Upload
pedro-guedes-de-oliveira
View
132
Download
0
Embed Size (px)
Citation preview
PowerPoint Presentation
INFANTS PERCEPTION OF STATEMENT AND QUESTION PROSODY: NATIVE vs. NON-NATIVE LANGUAGE INPUTSnia Frota, Joseph Butler, Shuang Lu & Marina VigrioNew Orleans, 2016
INFANTS PERCEPTION OF STATEMENT AND QUESTION PROSODY: NATIVE vs. NON-NATIVE LANGUAGE INPUT
Results of experiment 2. Symbols represent the proportion of infants with correct anticipatory looks. Red dots represent population averages for monolinguals, and blue diamonds represent averages for bilinguals. Regression lines are shown for both groups. (ABACKGROUND & GOALS- Infants ability to distinguish between forms of phonetic variation in speech that are relevant to meaning is essential for their language development - Learning a language >> a stronger commitment to the native language as development proceeds, modulated by perceptual assimilation and phonetic salience (e.g., Kuhl 2004, Safran et al. 2006, Best & Roberts 2003, Narayan et al. 2010)
My language?Frota et al., 2016
Across languages, pitch can mark prominence, edges, distinguish between sentence types, as in English or Portuguese intonation languages, or signal differences in word meanings, as in a tone language like Mandarin or a pitch accent language like Japanese.2
INFANTS PERCEPTION OF STATEMENT AND QUESTION PROSODY: NATIVE vs. NON-NATIVE LANGUAGE INPUT
Results of experiment 2. Symbols represent the proportion of infants with correct anticipatory looks. Red dots represent population averages for monolinguals, and blue diamonds represent averages for bilinguals. Regression lines are shown for both groups. (ABACKGROUND & GOALS
My language?LexicalLimited variationSegmental variabilityStress only after 6 mos & native (e.g., Skoruppa et al. 2013)Pitch accent as early as 4 mos, for Japanese learners (Sato et al., 2009) Tone as early as 4 mos, but only tone learners > 6 mos, unless very salient (Mattock & Burnham, 2006; Yeung et al, 2013; Liu & Kager, 2014) ? only after 6 months, native (Shi, 2010)
Frota et al., 2016
Pitch studies on infants discrimination abilities have revealed differences across domains of lexical prosody. In the presence of segmental variability (as in real speech), only pitch accent contrasts are perceived very early on, by pitch accent language learners.3
INFANTS PERCEPTION OF STATEMENT AND QUESTION PROSODY: NATIVE vs. NON-NATIVE LANGUAGE INPUT
Results of experiment 2. Symbols represent the proportion of infants with correct anticipatory looks. Red dots represent population averages for monolinguals, and blue diamonds represent averages for bilinguals. Regression lines are shown for both groups. (ABACKGROUND & GOALS
My language?Frota et al., 2016Developmental course of infants perception of pitch contrasts, in the presence of segmental variability >> the ability to extract and generalize the contrastive patterns. Portuguese-learning infants discriminate the statement (falling)/yes-no question (falling-rising) native intonation contrast as early as 5-months (Frota et al. 2014)
Non-native discrimination of the Portuguese contrast has different outcomes in English-learning and Basque-learning infants (Sundara et al., 2015; P.10.23 ICIS)
However, Little is known about the developmental course of infants perception of pitch contrasts, in the presence of segmental variability.Moreover, very little is know about how intonation is perceived. Recent studies have shown,4
INFANTS PERCEPTION OF STATEMENT AND QUESTION PROSODY: NATIVE vs. NON-NATIVE LANGUAGE INPUT
Results of experiment 2. Symbols represent the proportion of infants with correct anticipatory looks. Red dots represent population averages for monolinguals, and blue diamonds represent averages for bilinguals. Regression lines are shown for both groups. (ABACKGROUND & GOALS
My language?Frota et al., 2016Do Portuguese-learning infants also perceive the salient contrast in pitch (falling/low versus falling-rising/high) in segmentally varied non-native input? (Mandarin Chinese Tone1+Tone4 vs. Tone1+Tone2) Similar overall contour shapes predict early discriminationEffects of language experience predict NO discrimination
In this study, we address the question whether5
Frota et al., 2016
