Conceptualising the development of the native listener: What are infants perceptually attuning to in

1
Conceptualising the development of the native
listenerWhat are infants perceptually attuning
toin native speech?

• Catherine Best
• MARCS Auditory Laboratories
• University of Western Sydney
• Haskins Laboratories, USA

2
Conceptualising speech natural, 1st
communication medium for human language

• Speech has multiple layers of structure
• phonological units rules re lexical items
• phonetic details
• global general feel of a language, dialect,
  register
• local physical details of phonological units
• Speech perception the basis of language learning
• production guide to becoming a native speaker
• linguistic meaning lexical and syntactic meaning
  associated with recurring phonetic patterns
• phonological structure discovered in recurrent
  phonetic patterns of familiar lexical items

3
Cross-language speech perceptionwindow on
acquisition of native phonology phonetics

• Language-specific speech properties
• phonological inventories
• phonetic realizations of shared phonemes
• Native-language constraints in adult speech
  behavior
• foreign accent in production of non-native
  consonants and vowels
• foreign accent in perception of non-native
  consonant and vowel contrasts

4
Why are nonnative consonant contrasts
discriminated poorly by adults?

• selected developmental hypotheses
• critical period in sensory-neural tuning
  (Aslin Pisoni)
• auditory experience and development of auditory
  perceptual categories (Kuhl)
• linguistic experience and establishment of
  phonetic representations (McGurk, others)

5
Probing with more exoticnonnative consonant
contrasts (Best, McRoberts Sithole, 1988)

• isiZulu clicks POA
• dental
• lateral
• alveolar

• VOT (lateral click)
• voiceless aspirated
• voiceless unaspirated
• voiced

no linguistic experience little auditory
experience w/ oral clicks, thus no sensory-neural
tuning
90 M discrimination by English speakers All
three prior hypotheses failed
6
Perceptual Assimilation Model(Best, McRoberts
Sithole 1988Best 1995)
7
Uncategorized Speech sounds
Categorized as Native phonemes
Best, 1994, 1995 Best et al., 1988
Non-assimilated Nonspeech sounds
8
ASSIMILATION OF NONNATIVE CONTRASTS
Single-Category Assimilation
Two-Category Assimilation
Best, 1994, 1995 Best et al., 1988
Category Goodness Difference
Non-Assimilable Nonspeech Contrast
9
Test of PAM predictions regarding other nonnative
contrast assimilations Best, McRoberts
Goodell, 2001

• isiZulu Single Category assimilation (SC)
• bilabial stops /b/ - /?/

• isiZulu Category Goodness difference (CG)
• velar stops /kH/ - /k'/

• isiZulu Two Category assimilation (TC)
• lateral fricatives /L/ - //

10
Best, McRoberts Goodell, 2001
11
Underlying developmental questions
When, how, and why do native language biases
emerge in speech perception?
12
Initial, universal speech perception abilities

• Infants under 8 months can discriminate
• native consonant contrasts
• nonnative consonants that adults find difficult

(Werker et al., e.g., 1981, 1983, 1984)

• Hindi dental-retroflex stops

• Nthlakampx velar-uvular ejectives

yet by 10-12 months, infants have difficulty with
such nonnative contrasts
13
But how would older infants respond to nonnative
click consonants,given good discrimination in
adults?
(Best, McRoberts Sithole, 1988)

• isiZulu dental-lateral clicks

This click contrast continues to be discriminated
at least to 14 months
What explains the difference from Werkers
findings?
14
Infant testing procedures Conditioned head turn
vs. conditioned visual fixation habituation
15
fixation recovery in test block (sec)
Best, McRoberts, LaFleur Silver-Isenstadt, 1995
16
Why do many, but not all, nonnative consonants
show a decline at 10-12 months?

• selected hypotheses
• development of auditory perceptual categories
  Native Language Magnet (Kuhl)
• psychoacoustic properties of nonnative contrasts
  Fragile-Robust hypothesis (Burnham)
• perceptual assimilation to native phonological
  contrasts Perceptual Assimilation Model PAM as
  for adults (Best)
• articulatory properties of native phonetic
  distinctions Articulatory Organ hypothesis
  (Goldstein/Best)

17
The three nonnative, non-click contrasts from
isiZulu

• bilabial stops /?/ - /b/
• velar stops /kH/ - /k'/
• lateral fricatives /L/ - //

• all involve 1-2 nonnative phonetic categories
• all contrasts are psychoacoustically robust
• the 3 contrasts are assimilated differently by
  adults
• all use within-organ articulatory distinctions
• laryngeal contrasts

18
mean looking time (sec)
Best McRoberts, 2003
10-12 month olds differ from younger infants
AND adults
19
Why do all three become difficult by 10-12 months?

