37
Psychology of Language Development Etsuko Haryu Department of Educational Psychology 2008・6・19 Works indicated with this mark are owned by a third party. The reuse of this work or creation based on a derivative work requires the direct permission for use from the author.

Psychology of Language Development...Psychology of Language Development Etsuko Haryu Department of Educational Psychology 2008・6・19 1. Segmenting words from fluent speech • Head-turn

  • Upload
    others

  • View
    27

  • Download
    0

Embed Size (px)

Citation preview

Page 1: Psychology of Language Development...Psychology of Language Development Etsuko Haryu Department of Educational Psychology 2008・6・19 1. Segmenting words from fluent speech • Head-turn

Psychology of Language Development

Etsuko

HaryuDepartment of Educational

Psychology

2008・6・19

‡:Works indicated with this mark are owned by a third party. The reuse of this work or creation based on a derivative work requires the direct permission for use from the author.

Page 2: Psychology of Language Development...Psychology of Language Development Etsuko Haryu Department of Educational Psychology 2008・6・19 1. Segmenting words from fluent speech • Head-turn

When did you start talking?

Page 3: Psychology of Language Development...Psychology of Language Development Etsuko Haryu Department of Educational Psychology 2008・6・19 1. Segmenting words from fluent speech • Head-turn

How did you learn to talk?

Page 4: Psychology of Language Development...Psychology of Language Development Etsuko Haryu Department of Educational Psychology 2008・6・19 1. Segmenting words from fluent speech • Head-turn

Developmental Transition of Utterances by Children(Maeda・Maeda, from 1996)

1;0 ‘nen ne’ ネンネ(sleepy) 1;6 ‘kete’ ケテ(from akete=open)

1;7 ‘totta’ トッタ(got) ‘kitty’ キティ―

(kitty doll) ‘chan’チャン

(dear)1;8 ‘mochi’ モチ

(rice cake) ‘tabeta’ タベタ(ate) /‘kuchushita’ クチュッシタ

(socks) ‘haku’ ハク

(put on)1;9 ‘nana’ ナナ(doggy name) ‘mizu’ ミズ(water) ‘nonda’ ノンダ (drank)2;1 ‘~chan’ ~チャン(dear) (own name) ‘mo’ モ (too) ‘nomu’ (drink) ノム

2;3 ‘gyunyu’ ギューニュー(milk) ‘nonderundayo’ ノンデルンダヨ(drinking)‘○chan’ ○チャン

(dear) ‘wa’ ワ

(topic-marker) 2;9 ‘kaze’ カゼ(cold) ‘naotutara’ ナオッタラ(recover)

‘calupisu’ カルピス(brand name of drink) ‘nomareruno’ ノマレルノ

(allowed to drink?)2;10 ‘ah’ アー (oh) ‘oniichan’ オニイチャン(elder brother) ‘ga’ ガ(subject-marker) ‘ocha’ オチャ

(tea) ‘o’ ヲ(object-marker) ‘nonda’ ノンダ(drank) ‘na’ ナ

(didn’t he)

More w

ords

Longer sentences

Word segmentation

Page 5: Psychology of Language Development...Psychology of Language Development Etsuko Haryu Department of Educational Psychology 2008・6・19 1. Segmenting words from fluent speech • Head-turn

18m

12m

6m 1. Word Segmentation

First Words

Vocabulary SpurtTwo-WordUtterances

Multiple-Word Utterances

24m

2. Word Learning

3. Grammar Acquisition

Page 6: Psychology of Language Development...Psychology of Language Development Etsuko Haryu Department of Educational Psychology 2008・6・19 1. Segmenting words from fluent speech • Head-turn

1. Segmenting words from fluent speech

• Head-turn Preference Procedure (Jusczyk & Aslin, 1995)

The cup was bright and shiny. A clown drank from the red cup. . His cup was filled with milk.

