THE PRODUCTION OF ENGLISH s CLUSTERS IN INITIAL POSITION OF ENGLISH NON MAJORED STUDENTS AT CAN THO UNIVERSITY

SCHOOL OF EDUCATION ENGLISH EDUCATION DEPARTMENT THE PRODUCTION OF ENGLISH /S/- CLUSTERS IN INITIAL POSITION OF ENGLISH NON-MAJORED STUDENTS AT CAN THO UNIVERSITY B.A Thesis Field of s

SCHOOL OF EDUCATION ENGLISH EDUCATION DEPARTMENT

THE PRODUCTION OF ENGLISH /S/- CLUSTERS IN INITIAL POSITION OF ENGLISH NON-MAJORED

STUDENTS AT CAN THO UNIVERSITY

B.A Thesis Field of study: English Language Teaching

Class: NN0652A2 Student’s code: 7062957

Bài nghiên cứu này được thực hiện với mục đích khảo sát về sự phát âm chùm phụ

âm đầu [s] trong tiếng Anh của sinh viên trường Đại học Cần Thơ Cụ thể hơn, nghiên cứu này nhằm tìm hiểu xem giới tính có ảnh hưởng đến sự phát âm chùm phụ âm đầu [s] của những người học tiếng Anh hay không, những lỗi phát âm nào thường đi kèm với từng phái và giữa những chùm phụ âm [s] gồm hai phụ âm và những chùm phụ âm [s] gồm ba phụ âm, loại nào khó phát âm chuẩn hơn Đối tượng nghiên cứu được chọn một cách ngẫu nhiên trong các sinh viên trường Đại học Cần Thơ bao gồm hai mươi người: mười nam, mười nữ Những sinh viên này sẽ đọc một danh sách những từ và những câu

có chứa chùm phụ âm đầu [s] được soạn sẵn để tác giả thu âm lại Sau đó, những phần thu âm này sẽ được phân tích bằng chương trình mô tả số liệu và chương trình ANOVA một chiều Tất cả những người tham gia nghiên cứu đều có những lỗi sai về phát âm của các âm như [sp], [sk], [sl], [str], [st], [sm], [sn], [spr], [spl], [skr] và [skw] Tuy nhiên, chỉ

âm [sm] là có sự khác biệt lớn giữa nam sinh và nữ sinh Có ba loại lỗi cơ bản những sinh viên này gặp phải khi phát âm chùm phụ âm đầu [s] đó là không phát âm âm [s] đầu tiên, không phát âm những phụ âm sau âm [s] và chêm thêm một nguyên âm sau âm [s] Trong đó, không phát âm âm [s] là tiêu biểu nhất Một kết quả khác được rút ra đó là các sinh viên gặp khó khăn khi phát âm những chùm phụ âm [s] gồm ba phụ âm hơn những chùm phụ âm [s] gồm hai phụ âm Nhìn chung, nữ sinh có nhiều lỗi sai hơn nam sinh khi phát âm chùm phụ âm đầu khi đọc các từ này riêng lẻ nhưng khi những từ này được đặt trong ngữ cảnh, nữ sinh lại ít mắc lỗi về phát âm sai hơn nam sinh

Abstract

This study aims to investigate the production of English /s/- clusters in initial position in word list and in context of English non-majored students at CTU More specifically, it were done to examine if gender significantly affects the EFL Vietnamese learners of productions [s]- cluster sounds and the types of mispronunciations associated with gender as well as to know whether tri-literal clusters are more difficult than bi-literal clusters based on the analysis of the recording Twenty students including ten males and ten females from course 32 to 35 were randomly chosen to be the participants of this study The data was collected while the participants were reading a word list and a text contained the target sounds After that, the data were analyzed based on the descriptive statistics and One Way ANOVA program All of mispronunciations of [s]-cluster sounds initial position such as [sp], [sk], [sl], [str], [st], [sm], [sn], [spr], [spl], [skr] and [skw]were observed but gender had a significant effect on pronouncing [sm] sounds in isolated words There are three main type of mispronunciations that happened in the

production of s-clusters which are unpronounced the pre-initial consonant [s],

unpronounced the initial consonant like [p], [s], [t], [m], [n] [tr] of s-clusters and

Epenthesis Among them, the first type had a greatest percentage Besides, another conclusion was drawn was that longer clusters would be more difficult for students of non English major at CTU to produce correctly Overall, female students had more mispronunciations than their male counterparts for s-cluster sounds in word list However, in context, they had fewer errors than male participants

