Joseph Butler, Marina Vigário and Sónia Frota Universidade de

Joseph Butler, Marina Vigário and Sónia Frota Universidade de Lisboa

Laboratório de Fonética & Lisbon BabyLab (FLUL/CLUL)

Infants’ perception of intonation and its use in early word learning

Workshop of the project Development of Prosodic Structure and Intonation

June 27th, 2013Faculdade de Letras da Universidade de Lisboa

Overview

Early intonation perception– Introduction– Aims and research questions– Methodolody– Results– Discussion

Early word learning– Introduction– Methodology– Results– Discussion

Overall summary2

Introduction

Little is known about the developmental course of infants´perception of linguistic intonation

Previous studies on pitch contrasts focused onacquisition of lexical pitch– Lexical pitch accent, as in Japanese

Japanese-learning infants discriminate native pitch accent contrast from as early as 4 months, and maintained during the 1st year (Sato et al., 2009)

– Lexical tone, as in Mandarin, Cantonese Learners of tone languages show stable discrimination of

lexical tone in the 1st year, as early as 4 months (Mattock & Burnham, 2006; Mattock et al., 2008; Yeung et al., 2013) 3

Introduction

Previous studies on lexical pitch focused on pitchdirection/height only (e.g. Tones 25 vs. 33, Cantonese –Yeung et al, 2013). However, tones are also distinguished by temporal properties, like the temporal location of the pitch turning point (Jongman et al, 2006)

Recent research on infant perception suggests pitch andduration differences are processed differently– Infants at 7 months prefer pairs of syllables with variable pitch

when pitch height in the first syllable, but no preference when duration differs between syllables (Bion et al., 2011).

– Human infants and nonhumans group sound sequences into trochaic patterns based on pitch, but do not utilize duration as a cue to group sequences in iambic structures (de la Mora et al., 2012).

4

Introduction

Recent research on infant perception suggests pitch andduration differences are processed differently– Unlike for lexical pitch, word stress discrimination abilities

involve sensitivity gains during the first year of life, dependent on exposure to the native language (Skoruppaet al. 2009, 2013). Stress contrasts typically use to a greater extent phonetic cues other than pitch (namely, duration – Beckman, 1986)

By and large– Pitch discrimination dependent on perceptual abilities– Duration dependent on language experience

5

Introduction

Intonation is a prosodic dimension that varies across languages and impacts upon meaning.

Unlike lexical pitch, which is meaningful at the wordlevel, intonation conveys phrasal meanings, likesentence type and pragmatic distinctions (Ladd, 2008; Frota, 2002, in press)

Intonation languages (e.g., English, Portuguese) may use pitch height, pitch direction and pitchtiming (temporal location of pitch turning points) to convey phrasal meanings (statement, question, calling contour, information status, etc).

6

Aims and Research Questions

We investigated European Portuguese-learning infants´perception of two native pitch contrasts:– The statement/yes-no question distinction, marked

by a pitch direction contrast (rising/falling)– The broad/narrow focus distinction, signalled by a

pitch timing contrast (early/late fall within the syllable)

We asked (i) whether the developmental perceptual trajectory of the two contrast types was similar and (ii) how it related to previous reports on pitch perception by infants learning lexical pitch systems

7

The two pitch contrasts in EP

8

Illustrated with one-word utterances/intonational phrases (single prosodic word intonational phrases are the most frequent type of IP in Portuguese IDS -http://ww3.fl.ul.pt/laboratoriofonetica/babylab/CDS_EP_Prosody/CDS_EP_prosodic_words_prosodic_phrases.pdf )

Statement Yes-no question

Broad focus Narrow focus

What happened? Did they split?Context:(they got) married

(Did it) rain ?

Both contrasts have been shown to be perceived by adult native speakers (Falé & Faria, 2005; Frota, 2012)

(It has) rained

Methodology

Experiment 1 (Statement/Yes-No Question distinction)– 40 infants

20 younger (8 female, M = 5 months 29 days, range 5 months 3 days – 6 months 23 days)

20 older (10 females, M = 8 months 12 days, range 7 months 11 days-9 months 29 days)

9

Experiment 2 (Broad/Narrow Focus distinction)– 28 infants

12 younger (6 female, M = 6 months 30 days, range 5 months 30 days – 8 months 3 days)

16 older (8 females, M = 12 months 8 days, range 10 months 16 days-14 months 6 days)

Stimuli

Stimuli: Contrast Statement/Yes-no question (Exp1), Contrast Broad/Narrow Focus (Exp2) – Child directed speech

Declarative

Interrogative

Phonetic variation relevant to meaning > conveys important distinctions for processing the input language, for communication and social interaction