• because all are unexperienced acoustically?
• because all are unfamiliar phonological
  contrasts?
• because all are within-organ distinctions?
• native fricative voicing /z/ - /s/
• nonnative between-organ /p'/ - /t'/
• nonnative within-organ fricative voicing /L/ -
  //

20
mean looking time (sec)
Best McRoberts, 2003
? Articulatory Organ Hypothesis even for native
contrast!
21
How does native language experience affect
perception of nonnative vowels?
22
Early development of nonnative vowel perception
shifts by 6 months
Perceptual Assimilation Model (PAM) Single
Category (SC) assimilation Two Category (TC)
assimilation Category Goodness (CG) diff. in
assimilation Native Language Magnet
(NLM) prototype vs. non-prototype
asymmetries Natural referent vowel asymmetries
(RVA) peripheral (corner or point) vowels as
universal referents
23
Infant perception of nonnative vowels

• Norwegian 4 high vowels / i y u /
• Expt. 1 rounding-type contrast, high front
• in-rounded vs. out-rounded
• // - /y/

English-speaking adults Assimilation Two
Category (to Eng. /u/ - /i/) Discrimination
excellent ( gt 95)
24
Infant Vowel Experiment 1(Best et al., 1997)

• Experimental contrast /k/ - /ky/
• No-change control /ka/
• Experimental conditions a y
• y a
• Ages 3-5 mo 6-8 mo 10-12 mo
• 10 Ss/condition/age
• Task conditioned visual fixation habituation

25
(No Transcript)
26
Norwegian 4 high vowels / i y u /

• Expt. 2 rounding contrast, high front
• un-rounded vs. out-rounded
• /i/ - /y/

English-speaking adults Assimilation Single
Category (Eng. /i/ - /i/) Discrimination
relatively poor (74)
27
Infant Vowel Experiment 2(Best Faber, 2000)

• Experimental contrast /ki/ - /ky/
• No-change control /ka/
• Experimental conditions i a y
• y a i
• Ages 3-5 mo 6-8 mo 10-12 mo
• 10 Ss/condition/age
• Task conditioned visual fixation habituation

28
(No Transcript)
29
Norwegian same V contrast, different speaker
(gender, dialect) and C-contexts

• Exp. 3 rounding contrast, high front
• un-rounded vs. out-rounded
• /bi/ - /by/ and /bid/ - /byd/

English-speaking adults Assimilation Category
Goodness or Single Category (/i/ - /i/)
Discrimination fair in /bVd/ (79)
very good in /bV/ (89)
30
Infant Experiment 3(Best Faber, 2004)

• Experimental contrast /bi/-/by/ or /bid/-/byd/
• No-change control /bu/ or /bud/
• Experimental conditions i a y or y a i
• in CV or CVC
• Ages 3-5 mo 6-8 mo 10-12 mo
• 8 Ss/syllable/condition/age
• Task conditioned visual fixation habituation

31
NOTE no syllable effects
32
Comparing Experiments 3 2

• 3-5 mo i a y only Exp. 2
• y a i only Exp. 3
• 6-8 mo no discrimination Exp. 2 3
• 10-12 mo symmetrical discrim. Exp. 2
• no discr. (symmetrical) Exp. 3
• Same V contrast /i/ - /y/
• but different speakers (dialect, gender)

33
F2-F3 difference score
34
How does experience with native speech help young
children recognize words?

• articulatory (phonetic) details?
• abstract phonological patterns?

Cross-dialect comparison Connecticut (native)
vs. Jamaican (non-native)

• Familiar vs. unfamiliar word set preferences in
• 14-months (early word learners, lt 25 words)
• 19-months (vocabulary spurt, 50 words)

35
(No Transcript)
36
Conclusions

• Infant attunement to native speech
• phonetic tuning to native consonants by 10-12 mo,
  differs from adults
• phonetic tuning to native vowels by 6-8 mo.
• before morpho-syntax, i.e., rule-governed word
  combinations ( 18 months)
• before first words produced ( 12-14 months)
• native-dialect phonetic patterns 14 mo.
• native-language phonological functions 19 mo.
• proposal early attunement to articulatory
  gestures (phonetic details), later attunement is
  to gesture constellations (phonological structure)

37
Acknowledgements Colleagues Louis Goldstein,
Doug Whalen, Andrea Levitt, Alice Faber, Pierre
Hallé, Ocke-Schwen Bohn, Winifred Strange, Gerald
McRoberts Students Nomathemba Sithole
(Shepherd), Eliza Goodell, Janet Calderón, Jane
Womer, Rosemarie LaFleur, Jean Silver-Isenstadt,
Tiffany Gooding, and many others Funding Nationa
l Institutes of Health (USA), NATO Institutional
affiliations Haskins Laboratories, Wesleyan
University
38
Predominant assimilation patterns for nonnative
Xhosa stop contrasts
Calderón, 1996 Calderón Best, 1996
39
(No Transcript)
40
(No Transcript)
41
Significance

• Theoretical implications
• theories of speech perception
• interface of phonetics and phonology
• brain language system for speech perception
• Potential practical significance
• language acquisition problems (and reading)
• optimization of L2 learning
• multilingual communication issues

42
Research Directions

• Further infant studies
• Articulatory Organ hypothesis in infancy
• infant cross-dialect perception
• Studies of early multilinguals
• L2 learners
• Early bi/multi-linguals
• Sign language phonetic perception

43
Best, 1994, 1995 Best et al., 1988
44
Needs a recap on the 3 remaining questions on
inants perceptual learning of native speech, and
45
Possibly raise Calderons and other recent
findings on very early and persisting maternal
lang influences in bilingual adults and infants,
which are nonetheless malleable to attentional
manips (lang of instruction set), and need for
further research to tease apart causes
46
Another approach to studying development
Persistent influence of L1 in early sequential
bilinguals
Calderón, 1996 Calderón Best, 1996

• Spanish /b/-/p/ b-p
• English /b/-/p/ p/b-ph

47
Needs a summary of other recent findings on
bilingual adults and infants