Second, my regards are respectively sent to all the teachers of the English Department for their encouragement and guidance Especially, my sincere gratitude is sent to Ms Hong Thi Thanh Truc, an English teacher in the English Department, and Ms Elizabeth Hollingsworth, a native speaker from America for helping me detect the participants' mispronunciations Also, I would like to thank all of the participants and some of my friends, especially Nguyen Ngoc Cat Khuyen and Le Thi Van for their help during the time I collected the data for this study And I would like to acknowledge Ms Ngo Thi Trang Thao who gave me useful instructions to analyse the data using SPSS program

Next, I would like to send my special thanks to my close friends for being by my side and giving much encouragement to finish this thesis

Last but not least, I really want to thank my family and my boyfriend whose encouragement was very meaningful and very important to me during the time of doing this research

TABLE OF CONTENTS

ABSTRACT (Vietnamese) i

ABSTRACT (English) ii

ACKNOWLEGMENT iii

LIST OF TABLES iv

LIST OF FIRGUGES vi

Chapter 1 - INTRODUCTION 1

1.1 General statement of problem 1

1.1.1 The Sonority Sequencing Principle 2

1.1.2 The Markedness Differential Hypothesis 3

1.2 Background of the study 3

1.2.1 Initial consonant clusters 3

1.2.2 The initial s- clusters 4

 The initial bi-literal s- clusters 5

 The initial tri-literal s- clusters 5

1.2.3 Problems for pronouncing English /s/-clusters 6

1.3 Research Aim 6

1.4 The organization of the study 6

Chapter 2 – LITERATURE REVIEW 7

2.1 Studies that examined the production of English initial s-clusters in some countries around the word 7

2.2 Studies about English pronunciation of Vietnamese learners 12

2.3 Studies that investigated gender differences in language learning 13

Chapter 3 – METHODOLOGY 15

3.1 RATIONAL 15

3.2 RESEARCH QUESTIONS 15

3.3 HYPOTHESIS 15

3.4 PARTICIPANTS 16

3.5 RESEARCH INSTRUMENTS 16

3.6 RECORDING PROCEDURE 17

3.7 CODING 17

Chapter 4 – RESULTS 18

in initial position of English non-majored students at CTU 18

4.1.1 Descriptive statistics for [sp] in word list and in context 19

4.1.2 Descriptive statistics for [sk] in word list and in context 20

4.1.3 Descriptive statistics for [sn] in word list and in context 21

4.1.4 Descriptive statistics for [sw] in word list and in context 22

4.1.5 Descriptive statistics for [st] in word list and in context 23

4.1.6 Descriptive statistics for [sm] in word list and in context 24

4.1.7 Descriptive statistics for [sl] in word list and in context 25

4.1.8 Descriptive statistics for [spr] in word list and in context 26

4.1.9 Descriptive statistics for [spl] in word list and in context 27

4.1.10 Descriptive statistics for [skr] in word list and in context 28

4.1.11 Descriptive statistics for [str] in word list and in context 29

4.1.12 Descriptive statistics for [skw] in word list and in context 29

4.2 Types of mispronunciations of s-cluster sounds in initial position 30

4.3 Analysis of bi-literal s-clusters /sC/ versus tri-literal s-clusters /sCC/ clusters 32

4.3.1 Cluster /sp/ vs / spC/ 33

4.3.2 Cluster /sk/ vs / skr/, /skw/ 34

4.3.3 Cluster /st/ and / str/ 35

Chapter 5 – DISCUSSION, IMPLICATIONS, LIMITATIONS AND CONCLUSION 5.1 Discussion 37

5.1.1Pronunciation of English /s/- clusters in initial position between males and females 37