The cup was bright and shiny. A clown drank from the red cup. . His cup was filled with milk.

cup cup cup cup… dog dog dog…

Page 7: Psychology of Language Development...Psychology of Language Development Etsuko Haryu Department of Educational Psychology 2008・6・19 1. Segmenting words from fluent speech • Head-turn

1. Segmenting words from fluent speech

• Clues to word segmentation– Transitional Probability (Saffran et al.,1996)

– Stress(strongly pronounced syllable)

Frequent Words

「ミルクほしの?」

‘miluku hoshii no?’ (Want milk?)

「ミルクのむ?」

‘miluku nomu?’ (Drink milk?)

「ほらミルクよ」

‘hora miluku yo’ (Here is milk)

○ English…doctor,kingdom○ Japanese…ブ―ブ、buh-bu (car) くっく、

kuQku (shoe)

あんよ、aN-yo (leg)

○ Own name (Bortfeld et al.,2005)

(Jusczyk et al.,1999)

(Hayashi・Mazuka,2007)

○ Functors … grammatical particles in Japanese?

Page 8: Psychology of Language Development...Psychology of Language Development Etsuko Haryu Department of Educational Psychology 2008・6・19 1. Segmenting words from fluent speech • Head-turn

But…• Indeed, children must hear functors, or grammatical

particles, in speech input very frequently

• But, children’s early production are likely to omit those grammatical elements…

「ワンワン

イタ」

wan wan (doggy) ita (there)

「パパ

カイシャ」

pa pa (daddy) kaisha (office)

「クチュシタ

ハク」

kutushita (socks) haku (put on)

Are children really able to utilize grammatical particles to segment a word ? When do they recognize grammatical particles in the speech input?

Page 9: Psychology of Language Development...Psychology of Language Development Etsuko Haryu Department of Educational Psychology 2008・6・19 1. Segmenting words from fluent speech • Head-turn

Japanese infants’ recognition of the particle ‘ga’ (Kajikawa & Haryu, 2008)

0

5

10

15

1 2 3 4 5 6 7 8 9 1 2 3 4

↓’rume ga muwa-tteiru yo’

Test

?Test1

‘rume ki muwa-tteiru yo’ (‘ga’ replaced with a non-particle syllable)

Test2

‘rume muwa-tteiru yo’ (‘ga’ dropped)Test3

‘rume ga muwa-tteiru yo’ (standard)

Test4

‘wapu ga rite-tteiru yo’ (control)*Order of Test1-3 was counterbalanced across infants

Habituation Phase

*Modified Habituation Procedure

Page 10: Psychology of Language Development...Psychology of Language Development Etsuko Haryu Department of Educational Psychology 2008・6・19 1. Segmenting words from fluent speech • Head-turn

0

1

2

3

4

5

6

7

8

9

10

6mos 10mos 15mos

標準

脱落

置換

0

1

2

3

4

5

6

7

8

9

10

標準 脱落 置換

* *Listening Time

standard

dropreplacement

standard drop replacement

When habituated to a sentencethat does not include a grammatical particle

Page 11: Psychology of Language Development...Psychology of Language Development Etsuko Haryu Department of Educational Psychology 2008・6・19 1. Segmenting words from fluent speech • Head-turn

• Clues to word segmentation-Transition Probability (Saffran et al.,1996)-Stress(strongly pronounced syllable)

English…doctor, kingdom(Jusczyk et al.,1999)

Japanese…buh-bu,(car) kuQku (shoe) aN-yo(leg)

(Hayashi・Mazuka,2007)

-Frequent WordsOwn name(Bortfeld et al.,2005)

Functors ? (Kajikawa & Haryu,2008)At least, 15-mos recognize those elements in the speech input…

To be continued ….

