Patrick Chu, PhD
Research on Second language (L2) speech perception / word recognition
Bibliographies of Accent perception and adaptation / L2 speech research
Towards a Model of Second Language Word Production and Recognition in Mandarin
Patrick Chun Kau Chu University of New South Wales, Sydney, Australia
The production and recognition of Mandarin words by Cantonese speakers are influenced by the pronunciation correspondences between Cantonese and Mandarin in all sublexical levels (i.e., onset, rime and tone). A three-route second language word production and recognition model is proposed using these pronunciation correspondences. A concept route links the concept and the L2 Mandarin phonological representation directly, while a lexical route links the concept and the L2 Mandarin phonological representation through the L1 Cantonese phonological representation. A sublexical route activates the concept and/or the L2 Mandarin phonological representation through the mediation of L1 Cantonese phonological and sublexical representation using Cantonese-Mandarin pronunciation correspondences. Beginning learners of Cantonese mainly use the sublexical route in L2 Mandarin word production and recognition. Advanced learners gradually shift from the sublexical to lexical/concept route in producing Mandarin words while both lexical/conceptual and sublexical routes are still in active use to generate possible word candidates in L2 Mandarin word recognition. Evidence for the proposed model is drawn from a Mandarin word production task, a Mandarin disyllabic word transcription task, a Mandarin pinyin transcription task, and a character-sound matching task. Further testing of the model in the neurolinguistics and computational domain are also proposed.
The Influence of the L1 Lexical System on the Processing of Tones in L2
Patrick Chun Kau Chu and Marcus Taft University of New South Wales, Sydney, Australia
Systematic correspondences can be found between Cantonese and Mandarin tones. For example, most of the words pronounced with tone 2 in Cantonese are pronounced with tone 3 in Mandarin (e.g., 找 ‘find’). Therefore, Cantonese speakers may mistakenly think that words where the Mandarin tone pronunciation does not follow the dominant pronunciation relationship (e.g., 摸 ‘touch’, pronounced with Cantonese tone 2, but Mandarin tone 1) are pronounced in Mandarin with the dominant correspondence (i.e., tone 3). A Mandarin character-sound matching task using words which either employed a dominant (i.e., regular) or subdominant (i.e., irregular) tone relationship was carried out to examine whether the processing of Mandarin tones by Cantonese speakers was influenced by such relationships. The auditory stimulus either had a matched or a tone-mismatched pronunciation, where the mismatched pronunciations of irregular and regular words follow the dominant and subdominant tone relationships respectively. Results indicated that more mismatched pronunciations in irregular than regular Mandarin words were treated as the correct pronunciations for Cantonese speakers. The findings indicate that the processing of tones in an L2 is influenced by the L1 lexical system, not just the native phonological system.
Regularity and Congruency effect in Cantonese speakers’ phonological knowledge of Mandarin words
Patrick Chun Kau Chu and Marcus Taft University of New South Wales, Sydney, Australia
Correspondences exist between Cantonese and Mandarin at the sublexical level (i.e., onset, rime and tone) and it was proposed that Cantonese speakers employ both lexical and sublexical route in producing L2 Mandarin words. The involvement of the sublexical route was examined in a Mandarin pinyin transcription task using words which either employ a dominant (i.e. regular) or non-dominant (i.e. irregular) corresponding relationships in different sublexical units. More errors were found for irregular than regular words for all sublexical units. In addition, whether the L1 and L2 sublexical units of the same word are pronounced the same in the two languages (i.e. congruency) had an influence on error rate, though with a strength mediated by the knowledge of Mandarin phonotactic constraints. The results give support and add further constraints to the use of the sublexical route in L2 Mandarin word production
Do first-language (L1) phonemic categories play a role in the perception of second-language (L2) phonemic contrasts? A look from the perception of Cantonese codas by Mandarin speakers
Patrick Chun Kau Chu University of New South Wales, Sydney, Australia and Chinese University of Hong Kong
This study investigated whether the phonological system of L1 or universal hierarchy of perceptual similarities has a larger effect on Mandarin speakers’ perception of Cantonese phonemic categories. In Mandarin and Cantonese, there are differences in the distribution of nasals and plosives in the coda position. Both nasals (/m/, /n/ and /ng/) and plosives (/p/, /t/ and /k/) are possible codas in Cantonese while only nasals (/n/ and /ng/) can appear in the coda position in Mandarin. Non-native phonemic categories (e.g. /t/) will be assimilated to a native phonemic category (e.g. /n/) if L1 plays a larger role while assimilation to a non-native phonemic category (e.g. /p/ or /k/) may support the idea that universal hierarchy of perceptual similarities plays a more important role. An AX discrimination task is carried out to see whether Mandarin and Cantonese speakers can discriminate these six phonemic contrasts in Cantonese monosyllabic words. Results showed that Mandarin speakers have poorer discrimination ability in the later type (e.g. /t/ vs. /k/) than the former type (e.g. /t/ vs. /n/). This suggests that the Perceptual Assimilation Model may need to take universal hierarchies of perceptual similarities into accounts when predicting L2 learners’ discrimination ability in a second language.