5.1.2 Types of mispronunciations of s-cluster sounds in initial position 39

5.1.3 Bi-literal s-clusters /sC/ versus tri-literal s-clusters /sCC/ 41

5.2 Conclusion 41

5.3 Limitations of the research 42

5.4 Implications 42

5.5 Suggestion for further research 43

REFERENCES viii

APPENDIX xii

List of Tables

page

Table 1.1: Two-consonant clusters with pre-initial s 5

Table 4.1: Number of errors for [s]-clusters 18

Table 4.2: Descriptive statistics for [sp] in word list and in context 19

Table 4.3: Levene’s Result for [sp] in isolated word list and in context 19

Table 4.4: Descriptive statistics for [sk] in word list and in context 20

Table 4.5: Levene’s Result for [sk] in word list and in context 20

Table 4.6: Descriptive statistics for [sn] in word list and in context 21

Table 4.6: Descriptive statistics for [sn] in word list and in context 21

Table 4.8: Descriptive statistics for [sw] in word list and in context 22

Table 4.9: Levene’s Result for [sw] in word list and in context 22

Table 4.10: Descriptive statistics for [st] in word list and in context 23

Table 4.11: Levene’s Result for [st] in word list and in context 23

Table 4.12: Descriptive statistics for [sm] in word list and in context 24

Table 4.13: Levene’s Result for [sm] in word list and in context 24

Table 4.14: Descriptive statistics for [sl] in word list and in context 25

Table 4.15: Levene’s Result for [sl] in word list and in context 25

Table 4.16: Descriptive statistics for [spr] in word list and in context 26

Table 4.17: Levene’s Result for [spr] in word list and in context 26

Table 4.18: Descriptive statistics for [spl] in word list and in context 27

Table 4.19: Levene’s Result for [spl] in word list and in context 27

Table 4.20: Descriptive statistics for [skr] in word list and in context 28

Table 4.21: Levene’s Result for [skr] in word list and in context 28

Table 4.22: Descriptive statistics for [str] in word list and in context 29

Table 4.23: Levene’s Result for [str] in word list and in context 29

Table 4.24: Descriptive statistics for [skw] in word list and in context 29

Table 4.25: Levene’s Result for [skw] in word list and in context 30

Table 4.26: Types of mispronunciations of s-cluster sounds 31

Table 4.27: Number of errors of tri-literal s-clusters and bi-literal s-clusters made by participants in isolated words 32

Table 4.28: Number of errors of cluster /sp/ vs /spC/ by participants in isolated words and in context 34

Figure 4.1: Total rates of errors of tri-literal s-clusters and

bi-literal s-clusters made by participants in isolated words 33

Figure 4.2: Total rates of errors of cluster /sp/ vs /spC/

by participants in isolated words and in context 34

Figure 4.3: Total rates of errors of cluster /sk/ vs /skC/ by

participants in isolated words and in context 35

Figure 4.4: Total rates of errors of cluster /st/ and /str/ by

participants in isolated words and in context 36

Chapter One

INTRODUCTION

There are four parts in this chapter The first part is General statement of problem, which explains the reason why this research was carried out and introduces some related theories The second part is the background of the study, which presents what s-clusters are The next part is the research aims of this paper The last part will show the organization of this thesis

1.1 General statement of problem

English is a global language So, the number of English learners around the world has been increasing tremendously In Vietnam, people are rushing to foreign language centers to learn English because there is a big need for schooling and jobs opportunities However, there are many factors which influence language learners’ pronunciation According to Hinofitis and Baily (1980), pronunciation is one of the most important things of studying second language because it affects learners' communication competence Szynalski and Wojcik (2002) also stated that pronunciation is the first and most important aspect that speakers pay attention to when having a conversation Knowing grammar and vocabulary is important but it becomes useless if the speaker is unable to pronounce words correctly

However, learning pronunciation is not easy for many people because of different reasons Nguyen (2007) mentioned three main reasons First, it may be influenced by learners' mother tongue Second, it may be because of the study environment and third, it

is learners' motivation in studying pronunciation

Like some other languages, Vietnamese has phonetics that keeps native learners from pronouncing English like native speakers The limitation of Vietnamese word-initial consonantal cluster sounds and the frequency of English initial clusters errors, which are made by many Vietnamese learners, have caught attention to this area of this topic

At Can Tho University, all undergraduates of English non-major have to finish three courses of The materials for these courses are REWARD ELEMENTARY books which supply students basic knowledge about English grammar, pronunciation and vocabulary However, in non-majored classes, because of limited time of 4 periods per week in class, lessons usually focus on grammar and reading and therefore, teachers do not much focus on teaching English pronunciation or correcting students’ pronunciation mistakes Furthermore, in these courses of General English, students study different majors in different schools, they have different English levels and English is just a subsidiary subject Thus, having problems in pronouncing English is understandable

This paper will investigate the difficulties of English non-majored students at Can Tho University when dealing with English initial s-clusters This is considered one of the most significant problematic features of English learners The finding of this research will hopefully help English non-majored students at Can Tho University to become aware

of their systematic errors and carefully try to produce English initial s-clusters correctly

1.1.1 The Sonority Sequencing Principle

There is a universal (in all languages) tendency for sonority to gradually increase

in the onset, and decrease in the coda This is called the Sonority Sequencing Principle

The Sonority Sequencing Principle was identified in linguistics research during the last

