1

1

Phonological specificity of vowels and consonants

in 20-month-olds’word representations

Anders Højen

Center for Child LanguageDepartment of Language and Communication

University of Southern Denmark

2

Background

• Infants discriminate phonetic contrasts already before 4 months of age (e.g., Eimas, Siqueland, Jusczyk, & Vigorito, 1971).

• Discriminate both native and nonnative contrasts from the beginning

• Sensitivity to nonnative contrasts decline around 12 months (Werker & Tees, 1984)

• Sensitivity to native contrasts increases around 12 months

3

Discrimination of Hindi and Salish by English-learning infants6-8 mo. Discriminate contrasts in both languages8-10 mo. Half of the infants could discriminate10-12 mo. Few could discriminate (Werker, 1989)

4

Discrimination of the English /r/-/l/ distinction:

Sensitivity increases for American infants but decreases for Japanese infants between 6-12 mo.

(Kuhl et al., 2006)

5

Background

• High capability of perceiving native phonetic distinctions at around 1 year

• At this time word learning begins• Word learning requires encoding of sound

for later recognition• Expect infants to encode the sound

pattern of words accurately because of their perceptual abilities at 1 year of age

6

Background

• BUT an early report indicated that at around 1 year, infants confuse minimal pairs (Shvachkin, 1973)

• They could discriminate• …but could not associate the right word

with a referent• Access to phonetic detail hampered by

word-object association?

(c) Center for Børnesprog

2

7

Background

• The effect of word-object association on perception shown again later in habituation-switch task (Stager & Werker, 1997)

• 14-month-olds were taught novel “words”:– very different word pair: Lif – Neem– minimal pair: Bih – Dih

• In test phase, infants noticed word-object switch for Lif-Neem pair, but not Bih-Dih

• But infants could discriminate Bih-Dih8

Background

• Reduced phonological sensitivity at 14 mo. when associating words with objects?– I.e., words phonologically underspecified?

• Only novel words• Sensitivity to mispronunciations of well-

known words at 14 mo. shown with switch task (Fennel & Werker, 2003) and inter-modal preferential looking task or IPL (Swingley & Aslin, 2002)

9

Background

• Sensitivity to phonological detail of well-known words at 18-23 months (Swingley & Aslin, 2000)

• “Recovery” of sensitivity to phonological detail of novel words at 17-20 months (Werker et al. 2002)

Well-known Novel 14 mo. √ ÷ 20 mo. √ √

10

Background

Hypothesis (Fennel & Werker, 2003)

• Word-learning (novel) too computationally demanding at 14 mo. but not 20 mo. to encode words accurately

• Word recognition (well-known) less demanding – can be done already at 14 mo.

11

Background

• The just-reviewed studies examined sensitivity to consonants

• But vowels might have either stronger or weaker representation than consonants– Stronger: Vowel perception language specific sooner

than consonants perception (Kuhl, 1991; Kuhl et al. 1992; Polka & Werker, 1994; but cf. Polka & Bohn, 1996)

– Weaker: Vowels constrain lexical access less than consonants in adults (Cutler et al., 2000)

• More or less detail for vowels than consonants in early word representions?

12

Background

• Conflicting results• 20-month-olds showed sensitivity

mispronunciations in the consonant but not the vowel in novel words (Nazzi, 2005)

• 15 to 24-month-olds sensitive to mispronunciations in both vowels and consonants in well-known words (Mani & Plunkett, 2007)

(c) Center for Børnesprog

3

13

This studyTested sensitivity to vowel and consonant

mispronunciations in well-known words using IPL

Purpose• To pit sensitivity to vowel mispronunciations

against consonant mispronunciations to determine if vowels constrain lexical access less than do consonants

• To examine sensitivity to consonant mispronunciations in the first vs. last consonant in CVC words

14

Methods

Participants• Thirty-two 19 to 21-month-olds – 16 each

in experiment 1 and 2• All were healthy full-term infants born to

native speakers of Danish

15

Methods – Exp. 1

Auditory stimuli• Four correctly pronounced (CP) and four

mispronounced (MP) Danish CVC• MP words differed from the correct words in

either the initial consonant or the vowel.• Each MP stimulus could be “reconstructed”

to two of the correct words by changing the vowel or initial consonant

16

Methods – Exp. 1Table 1. Auditory stimuli, which were either correctly pronounced or mispronounced

