Language acquisition 11/18/11 Michael Frank Department of - - PowerPoint PPT Presentation

Language acquisition

11/18/11

Michael Frank Department of Psychology

Tuned towards communication

Hierarchical organization of language

/dəӚ dɔg snæpt/ phonological the dog snap+[past] morphological/lexical det N V NP VP S the dog snapped syntactic

Language development through YouTube

Babbling Single-word speech Telegraphic speech Multi-word speech with

• ccasional morphology errors

“annie big shoes” “I no like mustard” “ga” “cheese!” 1 year 1.5 years 2 years 3 years

Language development through YouTube

Babbling Single-word speech Telegraphic speech Multi-word speech with

• ccasional morphology errors

“annie big shoes” “I no like mustard” “ga” “cheese!” 1 year 1.5 years 2 years 3 years

Recognizing mom’s voice

DeCasper & Fifer (1980) Individual neonates’ preference for mother’s voice

• 2.5
• 2
• 1.5
• 1
• 0.5

0.5 1 1.5 2 2.5 Sucking preference More for mom Less for mom

Recognizing specific phrases

DeCasper & Spence (1986) Mothers read “The Cat in the Hat” to their babies over the last 6 weeks of pregnancy, then the resulting newborns were tested on their preference for the target story

142 DKASPER AND SPENCE

.60 IBI BIN Ct

• sec1

Figure

• 2. Mean

reinforcement ratios

• f the target

(hatched bars) and novel (open bars) stories for Experienced infants in the 161 <f condition (left side) and in the 161 >t condition (right side). The means are based

• n a total
• f 400 baseline

and 1040 reinforced interburst intervals.

during the baseline and reinforcement phases revealed no effect of Contingency, F(1) 10) < 1 .O, a significant effect of Bin, F(9,90) = 5.19, p< .OOl, and a signifi- cant Contingency x Bin interaction, F(9,90) = 3.48, p c .005. However, none

• f the follow-up

tests of simple effects were statistically reliable; the inter- action seemed to result from unsystematic variation in the difference scores of the two contingency conditions in Bins 1-5. Subsequent analysis of conditional probabilities confirmed that the pre- ceding interaction did not result from systematic effects of target-story rein- forcement. The baseline conditional probabilities

• f target and novel stories

did not differ, t( 11) < 1 .O; neither did their reinforcement ratios computed for the intervals O.Ot-0.4t and 1 .Ot-1.41. The mixed ANOVA with Contingency and Interval as factors revealed no reliable effects whatever, p values of all F statistics > .10 (Figure 3). A comparison

• f the reinforcement

ratios of matched-subject pairs re- vealed that experienced newborns had larger target-story ratios than their matched naive counterparts, t(l1) =2.68, p< .05, but that their novel-story ratios did not differ, t(1 1)~ 1 .O. DISCUSSION Three implications

• f the prenatal-experience

hypothesis were confirmed: (1) For experienced subjects the target story was more reinforcing than the novel

target novel experimental control High-amplitude sucking preference

Categorical perception

Reviewed in Kuhl (2004) Many phonemes differ only on some continuum.

• /b/ & /p/: voice onset time
• Question: /b/ & /d/ & /g/ differ on place of articulation.
• Categorical?

% Adult Judgments

Voice-onset time (VOT) Phonetic boundary at +30 msec VOT

BA PA

Infant categorical perception

Eimas et al. (1971) Kuhl et al. (1975)

Universal vs. specific

• /b/ vs. /p/ is close to universal
• But many sound contrasts are not

– Any Hindi speakers here?

beat lentil branch shield

Conditioned head-turn procedure

a a a a a a a a a a a a a a a a a a a a a a a b b b a a a a

Test trial “Control” trial

a = one Hindi sound b = another Hindi sound

Changes in speech perception

Werker & Tees (1984)

Language development through YouTube

Babbling Single-word speech Telegraphic speech Multi-word speech with

• ccasional morphology errors

“annie big shoes” “I no like mustard” “ga” “cheese!” 1 year 1.5 years 2 years 3 years

Baby’s first words

Table 4 Rank-Ordered Top 20 Words for Children Who Can Say 1–10 Words on CDI and Percentage of Children Producing Them, by Language

