controversial relationship Some language theorists postulate that the two processes- speech perception and speech production- are directly linked and one is the causative factor of the other For instances, the typical

argument of motor theorists (e.g., Liberman et al., 1967; Liberman and Mattingly,

1985; Liberman and Mattingly, 1989) goes that the sound perceived by listeners is stored in the form of its respective articulatory gestures, which govern the way that

12

sound is produced in later instances Based on this theory, there exists a specific phonetic segment that stores speech units in codes of articulatory gestures, thus governing both speech perception and production Articulatory gestures, admittedly, serve as the single mediator for speech perception and production, activating the conversion between these two processes In other words, the interconnection between these two processes is inextricable

Set apart from this stance are those theorists who lend more autonomy and separation to these two processes of speech communication Such attitude can be found in works of Steven and Blumstein (1981) or Diehl and Kluender (1989) These acoustic-auditory theorists suppose that speech perception acts upon an acoustic channel separated from the articulatory gestures regulating speech production However, this approach also presupposes that there exists an indirect connection between speech perception and production due to an interaction and inter-modification between acoustically-coded speech and auditory feedback mechanisms activating during producing speech Based on this theory, the stored information about acoustic properties of a perceived sound is retrieved for the produced sound to

be benchmarked against Hence, although speech perception does not exert a direct influence on the quality of speech production, the ultimate speech outcome does very much depend on the extent to which the listener-speaker can process the acoustic properties of sounds

A third theory, the direct-realist approach (Kiparsky & Menn, 1990) to speech perception, proposes that the perceived sounds are retained, first and foremost, not under the form of acoustically defined targets, but rather, in terms of gesturally defined targets At one end, listeners inadvertently perceive the articulatory gestures

of the sounds, which represent themselves as a set of distinctive gestural structures imparted into the acoustic system At the other end, this listener-speaker, during speech production, will aim to achieve these articulatory gestures by employing the

13

existent incorporated information Hence, the argument of the direct-realistic approach is that the strong cohesion between speech perception and production is deeply-stemmed from their common communicative goal In this way, the direct-realistic approach holds a shared perspective with the motor theory about the direct link between the two processes Yet, this approach does not imply a specific phonetic segment that determines the direct perception –production link Rather, it proposes that the link between speech perception and production is the result of the integration

of speech perception and production systems in order to achieve common communicative goal

Differences and disparity taken apart, these theorists have, either explicitly or implicitly, shared the thrust of argument about the mutual impact between language perception and production As far as this study concerns, these stances could represent different illustrations of different aspects and dimensions involved in the intricate process of language perception-production Might it be the case or not, the relationship between the two processes is inextricable and carries significant methodological implications Thus, it possibly is the case that the improvement in speech production can be fostered through perception refinement

Concerning the present study, the theoretical propositions have provided some guidelines for the improvement of the production of novel L2 sounds among Vietnamese adult learners Since the mere concentration on motor training skill does not prove sufficient to yield significant pronunciation enhancement, the mutual impact of speech perception and production can imply the inclusion of perceptual development in pronunciation training

2.1.2 Empirical evidence of perception-production relationship

In light of various theoretical works, the linkage between speech perception and production is well-justified, yet how this relationship exposes itself in learners‟ overt

14

linguistic performance is such spectrum that cannot be unfolded solely with theoretical hypothesis Thus, developed in parallel with the previous theoretical stances on the issue, empirical studies of the correlative performance between perception and production thrived in the 1980s, covering a wide range of language spectrum and target subjects (e.g Flege & Eefing,1987; Flege, Munto & Fox, 1994) These studies have helped provide forceful empirical facts and figures for language researchers and practitioners to gain revealing insight into the relationship between speech perception and production

