![]() |
ENQUIRE PROJECT DETAILS BY GENERAL PUBLIC |
Project Details |
Funding Scheme : | General Research Fund | |||||||||||||||||||||
Project Number : | 18601420 | |||||||||||||||||||||
Project Title(English) : | Third language (L3) phonological development for multilingual learners in the Chinese context | |||||||||||||||||||||
Project Title(Chinese) : | 在華人語境下多語言學習者的第三語言語音發展 | |||||||||||||||||||||
Principal Investigator(English) : | Dr Chen, Hsueh Chu | |||||||||||||||||||||
Principal Investigator(Chinese) : | ||||||||||||||||||||||
Department : | Department of Linguistics and Modern Language Studies | |||||||||||||||||||||
Institution : | The Education University of Hong Kong | |||||||||||||||||||||
E-mail Address : | hsuehchu@eduhk.hk | |||||||||||||||||||||
Tel : | 2948 7376 | |||||||||||||||||||||
Co - Investigator(s) : |
|
|||||||||||||||||||||
Panel : | Humanities, Social Sciences | |||||||||||||||||||||
Subject Area : | Education | |||||||||||||||||||||
Exercise Year : | 2020 / 21 | |||||||||||||||||||||
Fund Approved : | 747,960 | |||||||||||||||||||||
Project Status : | Completed | |||||||||||||||||||||
Completion Date : | 30-11-2023 | |||||||||||||||||||||
Project Objectives : |
|
|||||||||||||||||||||
Abstract as per original application (English/Chinese): |
‘兩文三語’即鼓勵香港市民會以中文和英文書寫,並使用廣東話,英語和普通話三語進行口語交流的政策在影響著香港中小學課程的設計和實施。自1997年香港回歸以來,該政策對香港的語言使用產生了影響 (Wang&Kirkpatrick, 2013)。大多數香港人已經意識到提升自己的第二語言和第三語言的水平並滿足日常交際和工作的迫切需要。但是,在我們的日常生活中通常可以觀察到,如果一位來自香港的粵語為母語的英語學習者(英語為第二語言),他/她 ‘three’ /θri:/ 這個單詞的發音可能會變成為 /fri:/。如果一位來自廣東省的粵語為母語的英語學習者(普通話為第二語言,英語為第三語言),‘three’ /θri:/ 這個單詞的發音可能會變為 /sri:/。人們似乎會注意到,當英語學習者說英文時,發音常常被當作最重要的考察因素之一。非母語發音幾乎可以立即被識別出,並且這種發音可能會產生一些不良後果 (Flege, 1995)。聽眾常常把不標準的非母語發音貼上“慢”,“需要幫助”以及可能會引起無效溝通的標籤。 關於華語環境下多語種的第三語言語音發展,以下兩個重要的問題應當被研究: 1)從第一語言到第二語言以及第三語言的語音特徵的跨語言轉換模式是什麼? 2)第二語言在語音特徵向第三語言的轉移的程度跟語言水平的關系? 該項目旨在通過以下方式回答上述兩個基本問題:a) 找到粵語學習者在英語和普通話的主要發音錯誤; b) 對母語為廣東話的學習者的第一語言,廣東話,在英語和普通話的習得的影響進行聲學研究; c) 確定多語言學習者習得第二語言以及第三語言語音特徵,特別關注第一語言或第二語言在第三語言學習過程中的影響。 該項目在兩方面創新:(a) 將設計更全面評估英語和普通話發音的任務。 上述任務將會被用於判定學習者學習英語以及普通話發音的困難;以及 (b) 將精心設計自我反思指引來幫助學習者對第二語言和第三語言語音學習過程和經驗進行自我反思。該項目將確定對語言學習者在第二語言以及第三語言語音習得影響最大的因素,從而為在多語言習得和多語言教育中的理論和實踐做出貢獻。該項目將設計考察語言學習者第二語言和第三語言發音的任務。研究結果可以用來開發語音教師培訓資源,以促進第二語言和第三語言語音教學,將英語和普通話教師聯繫在一起,來探索和解決母語為粵語的學習者的發音問題。 |
|||||||||||||||||||||
Realisation of objectives: | Objective 1 has been fully achieved. A set of comprehensive evaluations tasks has been constructed for identifying pronunciation features of Cantonese learners of English and Mandarin. A series of speech tasks were designed for assessing phonological and phonetic features of three languages. The Cantonese tasks included a reading of a list of monosyllabic words and a reading of a passage. In the list of monosyllabic words, all Cantonese vowels were involved in two types of words, one meaningful word and one pseudo word with the test vowel preceded by a glottal fricative /h/. The passage was the Cantonese version of ‘The North Wind and the Sun’, adapted to include all Cantonese consonants, vowels, and tones. The English tasks included the read-aloud of a list of meaningful words containing all English vowels and consonants, a list of meaningful and pseudo words (with /h/ as the initial consonant) containing all English vowels, and a passage ‘The Boy Who Cried Wolf’ containing all English vowels and consonants. A controlled free speech task was also included, to elicit authentic and spontaneous speech by providing the topic ‘an unforgettable trip’ and a series of guided questions. The Mandarin tasks included the read-aloud of a list of monosyllabic and multisyllabic meaningful words, a list of monosyllabic meaningful and pseudo words, a passage, and a controlled free speech task. In the first reading task, the test monosyllabic words include all Mandarin vowels, consonants, and tones; the test multisyllabic words contain the 20 permutations of tones, the tone sandhi of 一 ‘yi’, 不 ‘bu’, and