University of Groningen Faculty of Arts
Department of Applied Linguistics MA thesis
The influence of L1 dialects on L2 speech: Effects of Northern and Southern Vietnamese on English pronunciation
(Word count: 13.542)
Gwen Vu Van Quyen – S2458098
Supervisors: Charlotte Gooskens -‐ Nanna Haug Hilton Second reader: Wander Lowie
Acknowledgements
It is a great honour of me to have had a very chance to live and study in the city of Groningen, the Netherlands. This thesis has come into being thanks to a great deal of supporting and sharing from my friends and professors, whom I owe my sincerest gratitude.
To begin with the board of teachers, I would like to give my thanks to all of the professors and teachers who have been part of the program for their unfailing devotion, friendly attitude and willingness to support us. My thanks go to my supervisor, Dr. Charlotte Gooskens for her guidance, without which this work would not have come into being. I also thank Dr. Nanna Haug Hilton for her feedback.
My grateful acknowledgment is tendered to my dearest friends in Groningen; I thank Lisa, David, Maria and Hara for what we have been through together and for their comfort presence, what always keeps me brave and companied along the year. I also thank Leilani, Vaneide, Sanet, and all the classmates whose caring words have been continuously around me. My thanks also to my Vietnamese friends in Groningen: các chị Gia, Dương, Xuân Anh, Mai, Mia, Ngọc, anh chị Anh Phạm – Hoa, gia đình Ben Bun, Ben em, Kẹo, các anh Thế Anh, Dứt, Nhật, Lâm, Thịnh, Thái, Thắng, các bạn Dung, Minh, các em Lân, Ngọc Anh, Nhi, for their sympathies and the feeling of a big family.
I am also indebted to my family – my parents and my brother’s little family – for sending me here and creating the miracle that I could meet all of these beautiful people. My sincere thanks to is my two best friends Linh béo and Hoài Anh, for sharing my dreams. I thank Bờm for making me strong and independent.
Lastly, I want to thank my dear friends in Vietnam who do not hesitate to call for helps with my experiment. I also thank all the participants for their willingness.
List of abbreviations
L1 First language
L2 Second Language
SLA Second Language Acquisition
NVN Northern Vietnamese (dialect)
Table of Contents
Abstract ... 5
Chapter 1 -‐ Introduction ... 6
Chapter 2 -‐ Literature background ... 10
2.1 The Vietnamese Language ... 10
2.2 Vietnamese Dialects – Distinguishing Phoneme Systems ... 18
2.3 English Spoken by Speakers of the Different Vietnamese Dialects. ... 28
2.4 Research questions ... 37 Chapter 3 -‐ Methodology ... 39 3.1 Experiment ... 39 3.2 Recordings ... 39 3.2.2. Materials ... 40 3.3 Surveys ... 43 3.3.3. Analysis ... 45 Chapter 4 – Results ... 47
4.1 Survey 1 – Vietnamese Dialect Differentiation ... 47
4.2 Survey 2 – Vietnamese Dialectal Effects in English Spoken by Vietnamese ... 48
Chapter 5 -‐ Discussion ... 52
5.1 Can Vietnamese Identify Different Vietnamese Dialects? ... 52
5.2 Is There any Influence of Vietnamese Dialects on English Pronunciation? ... 56
5.3 Is English Spoken by Vietnamese Similar to English by Chinese and Thai? ... 59
Chapter 6 – Conclusion ... 61
Appendix A ... 63
Appendix B ... 64
Appendix C ... 67
Abstract
Chapter 1 -‐ Introduction
Due to the spread of English all over the world, the need to learn it as a foreign language arises in many countries, and especially in developing countries, such as Vietnam and Thailand, English is increasingly seen as highly important. The languages of these countries, however, do not belong to the same language family as English and are typologically highly different, providing difficulties in the acquisition and use of English by its learners. Vietnamese belongs to the Viet-‐Muong group, a language branch very distant from the Germanic branch of Indo-‐European language to which English belongs. Furthermore, Vietnamese is a monosyllable and tonal language, which, at least, differs from English in the most basic aspect of phonemic qualities. Due to these differences, the Vietnamese accent in L2 English production is highly noticeable, with Hwa –Froelich, Hodson and Edwards (2002) finding that Vietnamese English learners have difficulties with certain English sounds, such as /p/, /ʃ/, /tʃ/, /ʒ/, /dʒ/, /θ/ and /ð/, and tend to use substitutions of similar sounds.