2.1.2.1 Evidence from Spanish speakers of English

Studies on the Spanish subjects have undoubtedly accounted for quite a voluminous proportion in the growing body of literature on L2 perception and production In 1987, for instance, Flege & Eefting observed the disparity between two groups of subjects, one consisting of adult English monolinguals, the other comprising native English speakers, aged between 15 and 20, with about 6 years of English exposure, who in 90% speech instances produced the English /d/ and /t/ with

a Spanish accent The subjects in each group were due to take part in a perceptual decision task in which they made forced-choice decision on whether the presented auditory were English /da/ or /ta/ in the corresponding voice-onset time continuum Results revealed that, in correlation to their performance, the native Spanish speakers performed a relatively poorer perceptual outcome as compared to the English controls Based on these findings, the researchers arrived at two conclusions - firstly, there exists a parallel relationship between perception and production; and secondly, the internalization of L1 may influence the construction of the perceptual representations of L2 speech sounds Flege & Eefting further assumed that although the two consonants /d/ and /t/ were also contrastive in Spanish, the category boundary

of Spanish /d/ and /t/ was different from that of English Consequently, the Spanish

Discussion

By all means, the two studies above illustrate the stance that there remains a divergence in the classification of speech sounds into different categories among various languages In other words, two sounds proving to be contrastive in one language may be treated as belonging to one classification in another language Even

if the contrast remains a case in a language, the two sounds may display different

16

phonological distance and groupings L2 learners, after internalizing the sound system

of their L1, are conditioned to discriminate phonological features in a definite way and this would affect their construction of perceptual representations of their L2 sounds The result of these studies bears the implication that the inculcation of the L2 speech system, to some extent, can be facilitated by improving sound perception

Nevertheless, a deeper insight into these two experiments reveals that with the design of the study, it would be admittedly unjustified to assume a correlative relationship between speech perception and production On the one hand, the observations were implemented and the comparisons were made between two heterogeneous groups of subjects in terms of their language background In other words, with one group consisting of sheer native speakers and the other of non-native speakers, the disparity in speech perception outcomes between the two populations may be considered as the projection of their disparity in general language competence and can have nothing to do with their speech production capability On the other hand,

an outward correlation between speech perception and production cannot guarantee, though may imply, an inward direct interconnection between these two processes Chances are that there exists a third influential factor which can exert a consistent and uni-directional effect on both speech perception and production Thus, other researchers employed more elaborated experiments to venture on this perspective to gain a more justified conclusion

2.1.2.2 Evidence from adult learners of French

Other linguists, encouraged by the findings with Spanish speakers, began to specially adapt the research procedure into investigating the perception-production relationship in other L2 languages Schneiderman, Bourdages & Champagne (1988) conducted an experiment and analysed the disparity in the perception and production

of adult learners of French before and after an extensive training program on

17

pronunciation The training comprised 12 one-hour sessions, the first 6 of which were focused solely on perceptual coaching with accompanied oral tutelage beginning from the seventh session The subjects for the study, aged between 18 and 30 and at beginner and low-intermediate levels of French, were drawn from different L1 backgrounds These subjects were then segregated into experimental and control groups, each participating in the discrimination and production tests before and after the training sessions The production test examined their ability to imitate and copy French words, phrases and sentences in terms of segmental and suprasegmental features

The experiment yielded two important outcomes Firstly, there existed a strong correlation between the subjects‟ discrimination and production scores both before and after the specialized coaching Thus, chances are that perception and production are parallel at various stages of L2 learning Secondly, it was noted that the trained group registered a certain level of gains in both perception and production in all aspects with the exception of the discrimination of rhythmic patterns, whereas the untrained group did not exhibit any evidence of such improvement The researchers then came to the conclusion that the production performance of a second language learner is positively correlated, at least at the segmental level, with their perception ability at various stages of second language learning

in their production outcomes can be contributed mainly to the linguistic experience

18

gap created during the training sessions The findings of the study provide more direct implications for both the learning mechanism and methodological directions of L2 phonology learning In particular, the need for perceptual training in improving comprehensible pronunciation is further justified

2.1.2.3 Evidence from bilinguals of Asian languages

The theoretical formulations have served as an inspiration for consecutive research into the generality of the perceptual-based hypothesis of speech accuracy in some Asian languages, including Korean, Japanese and Chinese

Observations on Korean speakers

In one such pioneer attempts, Borden, Gerber & Milsark (1983), examined the perception, production and the relationship of the /r-l/ contrast in Korean adults learning English The subjects encompassed in the study included those aged between

