The raising effect of aspirated consonants on The raising effect of - PowerPoint PPT Presentation

The raising effect of aspirated consonants on The raising effect of aspirated consonants on F0 in Taiwanese F0 in Taiwanese Yuwen Lai Department of Linguistics, UBC Christopher Huff Joan Sereno Allard Jongman Dept Linguistics, University of Kansas 1 1

Consonant effect on F0 Consonant effect on F0 Prevocalic voicing triggers tonogenesis in Chinese and some Asian languages � Voiced onsets lower onset F0 � Well-documented in tonal and non-tonal languages Raised/lowered F0 can serve as a perceptual cue for consonantal feature � Higher F0 as a perceptual cue for aspiration in Cantonese (Francis et al. 2006) But, most interestingly… 2 2

Aspiration and onset F0 Aspiration and onset F0 C h < C C h > C Korean (Han 1967; Kim 1968) Hindi (Kagaya and Hirose 1975) Danish (Jeel 1975) Thai (Gandour 1974; Erickson 1975) Thai (Ewan 1976) Cantonese (Zee 1980) Mandarin (Xu and Xu 2003) Cantonese (Francis et al. 2006) Korean (Kim et al. 2002) Taiwanese (Lai 2004) Korean (Pearce 2007) C h = C Danish (Fischer-Jørgensen 1968) French unaspirated vs English aspirated stops (Hombert and Ladefoged, 1977) 3 3

General assumption General assumption Voiced and Voiceless Voiceless aspirated and Voiced and Voiceless Voiceless aspirated and unaspirated unaspirated after aspirated onset after aspirated onset after voiced onset after voiced onset � higher larynx higher larynx � lower larynx lower larynx � � (Ewan Ewan 1976) 1976) ( � less transglottal less transglottal � � faster airflow rate faster airflow rate � pressure pressure ( Ladefoged ( Ladefoged, 1967) , 1967) � increased laryngeal increased laryngeal � tension tension 4 4

Complexity Complexity Tonal or non- -tonal languages tonal languages Tonal or non Effect stronger on non- -tonal languages tonal languages Effect stronger on non � � (Hombert Hombert 1977) 1977) ( High or low tones High or low tones � Raising effect weaker on high tones Raising effect weaker on high tones � � Lowering effect weaker on low tones Lowering effect weaker on low tones � 5 5

6 6 Taiwanese tones Taiwanese tones

Methodology Methodology Stimuli Stimuli � CV(N) word pairs contrasting in prevocalic aspiration CV(N) word pairs contrasting in prevocalic aspiration � � 3 places of articulation (bilabial, alveolar, velar) 3 places of articulation (bilabial, alveolar, velar) � � 5 unchecked tones (1, 2, 3, 5, 7) 5 unchecked tones (1, 2, 3, 5, 7) � � 2 words per tone 2 words per tone � ± aspiration x 3 places x 5 tones x 2 words = 60 tokens ± aspiration x 3 places x 5 tones x 2 words = 60 tokens 7 7

Methodology Methodology Participants Participants 5 female and 5 male native speakers 5 female and 5 male native speakers Procedure Procedure � Self Self- -pased pased reading list reading list � � 3 repetitions x 60 unique words = 180 tokens per 3 repetitions x 60 unique words = 180 tokens per � speaker speaker 8 8

Methodology Methodology Acoustic Measurements Acoustic Measurements � F0 (onset and every 10% time steps) F0 (onset and every 10% time steps) � � VOT, Rime duration VOT, Rime duration � 9 9

Statistical Analyses Statistical Analyses Four- -way ANOVA way ANOVA Four � Aspiration (aspirated, unaspirated) Aspiration (aspirated, unaspirated) � � Gender (female, male) Gender (female, male) � � Tone (1, 2, 3, 5, 7) Tone (1, 2, 3, 5, 7) � � Place (bilabial, alveolar, velar) Place (bilabial, alveolar, velar) � 10 10

Main effect - - Onset F0 Onset F0 Main effect (*) Aspiration (aspirated > unaspirated) (*) Gender (female > male) (*) Tone (2 > 1 > 3, 7 > 5) (*) Place (velar > bilabial > alveolar) 11 11

F0 measurement across the contour F0 measurement across the contour 12 12

13 13 Results by Gender Results by Gender

Aspiration x Gender on onset F0 ( * ) Aspiration x Gender on onset F0 ( * ) Onset F0 after aspirated is higher only in females. 14 14