20 years or so It dictates that onsets (word-initial sounds) must rise in sonority and codas (ending sounds) must fall in sonority Sonority is the inherent loudness of sounds relative

to one another Many syllables of English conform to this universal tendency

However, some do not In particular those consonant clusters that start with /s/ slightly violate the universal tendency For this reason, L2 learners of English often have difficulty with these clusters They tend to make them more natural by inserting, or deleting segments:

1.1.2 The Markedness Differential Hypothesis

Ecknam (1993) based on a theory of markedness in phonology to hypothesize that more frequent sounds in many languages are called unmarked while the sounds which are not popularly used among languages in the world are marked sounds Gass & Selinker (2001) claim that sounds related to voicing contrast in initial, medial and final position are marked In this case, Vietnamese speakers of English may find it challenging to pronounce some s-clusters in initial position because they are not pronounced in Vietnamese

1.2 Background of the study

The pronunciation of English syllable codas by second language (L2 learners of English, especially whose native languages (L1s) do not have phonetic features similar to English, has received much attention in L2 research Vietnamese, for example, does not have consonant clusters which are allowed in English in various word-positions Vietnamese L2 learners of English have been found to have problems with pronouncing English consonant clusters

1.2.1 Initial consonant clusters

Cluster formation is one of the most interesting topics of various aspects of a language’s phonology This is attested in the variety of cluster formation in language acquisition (L1), language learning (L2), as well as the dialectal variants of the norm The fundamental claim governing cluster formation is that the bigger the distance between the members of a cluster on the sonority scale is, the better structured the cluster is (Clements, 1983)

A consonant cluster is described as being two or more adjacent consonants in the same syllable (e.g., /br/ in brush, /skw/ in square, and /nt/ in paint) The consonants that constitute a cluster are referred to as cluster elements in this paper Consonant clusters are very commonly used in English words McLeod et al (2001) reported that one third of single syllables in English start with word-initial clusters Many English words end in consonant clusters due to word-final grammatical morphological structures which can, but do not always, create word-final clusters For example, the phonemes /s, z, t, d/ are added in word final position for the possessive form of pronouns (e.g, it’s, mum’s), plural

nouns (e.g., dogs, lips) while the phonemes /sp, sl, / in “speak”, “slang” are consonant clusters in initial- position

Concerning the possible combinations of initial consonants, Brinton (2000, p 55) lists the following non-permissible sequences of consonants:

- stop + stop, such as in /pt/;

- stop + nasal, such as in /pn/;

- nasal + stop, such as in /np/;

- stop + fricative, such as in /ts/;

- fricative + stop, such as in /ft/, except where the fricative is /s/

A classification of the possible English clusters taken from Avery & Ehrlich (1992, p.55-58) is provided below together with their example of each cluster:

(1) initial bi-literal (two-consonant) clusters beginning with a stop include /pl,

1.2.2 The initial s- clusters

Sequences of two or three consonants are called clusters, which constitute an important aspect of restrictions on syllable types, since there is a limited number of possible combinations of segments in both initial and final positions Clusters with initial /s/, for instance, are the only instances of onsets where the second consonant may be an obstruent and where the onset may be formed by three consonants instead of one or two

In order to represent the particular category of /s/ clusters, Selkirk (1982) provides an auxiliary template, considering that “/s/ plus obstruent may qualify as a single obstruent

in English”

 The initial bi-literal s- clusters

In the book “English Phonetics and Phonology”, Roach (2008, 78) wrote “One sort

is composed of s followed by one of a set of about eleven consonants” It was a combination of [s] with one possible consonant, so it was called [sC] He listed some

examples such as “sting”, sway, smoke The /s/ in these clusters is called the pre-initial

consonant and the other consonants (t, w, m in the above examples) the initial consonant

These clusters are showed in table below

Table 1.1: Two-consonant clusters with pre-initial /s/

Initial

Pre-s + p t k b d g f θ s ∫ h v ð z ʒ m n η

e.g spin stik skin - - - sfiə - - - - - - - - smel snou -

Roach (2008) also noted that two-consonant clusters of s plus l, w, j were also possible

(e.g slip, swiη, sju:)

 The initial tri-literal s- clusters

Besides, Roach (2008) also mentioned about – They are [sCC] Example of three-consonant clusters are “split” /split/, “stream” /stri:m/, “square”/skweə/ The s is the pre-initial consonant, the p, t and k that follow s in the three example words are the initial consonant and the l, r and k are post-initial These clusters are showed in table below

l r w j

p splay spray - spew

S PLUS INITIAL t - string - stew

k sclerosis screen squeak skewer

However, in their studies, Rebello (1997) and Carlisle (1997, 1992, 1997) stated that there are six English two-segment /s/- clusters (/sp, st, sk, sm, sn, sl/), and five three-segment /s/-clusters (/spr, str, skr, spl, skw/) I follow this point of view to do my study