19 and 48 years old, with length of residence in America from 1 month to 3 years and had different years of exposure to English use This target group then underwent different tests before and after an intensive program of English speech training The training session included exercises on three aspects, which pivoted around the production of English /r/ and /l/, forced-choice identification of the same contrast and discrimination of pair stimuli comprising /r/ and /l/ After the program, their perceptual and production performance were re-evaluated The post-training perception test was of the forced-choice identification type that required the subjects

to respond to the perceived word stimuli as belonging to /r/- or /l/-containing group Consistent with Schneiderman, Bourdages & Champagne‟s (1988) findings, results indicated a strong correlation between production and perception and in some cases, the production errors appeared to parallel the perceptual errors The researchers then

19

further suggested that improvement in perception apparently acts as a prerequisite to that in speech production

Observations on Japanese speakers

In other attempts to examine Japanese speakers, Best & Strange (1992) and MacKain, Best & Strange (1981), in a force-choice identification task, investigated the perception of English contrast /r-l/ between American English speaking subjects and Japanese learners of English The researchers then separated the Japanese subjects into experienced and non-experienced divisions The former were those who reported having indulged in prior extensive speech training with native American speakers while the latter did not share the privilege The tested consonants /r/ and /l/ were basically selected based on the observation of their non-native production of these two phones The subjects were due to report on how they process the perceived sound by attributing to the initial consonant of monosyllabic word they heard to either /r/ or /l/ Based on the result, it was suggested that their perception outcomes remained parallel with their production performances Thus, the experienced group proved to outdo the non-experienced division in the perception This experiment has added to the existent abundant literature in confirming the non-detachable bond between speech perception and production, though this study apparently serves to assess the impact of production

on perception The challenges posed by the two phones /r/ and /l/ to Japanese speakers can be traced and explained by the fact that they are not contrastive in Japanese Following the experiment procedures of Schneiderman, Bourdages & Champagne(1988), Bradlow, Pisoni, Akahane-Yamada & Tohkura (1997), based on the result of their study, emphasized that the enhancement of Japanese subjects after several intensive perception training session undeniably lead to the improvement in their pronunciation outcomes L2 native speakers of Japanese, aged from 19 to 22, with relatively similar amount of exposure to English, participated in a four-week

20

perception training module The input for this training was naturally-spoken tokens of words, phrases and sentences After the training, with a two alternative forced-choice test, the subjects‟ perception performances were assessed and proved to exhibit a significant improvement Additionally, although there had been no direct training or input for their production, these Japanese speakers were also judged by various native speakers to have achieved greater intelligibility in their speech tokens Bradlow et al then insisted on a more causative relationship between speech perception and production, which imposed an important impact on the subsequent practices of English teaching to Japanese learners

Another observation on Korean and Japanese speakers

Despite the previous findings, in combination with other cumulative positive evidence, some language researchers (e.g Flege, Takagi & Mann, 1996; Flege, Bohn

& Jang, 1997) still cast some doubt on the validity of these results with the argument that there had been little control, if no at all, over some confounding affective variables such as the subjects‟ starting age of L2 learning, length of L2 experience and intensity of native speaker contact In response to this reaction, in 1997, Ingram & Park carried out a carefully-controlled experimental design among Japanese and Korean learners of English The study was meant to investigate the perception of non-native vowels of these subjects, who were observed to produce the English /`/ and /@:/ without any distinction In order to alleviate the effect of some confounding variables, these researchers grouped the participants into different divisions with different length and intensity of English exposure The experienced group had resided

in Australia for more than 5 years, thus having extensive exposure to English language, whereas the non-experienced group had underwent only one-year of residence in Australia, thus having limited contact with native language The researchers adopted a modified forced- choice perception test In listening to a series

of input, the participants were asked to tick out one of the five English words realized