whose area stretches from the border with China to the north of the Nghệ An Province. Meanwhile, SVN is spoken in the area south of the Huế Province, next to the habitat of Vietnamese minor ethnic groups of Khmer and Cantonese Chinese (See Appendix A to visualize this division on a map of Vietnam). Geographically, both NVN and SVN are influenced by different cultures and languages. Linguistically, the two dialects polarize in terms of the number and quality of consonants, vowels and combinations of some certain phonemes (Nguyen, 1997; 2009; Thompson, 1959; 1987). Especially these linguistic differences between the two dialects lead to the expectation that the difference in pronunciation is noticeable, for Van Bezooijen and Gooskens (1997) found that, although language variety identification is dependent of linguistic cues such as prosody (patterns of stress and intonation) and verbal information, the identification of language varieties within the United Kingdom and the Netherlands was based mainly on pronunciation. However, so far, no empirical evidence has supported the expectation that the different dialects can be identified correctly, leading to the first research question of this thesis: Can Vietnamese people distinguish Vietnamese Northern and Southern dialects?
Japanese dialect and Chinese Mandarin dialects on English speech. However, more empirical evidence is needed to investigate the influence of regional dialects on the L2, for the aforementioned studies have only focused on the influence of one dialect on the L2 production rather than investigating whether dialect varieties are retained and still noticeable in the L2 and they have focused on the production rather than the perception.
From my own experience as a native speaker of Vietnamese listening to Vietnamese L2 speakers of English, it appeared to me that, in the use and exchange of information in English, it is noticeable, not only that they were Vietnamese, but also from which part of the country they are. Therefore, for this thesis, it is predicted that speakers of the different Vietnamese dialects retain their regional pronunciation habits in foreign language speech. To be more specific, these habits appear to cause differences in pronunciation of L2 oral production that could be either correct or incorrect. The second research question seeks to find evidence for this prediction: Can Vietnamese people recognize different Vietnamese dialects in Vietnamese-‐influenced English speech?
Chapter 2 -‐ Literature background 2.1 The Vietnamese Language
For a long time, there were controversies around the categorization of the Vietnamese language, and it wasn’t until 1954, when the French linguist Haudricourt set a firm foundation, that a widely excepted categorization emerged. By decoding the history of the tones in Vietnamese, a member of the non-‐tonal Mon-‐Khmer language family, he conditioned other specialists to place Vietnamese in the same group with the Muong (Mường) language, named Viet-‐Muong, belonging to the Vietic branch of the Austro-‐Asiatic (South Asian) family (Nguyen, 1997 & 2009). Despite being located within the South Asian family, Vietnamese is the only language in the region that owns a writing system of Latin orthography.
Vietnamese national-‐language graphemes were transcribed for the first time by Catholic missionaries in the period from the 15th to the 16th century, who attempted to put the language in Latin graphic writings. In 1651, the first published trilingual dictionary of Vietnamese-‐Latin-‐Portuguese, written by the French missionary Alexandre de Rhodes, caused a huge impact on the nation’s language, especially the writing system. From then on, modern Vietnamese in Latin characters became increasingly popular and convenient to scholars in particular and the speakers in general. The characters were gradually accepted and recognized as the Vietnamese national language (chữ Quốc Ngữ). For this study, the concept of Vietnamese should be conventionally understood as the modern Vietnamese national language.
modern orthography originates from the earliest complete description of the language by de Rhodes (1651). This attempt describes Vietnamese as a monosyllabic language, formed with 36 phonemes (24 consonants and 12 vowels – see Table 2.1 & Table 2.2, adapted from Thompson, 1987). Vietnamese consonants are formed by single letters or combinations of letters, which are mostly identical to English consonants, with the exception of the consonant “Đ”, which is marked with a diacritic. In other words, Vietnamese consonants exist in the form of single, double and triple consonants. There are 27 different orthographically formed consonants, or consonant graphemes, listed by Thompson. However, the number of phonemic consonants is lower, due to identical pronunciation of consonant groups such as C/K/QU, D/GI NG/NGH and G/GH (Nguyen, 1997). In terms of vowels, Vietnamese vowels are also classified into 3 categories, namely monophthongs, diphthongs and triphthongs. These are very similar to those in English, with monophthongs being single vowels, like /ɔ/ in corn /kɔ:n/; diphthongs
being the glide from one vowel to another, like /ɔɪ/ in toy /tɔɪ/, and sometimes to the
third vowel as in triphthongs, such as /ɔɪə/ in loyal /lɔɪəl/. Table 2.1
Vietnamese consonant graphemes adapted from Thompson (1987)
B C CH D Đ G/GH GI H K KH
L M N NG/NGH NH P PH QU R S
T TH TR V X
Table 2.2
Vietnamese vowel alphabet adapted from Thompson (1987)
A Ă Â E Ê I O Ô Ơ U
Words in Vietnamese are usually combinations of one or more “syllabemes” (syllable + morpheme, or single written syllables – Pham, Bolger & Baayen, 2012). These syllabemes are formed by consonants and vowels in an order, which is bound to the existence of at least one vowel and an obligatory tonic diacritic (Nguyen, 1997 & 2009). Nguyen (1997) created a formula for Vietnamese syllabeme formation that is best illustrated by the rule: (C1) (w) V1 (V2) (C2), in which C, w, and V respectively stand for consonant, semivowel (a vowel not functioning as a nucleus of a syllable) and vowel. This formula holds that a syllabeme unit in Vietnamese does not necessarily need any consonants. Thus, the smallest syllabeme unit can be formed by a unique vowel and a tonic diacritic (e.g. ả ‘she’). The largest syllabeme unit can be as big as a 7-‐letter combination (e.g. nghiêng ‘askew’). Nguyen’s formula has confirmed the non-‐existence of pure consonant clusters in Vietnamese syllabeme structures, making each syllabeme unit a separate syllable in an utterance.