Tone 3 (dipping tone), and r-colouring. In the second reading task, all Mandarin vowels are included in one meaningful word and one pseudo word with the test vowel preceded by a glottal fricative /h/. The passage was the Mandarin version of ‘The North Wind and the Sun’, which includes all Mandarin vowels, consonants, and tones. In the controlled free speech task, the topic and guided questions are the same as those for eliciting English speech data. The evaluation tasks were used for collecting speech data in three phases, to assess the participants’ pronunciation features longitudinally. The speech data were analyzed to create a diagnostic profile of Cantonese learners' pronunciation errors in English and Mandarin. This profile aids multilingual education in the Greater Bay area and highlights negative transfer between the languages, linking cross-linguistic transfer to changes in language proficiency. Objective 2 has been fully achieved. The most frequently occurring pronunciation errors in a language identified by the phonological analysis were selected, along with the relevant sounds in the other two languages, to be the targets for the acoustic analysis across the three phases. Six categories of sounds were acoustically analyzed to investigate the possible cross-linguistic influences among Cantonese, English, and Mandarin. The first category concerned vowel tensing, since the errors of /ɪ/ pronounced as /i/ and /ʊ/ pronounced as /u/ were frequently found in the English speech for the participants. Therefore, the F1 and F2 values of /i/ and /u/ in the three languages and /ɪ/ and /ʊ/ in English were analyzed acoustically using Praat and compared in the F1/F2 plane. The second category is about vowel raising, because /æ/ was frequently pronounced as /e/ in English. /e/ in both English and Cantonese and /æ/ in English were analyzed acoustically for F1 and F2 values. The third category concerned vowel centralization, specifically, Mandarin /ɣ/ pronounced as the centralized vowels /ə/, /ɜː/, or /œ/. F1 and F2 values of /œ/ in Cantonese, /ə/ or /ɜː/ in English, and /ɣ/ in Mandarin were measured and compared in the F1/F2 plane. The fourth category concerned the lack of r-coloring for /ɚ/. The rate of change in F3 for /ɚ/ in Mandarin and /ər/ or /ɜːr/ in English were computed and compared. The fifth category considered the aspiration of /k/ in English and unaspiration and voicing of /ph/ in Mandarin. The last category is about the lack of neutral tone in Mandarin. The change in fundamental frequency (F0), the size of phrasal pitch span, and duration ratio for specific phrases in the three languages. There were two types of phrases in Mandarin, the first type contained a function word carrying the neutral tone preceded by content words (e.g., 拼命地 ‘desperately’ /pʰɪn55 mɪŋ51 tɤ0/), and the second type involved a content word with the second syllable carrying the neutral tone (e.g., 太陽 ‘the sun’ /tʰaɪ51 jɑŋ0/). The acoustic data were analyzed across languages and phases to explore dynamic cross-linguistic transfer. Additionally, a comparison of acoustic differences between participants and native speakers was conducted to further understand these variations. Objective 3 has been fully achieved. Immediately after the speech tasks, a series of verbal protocols adapted from Wrembel’s (2015) think-aloud protocols (TAPs) were conducted to elicit meta-phonological awareness from the participants. The verbal protocols include analytic listening, retrospective protocol and introspective protocol. During the analytic listening, the participants listened to the audio recordings of their L2 and L3 speech, compared their own pronunciation with that of native speakers, and noted down the pronunciation errors. This process allowed participants to identify discrepancies between their pronunciation and that of native speakers, fostering self-awareness. For the retrospective protocol, the