Positive and Negative Transfers in the Pronunciation Learning of Japanese Onyomi Kanji by Native Cantonese Speakers
Patrick Chun Kau Chu Chinese University of Hong Kong
Most of the L2 learners realize that the pronunciations of Japanese onyomi kanji are similar to the Cantonese pronunciation but they do not know the rules explicitly. Lee (1992) proposed the onset, rhyme and coda corresponding rules that map the pronunciation of Cantonese characters and Japanese onyomi kanji characters. For example, Cantonese characters with a coda /t/ (e.g. 突 /tsɯ/) or /k/ (e.g. 特 /kɯ/) correspond mainly to the
Japanese mora /tsɯ/ (e.g. とつ /to.tsɯ/) and /kɯ/ (e.g. とく/to.kɯ/) respectively. This study investigates whether these Cantonese-Japanese corresponding rules can facilitate the pronunciation learning of new Japanese onyomi kanji vocabularies by native Cantonese speakers.
A questionnaire with multiple choice questions was distributed to 76 participants to test their knowledge on the pronunciation of some Japanese vocabularies and their corresponding Cantonese characters. The participants were divided into four groups according to their proficiency in Japanese. They were encouraged to guess and choose the most appropriate pronunciation if they had not learnt the Japanese vocabularies before. Accuracy rates and error distributions for each item were computed and an error analysis was conducted to identify L1 transfer effects.
We observed that the learners use the corresponding rules implicitly when deducing the pronunciation of unfamiliar Japanese vocabularies. However, we also observed some instances of negative transfer from Cantonese. Some Cantonese speakers had difficulties in distinguishing Chinese characters with /n/-/ŋ/ pairs and /t/-/k/ pairs in the coda position. These problems were also reflected in their Japanese errors. For example, learners who erroneously classified a Cantonese character with a coda /t/ as /k/ (e.g. 突) would pronounce the Japanese character as と く /to.kɯ/. Pedagogical issues related to the use of the corresponding rules as a pronunciation learning strategy will be discussed.
Reference:
Wood Hung Lee. (1992). A Guide to Japanese Pronunciation for Cantonese Speakers. Hong Kong: The Chinese University Press.
Japanese mora /tsɯ/ (e.g. とつ /to.tsɯ/) and /kɯ/ (e.g. とく/to.kɯ/) respectively. This study investigates whether these Cantonese-Japanese corresponding rules can facilitate the pronunciation learning of new Japanese onyomi kanji vocabularies by native Cantonese speakers.
A questionnaire with multiple choice questions was distributed to 76 participants to test their knowledge on the pronunciation of some Japanese vocabularies and their corresponding Cantonese characters. The participants were divided into four groups according to their proficiency in Japanese. They were encouraged to guess and choose the most appropriate pronunciation if they had not learnt the Japanese vocabularies before. Accuracy rates and error distributions for each item were computed and an error analysis was conducted to identify L1 transfer effects.
We observed that the learners use the corresponding rules implicitly when deducing the pronunciation of unfamiliar Japanese vocabularies. However, we also observed some instances of negative transfer from Cantonese. Some Cantonese speakers had difficulties in distinguishing Chinese characters with /n/-/ŋ/ pairs and /t/-/k/ pairs in the coda position. These problems were also reflected in their Japanese errors. For example, learners who erroneously classified a Cantonese character with a coda /t/ as /k/ (e.g. 突) would pronounce the Japanese character as と く /to.kɯ/. Pedagogical issues related to the use of the corresponding rules as a pronunciation learning strategy will be discussed.
Reference:
Wood Hung Lee. (1992). A Guide to Japanese Pronunciation for Cantonese Speakers. Hong Kong: The Chinese University Press.