These features of syllabeme formation could well explain Vietnamese as a monosyllabic language. Besides, tonic factors, mandatorily appearing in syllabeme units, add to the complexity of this tonal language, which causes interesting variation in the quality of the phonemes, when compared to English.
The Vietnamese and English alphabets are almost identical. However, the Vietnamese alphabet (Table 2.3) does not include several letters that do appear in the English alphabet, such as F, J, W and Z, which are all consonants, causing Vietnamese to have less alphabetic consonants than English. Nevertheless, the Vietnamese alphabet includes a considerable number of vowels with diacritics, which do no occur in the English alphabet to compensate. Specifically, six more vowels are added to the five that also exist in the English alphabetic system.
Table 2.3
Vietnamese alphabets.
A Ă Â B C D Đ E Ê G
H I K L M N O Ô Ơ P
Q R S T U Ư V X Y
Vietnamese graphemes are normally reserved in the writing system of the language, regardless of any dialectal variations. However, pronunciation quality of these graphemes in Vietnamese surprisingly varies widely across dialects, which even leads to different phonemes representing the same graphemes. This phenomenon of differences in the pronunciation of Vietnamese consonants is illustrated in Figure 1 (Nguyen, 1997). For example, the grapheme r-‐ can be either pronounced as /z-‐/ in the North or /ʒ/ in the South (Nguyen). This big difference in pronunciation among Vietnamese dialects causes difficulty in producing a representative model of general phonemic consonants. Further discussion about this can be found in section 2.2.
Figure 1 -‐ Vietnamese consonants and their different pronunciations -‐ Nguyen (1997; 20)
noted that the phonemes in this table are not following IPA (International Phonetic Alphabet); the phonemic symbols for English consonants are more clearly described in Figure 3.
Figure 2 – English consonants (Department for Education and Skills, 2007)
Figure 3 – English IPA consonants (http://cmed.faculty.ku.edu/ipafolder/vowels.html)
With regard to vowels, although Vietnamese consists of significantly more alphabetic vowels than English, the latter shows more diversity in the vowel pronunciation. For Vietnamese, phonetic vowels are formed into monophthongs, diphthongs and triphthongs. 11 monophthongs are derived from 12 alphabetic vowels (Figure 4), with I and Y being pronounced identically. In total, three Vietnamese diphthongs are described by Nguyen (1997), which are the smooth movements of sounds from the sonorous [i], [u], and [ư] to the less sonorous [â̯]. The three diphthongs
position, and -‐iê-‐, -‐uô-‐, -‐ươ-‐ if they are followed by a final consonant. All other cases of two consecutive vowels occurring within the same syllable are constructed with a nuclear vowel (either of unrounded vowels) and a “medial vowel” (Nguyen), or semivowel /w/, spelled -‐u-‐ or -‐o-‐. However, Nguyen does not mention the case of /i/ occurring in the final position of a vowel cluster. In the same effort to describe Vietnamese vowels, Thompson (1987) accepted all cases of two-‐vowel clusters with one factor more prominent than the other as diphthongs. However, he proposed that Nguyen’s (1997) diphthongs, those written as -‐iê-‐, -‐uô-‐ and -‐ươ-‐, should only be called “vocalic sequences”, since it is difficult to decide which of the factors is more distinguished. In terms of triphthongs, neither Thompson (1987) nor Nguyen (1997) have actually mentioned Vietnamese triphthongs. However, based on Nguyen’s finding, it could be safe to describe triphthongs in Vietnamese as a combination of diphthongs and the semivowel /w/.
Figure 4 -‐ Vietnamese vowels (Nguyen, 2009)
shape of articulators like tongue heights and lip positions, highlight the complexity of English vowels, in comparison to Vietnamese vowels.