participants provided commentaries on their L2 and L3 phonetic performances following the guided questions, prompting them to conduct self-analysis of pronunciation using meta-languages. For the introspective protocol, the participants performed self-reflection on their own L2 and L3 acquisition experiences , shedding light on cross-linguistic influences and the array of learning strategies employed. Their verbal performances of the protocols were audio recorded, transcribed, and coded. The retrospective protocol data were analyzed in terms of instances of successful and unsuccessful L2 and L3 self-repair and levels of complexity of reported phonological awareness. The introspective protocols were analyzed from two aspects, cross-linguistic influence and learning strategies. The latter aspect was further categorized into memory, cognitive, compensatory, metacognitive, effective, and social (Oxford, 1990). L2 listeners’ perceived comprehensibility and attitudes towards speech produced by L3 English learners from China were also collected and analysed. After a year of English and Mandarin learning, the participants showed notable improvement in high level awareness of meta-phonology, as they could use metalanguage to describe or explain features in their L2/L3 pronunciation. Both HK and GD groups reported progressive and regressive transfers in pronunciation among the three languages. HK participants used both L1 Cantonese and L2 English to facilitate the learning of L3 Mandarin, while most GD participants predominantly used their L1 Cantonese for facilitating the learning of L2 Mandarin. Additionally, both groups reported regressive transfer from L2/L3 Mandarin to L1 Cantonese resulting from the increasing use of Mandarin. The findings highlighted the complex and bidirectional nature of language interface for multilingual speakers, and provided more support and explanations for the language interaction patterns. | |||||||||||||||||||||
Summary of objectives addressed: |
|
|||||||||||||||||||||
Research Outcome | ||||||||||||||||||||||
Major findings and research outcome: | The findings highlight key pronunciation issues for Cantonese learners in English and Mandarin and how these errors change over time. For both Hong Kong (HK) and Guangdong (GD) speakers, common English pronunciation errors include substituting /æ/ with /e/ and /ɪ/ with /i/. Among HK speakers, the error of substituting /æ/ with /e/ decreased, while GD speakers increased in substituting /ɪ/ with /i/. Common consonant errors include substituting /ð/ with /d/ and /z/ with /s/, with the latter error decreasing among GD speakers. In Mandarin, common errors include issues with ‘erhua’ and the substitution of /ʂ/ with /ɕ/ among HK speakers. Tone errors are prevalent, especially with incomplete Tone 3, which is more common in GD speakers but has decreased over time. HK speakers frequently mispronounce Tone 3 as Tones 1, 2, or 4, and both groups often use lexical tones instead of the neutral tone in di- and multi-syllabic words, with a higher error rate in GD speakers. The acoustic analysis identified four potential cross-linguistic influence (CLI) patterns: L1→L3, L2→L3, L1+L2→L3, and L3→L2. CLI from L1 Cantonese to L2/L3 English was observed in the pronunciation of the English vowel [u], and a combined CLI from L1 Cantonese and L2/L3 English to L2/L3 Mandarin [u] was noted for both groups. Regressive CLI from L3 English to L2 Mandarin was detected for Mandarin [ɤ] in GD speakers. Among HK speakers with advanced English, CLI was observed from L2 English to L3 Mandarin. GD speakers exhibited a CLI from L2 Mandarin to L3 English in increased pitch span size, while HK speakers with advanced Mandarin showed a regressive CLI from L3 Mandarin to L2 English, resulting in increased pitch span size. Finally, meta-phonological awareness protocols revealed interaction patterns among the three languages. In Phase Three, both HK