Direction Alignment of the pitch fall

10

Acoustic analysis

11

Exp 1: Statement/Yes-No Question

Exp 2: Broad/Narrow Focus

Statements Questions t-test

F0 range 1st syll (Hz) 67 66 0.12, p = .9

F0 range 2nd syll (Hz) -25 192 23.46, p<.001

Final F0 (Hz) 163 380 23.61, p<.001

Duration (ms) 529 765 11.91, p<.001

Focus Neutral t-test

F0 peak (Hz) 249.79 230.8 7.4, p < .001

F0 low (Hz) 160.26 161.53 1.05, p = .31

Duration (ms) 0.59 0.63 3.29, p < .01

Timing of the fall (ms) 0.27 0.15 11.15, p < .001

Procedure

Modified version of the visual habituation paradigm (Stager & Werker, 1997)

Habituated with lists of utterances– Half the infants habituated with declarative (Exp1)/broad

focus (Exp2)– Other half habituated with question (Exp1)/narrow focus

(Exp2) Habituation continued until pre-defined criteria reached

– Avg looking time to last 4 habituation trials <60% the avg looking time to the first 4 habituation trials

Test phase presented lists of different utterances in both declarative/question intonation (Exp1)/ broad/narrow focus intonation (Exp2)

12

Procedure

Each trial lasted 16 seconds Presentation of test stimuli counterbalanced

(same/switch trial first) Look software was used (Meints & Woodford, 2008) Looking times recorded and compared If sensitive to the intonational contrast, longer

looking times to the switch trials should be evident

13

Visual displayAttention getter

video

14

Results

Exp 1: Statement/Yes-No Question distinction

15

- Significant difference between same and switch test trials (F(1,36) = 54.18, p < .001, η2 = .6)

- No effect of age group (F(1,36) = 2.13, p = .15, η2 = .06)

- No interaction between trial type and age group (F(1,36) = 3.29, p = .08, η2 = .08).

- Paired T-tests: significant difference between same and switch trials for younger (t(19) = 6.1, p < .001, d = 1.474) and older (t(19) = 4.42, p < .001, d = 0.816) groups.

(Frota et al, submitted)

Results

Exp 2: Broad/Narrow Focus distinction

16

Preliminary analysis:- Significant interaction between trial type and age group (F(1,24) = 5.81, p < .05, η2 = .2)

- No effect of trial type (F(1,24) < 1).

- No effect of age group (F(1,24) < 1).

- Paired T-tests: younger group no significant difference between same and switch trials (t(11) = 1.35, p = .21), older significant difference (t(15) = 2.29, p < .05).

Summary

Infants learning European Portuguese demonstrate a discrimination ability for the statement/question prosodic contrast as early as 5 months, and maintain this ability throughout the first year

However, for the broad/narrow focus contrast, infants only demonstrate discrimination by 12 months

17

Discussion

Our findings suggest that the perceptual trajectory of intonation categories depends on the primary cues involved (pitchdirection, pitch timing).

Supports earlier results that show a protracted development of the perception of timing relative to pitch height/direction (Bion et al., 2011)

18

Discussion

Our findings suggest that perception of intonation categories based on a pitch direction contrast may be as precocious as lexical tone/lexical pitch accent perception.

Pitch-dominant contrasts are those that show a precocious development of discrimination abilitiesversus stress and other duration/timing-relatedcontrasts that seem more dependent on languageexperience (Skoruppa et al., 2013; Frota et al., submitted)

19

Interpretation of phonetic variationin word learning

Acquisition of phonology requires learning to interpret phonetic variation

Across languages, prosodic properties may vary in their acoustic correlates and the phonological domains they signal (segments distinguish words): Stress signals word contrasts in Spanish, English or EP but

not in French (phrasal level); Pitch signals word contrasts in Chinese or Japanese but not in English or EP (phrasal level)

Pitch is an important cue to stress in Spanish (Hualde 2005), but not in European Portuguese

The task for the young learner: which variation is meaningless?; which variation conveys meaning, at what prosodic level (word level/phrasal level)? 20


European Portuguese (EP) is an intonation languagewith lexical stress

Stress is a word level property Pitch is a property of phrase level phonology We tested whether EP learners were sensitive to

stress and pitch contrasts in a word learning task, using an eyegaze-based procedure (and comparedthis with a vowel contrast). Will young learners notice stress differences and/or intonation

differences in ‘new words’ ? When do young learners interpret phonetic variation at the

appropriate levels according to the native language?21


Eyegaze-based procedure (similar to Quam & Swingley 2010) where visual fixation to the labelled picture is theresponse variable

3500msx5

3000msx10

4000msx12

Animation Ostensive-labelling

Test


Auditory stimuli: Phonetic propertiesStress contrast: penult / final [milu] / [milu]

– Stress location cued by relative duration (stressed > 58-132 ms) and by the alignment of the pitch fall (through the stressed syllable)

Pitch contrast: declarative / interrogative (H+L* L% / H+L* LH%)– Intonation contrast cued by the low versus rising boundary and by the

longer duration of the final syllable in interrogatives (117-165ms)

23

Vowel contrast: change in height andbackness [milu] / [malu]


146 children between 1;0 and 4;9 were tested, allfrom monolingual EP homes. 72 children included inthe analysis (learned the trained word: >50% fixation to thelabelled picture; matching numbers across conditions)

24

Interpretation of phonetic variation in word learning

25

Younger: 1-year olds and 2-year olds

Older: 3-year olds and 4-year olds

*** ***

*** ***

** ** **

******

*

** * **

*

*

*** **

**

Prosodic changes: stress (SC) intonation (IC)

Vowelchange

Interpretation of phonetic variation in word learning

Our results show that

– Pitch contour variation is regarded as relevant in new words by 1-year olds and 2-year olds, at odds with native language phonology.