Danish IPA Enlish

Correct sut kat bil ful

sut kt bil ful

“pacifier” “cat” “car” “bird”

Mispronunced sat kut bul fil

st kut bul fil

17

Methods – Exp. 1

Example of MP trial

Kut

Change consonant

”Sut”

Change vowel

”Kat”

(Consonant target) (Vowel target)

(Monitor)

18

Methods – Exp. 1

Procedure• Infant sits on parents lap in dim booth• On each trial, two pics shown for 5000 ms• At 2250 ms, word stimulus (CP or MP) is

presented in carrier sentence: Se, en ___• Offline examination of picture fixation after

word stimulus, frame by frame from video recordings (show video)

• Observation window: 1500 ms

(c) Center for Børnesprog

4

19

Methods – Exp. 1

Procedure• In CP trials, the picture named by the

auditory stimulus was the target • In MP trials, both pictures were possible

targets. Auditory stimulus e.g., kut– Consonant target: Kat-picture (correct

consonant, wrong vowel)– Vowel target: Sut-picture (correct vowel wrong

consonant)

20

Methods – Exp. 1

Procedure• Preference for looking at the consonant

target in MP trials = consonants constrain lexical access more than vowels (weaker specification of vowels)

• Preference for looking at the vowel target = vowels constrain lexical access more than consonants

• 16 test trials (8 CP, 8 MP)

21

Results – Exp. 1

TargetminusDistractor

0

100

200

300

400

500

Δ lo

okin

g tim

e, m

s

Cons. targetminusvowel target

Difference score (looking time at target minus distractor in MP trials) shows target preference in CP trials

Small positive difference score in MP trials (looking time a consonant target minus vowel target), not signific. different from 0

Non-significant preference for consonant target 22

Discussion

• Preference for target in CP trials = experiment works

• Non-significant preference for consonant target = no evidence that consonants and vowels differ in constraining lexical access

• Does not suggest a difference in specificity for vowels and consonants in well-known word representations at 20 months

23

Methods Exp. 2

Examined sensitivity to mispronunciation in vowel vs. final consonant in CVC words

Stimuli• Four CP words and four MP words,

analogous to experiment 1 stimuli• Each MP stimulus could be reconstructed

to two CP stimuli, e.g., Kot– Consonant target: Kat-picture– Vowel target: Kop-picture

24

Results – Exp. 2Difference score (looking time at target minus distractor in MP trials) shows target preference in CP trials

Small negative difference score in MP trials (looking time a consonant target minus vowel target), not signific. different from 0

Non-significant preference for vowel target

TargetminusDistractor

-300-200-100

0100200300400500600

Δ lo

okin

g tim

e, m

s

Cons. targetminusvowel target

(c) Center for Børnesprog

5

25

Discussion

• Non-significant preference for vowel target = no evidence that consonants and vowels differ in constraining lexical access

• Does not suggest a difference in specificity for vowels and consonants in well-known words at 20 months– Supports Mani & Plunkett (2007)

26

Discussion

Interpretation problem• Lack of preference for either consonant or

vowel target upon hearing MP word, e.g., kot, could mean two things:

1. Rejection of kot as either “kat” or “kop”– vowels and consonants both well-specified

2. Accept of kot as both “kat” and “kop”– vowels and consonants both phonologically

underspecified

27

Discussion

Interpretation problem• Under specification of both vowels and

consonants in well-known words at 20 mo. is unlikely– evidence of substantial detail for vowels and

consonants in English infants aged 14-15 mo.– if both vowel and final consonant are

underspecified, it should be hard to tell katfrom kop in CP trials

• Convinced?28

Discussion

• Possible effect of position of mispronunciations

TargetminusDistractor

0

100

200

300

400

500

Δ lo

okin

g tim

e, m

s

Cons. targetminusvowel target

TargetminusDistractor

-300-200-100

0100200300400500600

Δ lo

okin

g tim

e, m

s

Cons. targetminusvowel target

Non-significant preference for cons. target

Non-significant preference for vowel target

Initial consonant vs. vowel Final consonant vs. vowel

29

Discussion• Suggests that infants process words from

linearly, “from one end”– If they hear kut, they look more for a picture whose

name begins in /k/ (kat) than a name whose rhyme is right (sut)

– If they hear kot, they look more for kop than kat• Hypothesis: lexical access constrained more by

“early” sounds in words than late sounds– But they eventually discard MP forms resulting in

lower looking times than in CP trials

30

Discussion

• Could this effect solve interpretation problem?