The Perception of Cantonese Codas by Native Speakers of Cantonese and Mandarin
Patrick Chun Kau Chu Chinese University of Hong Kong
This study explores the perception of Cantonese codas by both Mandarin and Cantonese speakers through the use of a sound-character discrimination task. Previous studies by Winters (2003) showed that both native speakers of English and Dutch have similar hierarchy of perceptual similarities when perceiving phonemic contrasts in English. This may suggest that there are some universals in speech perception no matter whether the language is a first or second language of the speakers. In Mandarin and Cantonese, there are differences in the distribution of plosives and nasals in the coda position. We would like to see whether Mandarin and Cantonese speakers will observe the same hierarchy of perceptual similarities when perceiving phonemic contrasts in Cantonese codas. Moreover, we are interested to see whether the phonological system of the native language will have any effect on speakers’ perception of another language. We have used error analysis and analysis of perceptual distance in analyzing the responses made by the participants in the discrimination task. Some similarities and differences on hierarchy of perceptual similarities between Mandarin and Cantonese speakers were observed in the experiment. Results showed that both universal hierarchies on perceptual distance and
influence from native language would need to be considered in order to explore the competence of the phonological system of a second language speaker. However, as the results also showed that native Cantonese speakers also encountered difficulties in distinguishing phonemic contrast in Cantonese codas, an effect of input may also need to be considered when accounting for the errors made by Mandarin speakers.
Reference:
Stephen Winters. (2003). Empirical investigations into the perceptual and articulatory origins of cross-linguistic asymmetries in place assimilation. Ph.D. Dissertation, Ohio State University.
influence from native language would need to be considered in order to explore the competence of the phonological system of a second language speaker. However, as the results also showed that native Cantonese speakers also encountered difficulties in distinguishing phonemic contrast in Cantonese codas, an effect of input may also need to be considered when accounting for the errors made by Mandarin speakers.
Reference:
Stephen Winters. (2003). Empirical investigations into the perceptual and articulatory origins of cross-linguistic asymmetries in place assimilation. Ph.D. Dissertation, Ohio State University.
The Pronunciation Learning Strategy of Japanese Onyomi Kanji by Native Cantonese Speakers: a Look from the Cantonese coda
Patrick Chun Kau Chu Chinese University of Hong Kong
Lee (1992) proposed the onset, rhyme and coda transformation rules that map the pronunciation of Cantonese characters and Japanese onyomi kanji characters. For example, Cantonese characters with a coda /t/ (e.g. 突 /tsɯ/) or /k/ (e.g. 特 /kɯ/) correspond to the Japanese mora /tsɯ/ (e.g. とつ /to.tsɯ/) and /kɯ/ (e.g. とく /to.kɯ/) respectively. Most of the L2 learners realize that the pronunciations of Japanese onyomi kanji are similar to the Cantonese pronunciation but they do not know the rules explicitly. This study investigates whether native Cantonese speakers use the Cantonese-Japanese onyomi kanji pronunciation transformation rules implicitly when they are learning the pronunciations of new Japanese vocabularies.
A questionnaire with multiple choice questions was distributed to 76 participants to test their knowledge on the pronunciation of some Japanese vocabularies and their corresponding Cantonese character. The participants were divided into four groups according to their proficiency in Japanese. They were encouraged to guess and choose the most appropriate pronunciation if they had not learnt the Japanese vocabularies before. Accuracy rates and error distributions for each item were computed and an error analysis was conducted to identify L1 transfer effects. We observed that the learners use the transformational rules implicitly when deducing the pronunciation of unfamiliar Japanese vocabularies. However, we also observed some instances of transfer from Cantonese. Some Cantonese speakers had difficulties in distinguishing Chinese characters with /n/-/ŋ/ pairs and /t/-/k/ pairs in the coda position. These problems were also reflected in their Japanese errors. For example, learners who erroneously classified a Cantonese character with a coda /t/ as /k/ (e.g. 突) would pronounce the Japanese character as とく/to.kɯ/. Pedagogical issues related to the use of the transformation rules as a pronunciation learning strategy will be discussed.
Reference:
Wood Hung Lee. (1992). A Guide to Japanese Pronunciation for Cantonese Speakers. Hong Kong: The Chinese University Press.
A questionnaire with multiple choice questions was distributed to 76 participants to test their knowledge on the pronunciation of some Japanese vocabularies and their corresponding Cantonese character. The participants were divided into four groups according to their proficiency in Japanese. They were encouraged to guess and choose the most appropriate pronunciation if they had not learnt the Japanese vocabularies before. Accuracy rates and error distributions for each item were computed and an error analysis was conducted to identify L1 transfer effects. We observed that the learners use the transformational rules implicitly when deducing the pronunciation of unfamiliar Japanese vocabularies. However, we also observed some instances of transfer from Cantonese. Some Cantonese speakers had difficulties in distinguishing Chinese characters with /n/-/ŋ/ pairs and /t/-/k/ pairs in the coda position. These problems were also reflected in their Japanese errors. For example, learners who erroneously classified a Cantonese character with a coda /t/ as /k/ (e.g. 突) would pronounce the Japanese character as とく/to.kɯ/. Pedagogical issues related to the use of the transformation rules as a pronunciation learning strategy will be discussed.
Reference:
Wood Hung Lee. (1992). A Guide to Japanese Pronunciation for Cantonese Speakers. Hong Kong: The Chinese University Press.