21

in the form of /h-d/ where the blank was embedded with one of the vowels /i:/, /1/, /e/, /@:/ or /`/ In line with their production, these subjects of both groups, with their non-native pronunciation, exhibited great difficulties in distinguishing between the two phones /@:/ and /`/, however different their intensity of language exposure might prove However, in consistent with their native production of the three vowels /i:/, /1/and /e/, these speakers showed no difficulties in perceiving these phones These experiment outcomes and observations yield two implications: first, speech production, to some extent, always stays in parallel with speech perception and secondly, those categories of sound contrastiveness in L1 play an important role in the process of acquiring novel phones in L2 Apparently, these subjects underwent no pains in dealing with the three sounds /i:/, /1/ and /e/ since they are also contrastive in their language However, due to the non-contrastiveness of /`/ and /@:/ in reference to Japanese and Korean language, these subjects showed difficulty in tuning in to the distinction The researchers then further proposed that they two phones can be stored

in the same sound category which forms the same basis for their production

Observations on Chinese speakers

Flege (1989), among the scarce studies on Chinese subjects, adopted a experimental design to investigate the cross-language consonant perception of Chinese subjects, the average age of whom was about 30 years and who had about two-year residence in the US A close observation revealed that these speakers usually had great difficulty in distinguishing the English contrast /t-d/at the word-final position when a burst release (a cue to signal /t-d/contrast in Chinese) was removed The non-contrastiveness of these two phones at the final position in their native languages was also considered as a plausible explanation for such transfer effect

quasi-In summary, these investigations into a wide range of language spectrum and target subjects have partially unveiled the correlative nature of language perception and production In another aspect, they have contributively reinforced the belief that

22

this relationship remains a popular phenomenon across different language registers

However, other researchers find their experimental results yielding in the opposite direction and thus claim the invalidity of this correlation

2.1.3 Counter-evidence of perception-production relationship

2.1.3.1 Counter-evidence from Swedish and Finnish learners of English

Although the correlation between speech perception and production has been evidenced in many studies and experiment, some other contemporary studies have pledged different results For instance, in 1985, Flege and Hillenbrand investigated a group of native and non-native speakers of English in their perception performance of the voicing feature of word-final fricative /s/ and /z/ The non-native participants were consisted of French, Swedish and Finnish speakers, who all started to experience contact with French at junior high school The French subjects were treated as the controls since the pair /s-z/‟s contrastive feature of voicing is prevalent in French, as with English, while such contrast does not exist in Swedish or Finnish The participants, in a given two-alternative forced-choice test, were due to identify the voicing feature of a series of words by marking the final sounds as either under /s/ or /z/ heading Surprisingly enough, although the /s-z/contrast does not prove a case in Finnish or Swedish and these subjects cannot acquire a native-like pronunciation for these two phones, they were still able to discriminate, at an acceptable level, between these two foreign sounds The researchers then drew a conclusion that speech perception and production can be two independent processes operated by distinctive mechanisms and factors, thus rejecting any linkage between these two channels However, a thorough investigation into this study reveals that an outward disparity between the perception and production performances does not guarantee the validity of the conclusion The causal interaction between speech perception and production may undergo some intermediary mechanisms and can be influenced by

23

some intervening factors Thus, the observed problems with the production intelligibility of Swedish and Finnish speakers might not be traced back at the perceptual level but rather, at the mapping rules between input and output phonological lexicon

2.1.3.2 Counter-evidence from Japanese speakers

Further supporting evidence for the conception of independent speech perception and production can be found in Sheldon & Strange‟s work (1982) In this observation, Japanese learners of English, aged from 19 to 31 and with length of residence in American from 20 to 30 months, were investigated on their perception and production performance of the English contrast pair /r-l/ in different word positions In the production test, these subjects were required to read out words containing /r/or /l/ and their outcomes were assessed by American native speakers At the same time, their perception was tested with a two-alternative forced-choice test Contrasting to their expectations, those Japanese subjects who had acquired native-like pronunciation of these two sounds still made significant errors in perceiving the contrast With this result, the researchers casted doubt on the alleged relationship between speech perception and production and stated that perception might not be a prerequisite for production Nevertheless, other observers (e.g Flege, 1991) argued in the later commentary that the selected subjects were not trustworthy representatives of the whole population The completely intelligible production of these two foreign phones may be the result of their previous extensive training on the production domain, thus the production of these subjects can be attributed to a different group and nature as compared to other L2 learners

In short, this section has provided different and even contrasting insights into the nature of language perception-production relationship Though some experimentalists have proposed apparently convincing argument against this

24

relationship, their counter-evidence proves ungrounded since they did not take into consideration different factors in choosing the target subjects In light of this, their results stand little chance of proving applicable to the whole population