1.2.3 Problems for pronouncing English /s/-clusters

According to Selkirk (1982), the /s/-clusters are equally problematic for theoretical accounts of English and other languages with [s]-clusters First, [s] stop clusters violate the sonority sequencing principle, in that they have a falling sonority slope Secondly, clusters [sl-], [sn-] and [st-] violate a phonotactic constraint on initial clusters in English which prohibits homorganic clusters Thirdly, [s] is the only sound that may be followed

by a nasal or a stop in initial clusters Finally, [s] is the only sound that may occur at the beginning of a three-element cluster such as [str-], [spr-], [skw-] or [spl] All these facts reveal the special status of /s/-clusters

1.3 Research Aim:

The purpose of this research is to examine the production of s- clusters in initial

position of English non-majored students at Can Tho University (CTU) Especially, this

study aims to investigate if male and female speakers pronounce those sounds

significantly differently

1.4 The organization of the study

This thesis is divided into five chapters Chapter 1 introduces the importance of pronunciation in studying languages and presents some related theories of this study Chapter 2 summarizes some studies that examined the production of English initial s-clusters in some countries and some other studies about Vietnamese EFL learners' pronunciation of the target sounds Chapter 3 is about the methodology through which the study is conducted Chapter 4 presents the results of the study And, chapter 5 discusses the research findings, limitations and conclusion

Chapter Two

Literature Review

This chapter is categorized into three parts Part 1 summaries some studies that examined the production

of English initial s-clusters in some countries Part 2 presents some other studies about Vietnamese EFL learners' pronunciation of the target sounds In addition, part 3 includes some studies that investigated gender differences in language learning

1 Studies that examined the production of English initial s-clusters in some countries around the world

There have been many studies about the production of English initial s-clusters in some countries around the world such as in Australia, in Brazil, in China and so on Some of those are mentioned below

Silveira (2002) did a research to investigate the relationship between perception and production in the acquisition of word-initial /s/ clusters in the interlanguage of Brazilians learning English as a foreign language While testing this relationship, he also contributed to the controversial issue of the role of markedness and the first language (L1) interference on the production of word-initial /s/ cluster In addition to the interface between perception and production, other factors that might influence both mental processes such as faulty perception and L1 interference, or a combination of these factors, were taken into account This was tested by collecting data assessing subjects’ perception and production of English initial /s/ clusters in two different sessions: production test and perception test In production test, the subjects were required to record the translation from Portuguese into English of 14 sentences, where the English

version was expected to contain at least one word with an initial /s/ cluster The

perception test was carried out two weeks after the production test The subjects received

a short training by repeating s-cluster words provided by an American native speaker

Immediately after the training, they were tested again by listening to the tape containing

the 26 sets of sentences The results partially supported studies which proposed that

perception influenced production There was also support for the power of L1

interference over Markedness and Universal Canonical Syllable Structure Epenthesis

was the strategy of syllable simplification present in all cluster types Some subjects,

though, resorted to a short epenthetic vowel (/I/), thus indicating that they might be

developing a separate category for initial /s/ clusters

Like Silveira, Cardoso, French & John (2008) also based on a theory of

markedness to do a research to investigate the effects of markedness and input frequency

in the development of English homorganic /sl/, /sn/ and /st/ among Brazilian Portuguese

native speakers learning English in a classroom environment It provided a

multidisciplinary analysis for the acquisition of /s/ + consonant onset clusters (sC) in

second language phonology, adopting a variationist approach for data collection and

analysis that included from a variety of linguistic disciplines, including theoretical

phonology, psycholinguistics, second language acquisition, and sociolinguistics The results of an oral production study indicated that English learners acquire /sl/ and /sn/

clusters earlier than the most marked /st/, corroborating the hypothesis that markedness,

and not input frequency, determines the order of acquisition of bi-literal s-clusters in

second language production

To consider the acquisition of /s/-clusters relative to other clusters of English,

many studies were done by Smit (1973), Barlow (1997), Gierut, (1999) In Smit’s study