Figure 5 -‐ English IPA vowels (University of Kansas, 2003)
2.2 Vietnamese Dialects – Distinguishing Phoneme Systems
distinctions such as pronunciation at word and phrase level, however, need more empirical evidence into the linguistic aspects to be specified.
Despite the difference in pronunciation of Vietnamese words varying from region to region, all varieties refer to the same standard construct of orthography. The difference in pronunciation can be found at all levels: sentence level (intonations), word level (tones) and phonetic sound level (phonemes). The following part will focus on the difference in phonemes between the two biggest and most important dialects of Vietnam. Since Hanoi, as the cultural center, and Saigon (renamed Ho Chi Minh city), as the industrial center of Vietnam, have long attracted the biggest populations, both cities have close association with the two different dialects. This study focuses on the Vietnamese Northern dialects, regionally centering around Hanoi, and the Vietnamese Southern dialects, around Saigon.
2.2.1 Vietnamese Northern and Southern consonants. Nguyen’s (1997 & 2009) formula for Vietnamese syllabemes revealed two positions for consonants in the language: initial and final positions. Consonants and semivowels at initial position are called initials. Similarly, finals refer to consonants and semivowels at final position of a syllable. Although all Vietnamese consonants can occur initially, not every consonant can occur in the final position. Studies conducted by Thompson (1987), Nguyen (1997 & 2009) and Kirby (2011) agreed that the possible positions of Vietnamese consonants at the boundary of a syllable, determines their quality of pronunciation. Moreover, dialectal varieties also contribute to the differences of phoneme quality, since each dialect has its own consonant property regarding number and pronunciation.
2.2.1.1 Initials. Nguyen (2009) listed 19 NVN consonants, plus a glottal stop /ʔ/,
picture about NVN dialect (Figure 6). Several features were reconsidered, re-‐compared and re-‐described. In Kirby’s work, he denies /p/ as a Vietnamese consonants while Nguyen (1987) accepts this consonant, even though it only occurs in foreign or borrowed words. Moreover, Kirby (2011) added the labial approximant /w/ in his table when it is seen as a medial sound or semivowel. Regarding the most recent description of NVN consonant system, Kirby’s work has been chosen to be representative, since his paper follows IPA rules and is accessible for a broader number of readers.
Figure 6: Hanoi phonemic consonants (Kirby, 2011)
Despite the fact that 27 graphemes exist in orthography, NVN tends to group two or three graphemes into one, similar to the pronunciation. A notable example is sound description of consonants spelled CH and TR; these graphemes are identically articulated in NVN, as an unaspirated, unaffricated palatal stop (read /c/ in Thompson, 1987 and Nguyen, 2009, for they do not follow IPA), however, the sound is realized as a palatal affricate /tɕ/ (Harris, 2006). Secondly, consonant graphemes G and GH,
described by Thompson (1987) as pronounced in Vietnamese either as /ɠ/ (IPA – velar,
Meanwhile, SVN phonemes, especially consonants, involve different quality of pronunciation. As stated in Thompson (1987), “Saigonese consonants are closer to those of the writing system than their Hanoi equivalents: they show the distinction between plain and retroflexed sounds formed by the tip of the tongue” (p. 88). Accordingly, SVN consonants involve 6 positions of articulators (Figure 7), which mostly correlate with movement of different parts of the tongue: the tip (apical), the front (laminal-‐retroflex, -‐ nonretroflex) and the body (dorsal), while NVN does not show any retroflex quality (Figure 8). Besides differences in qualities of articulation, there are differences in the distinguishing consonant between the two dialects as well; while NVN tends to synchronize pronunciation of graphemes CH and TR into /tɕ/, S and X into /s/, GI, D and
R into /z/, SVN distinguishes them with different phonemes. While /tɕ/ is used to
express CH, TR is realized as a fortis stop, apical retroflex /ʈ/ (or /t̩/). Furthermore, S is
expressed as a fortis continuant, apical retroflex /ʂ/and X as /s/. Finally, GI and D are both realized as a fortis continuant, frontal /j/, while R as a lenis, apical retroflex /r/. In spite of the fact that the SVN consonants system has some phonemes that are absent in NVN, two consonants /z/ and /v/ do not exist in SVN.
Figure 7 -‐ Saigon (SVN) consonant description (Thompson, 1987)
Figure 8 -‐ Hanoi (NVN) consonant description (Thompson, 1987).
2.2.1.2 Finals. As already mentioned, Vietnamese consonants occur in two
positions: initial and final, with most of the consonants only being able to occur in the syllable initial position. Not only does Vietnamese have fewer phonemic consonants than English, it also possesses a fewer number of consonants that can occur as a syllable final. These final consonants, as described by Thompson (1987), are unreleased in pronunciation.