• If the effect of position of mispronunciation is significant, how could both vowels and consonants be underspecified?

(c) Center for Børnesprog

6

31

Conclusion

• Vowels and consonants do not differ in constraining lexical access to well-known words at 20 months

• Suggests that vowels are as phonologically well-specified as consonants are in word representations at 20 months

• Future: difference between vowels and consonant specificity at 14 months?

32

Thank you!

• Presentation will soon be available from– www.andershojen.dk– www.sdu.dk/cfb

• These people contributed to the research:– Thomas O. Madsen, Werner Vach, Torkil Østerbye,

Karina Faber, Hans Basbøll, Sueli Caporali, DortheBleses

• Research sponsored by– University of Southern Denmark– Widex A/S

33

ReferencesCutler, A., Sebastián-Gallés, N., Soler-Vilageliu, O., & van Ooijen, B. (2000).

Constraints of vowels and consonants on lexical selection: Cross-linguistic comparisons. Memory and Cognition, 28(5), 746-755.

Eimas, P. D., Siqueland, E. R., Jusczyk, P. W., & Vigorito, J. (1971). Speech perception in infants. Science, 171, 303-306.

Fennel, C. T., & Werker, J. F. (2003). Early word learners' ability to access phonetic detail in well-known words. Language and Speech, 46(2/3), 245-264.

Kuhl, P. K. (1991). Human adults and human infants show a 'perceptual magnet effect' for the prototypes of speech categories, monkeys do not. Perception and Psychophysics, 50(2), 93-107.

Kuhl, P. K., Stevens, E., Hayashi, A., Deguchi, T., Kiritani, S., & Iverson, P. (2006). Infants show a facilitation effect for native language phonetic perception between 6 and 12 months. Developmental Science, 9(2), F13-F21.

Kuhl, P. K., Williams, K. A., Lacerda, F., Stevens, K. N., & Lindblom, B. (1992). Linguistic experience alters phonetic perception in infants by 6 months of age. Science, 255(5044), 606-608.

Mani, N., & Plunkett, K. (2007). Phonological specificity of vowels and consonants in early lexical representations. Journal of Memory and Language, 57, 252-272.

Nazzi, T. (2005). Use of phonetic specificity during the acquisition of new words: differences between consonants and vowels. Cognition, 98, 13-30.

Polka, L., & Bohn, O.-S. (1996). A cross-language comparison of vowel perception in English-learning and German-learning infants. Journal of the Acoustical Society of America, 100(1), 577-592.

34

ReferencesPolka, L., & Werker, J. F. (1994). Developmental-Changes in Perception of Nonnative

Vowel Contrasts. Journal of Experimental Psychology-Human Perception and Performance, 20(2), 421-435.

Shvachkin, N. K. (1973). The development of phonemic perception in early childhood. In C. A. Ferguson & D. I. Slobin (Eds.), Studies of child language development (pp. 92-127). New York: Holt, Rinehart & Winston.

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.

Swingley, D., & Aslin, R. N. (2000). Spoken word recognition and lexical representation in very young children. Cognition, 76(2), 147-166.

Swingley, D., & Aslin, R. N. (2002). Lexical neighborhoods and the word-form representations of 14-month-olds. Psychological Science, 13(5), 480-484.

Werker, J. F. (1989). Becoming a native listener. American Scientist, 77(1), 55-59.Werker, J. F., Fennel, C. T., Corcoran, K. M., & Stager, C. L. (2002). Infants’ ability to

learn phonetically similar words: Effects of age and vocabulary size. Infancy, 3(1), 1-30.

Werker, J. F., & Tees, R. C. (1984). Cross-language speech perception: Evidence for perceptual reorganization during the first year of life. Infant Behavior and Development, 7(1), 49-63.

(c) Center for Børnesprog