(1973), the errors on word-initial consonant clusters made by children in the Iowa were

tabulated by age range and frequency The error data show considerable support in the

acquisition of clusters: the youngest children show cluster reduction, somewhat olderchildren show cluster preservation but with errors on one or more of the cluster elements, and the oldest children generally show correct production These stages extended to three-element clusters as well Typical cluster reduction errors were (a) reduction to the obstruent in obstruent-plus-approximant clusters and (b) reduction to the second element

in /s/-clusters Whenclusters were preserved, but one member was in error, the error was typically the same as for the singleton consonant Cluster errors are discussed in terms of theories of phonologic development, including open genetic programs and feature geometry Meanwhile, Barlow (1997) stated that normal and disordered phonological development have evidenced an asymmetry in the development of clusters in English, particularly the s-clusters relative to other clusters She explained that this asymmetry was found by appealing to differing structural representation Some clusters may be represented as a complex onset, while others like s-clusters may be represented as an adjunct of clusters Having the same idea about s-clusters as adjuncts with Barlow, Giuret (1999) did a research on a presumed universal that governs the permissible sequences of consonants within syllables In two single-subject experiments, he evaluated this principle as applied to the acquisition of onset clusters and adjuncts by children exhibiting functional phonological delays Experiment 1 tested the hypothesis that children abide by the Sonority Sequencing Principle (a universal tendency for sonority to gradually increase in the onset, and decrease in the coda) in development, such that the occurrence and use of marked true clusters implies unmarked clusters Others who were taught unmarked clusters exhibited limited learning characteristic of within-class generalization, with apparent gaps in sonority sequencing Experiment 2 examined the role of adjunct sequences /sp, st, sk/, whose markedness status is questionable given their violation of the Sonority Sequencing Principle Results indicated that children learned adjuncts consistent with patterns of within-class generalization, thereby supporting the view that these sequences are unmarked in structure

Another study was done by Rauber (2002) to examine the production of English initial /s/ clusters by Portuguese and Spanish EFL Speakers In order to compare the production of speakers of both native languages the same corpus was used By basing on the analysis on the Markedness Differential Hypothesis, he tried to confirm whether the difference in the structure of Portuguese and Spanish syllables results in greater difficulties in different /s/ clusters in different environments or not The participants chosen for this experiment were nine native Spanish speakers from Argentina and ten native Portuguese speakers from Brazil They were asked to read and were recorded the sentences done by the researcher After analyzing, the results revealed that epenthesis (the addition of an extra vowel) to the initial clusters is the usual strategy for dealing with syllable structure difficulty, which can be a result of native language interference and linguistic universals He concluded that learners’ difficulties established by the Markedness Differential Hypothesis and the Structural Conformity Hypothesis were borne out, since longer initial /s/ clusters caused a greater rate of epenthesis than shorter ones for both Spanish and Portuguese speakers

Although few studies have investigated the phonological context where errors occur, another aspect that should be taken into account regarding epenthesis and consonant deletion for syllable simplification is the influence of environment in Spanish/English and Portuguese/English interphonology Carlisle carried out several studies involving native Spanish-speaking learners of English as a second language, who were asked to read a number of topically unrelated and randomly ordered sentences containing initial /s/-clusters in two types of environment Carlisle (1991) examined epenthesis before three word-initial onsets in English: /sk/, /st/ and /sp/; Carlisle (1991,

1992, ) investigated the frequency of epenthesis before the word-initial onsets /sl/ and /st/; Carlisle (1992) investigated the production of epenthesis before the word-initial onsets /sl/, /sm/ and /sn/; and Carlisle (1997) compared the production of /sC/ vs /sCC/ clusters All the studies controlled the environments before the onsets and the sonority

relationships among the consonants in the onsets His findings in the four studies revealed that vowel epenthesis was significantly more frequent after consonants than after vowels Concerning the variable length examined in Carlisle (1997), his findings revealed that tri-literal clusters, whose structure is more marked, were more frequently modified than bi-literal clusters, confirming that language universals “influence the structuring of interlanguage phonology” (p 327)

Similar results were obtained by Abrahamsson (1997, 1999) In his longitudinal study, he investigated a male native speaker of Spanish from Bolivia, who was a guest student at Stockholm University, and an absolute beginner of Swedish Data were collected during eight months, and consisted of recordings where the participant and a native speaker of Swedish talked about various topics Besides free speech, activities such as object or picture description and picture story retelling were adopted to elicit specific grammatical/discourse patterns and communication strategies His findings corroborated those of Carlisle, since longer clusters were more frequently modified, and consonantal environments caused more epenthesis than silence and vocalic environments Concerning the longitudinal aspect of the study, Abrahamsson concluded that the frequency of epenthesis increased in the beginning of the acquisition process, since the participant was starting to develop his target language fluency