Vietnamese three-‐element nuclei, due to the lack of consonant clusters, include a vowel cluster (diphthong or a vowel sequence) and any of the finals. Observation of the tables of two-‐element nuclei and three-‐element nuclei occurring in both NVN and SVN (Figure 9, 10, 11 & 12) showed that consonant finals in three-‐element nuclei do not occur apart from those in two-‐element nuclei.
On the one hand, NVN finals are noted to consist of all 4 semivowels and 8 consonants. As listed in Figure 9, a combination made by choosing one of the 11 vowels (vertical) with any of the 12 finals (horizontal) constructs two-‐element nuclei in NVN. Pronunciation of these consonants as finals is reserved, which entails that they are not pronounced differently depending on their position.
Figure 9 -‐ NVN consonant finals in two-‐element nuclei – Thompson (1987)
On the other hand, SVN syllables can end with a smaller number of finals. Besides the 4 semivowels that can be finals, only 6 more consonants can be placed in this position (Figure 11). The absence of [ch] (read /tɕ/ -‐ following IPA) and NH (read /ɲ/ -‐ following IPA) in comparison with NVN, however, is not the mere difference. In fact, grapheme CH, pronounced /tɕ/ as initial, is pronounced as an apical stop, unreleased /t/ when it occurs in the final position: for instance, ếch (frog) is pronounced as [ết]. NH, similarly, transforms from /ɲ/ as initial to nasal /n/: lênh bênh (unstable) as [lên bên]. In addition, grapheme N occurring in the final position, becomes /ŋ/ as in bùn (mud) [bùŋ]; and T becomes /k/ as in ớt (pepper) [ớk]. These changes in pronunciations of certain graphemes occurring in the final position of a syllable in SVN have highlighted the basic difference between the two dialects.
Figure 12 -‐ SVN consonant finals in three-‐element nuclei -‐ Thompson (1987)
Figure 13 -‐ NVN vowel system -‐ Thompson (1987)
Figure 14 -‐ SVN vowel allophones -‐ Thompson (1959)
phenomenon that can be observed in Figure 15 regarding NVN is the pronunciation of the syllable spelled rượu; although NVN retains the number of vowels in the sequence, it changes the pronunciation of the nuclei dramatically from [ượw] into [iệu]. The final difference between the two dialects is the dissimilarity in treatment of semivowel [j] as a final following grapheme A. Semivowel [j] is spelled either as I or Y in Vietnamese. However, both these graphemes contribute to a different pronunciation quality when it comes after grapheme A. In NVN, the two nouns of tai (ears) and tay (arms) is respectively pronounced as [taj] and [tăj]. Thus, the change in appearance of this semivowel leads to a transformation in the articulation of previous grapheme A. On the contrary, SVN gives these two nouns the same quality of treatment; both syllables are pronounced [taj] (ears), which might cause serious misunderstanding in some situations.
Figure 15 -‐ Reduction of three-‐element to two-‐element nuclei (Thompson, 1987)
2.3 English Spoken by Speakers of the Different Vietnamese Dialects.
well. However, within the limit of this study, it was decided to only consider pronunciation at phoneme level due to the insufficient scientific concern about this aspect in identification of dialects and the priority given to the transfer of prosodic features from tonal languages (see 2.3.2).
the pronunciation of the new sounds. If the transfer is positive, which means identical sounds in both languages; it is more likely for L2 learners to correctly pronounce L2 words. On the contrary, negative transfer of distinguishing or new sounds from one language to the other, may lead to pronunciation errors. Being aware of the cause of L2 pronunciation errors, by tracing back the L1 phonology, is necessary to decode the differences in L2 oral speech produced by different speakers.
In Vietnam, English is included in the official curriculum as a mandatory foreign language subject at all levels, ranging from elementary to tertiary education. Thus, studying the effects of the Vietnamese language on English speech is part of the SLA field that focuses on the interference of the mother tongue generally and phonetic transfer particularly. This study field could contribute to a new basis for language pronunciation design so as to minimize errors in L2 speech caused by distinguishing phonology properties of different dialects.) According to Avery and Ehrlich (1995), phonological patterns of the native language might affect SLA because of 3 aspects: sound inventory, sound combination and intonation patterns. However, out of these three aspects, this study will only deal with the first two sources from the framework, as most of the previous studies have focused on Vietnamese acoustic properties (Brunelle, 2009a; 2009b; Brunelle, Ha & Grice, 2012; Vu, 1982).
2.3.1.1 Influence of sound inventory. One influence that might provide
Vietnamese speakers of English, it appears that it is either the similarity or the dissimilarity that influences the pronunciation, rather than a combination of both.