– 3-year olds already regard pitch variation as NOT relevant: Stress variation is regarded as meaningful, both by the younger and older age groups (similar to segmental variation)

– The effect of segmental variation is stronger than stress variation in the older group (mutual exclusivity)

– Only at 3;0 do young learners interpret phonetic variation atthe appropriate levels according to the native language

– Pitch is a very salient property in more demanding tasks earlier in development

26

Overall summary

The development of the perception of different intonational categories varies dependent on the primary cues involved.

Pitch contour variation is initially regarded as relevant in early word learning, and is only interpreted in line with native phonology later in development.

27

28

ObrigadoThank you

DEPE: PTDC/CLE-LIN/108722/2008

Thanks to Cátia Severino and Susana Correia

And especially to all the families and nurseries that have taken part in these studies.

References Beckman, M.E. (1986). Stress and non-stress accent. Dordrecht: Foris. Bion, A.H., Benavides-Varela, S. & Nespor, M. (2011). Acoustic markers of prominence influence infants‘ and adults‘

segmentation of speech sequences. Language and Speech, 54(1), 123-140. de la Mora, D.M., Nespor, M. & Toro, J.M. (2012). Do human and nonhuman animals share the grouping principles of

the iambic-trochaic law? Attention, Perception & Psychophysics, 75(1), 92-100- Falé, I., Hub Faria, I. (2005). Intonational contrasts in EP: a categorical perception approach. INTERSPEECH, 1705-1708. Frota, S. (in press): The intonational phonology of European Portuguese. In Sun-Ah (ed.) Prosodic Typology II. Oxford:

Oxford University Press, Chapter 2. Frota, S. (2012). A focus intonational morpheme in European Portuguese: Production and Perception. To appear in G.

Elordieta & P. Prieto (eds). Prosody and Meaning. Berlin/New York: Mouton de Gruyter, pp. 163-196. Frota, S., Butler, J. & Vigário, M. (submitted). Infants‘ perception of intonation: Is it a statement or a question? Hualde, J. I. (2005). The Sounds of Spanish. Cambridge: CUP Jongman, A., Wang, Y., Moore, C., Sereno, J., 2006. Perception and production of Chinese tones. In: Li, P., Tan, L.H.,

Bates, E., Tzeng, O.J.L. (Eds.), Handbook of East Asian psycholinguistics (Vol. 1: Chinese). Cambridge University Press

Ladd, D.R. (2008). Intonational phonology. Cambridge: Cambridge University Press. 2nd Edition. Mattock, K., Molnar, M., Polka, L., & Burnham, D. (2008). The developmental course of lexical tone perception in the

first year of life. Cognition, 106(3), 1367–1381. Mattock, K., & Burnham, D. (2006). Chinese and English infants’ tone perception: Evidence for perceptual

reorganization. Infancy, 10(3), 241-265. Meints, K. & Woodford, A. (2008). Lincoln Infant Lab Package 1.0: A new programme package for IPL, Preferential

Listening, Habituation and Eyetracking. [WWW document: Computer software & manual]. URL: http://www.lincoln.ac.uk/psychology/babylab.htm

Quam, K. & Swingley, D. (2010). Phonological knowledge guides two-year-olds' and adults' interpretation of salient pitch contours in word learning. Journal of Memory and Language, 62, 135-150.

Sato, Y., Sogabe, Y. & Mazuka, R. (2009). Development of hemispheirc specialization for lexical pitch-accent in Japanese infants. Journal of Cognitive Neuroscience, 22(11), 2503-2513.

Skoruppa, K., Pons, F., Bosch, L., Christophe, A., Cabrol, D. & Peperkamp, S. (2013). The development of word stress processing in French and Spanish infants. Language Learning and Development, 9(1), 88-104.

Skoruppa, K., Pons, F., Christophe, A., Bosch, L., Dupoux, E., Sebastián-Gálles, N., Limissuri, R.A: & Peperkamp, S. (2009). Language-specific stress perception by 9-month-old French and Spanish infants. Developmental Science, 12(6), 914-919.

Stager, C.L. & Werker, J.F. (1997). Infants listen for more phonetic detail in speech perception than in word learning tasks. Nature, 388(6640), 381-382.

Yeung, H.H:, Chen, K.H. & Werker, J.F. (2013). When does native language input reorganize phonetic perception? The precocious case of lexical tone. Journal of Memory and Language, 68(2), 123-139.

29

Documents

Joseph Butler, Marina Vigário and Sónia Frota Universidade de