The studies above show that acquiring the s-clusters in initial position is not an easy task for many learners of English in some parts of the world Mostly, difficulties in pronouncing [s]-cluster sounds established by the Markedness Differential Hypothesis Epenthesis was a popular error among Brazilians, Portuguese and Spanish when pronouncing s-cluster sounds (Rauber,2002) They inserted a vowel like [ə] or [i] after the sound [s] to make s-clusters more easily to produce In the context of Vietnam, problems about the acquisition of English pronunciation have not been paid enough attention, as they should be It is the fact that Vietnamese learners of English have much

difficulty in English pronunciation Similarly, these s-cluster sounds also cause

difficulties for English learners that were specified in some studies below

2 Studies about English pronunciation of Vietnamese learners

Whereas much research with a focus on syllable codas has been done on English

as a second language (ESL), native speakers of Portuguese, Arabic, and Chinese, not

many researchers have examined native speakers of Vietnamese in this regard Not until

the 1980s, when there was an influx of Vietnamese immigrants to the United States after

the Vietnam War, did researchers such as Sato (1984, 1985) and Benson (1986) begin to

investigate the L2 phonology of Vietnamese immigrants upon noticing their erroneous

production of syllable onsets and codas Findings from these studies have confirmed that

consonant clusters are one area of difficulty Vietnamese ESL learners have

In this regard, Hwa-Froelich, Hodson, and Edwards (2002) did a rather

comprehensive investigation into Vietnamese phonology, comparing and contrasting its

phonological characteristics with English, in order to see how these features may be

carried by Vietnamese learners into their L2 English phonology They also suggested that

Vietnamese learners might have problems with both English consonant clusters in initial

position and final position

Tang (2007) used survey and ethnographic methods to provide a linguistic basis

for promoting first language maintenance of Vietnamese in a larger United States context

and to encourage future research in language acquisition of Vietnamese-English speakers

This article also presented the social and linguistic issues related to language maintenance

among Vietnamese Although Vietnamese and English share certain sounds, there are

multiple consonant sounds specific to each language Especially, Tang (2007) listed

consonant clusters in initial position that are absolutely foreign to the Vietnamese

language such as s-clusters (sk, scr, sm,sn ), r-clusters (br, cr, dr ), l-clusters (bl,cl )

and so on However, this material is quite detailed and should be appreciated It is easily

seen from this data that English have a number of consonants and consonant clusters that

do not exist in Vietnamese rather than vice versa As a result, he concluded that

pronouncing English consonant clusters properly is one of the most difficult things that

learners have to face from the beginning

In summary, like other English learners in some countries mentioned above, Vietnamese learners can not avoid some troubles when they pronounce the initial s-clusters include six English two-segment /s/- clusters (/sp, st, sk, sm, sn, sl/), and five three-segment /s/-clusters (/spr, str, skr, spl, skw/) Yet, the main purposes of this study are to investigate the production of English /s/- clusters in initial position of English non-majored students at Can Tho University and find out if three-segment /s/-clusters are more difficult than two-segment /s/- clusters for English non-majored students at Can Tho University

3 Studies that investigated gender differences in language learning

English learning is influenced by many factors such as learners’ motivation, studying environment, frequency of practicing and so on, which affect directly and indirectly the students’ English acquirement Among these, the learner himself or herself

is the most direct and important That some researchers have been eager to find out the relation between gender and language learning However, no one examined the gender difference in pronunciation of Vietnamese learners of English, especially in pronouncing the clusters

Li (2004) conducted a study about the English vowels pronounced by Taiwanese learners in order to find out if there were similar linguistic features that impeded or facilitated English vowels and to examine whether gender had effect on vowels perception The results show that both male and female Taiwanese speakers of English had difficulty on pronouncing vowels However, female speakers were reported to be better at pronouncing vowels than their counterparts

Many other studies reported the superiority of females over males in language learning For example, a study conducted by Carooll and Sapon (1967) showed that girl pupils were better than their boy counterparts in learning foreign languages in grades kindergarten to 12 Having the same purpose, in his research, Farhady (1982) finally found out some evidence suggesting that females were better than males in listening

comprehension tasks In another study, Boyel (1985) conclude that female were typical superior to males in all aspects of language leaning, except listening vocabulary

These studies involving gender differences in learning language once again revealed that gender had effect in learners’ learning languages Because no one examined the gender difference in pronouncing the clusters, hopefully, with the purpose of examining the influence of gender on the production of [s]-cluster sounds, this study can investigate something new for Vietnamese learners of English The methodology of the study will be presented in the next chapter

study aims to investigate if male and female speakers pronounce those sounds

significantly differently In addition, the study aims to find out whether tri-literal s- clusters or bi-literal s-clusters are more difficult for students of non English major at CTU

to produce correctly

3.2 RESEARCH QUESTIONS

This study attempts to answer the three following questions:

Question 1: Does gender significantly affect the CTU English non-majored students’

production of s-clusters in initial position in word list and in context?