1. Segmenting words from fluent speech

Page 12: Psychology of Language Development...Psychology of Language Development Etsuko Haryu Department of Educational Psychology 2008・6・19 1. Segmenting words from fluent speech • Head-turn

18m

12m

6m 1. Word Segmentation

First Word

VocabularySpurt

Two Word Utterances

Multiple Word Utterances

24m

2. Word Learning

3. Grammar Acquisition

Stressed SyllableTransitional Probability

Frequent Words

Page 13: Psychology of Language Development...Psychology of Language Development Etsuko Haryu Department of Educational Psychology 2008・6・19 1. Segmenting words from fluent speech • Head-turn

産出語彙数

050

100150200250300350400450

10m 12m 14m 16m 18m 20m 22m 24m 26m 28m

2. Word learning• Number of productive words

Ogura(2000)

Why do children progress so slowly in their early stages?

Page 14: Psychology of Language Development...Psychology of Language Development Etsuko Haryu Department of Educational Psychology 2008・6・19 1. Segmenting words from fluent speech • Head-turn

rabbit

rabbit(category)

The bunny’s name

long ears!

carrot

looks fluffy

When a child hears a novel word…

What does the word ‘rabbit’ mean?

Page 15: Psychology of Language Development...Psychology of Language Development Etsuko Haryu Department of Educational Psychology 2008・6・19 1. Segmenting words from fluent speech • Head-turn

9 months

10months

11months

12months

13months ~

(white dog stuffed toy) (white dog in storybook)

(a real white dog) (white woolen puff) (black rope tuft)

(Dogs in general) (cats) (white yarn・blanket) (white wall)白い壁

(lion) (tiger) (white bear)(shoe with white yarn)

/nan nan/ (dog)/na-n/ (cat)/mo-/ (cow)/zo-/ (elephant)/kun chan/ (bear) ニャンニャンクック

/nyan nyan kuQku / (white woolen shoe)ニャンニャンチョッキ/nyan nyan choQki/( white woolen vest)

Okamoto(1982)

A child’s use of the word /nyan nyan/ :

Page 16: Psychology of Language Development...Psychology of Language Development Etsuko Haryu Department of Educational Psychology 2008・6・19 1. Segmenting words from fluent speech • Head-turn

rabbit

rabbit(category)

bunny’s name

long ears!

carrot

Word-learning ・mutual exclusivity bias・whole-object bias・taxonomic bias

(shape bias)

⇒vocabulary spurt

looks fluffy

Page 17: Psychology of Language Development...Psychology of Language Development Etsuko Haryu Department of Educational Psychology 2008・6・19 1. Segmenting words from fluent speech • Head-turn

産出語彙数

050

100150200250300350400450

10m 12m 14m 16m 18m 20m 22m 24m 26m 28m

• Number of productive words

Ogura(2000)

Word-learning biases

vocabulary spurt

Page 18: Psychology of Language Development...Psychology of Language Development Etsuko Haryu Department of Educational Psychology 2008・6・19 1. Segmenting words from fluent speech • Head-turn

18m

12m

6m 1. Word Segmentation

First Word

VocabularySpurt

Two Word Utterances

Multiple Word Utterances

24m

2. Word Learning

3. Grammar Acquisition

Stressed SyllableTransitional Probability

Frequent Words

Word Learning Biases

Efficient Word Learning

Page 19: Psychology of Language Development...Psychology of Language Development Etsuko Haryu Department of Educational Psychology 2008・6・19 1. Segmenting words from fluent speech • Head-turn

Is that enough?…All the words children must learn are

NOT object labels.“Ah, NEKE* ga iru!”(Ah, here is NEKE* )「あ、ネケがいる!」

“Ah- NEKE* -tte iru!”(Ah- it is NEKE* -ing!)「あー

ネケっている!」

Word’s form class

Kinds of meaning characteristic of the form class* ‘NEKE’ is a nonsense word in Japanese, which was used to examine whether children were able toinfer its meaning, an object label or a verb denoting the action, by attending to the sentence frame in which the word appeared.