15 15 Tone 2 Tone 2

16 16 Tone 3 Tone 3

17 17 Tone 5 Tone 5

18 18 Tone 1 Tone 1

19 19 Tone 7 Tone 7

Summary Summary Prevocalic aspiration – – raiser for F0 in Taiwanese raiser for F0 in Taiwanese Prevocalic aspiration Effect tailors off between 30%- -40% of the tonal contour 40% of the tonal contour Effect tailors off between 30% Gender difference was found Gender difference was found Cross- -over effect found in tone 1 (high level) and tone 7 over effect found in tone 1 (high level) and tone 7 Cross (low level) (low level) 20 20

Aspiration - - a raiser or a depressor? a raiser or a depressor? Aspiration Language matters (Downing and Language matters (Downing and Gick Gick 2001) 2001) Individual differences may exist Individual differences may exist Thai (Erickson 1975) Thai (Erickson 1975) � raising in 8 of 11 subjects, lowering in 3/11 raising in 8 of 11 subjects, lowering in 3/11 � Measurement matters Measurement matters � Raising at F0 onset Raising at F0 onset � � Lowering the overall F0 of the syllable Lowering the overall F0 of the syllable � Gender matters Gender matters 21 21

22 22 But why? But why?

Aspiration, F0, and VOT Aspiration, F0, and VOT Madurese (Cohn and Lockwood 1994) Madurese (Cohn and Lockwood 1994) Voiceless aspirated onset lower F0 Voiceless aspirated onset lower F0 Voiceless aspirated stops longer in Voiceless aspirated stops longer in duration than voiced and voiceless duration than voiced and voiceless unaspirated stops stops unaspirated 23 23

VOT VOT male female 90 80 70 VOT (ms) 60 50 40 30 20 aspirated unaspirated (Lai, 2004) 24 24

Airflow rate, aspiration and VOT Airflow rate, aspiration and VOT Air flow rate 25 25

Correlation Correlation VOT difference = VOT (aspirated) – VOT (unaspirated) Onset F0 difference = OnsetF0 (aspirated) – OnsetF0 (unaspirated) diff_vot 40 Linear (diff_vot) differences in onset F0 (Hz) 30 20 10 0 0 20 40 60 80 100 120 140 -10 -20 -30 -40 differences in VOT (ms) • A significant correlation between VOT and onset differences • As the difference in VOT between aspirated and unaspirated onsets increases, the difference in OF decreases. 26 26

Discussion Discussion Determinants for VOT Determinants for VOT � Place of articulation Place of articulation � � Tone Tone � � Speaking rate Speaking rate � Gender effect Gender effect � Characteristics of vocal folds make females Characteristics of vocal folds make females � more susceptible to transglottal air pressure more susceptible to transglottal air pressure differences differences � Sociolinguistic effect Sociolinguistic effect � 27 27

Discussion Discussion Earlier research lack phonetic analyses and Earlier research lack phonetic analyses and statistical significance report statistical significance report Method of measurement Method of measurement � Tang (2008) implosives in Bade Tang (2008) implosives in Bade � raise F0 at vowel onset raise F0 at vowel onset lower the overall F0 for the syllable lower the overall F0 for the syllable Gender balance, sample size, speaking rate Gender balance, sample size, speaking rate 28 28

Aspiration and onset F0 Aspiration and onset F0 C h < C C h > C Korean (Han 1967; Kim 1968) Hindi (Kagaya and Hirose 1975) Danish (Jeel 1975) Thai (Gandour 1974; Erickson 1975) Thai (Ewan 1976) Cantonese (Zee 1980) Mandarin (Xu and Xu 2003) Cantonese (Francis et al. 2006) Korean (Kim et al. 2002) Taiwanese (Lai 2004) Korean (Pearce 2007) Taiwanese (Lai et al. 2008) C h = C Danish (Fischer-Jørgensen 1968) French unaspirated vs English aspirated stops (Hombert and Ladefoged, 1977) 29 29

Acknowledgements Acknowledgements Experiment participants Experiment participants Praat scripts by scripts by Mietta Mietta Lennes Lennes and Yi and Yi Xu Xu Praat Members of University of Kansas Members of University of Kansas Phonetics and Psycholinguistics Phonetics and Psycholinguistics Laboratory Laboratory 30 30

31 31 VOT VOT

Previous attempt Previous attempt Participants Participants � 4 subjects: 2 males, 2 females 4 subjects: 2 males, 2 females � Stimuli Stimuli � All 7 tones, 3 places of articulation and affricates All 7 tones, 3 places of articulation and affricates � Results Results � Raising effect after aspirated stops found in female Raising effect after aspirated stops found in female � speakers, but not in male speakers speakers, but not in male speakers � No significant interaction between place, tone and No significant interaction between place, tone and � aspiration aspiration (Lai, 2004) 32 32