2.1.4 Discussion

Different studies have been carried out to examine the association of the perception and production of non-native sounds across different languages In these studies, different research designs have been employed varying from non-experimental, quasi-experimental to experimental register Most of the observations agree upon the existence of the relationship between speech perception and production, though some conflicting results have also been reported

The examination on those conflicting results has lent special significance on the control of various confounding variables as planning the design for the study Irrefutably, the process of L2 speech production is affected by a complicated mechanism under the influence of different factors Hence, the result of the study can

be distorted should some main influential factors be not carefully weighed Thus, in this study, the relationship was re-investigated in the context of adult Vietnamese learners with careful consideration of various affecting factors It was hoped that the result of this study would provide more evidence to dissolve the conflicts and prove more valid in Vietnam context

In order to evaluate the relationship between speech perception and speech production among Vietnamese adult learners, the present study observed the changes

in sound production in correlation with the changes in their perception The study employed a specific perceptual training technique to create perception modification after a certain period of time Thus, an insight into the nature of L2 speech perception should be gained so as to select a proper training method and which L2 sounds deserve the training The next part of the literature addresses this particular issue

25

2.2 THE NATURE OF L2 SPEECH PERCEPTION

Different researchers have constructed models and hypotheses of the corresponding mechanism in L2 speech perception Among such proposals, models

by Flege (1991), Best & Strange (1992) and Hancin-Bhatt (1994) have been the most influential and widely cited frameworks

2.2.1 Speech Learning Model

In Flege (1991)‟s argument, LI phonological acquisition is viewed as a

“bottom-up” process of learning, while L2 phonological learning is described as a

“top-down” process Basically, this argument implies that adult L2 learners need to establish correspondences between phonic elements in L1 and L2 He also proposes that L2 speech accents are largely perceptually based, which means that speech production does very much depend on the quality of the perceived sounds Cited in Bongaerts, Van Summeren, Planken & Schils (1997), Flege coins Wode's (1993)

terms - continuous and categorical modes of perception to further explain the

disparity in L1 and L2 speech learning The continuous mode of perception allows one to focus on the subtle differences between speech sounds, for example, different realizations of /t~/ by different speakers, while the categorical mode operates on the overall sound properties with which one only attends to those acoustic cues that signal contrasts between different sound categories, for example, /t~/ and /d2/ It is the mutual interaction between these two modes of perception that allows children, in their L1 phonological acquisition, to create their L1 sound categories This is accomplished by first using the continuous mode to attend to variations of sounds to a meaning and then forming the sound categories gradually With the solid and deeply-rooted cultivation of their existing L1 sound systems, L2 learners tend to fit the new L2 phones under L1 sound categories The greater the phonetic similarity between the

26

L1 and L2 sound, the more likely that the minute differences between them are not noticed The over-reliance on the categorical mode of perception by L2 learners, together with the previously established L1 sound categories, contribute to the hindrance of correct L2 sound category formation Without the correct perceptual basis, their subsequent L2 productions would be deeply influenced

Upon predicting the challenges faced by L2 learners, Flege (1991) proposes a hypothesis to account for three operations If an incoming L2 speech sound is highly

dissimilar one to their L1 categories, learners might undergo an initial difficulty in

categorizing it However, more L2 exposure will facilitate learners to gradually form a distinct L2 perceptual representation for this sound Alternatively, when a new L2

sound is identical to their L1 sound, the L2 speakers perceive it easily because it fits

into their existent L1 sound categories Meanwhile, when the speech sounds are

similar but not identical to the L1 categories, L2 learners will classify these similar

sounds into the most similar L1 sound categories Generally, with the third operation,

it is necessary to employ a special teaching approach and a distinct training scheme so

as for the new sound system to be inculcated into the existent one Thus, the recognition of the right operation in teaching segmental phonology is of vital role

Concerning the present study, the Speech Learning Model has provided basic clarification of the hindrances challenging adult learners as they experience a new language, especially when it comes to pronunciation mastery The operation of the perceptual continuous mode curbed after the complete internalization of L1 phonological system, these learners have to juggle between maintaining the fixed L1 phonetic categories and inculcating new phones into this system This phenomenon, simply enough, has created the unresolved paradox which results in adults‟ failure to achieve highly intelligible pronunciation of non-native contrast sounds Thus, if the relationship between the speech perception and production is to be proved and the quality of produced sound does very much depend on the quality of the perceived