Page 20: Psychology of Language Development...Psychology of Language Development Etsuko Haryu Department of Educational Psychology 2008・6・19 1. Segmenting words from fluent speech • Head-turn

(Imai, Haryu, & Okada,2005)

“Hora、NEKE* ga aruyo”(Look, here is (a) NEKE)「ほら、ネケがあるよ」

“NEKE* ga aru no ha dotchi?”(In which is NEKE?) 「ネケがあるのはどっち?」

Mutsumi Imai, Etsuko Haryu

‘Construction of the Lexicon: How Children Learn Words and Concepts’(Iwanami 2007)

Drawing 5.3 Imai et al.(2005)

Noun Condition

Standard scene

Action-Same-Object-Change sceneObject-Same-Action-Change scene

Page 21: Psychology of Language Development...Psychology of Language Development Etsuko Haryu Department of Educational Psychology 2008・6・19 1. Segmenting words from fluent speech • Head-turn

“NEKE*- tteiru no ha dochi?”(In which is she NEKE-ing?) 「ネケっているのはどっち?」

(Imai, Haryu, & Okada, 2005)“Hora、NEKE* -tteiru”(Look, it is NEKE-ing)「ほら、ネケっているよ」

Standard scene

Object-Same-Action-Change scene Action-Same-Object-Change scene

Verb ConditionMutsumi Imai, Etsuko Haryu

‘Construction of the Lexicon: How Children Learn Words and Concepts’(Iwanami 2007)

Drawing 5.3 Imai et al.(2005)

Page 22: Psychology of Language Development...Psychology of Language Development Etsuko Haryu Department of Educational Psychology 2008・6・19 1. Segmenting words from fluent speech • Head-turn

010

2030

4050

6070

8090

100

3-year-olds 5-year-olds

noun

verb

(%)M

ean

prop

ortio

n of

the

Act

ion-

Sam

e ch

oice

(%) (Imai, Haryu, & Okada,2005)

⇒Distinction of word types Mapping to an appropriate concept

Page 23: Psychology of Language Development...Psychology of Language Development Etsuko Haryu Department of Educational Psychology 2008・6・19 1. Segmenting words from fluent speech • Head-turn

18m

12m

6m 1. Word Segmentation

First Word

VocabularySpurt

Two Word Utterances

Multiple Word Utterances

24m

2. Word Learning

3. Grammar Acquisition

Stressed SyllableTransitional Probability

Frequent Words

Word learning biases

Efficient Word Learning

Distinction of Word Types

Understanding of Type of Concepts Characteristic of

the Word Types

Elaboration of Word Learning Processes

Page 24: Psychology of Language Development...Psychology of Language Development Etsuko Haryu Department of Educational Psychology 2008・6・19 1. Segmenting words from fluent speech • Head-turn

Many verbs can be applied to this scene…

Ta-tteiru (standing)立っている

Mo-tteiru (gripping)持っている

Utsumui-teiru (looking down)うつむいている

Kakae-teiru (carrying)かかえている

Osae-teiru (holding)おさえている

Fumishime-teiru (standing firmly)ふみしめている

Dama-tteiru (staying silent)黙っている

SOV

verbs SV

verbs

Page 25: Psychology of Language Development...Psychology of Language Development Etsuko Haryu Department of Educational Psychology 2008・6・19 1. Segmenting words from fluent speech • Head-turn

From when are children able to infer the meaning of a novel verb by utilizing the syntactic structures in which the verb appears ?

Page 26: Psychology of Language Development...Psychology of Language Development Etsuko Haryu Department of Educational Psychology 2008・6・19 1. Segmenting words from fluent speech • Head-turn

video 1 video 2

Transitive Condition “Look, the duck is gorping* the bunny!”

Intransitive Condition

“Look, the duck and the bunny are gorping!”

Do children utilize syntactic structuresto figure out the meaning of a novel verb?