EXPERIMENT 15-6 month-olds, n=20: Mean age= 5 mos 29 days; 8 girlsRange= 5 mos 3 days 6 mos 23 days8-9 month-olds, n=20:Mean age= 8 mos 12 days; 10 girlsRange= 7 mos 11 days 9 mos 29 daysParticipants16 segmentally varied, single-prosodic word utterances produced in IDSINFANTS PERCEPTION OF STATEMENT AND QUESTION PROSODY: NATIVE vs. NON-NATIVE LANGUAGE INPUT
Materials
Declarative intonationQuestion intonationNative pitch contrast: falling (statement) versus falling-rising (yes-no question) intonation (Frota et al. 2014)
Exp 1 examined the native6
Frota et al., 2016
EXPERIMENT 1Native pitch contrast: falling (statement) versus falling-rising (yes-no question) intonationINFANTS PERCEPTION OF STATEMENT AND QUESTION PROSODY: NATIVE vs. NON-NATIVE LANGUAGE INPUT
VISUAL HABITUATION PROCEDUREHabituationmalo, lemo, mela, nirra
Testluma, milo, rina, lamo SameSwitch
Different pseudo-words used for the habituation and test phase
Frota et al., 2016
EXPERIMENT 1Native pitch contrast: falling (statement) versus falling-rising (yes-no question) intonationINFANTS PERCEPTION OF STATEMENT AND QUESTION PROSODY: NATIVE vs. NON-NATIVE LANGUAGE INPUT
Results
Listening time (s)
The results for the native.... Showed Successful discrimination both for younger and older infants
8
Frota et al., 2016INFANTS PERCEPTION OF STATEMENT AND QUESTION PROSODY: NATIVE vs. NON-NATIVE LANGUAGE INPUT
Participants5-6 month-olds, n=20: Mean age= 5 mos 25 days; 8 girlsRange= 5 mos 2 days 6 mos 19 days8-9 month-olds, n=20:Mean age= 8 mos 21 days; 10 girlsRange= 7 mos 13 days 10 mos 8 daysMaterials
Tone1+Tone4 declarative intonationTone 1 + Tone 2 question intonationWang et al. 2001
Non-native contrast: Tone 1+Tone 4 and Tone 1+Tone 2 in Mandarin Chinese (also a falling versus falling-rising pitch contrast)
EXPERIMENT 2Tone 1 + Tone 4Tone 1 + Tone 2
Exp.2 examined a non-native contrast: the Lexical distinction betweenThe materials consisted of 16 segmentally varied pseudo-words or actual words in Mandarin Chinese in IDSThe combination.. Has an overall contour shape similar to.
9
Frota et al., 2016INFANTS PERCEPTION OF STATEMENT AND QUESTION PROSODY: NATIVE vs. NON-NATIVE LANGUAGE INPUT
ResultsMaterialsNon-native contrast: Tone 1+Tone 4 and Tone 1+Tone 2 in Mandarin Chinese (also a falling versus falling-rising pitch contrast) EXPERIMENT 2Tones14/12Stat/Questt-test Man/EPF0 patterns 1st syllH / HHL / HL-F0 patterns 2nd syll HL / HLHL / LH- F0 range 2nd syll (Hz)103/3525/192-Duration 1st syll (ms)270/279310/397p = .07 / p < .001 Duration 2nd syll (ms)493/522310/437p < .01 / p < .01
Similar overall contour shapes, but differences in how the contours relate to the textSame procedure
Listening time (s)
Alongside the similar contours, there are differencesResults showed NO discrimination of the non-native contrast10
INFANTS PERCEPTION OF STATEMENT AND QUESTION PROSODY: NATIVE vs. NON-NATIVE LANGUAGE INPUT
Results of experiment 2. Symbols represent the proportion of infants with correct anticipatory looks. Red dots represent population averages for monolinguals, and blue diamonds represent averages for bilinguals. Regression lines are shown for both groups. (ADISCUSSION
Frota et al., 2016Infants discriminated the native intonation contrast at 5-6 and 8-9 mos Infants failed to discriminate the non-native tone contrast at both ages
Results suggest that the similar contour shapes of lexical pitch were not *similar* enough to intonation to be perceived as native (segmental inventory controlled to be native-like).
Tune vs. Tone: utterance domain vs. syllable domain My language!Language-specific perception for pitch, and for the tone/intonation distinction, emerges as early as 5 months of age (earlier than for vowels, consonants or stress)
So,. (importantly, Segmental content was controlled to be native-like, and thus the failure seems to be due to prosodyThe pitch differences found in the stimuli are related to the tune vs. tone distinction, with different distributions of the falling and rising pitch across the bisyllabic utterances.Our findings suggest that - 11
Further research into infants precocious sensitivity to pitch across languages is needed (ongoing work) More:Speech Prosody 2016, BostonMore:LCICD, Lancaster,August 2016
Coloring the worldLisbon Baby LabThank you for watching
EXCL/MHC-LIN/0688/2012 [email protected]