27

sound, then the mere motor skill training in pronunciation teaching does not stand highly effective in modifying the articulation patterns among adult learners This issue further justifies the role of perception ability in forming comprehensible pronunciation

2.2.2 Perceptual Assimilation Model

Best & Strange (1992), sharing a similar view as Flege, further proposed the Perceptual Assimilation Model to predict and explain learners‟ difficulties in perceiving L2 contrasts They state that L2 learners tend to assimilate non-native speech sounds to the native categories on the basis of their perceived articulatory similarities to the native sounds According to the model, four perceptual assimilation patterns were suggested:

(i) when the two phones of the L2 contrast are assimilated into two separating categories in the native sound system,

(ii) when both L2 phones are assimilated equally well into a single category,

(iii) when both phones are assimilated into a single category, yet unequally,

demonstrating a category goodness difference in their inculcation into the native

28

The model of Best & Strange is the further elaboration of the Speech Learning model proposed by Flege (1991) Except for assimilation pattern [4], non-native contrast pairs are usually perceived in accordance with L1 categories, which can affect the quality of the produced sounds, especially among adult learners Thus, the assumption that pronunciation training of L2 novel sounds should take into consideration perception process is further justified

With the discussed Speech Learning Model and the Perceptual Assimilation Model, the problems of L2 adult learners in perceiving and producing the non-native contrasts have been explicated Also, they have proved the insufficience of current pronunciation teaching approach However, it is still difficult to decide which L2 sounds should be considered novel, or non-native, and into which L1 phonetic categories non-native sound tend to be assimilated Concerning the present study, this decision helps decide which L2 contrast pair should be investigated and in what direction the perceptual training should be directed This issue is discussed in the Feature Competition Model proposed by Hancin-Batt (1994)

2.2.3 Feature Competition Model

In consideration of the impact of L1 sound categories on L2 phonological acquisition, Hancin-Bhatt (1994) observed that the phone substitution in L2 speech learning is not arbitrary, but represents a number of disciplines and constraints She proposed the Feature Competition Model to explain how L2 sounds are positioned onto different specific L1 sound categories Subsequently, a specific method for calculating L2 learners‟ difficulty in learning different L2 contrasts was devised The Feature Competition model was based on the idea of Chomsky & Halle's (1968) Phonological Theory of Distinctive Features This theory holds that speech sounds are

characterized in terms of patterns of articulatory and acoustic features Thus, for

instance, the English consonant /p/ is classified as [+ consonantal], [+ anterior], etc,

29

which are called distinctive features of a phone Hancin-Batt specifically pointed out that, during speech perception, it is those distinctive features that determine the underlying representation of the phonological form However, the attributive features

of a sound are not necessarily of equal prominence They possess different value of significance in the input Those features of higher prominence in a language are more salient and are thus noticed over those which are less salient Once the pattern is established, the prominent features will be used to construct the phonetic plan for the realizations of the phones Thus, in hearing an L2 English sound, Vietnamese learners will covertly perceive the sound in terms of different articulatory and acoustic features which are prominent in L1 and categorize that sound according to those prominent features Based on the comparison of the English and Vietnamese sound systems, it can be predicted which English sounds will cause great troubles for adult learners and thanks to that, the selection of a sample contrast pair can be decided

In extending her research into the salience of phonetic distinctive features, Hancin-Bhatt continued to propose a specific method to determine the feature prominence for a language This is accomplished by defining the phonemes of the language according to feature compositions based on Chomsky & Halle's (1968) system and selecting only those features which are distinctive for a given phoneme Finally the ratio of the feature prominence (P) in a language is calculated by using the formula:

Number of phonemes for which feature is specified

P =

Total number of phoneme For example the feature of [sonorant] in English consonants has a feature

prominence ratio of 333 as there are 8 consonants distinguished by the presence of

[sonorant], namely /m/, /n/, /7/, /r/, /l/, /h/, /w/, /j/, divided by a total of 24 consonants

in English (Gimson, 1989)