(Naigles,1990)

duck

bunny

‡ Drawings were borrowed from Naigles et al, “Children Use Syntax to Learn Verb Meanings.”, Journal of Child Language (Cambridge University Press), vol.17 No.2 p357-374 (1990)、Fig.2http://journals.cambridge.org/

*‘gorp’is a nonsense word that was used to examine children’s use of sentence frame inInterpreting a novel verb

Page 27: Psychology of Language Development...Psychology of Language Development Etsuko Haryu Department of Educational Psychology 2008・6・19 1. Segmenting words from fluent speech • Head-turn

+

Page 28: Psychology of Language Development...Psychology of Language Development Etsuko Haryu Department of Educational Psychology 2008・6・19 1. Segmenting words from fluent speech • Head-turn

video 1 video 2

Where is gorping now?Find gorping !

‡ Drawings were borrowed from Naigles et al, “Children Use Syntax to Learn Verb Meanings.”, Journal of Child Language (Cambridge University Press), vol.17 No.2 p357-374 (1990)、Fig.2http://journals.cambridge.org/

Causative event Non-Causative event

Page 29: Psychology of Language Development...Psychology of Language Development Etsuko Haryu Department of Educational Psychology 2008・6・19 1. Segmenting words from fluent speech • Head-turn

0

0.5

1

1.5

2

2.5

3

3.5

他動詞構文 自動詞構文

因果事象

非因果事象non-causativecausative

Transitive Condition Intransitive Condition

Look

ing

time

(s)

Page 30: Psychology of Language Development...Psychology of Language Development Etsuko Haryu Department of Educational Psychology 2008・6・19 1. Segmenting words from fluent speech • Head-turn

18m

12m

6m 1. Word Segmentation

First Word

VocabularySpurt

Two Word Utterances

Multiple Word Utterances

24m

2. Word Learning

3. Grammar Acquisition

Stressed SyllableTransitional Probability

Frequent Words

Word learning biases

Efficient Word Learning

Distinction of Word Types

Understanding of Type of Concepts Characteristic of

the Word Types

Elaboration of Word Learning Processes

Page 31: Psychology of Language Development...Psychology of Language Development Etsuko Haryu Department of Educational Psychology 2008・6・19 1. Segmenting words from fluent speech • Head-turn

But…Japanese language allows ellipsis of arguments.

“A, ninjin tabe-teiru ne”(Ah, ___ eating carrot, is it not)「あ、にんじん

食べてるね」

“Ahー

usagi san tabe-teiru yo”

(Ah, bunny is eating ___.)「あー

うさぎさん

食べてるよ」

“Ahー

tabe-teiru !”

(Ah, ____ ____ eating!)「あー

食べてる」

⇒ Do Japanese children also use syntactic structures to infer the meaning of a novel verb?

Page 32: Psychology of Language Development...Psychology of Language Development Etsuko Haryu Department of Educational Psychology 2008・6・19 1. Segmenting words from fluent speech • Head-turn

Do Japanese Children Infer the Meaning of a Novel Verb on the Basis of Sentence Structures ? (Imai・Haryu, 2007)

Transitive Condition:「ウサギ(さん)が

クマ(さん)を

ヘクっているのは、どっち?」

“usagi(san) ga kuma(san) o HEKU*- teiru no ha, dotchi?”(In which one is the bunny HEKU*-ing the bear?)

Intransitive Condition:「ウサギ(さん)と

クマ(さん)が

ヘクっているのは、どっち?」

“usagi(san) to kuma(san) ga HEKU*-tteiru no ha, dotchi?”(In which one are the bunny and the bear HEKU*-ing?)

Mutsumi Imai, Etsuko Haryu‘Construction of the Lexicon: How Children Learn Words and Concepts’(Iwanami 2007)

Drawing 5.7 Can Japanese children map a novel verb on the basis of syntactic structures of the sentence in which the verb appears ? : Stimulus materials (movies) used in the experiments done with Japanese children. (Imai, Haryu et al., in preparation)

Causative event Non-Causative event

* ‘HEKU’ is a nonsense word in Japanese, which was used to examine whether childrenwere able to infer its meaning, an objectlabel or a verb denoting the action,by attending to the sentence frame in which the word appeared.