In the first place, having a number of phonemes with relatively the same qualities often results in the “Kellerman’s effect” for Vietnamese speakers, called “homoiophobia” (Kellerman, 2002), when they speak in English. “Homoiophobia” can be defined as the fear or caution that people experience when using equivalent structures in a foreign language. Structures here mean all possible linguistics aspects, such as syntax, lexical and phonology. Therefore, Vietnamese speakers of English producing sounds of the same quality are expected to avoid and be overly cautious in their production. However, to my knowledge, no evidence of this phenomenon in Vietnamese has ever been recorded in empirical research or experiments.
consonants and vowels all combined lead to difficulties for Vietnamese speakers with English words like “food” /fuːd/ vs “foot” /fʊt/ and “should” /ʃʊd/ vs “shoot” /ʃuːt/.
In Figure 16, English phonemes that became problematic for Vietnamese speakers and common mispronunciation of those phonemes are presented. However, it is unclear whether there is a difference between the errors made by speakers of the two dialects, and if so, which specific errors this entails. Although differences in phonology of NVN and SVN dialects are clearly perceived by linguistics in comprehension studies by Thompson (1959 & 1987) and Nguyen (1997 & 2009), there is still need for further research concerning the dialectal effects on English pronunciation.
Figure 16 -‐ English phonemes that may be difficult for Vietnamese speakers -‐ adapted from Center for Applied Linguistics (1976) by Hwa-‐Froelich, Hodson & Edwards (2002).
2.3.1.2 Influence of sound combination. The second possible L1 influence on L2
expressing tones), and optional consonants as initials or finals: (C1) (w) V1 (V2) (C2) (Nguyen, 1997 & 2009). Thus, Vietnamese does not consist of consonant clusters, but only vowel clusters. Meanwhile, the smallest unit of sound combination in English is composed of at least one syllable, which possibly includes consonant clusters at the border between syllables. For example, the word “contrast” has 2 syllables and such a consonant cluster: /st/. The presence of this factor in English, unknown to Vietnamese speakers, is assumed to be difficult to pronounce for Vietnamese speakers of English (Avery and Ehrlich, 1995).
Moreover, the habit of releasing the ending sounds in English is also often an obstacle for Vietnamese, which could be explained by the lack of released ending sounds in Vietnamese. Despite being a monosyllabic language, syllabemes are usually not uttered separately in Vietnamese. Instead, they form the smallest meaningful units of words, which are clusters of single syllables (Thompson, 1987). Furthermore, most of Vietnamese consonant finals are unreleased, made by the closing of the lips. Resulting in Vietnamese speakers tending to skip the ending sounds in English, due to the transfer of the handling of finals from the mother tongue. For example, “church” /tʃɜːtʃ/ is often
cut down, and pronounced as /tʃɜːt/, with the ending /t/ unreleased.
put in the context or quantifiers; for example in mấy/mọi/những/các cái bàn (equivalent to some or several). Similarly, words remain unchanged across tense and gender in Vietnamese, like in tôi ăn kem còn anh ấy ăn bánh (I eat ice-‐cream and he eats (a piece of) cake) and “hôm qua, tôi (đã) ăn kem còn anh ấy (đã) ăn bán (yesterday, I ate ice-‐ cream and he ate (a piece of) cake). The latter sentence is in past tense, which is marked by the adverbial clause of time: hôm qua (yesterday) and an optional functional word of time đã. Thus, changing English words by adding morphemes is new for Vietnamese speakers and often induces much confusion, resulting in over-‐pronunciation and redundant use of the ending /s/ and /z/. Moreover, because the two alveolar fricatives do not exist as finals in Vietnamese, when they occur as the same morpheme -‐s or -‐es, they are usually treated by speakers of Vietnamese as one /s/. As can be seen in Figure 17, /z/ as final is often substituted by /s/ by Vietnamese speakers. Furthermore, the nine English finals /b/, /d/, /t/, /v/, /s/, /ʃ/, /z/, /tʃ/ and /l/ are usually mispronounced and replaced by the phonemes in the right column. Again, Vietnamese dialectal factors affected in English consonant finals’ pronunciation have not received much interest from researchers so far.
comparisons have demonstrated that the folk employ factors other than linguistic differences in constructing their mental maps” (p. 307).