United States (n 264) Hong Kong (n 367) Beijing (n 336) Daddy (54) Daddy (54) Mommy (87) Mommy (50) Aah (60) Daddy (85) BaaBaa (33) Mommy (57) Grandma—Paternal (40) Bye (25) YumYum (36) Grandpa—Paternal (17) Hi (24) Sister—Older (21) Hello?/Wei? (14) UhOh (20) UhOh (Aiyou) (20) Hit (12) Grr (16) Hit (18) Uncle—Paternal (11) Bottle (13) Hello?/Wei? (13) Grab/Grasp (9) YumYum (13) Milk (13) Auntie—Maternal (8) Dog (12) Naughty (8) Bye (8) No (12) Brother—Older (7) UhOh (Aiyou) (7) WoofWoof (11) Grandma—Maternal (6) Ya/Wow (7) Vroom (11) Grandma—Paternal (6) Sister—Older (7) Kitty (10) Bye (5) WoofWoof (7) Ball (10) Bread (5) Brother—Older (6) Baby (7) Auntie—Maternal (4) Hug/Hold (6) Duck (6) Ball (4) Light (4) Cat (5) Grandpa—Paternal (4) Grandma—Maternal (3) Ouch (5) Car (3) Egg (3) Banana (3) WoofWoof (2) Vroom (3)

Tardiff et al. (2008) First words are high frequency, grounded in social context and routine

The meanings of the first words

Underextensions “doggie” only means Fido Overextensions “ball” means anything round (ball, balloon, moon, apple, egg…)

Words produced

Yet kids learn tons of words!

Fenson et al. (1994) Words understood

The importance of early experience

Hart & Risley (1995) Total words heard Total words known

Associative word learning

Smith & Yu (2008)

• Raise your left hand when you know

what a hiftam is

• And your right when you know what a

gensim is

Intentions, not cues

1 2 Baldwin (1993)

• Coincide: child

looks at 1 after hearing 1’s name

• Conflict: child

looks at 2 after hearing 1’s name

25 50 75 100 "Which one is the modi?" "Which one is your favorite?" Percent Correct

Coincide Conflict

Children represent speakers’ intentions!

Language development through YouTube

Babbling Single-word speech Telegraphic speech Multi-word speech with

• ccasional morphology errors

“annie big shoes” “I no like mustard” “ga” “cheese!” 1 year 1.5 years 2 years 3 years

English morphology

• English past tense

– Mostly regular: walk -> walked – Occasionally (~100 forms) irregular: go

• > went, run -> ran, sing -> sang
• English plural

– Almost entirely regular: book -> books – Very few irregulars (~10 forms): mice, geese, teeth, feet, cacti, children, men, etc.

Morphological generalization

Berko (1958)

Lun Tor Tass Gutch +/z/ +/Əәz/ Today I spling. Yesterday I ___ Today I gude. Yesterday I ___ Today I scride. Yesterday I ___

Overregularization

Marcus et al. (1992)

• 1. Some irregulars learned by rote

– Most irregulars tend to be very high frequency

• 2. Then over-generalization of predominant pattern
• 3. Finally, correct performance on regulars

Development of a “blocking rule”? Or consequence of gradually increasing evidence?

Theoretical positions

Principles and Parameters

• Principles of syntax are

innately given

• A small set of

parameters vary cross- linguistically

• Children determine

these parameter settings from the input

• Developmental errors

are caused by the course of maturation

Item-based acquisition

• Syntax is learned

piecemeal through individual words

• Cross-linguistic

similarities caused by cognitive similarities

• Children learn and

change all parts of languages

• Developmental errors

are caused by mistakes in generalization

Transitive vs. intransitive

• Transitive takes an agent/subject and a

patient/object

– Pat ate snails. – Kim drew spirals. – Alex likes running.

• Intransitive just takes a subject

– Pat ate. – Kim drew. – * Alex likes.

• Many verbs can move between the two

but not all.

Gradual generalization of syntax

Tomasello & Brooks, 1999

25 50 75 100 2 2.5 3 3.5 4 4.5 5 Percent correct transitive generalizations Age (years)

Here’s fudding! Look at fudding! Kiwi is fudding! What is kiwi doing to the dough? Kiwi is fudding it! (transitive)

Early abstract knowledge

Gertner, Fisher, & Eisengart, 2006 Duck is gorping bunny! Success at 2 years, even younger for “the girl is gorping the boy.”

Learning to sign without native input

Young deaf children who were unable to acquire oral language naturally and had not been exposed to a conventional manual language were found to use spontaneously a gesture system that has some of the structural characteristics of early child language.

Goldin-Meadow & Mylander, 1983

Production probability T P I T P I Mom Child T = transitive actor, I = intransitive actor, P = patient You eat (I) You eat [chocolate] (T) [You] eat chocolate (P) You eat chocolate (T P)

Conclusions

• Early perceptual abilities

– Learning rhythms, discriminating sounds

• Building a vocabulary through social

word learning

– Importance of input – And of mechanisms of learning

• Then gradual generalization of

syntactic structure

– Early comprehension, later production