and GD participants demonstrated increased meta-phonological awareness compared to Phase One. Both groups reported experiencing progressive and regressive transfers in pronunciation across Cantonese, English, and Mandarin. HK participants utilized different strategies from both L1 Cantonese and L2 English when learning L3 Mandarin, while GD participants primarily relied on L1 Cantonese when learning Mandarin as their L2. Both groups noted that their L1 Cantonese was influenced by regressive CLI from Mandarin as its usage increased, regardless of whether Mandarin was their L2 or L3. | |||||||||||||||||||||
Potential for further development of the research and the proposed course of action: |
The project can be further developed from three aspects – L3 production training, L3 perception, and L3 phonology in other language contexts. Based on the findings in the current project, targeted training programs can be developed to improve pronunciation accuracy for the two groups of language learners. The effectiveness of such training program will be evaluated. It will be promoted through knowledge transfer activities if it is proven effective. Despite the examination of L3 production in this project, an investigation of L3 perception can be conducted in the future. Cantonese L1 speakers in Hong Kong and Guangdong and monolingual speakers of Cantonese, English, and Mandarin may be invited to perform a series of perception tasks using the production data in the current project, such as identification, discrimination, and dictation of target sounds, and ratings of comprehensibility and intelligibility of speech data. The perception data can be compared with production data in this project. Moreover, the examination of L3 phonology can be expanded to other linguistic contexts in Hong Kong, especially for the ethnic minority groups, who have languages such as Filipino, Hindi, Urdu as their L1, English as the L2, and Cantonese as the L3. | |||||||||||||||||||||
Layman's Summary of Completion Report: | In the trilingual education system in Hong Kong, most learners speak Cantonese as their L1, English as their L2, and Mandarin as their L3. The acquisition of these languages is complex and dynamic, with significant cross-linguistic influences. A trilingual speaker’s pronunciation in one language can be positively or negatively impacted by their proficiency in the others. This project aimed to explore these cross-linguistic influences for Cantonese L1 speakers by a) identifying major pronunciation features in English and Mandarin through evaluation tasks; b) examining how Cantonese affects the acquisition of English and Mandarin, and investigating possible phonological transfers from L2 to L3 using acoustic analysis; and c) understanding how multilingual learners acquire L2 and L3 pronunciation, particularly how their L1 or L2 interacts with L3 during the learning process, through meta-phonological awareness protocols. The significance of this project is threefold. First, it provides empirical evidence to enhance language acquisition theories for L2 and L3. Second, it helps establish diagnostic profiles for pronunciation in L1, L2, and L3, supporting more effective lesson planning and teaching strategies. Lastly, the project's impact is amplified through visibility in open-access platforms, international conferences, and prestigious journals, contributing to the global discourse on language education. | |||||||||||||||||||||
Research Output | ||||||||||||||||||||||
Peer-reviewed journal publication(s) arising directly from this research project : (* denotes the corresponding author) |
|
|||||||||||||||||||||
Recognized international conference(s) in which paper(s) related to this research project was/were delivered : |
|
|||||||||||||||||||||
Other impact (e.g. award of patents or prizes, collaboration with other research institutions, technology transfer, etc.): |
SCREEN ID: SCRRM00542 |