Studies about accent identification have attracted the interest of a number of linguists (Amino and Osanai, 2014; Bezooijen & Gooskens, 1999; Wu, Duchateau, Martens and Compernolle, 2010). There are two main methods to classify speakers, automatic applications and human listeners that are being used by scholars. For example, Wu, et al. (2010) developed automatic methods to identify L1 accents, based on their experiment with the Flemish language (see Wu, et al. for an overview of these methods). In their study, Amino and Osanai (2014) compared automatic methods to the identification by human listeners, and found that the human listeners were more reliable. Therefore, for the current study, it was decided to use human judges. However, it should be noted that Amino and Osanai add that, despite their advantage in discriminating between native and non-‐native speakers, human listeners are unable in identifying foreign accents; in other words, they are not able to identify the L1 of the speaker in their L2 production.
pronunciation can only be used to identify dialects on, at least, a national scale. Therefore, within a language, it is necessary to investigate pronunciation of phonemes and phoneme features; “to focus on the salient differences between native accents, selecting accent relevant features on a phonemic specific basis becomes crucial” (Wu, et al., 2010, p. 84). However, as for a tonal language that bears huge differences between its two main dialects, like Vietnam, it could be unsafe to conclude the same.
Recently, there is a growing body of studies interested in the influence of L1 dialects on L2 speech, such as studies of Yoshizawa (2012), and Qin and Tremblay (2013). Yoshizawa investigated the influence of Japanese Fukisima dialect intonation on English pitch range and found a relation between this native dialect and English intonation and stress, and that speakers with higher ability of discriminating phonemes reveal to be more advanced in English speech. Of the same effort to look at the effects of dialects of mother tongue on L2 speech, Qin and Tremblay (2013) compared the prosodic transfer of Standard Chinese Mandarin and Taiwan Mandarin in L2 English production. The study of Qin and Tremblay suggested the presence of different effects on L2 speech from different dialects by analyzing the ability of speakers of these dialects to use prosodic cues in L1 to encode L2 stress. Though both studies have paid attention to influence of L1 dialects on L2 speech, they looked at the L1 transfer existing in L2 production rather than deal with the perception of L1 properties retained in L2 speech, and prosodic features rather than phoneme properties.
2.4 Research questions
Vietnamese dialects is tested, both in the L1 (Vietnamese) and the L2 (English). The purpose of the thesis is to answer two main research questions:
1. Is it possible for Vietnamese people to correctly identify the two different Vietnamese accents?
2. Is it possible for Vietnamese people to recognize the different Vietnamese dialects in Vietnamese L2 English production?
Due to the clear differences between the phoneme systems of the two dialects (NVN and SVN), it is expected that both in the Vietnamese and English production, Vietnamese people will be able to identify the dialect of the speaker.
Chapter 3 -‐ Methodology 3.1 Experiment
This experiment entails the recordings of a narrative in English by Vietnamese, Thai and Chinese speakers, and the recordings in Vietnamese of a narrative by Vietnamese speakers. Subsequently, two surveys were created, with the first one seeking for the ability to identify Vietnamese dialects spoken by Vietnamese speakers from different regions of the country, and the second survey examining the ability of the same group of listeners to still identify the Vietnamese dialect factors in the English narratives. All participants taking part in the experiment were selected from a questionnaire passing through the conductor’s network. Responses received from Listeners were analyzed by SPSS. The participants were asked to join a post-‐test interview for further details about themselves and their performance.
3.2 Recordings
1987). These foreign accents would play an important role in assuring realization of Vietnamese among themselves.
Table 3.1
Information of Speakers.
Origin Speaker Gender Age Profession English achievement
NVN 1 Male 23 Engineer Only in school
NVN 2 Female 24 Bank Officer TOEIC 660
NVN 3 Female 23 Accountant Only in school
NVN 4 Male 27 Teacher of Physics Only in school
NVN 5 Male 22 Accountant Only in school
SVN 6 Female 22 Student IELTS 6.0
SVN 7 Male 30 Teacher of Math Only in school
SVN 8 Female 20 Student Only in school
SVN 9 Male 27 Marketing Officer IELTS 6.0
SVN 10 Male 22 Student Only in school
Thai 11 Male 26 Student Only in school
Thai 12 Female 24 Student Only in school
Chinese 13 Male 25 Student Only in school
Chinese 14 Female 23 Student Only in school
Chinese 15 Female 23 Student Only in school
3.2.2. Materials
3.2.2.1. Background questionnaire. A questionnaire was designed to collect all
questionnaire included 12 questions, with 5 open questions and the other 7 questions being multiple-‐choice (See Appendix B). First, participants were asked to give basic information, such as their name and email address. This information would later be used to contact each of them privately. Second, participants were asked to reflect on what dialect they were speaking. Based on the fact that many Vietnamese families have moved around to work or to settle their lives in a different part of the country after Vietnam became reunited in 1975, many Vietnamese people nowadays no longer speak the dialects of their hometowns. Thus, the question was posed as an important clue to classify different-‐dialect speakers among the respondents. Third, a range of additional questions was asked about the English background and experience of the speakers. Among these questions, they had to indicate the approximate time of learning English, the highest level of English proficiency they had achieved and their current state of usage and exposition to English. This information was to spot people, whose English is close to represent Vietnamese-‐influenced English spoken by the major number of the population. All responses were collected in a Microsoft Excel file so that it was easier to classify the participants.