Page 33: Psychology of Language Development...Psychology of Language Development Etsuko Haryu Department of Educational Psychology 2008・6・19 1. Segmenting words from fluent speech • Head-turn

0

10

20

30

40

50

60

70

80

90

100

2yrs 3yrs 5yrs

自動詞

他動詞

**

Mean Prop

ortion

 of C

orrect Respo

nses (

%)

Correct Responses:Transitive Condition: Mapping the novel verb to a causative

event

Intransitive Condition: Mapping the novel verb to a non‐causative event

intransitive

transitive 

Page 34: Psychology of Language Development...Psychology of Language Development Etsuko Haryu Department of Educational Psychology 2008・6・19 1. Segmenting words from fluent speech • Head-turn

How do Japanese children determine semantic roles of noun phrases in transitive sentences?

○ word order …Subjects are likely to come first. ○ Case-marking particles mark the semantic role

of the preceding noun-phrase.

Japanese two-year-olds already know that causative verbstypically appear in transitive frames whereas non-causative verbs are likely to appear in intransitive frames, and utilize this knowledge to learn verbs.

うさぎ を

くま

押す

The rabbit object-marker the bear subject-marker push

くま

が うさぎ

押す

The bear subject-marker the rabbit object-marker push

***

⇒The bear pushes the bunny

Page 35: Psychology of Language Development...Psychology of Language Development Etsuko Haryu Department of Educational Psychology 2008・6・19 1. Segmenting words from fluent speech • Head-turn

‘subject+object’ word order condition「クマさんが

ウサギさんを

ネケっている

のはどっち?」

‘kuma san GA usagi san O NEKE*-tteiru no wa dochi?’(bear) (rabbit) (which one)(In which is the bear NEKE*-ing the rabbit?)

‘object +subject’ word order condition「ウサギさんを

クマさんが

ネケっている

のはどっち?」

‘usagi san O kuma san GA NEKE*-tteiru no ha dochi?’(rabbit) (bear) (which one)(In which is the bear NEKE*-ing the rabbit?)

* ‘NEKE’ is a nonsense word in Japanese, which was used to examine whether childrenwere able to infer its meaning, an objectlabel or a verb denoting the action,by attending to the sentence frame in which the word appeared.

Page 36: Psychology of Language Development...Psychology of Language Development Etsuko Haryu Department of Educational Psychology 2008・6・19 1. Segmenting words from fluent speech • Head-turn

0

10

20

30

40

50

60

70

80

90

100

2yrs(30m) 3yrs(43m) 5yrs(68m)

動作主->動作対象

動作対象->動作主

subject +object+

**

Why do five-year-olds fail to correctly respond to the ‘object+subject’ sentencewhereas three-year-olds succeed?

(Haryu et al., 2007)

Pragmatic Knowledge ?

object + subject

Page 37: Psychology of Language Development...Psychology of Language Development Etsuko Haryu Department of Educational Psychology 2008・6・19 1. Segmenting words from fluent speech • Head-turn

18m

12m

6m 1. Word Segmentation

First Word

VocabularySpurt

Two Word Utterances

Multiple Word Utterances

24m

2. Word Learning

3. Grammar Acquisition

Stressed SyllableTransitional Probability

Frequent Words

Word learning biases

Efficient Word Learning

Distinction of Word Types

Understanding of Type of Concepts Characteristic of

the Word Types

Statistical Analyses of Input

Information Easily Accessible in Sensory and Cognitive Level

●phoneme●objects●segments of event

Language Acquisition

Pragmatic Knowledge

Knowledge of syntax- semantic

correspondences