Question 2: What types of mispronunciations occur frequently?

Question 3: Do students of non-English major at CTU have more difficulty pronouncing

the tri-literal s- clusters or bi-literal s-clusters?

3.3 HYPOTHESES

Based on the findings of Carooll and Sapon (1967), Farhady (1982) and my own observation, it was hypothesized that gender would significantly affect the CTU English

non-majored students’ production of s-clusters in initial position in word list and in context Another hypothesis was that longer clusters would be more difficult for these participants to produce

3.4 PARTICIPANTS

Twenty students including ten males and ten females from course 32 to 35 were randomly chosen to be the participants of this study All of them are students at Cantho University However, their majors are different: six of them come from School of Education, 3 students from College of Agriculture and Applied Biology, 4 students from School of Engineering Technology, 1 student from School of Social Sciences and Humanities, 5 students from School of Economics – Business Administration, 2 students from College of Aquaculture and the last one from School of Law

Most of them had studied English for six years at their high schools At Cantho University, some of them finished three courses in General English while others are studying these courses Eighteen over twenty students have taken some English courses

in some Foreign Language Centers in Can Tho City Although they had different proficiency of English, it is certain that the participants already knew these sounds and the way to produce

3.5 RESEARCH INSTRUMENTS

The instrument of this study consisted of 2 parts Part 1 was a list containing 24 individual content words in which there are two words carrying the target [s]-cluster

sounds concluding: [sp], [st], [sk], [sw], [sm], [sn], [sl], [spl], [spr], [str], [skw] and

[skr] The second part was 24 sentences which written based on those 24 This

instrument could make the participants' production more reliable because they would be very conscious when reading the isolated words (Abrahamsson, 1999) The instrument was examined and developed with the help of a lecturer who got a Master's degree in the United States and especially he has much experience in teaching pronunciation

3.6 RECORDING PROCEDURE

The participants were recorded individually First, they were given the paper including the word list and the sentences that followed They had about five minutes prior to reading the words and then all the sentences The subjects allowed to ask any questions related to the meaning of any unknown words However, they were not given any explanation about the pronunciation of any words They are all carefully given the instructions before recording It took them from 2 to 3 minutes to finish reading the word list and the sentences

The recording was carried out in discussing room in Learning Resource Center where is rather quiet The fans were turned off to make sure that there was not any noise All samples were collected using a laptop with high quality microphones and specially designed program called Aurora Media Workshop

3.7 CODING

All of the recordings were listened to by three raters The first rater was the researcher, another rater was a Vietnamese teacher who is working in the English Department and the final rater was an America speaker and was teaching pronunciation for English majored students at CTU

Errors were counted basing on the accuracy of the participants’ production The participants were considered to have an error when they mispronounced or deleted the target sounds which were both the sound [s] and the consonants followed it Besides the targets sounds, any error made by the participants was not counted

Table 4.1 Number of errors for [s]-clusters

Context Male Female

4.1.1 Descriptive statistics for [sp] in word list and in context

Table 4.2: Descriptive statistics for [sp] in word list and in context

word list context Gender No Mean Std Deviation No Mean Std Deviation

Table 4.3: Levene’s Result for [sp] in word list and in context

word list Context F1(1,18) p-value F2(1,18) p-value

5.27 0.025 1.531 0.232

Note: Alpha=0.05

The ANOVA analysis was later used in which gender served as the independent variable and the participants’ mispronunciations of the sound [sp] functioned as the

dependent variable The ANOVA analysis showed that in word list and in context, the two gender groups had no significant effect

4.1.2 Descriptive statistics for [sk] in word list and in context

The result of the descriptive analysis for the observed mispronunciations of [sk] in word list and in context were calculated and presented in table 5

Table 4.4: Descriptive statistics for [sk] in word list and in context

word list context Gender No Mean Std Deviation No Mean Std Deviation

Again the Levene’s test homogeneity variance was observed at alpha = 0.05 Table 4.5 showed the result of Levene’s test

Table 4.5: Levene’s Result for [sk] in word list and in context

word list context F(1,18) p-value F(1,18) p-value

0.42 0.52 1.10 0.31

Note: Alpha=0.05

It was obvious that no significant effect was found both in the list of words F(1,18) = 42,

p = 0.52; and in context, F(1,18) = 1.10, p = 0.31