3.2.2.2. Retelling story activity. “The Tortoise and the Hare”, a famous fable, was
chosen to be the content for the story-‐retelling activity. Due to its popularity, it was expected that the story was already well known to all participants taking part in the study. 10 screen-‐shots were taken from a 2-‐minute video found on YouTube (2014). They were arranged chronically in order to help the participants remembering the content of the story. In addition, a short glossary with several key words was also distributed (See Appendix C).
Tortoise to take part in a running race. The Hare believes that he will definitely be the winner so he agrees to set up the race. At the beginning of the race, the Hare ran far ahead of the Tortoise. He looks back and sees the Tortoise trying his best, moving step by step, very slowly. The Hare thinks that it would be more fun if he went pass, and win from the Tortoise once he nearly reaches the finish line. So, he decides to take a nap while waiting for the Tortoise to come. However, the Hare oversleeps, and only wakes up when the Tortoise is a few steps away from the finish line. The Hare gets up and runs as fast as he can, but fails to win from the Tortoise.
3.2.2.3. Recording devices. Interviews were implemented 100% online,
supported by any of the following online calling applications: Skype, Viber, WhatsApp and Facetime. Speakers were free to choose which tool was most convenient for them. Conversations were recorded with an OS-‐supported application named eXtra Voice Recorder Lite. This application was run simultaneously with one of the calling applications during the interviews. All conversation recordings were later edited so that they only included the storytelling of the speaker. The surrounding conversations and background noises were deleted by using Fluctus -‐ another OS-‐supported application.
Speakers were asked in an interview with the conductor to retell the familiar story of “the Tortoise and the Hare” after looking at the 10 screenshots (Appendix B). In most of the interviews, the Speakers indeed showed comfort and familiarity with the story. To support them in their English, a short glossary was also distributed, which required the to use words, such as “tortoise” and “hare” to describe 2 main characters. However, they found “turtle” and “rabbit” to be more accustomed to use. The pictures were only as a reminder of the story’s content. The Speakers were asked to retell the story in their own words in English. Vietnamese speakers were also required to tell another version in Vietnamese. Interviews were made online and conversations were recorded without interviewees noticing.
3.3 Surveys
3.3.1. Participants. 60 people were asked to participate in the second part of the experiment by filling in two surveys. However, some completed both surveys while others were only interested in doing one. As a result, different numbers of responses were received for both surveys: 38 for Survey 1 and 34 for Survey 2. They were all between 20 and 30 years old, and all of the participants were chosen for having at least a tertiary education and having high contact with English (several times a week), in the setting of Vietnam (see Table 3.2).
Table 3.2
Information of Listeners.
Origin Age Education English Use
Frequency
NORTH SOUTH CENTER
Survey 1 18 9 11 20-‐30
tertiary and post-‐ tertiary
Several times per week
Survey 2 17 9 8 20-‐30
tertiary and post-‐ tertiary
3.3.2. Materials and Procedure. From the initial 80 people who had completed the questionnaire, another 60 participants were chosen, based on both the quantity and quality of contact with English, and their English proficiency. These 60 Vietnamese people made a group of Listeners, who had a higher level of English proficiency, a more frequent use and more contact with the English language than the Speakers. They were people from all three major parts of Vietnam: North – Centre – South. It was expected that this Listeners group would be able to identify differences in English spoken by different-‐dialect speakers, due to their frequent use and high contact with the language.
The edited recordings were uploaded to SurveyGizmo (2014), a website allowing the production of surveys and the collection of data in the form of PDF files. Two separate surveys, requesting the answering of 1 question per audio file, were made. All audio files were complete narratives told by the Speakers, with the length of up to 1 minute. They were arranged in a random order and set anonymous. Survey 1 was about ability to identify dialects of Vietnamese. It is composed of 10 recordings in Vietnamese, each of which was enclosed with a question designed as multiple choices of 4 options:
Where does the speaker come from? a. North Vietnam
b. South Vietnam
c. Neither North nor South Vietnam d. I don’t know
Similarly, Survey 2 conveyed the idea of the Listeners regarding the origin of the speaker, after they listened to English narratives. It was designed with 15 recordings spoken in English by both Vietnamese and non-‐Vietnamese (Thai and Chinese). Each recording came along with a question of 4 options as well.
