A Usage Based Approach To
-
Upload
nirmala-last -
Category
Technology
-
view
2.173 -
download
2
Transcript of A Usage Based Approach To
A Usage-Based Approach toEarly Syntactic Development
Michael Tomasello
Max Planck Institute for Evolutionary AnthropologyLeipzig, Germany
Elena Lieven Kirsten Abbot-SmithHeike Behrens Thea Cameron-FaulknerHolger Diessel Ceri SavageAngelika Wittek Kai KiekhöferSabine Stoll Robert MaslenNenagh Kemp Ben Ambridge
Stefanie Brosda
Wanna ride horsie
•Infinitival Complement
(deleted subject)
•Wanna X
•Wanna ride X
•Holophrase
•Etc.
__falldown
__kick __
__give__ __
__ running
--- breaks
Verb Island Constructions at About 2 Years of Age
Throw__
“YOU CANT GET THERE FROM HERE“
There = Formal Grammar
Here = Item-Based Constructions
Usage-Based Models
•Cognitive Grammar•Construction Grammar•Grammaticalization Theory
Learning
•Intention-Reading•Pattern-Finding
Statistical/Distributional An.AnalogyConstraints
Two Questions
• The nature of children’s linguistic representations
• The processes by which these representations change over developmental time
THE NATURE OF EARLY LINGUISTIC
REPRESENTATIONS
MARATSOS (1987)[5 year olds +]
Adult Model Always Intransitive:It fudded. The dough won‘t fud. The dough is fudding in the machine.
Transitive Biasing Question:What are you doing (to the dough)? [encouraging: I‘m fudding it]
PINKER ET AL. (1987)[4-5 year olds]
Adult Model Always Passive:It’s being floosed by the horsie. It‘s being floosed.
Active Biasing Question:
What‘s the horsie doing (to it)? [encouraging: He‘s floosing it.]
TOMASELLO & BROOKSCognitive Linguistics (1998)
Adult Model Always Intransitive:
It fudded. The dough won‘t fud. The dough‘s fudding in the machine.
Transitive Biasing Question:
What are you doing (to the dough)?[encouraging: I‘m fudding it]
Results
4 out of 32 two-year-old children (12.5%) produced a transitive SVO utterance
BROOKS & TOMASELLODevelopmental Psychology (1999)
Adult Model Always Passive:
It’s being floosed by the horsie. It‘s being floosed.
Active Biasing Question:
What‘s the horsie doing (to it)?[encouraging: He‘s floosing it]
Results
12 out of 48 three-year-old children (25%) produced a transitive SVO utterance
Figure 1. Percentage of children that produce transitive utterances using novel verbs in different studies. (TOMASELLO, Cognition, 2000)
0
10
20
30
40
50
60
70
80
90
2,0 2,6 3,0 3,6 4,0 4,6 5,0 8,0
CHILDREN ARE PRODUCTIVE WITH NOVEL NOUNS
Tomasello & Olguin (1993) Cognitive DevelopmentTomasello et al. (1997) Journal of Child Language
Model: “Look, a wug! A wug!“
Result: Almost all 22-month-old children say:
“More wug”, “Kiss wug”, “Wug gone”, etc.
CHILDREN USE NOVEL VERBS TRANSITIVELY WHEN THEY HAVE HEARD THEM USED THAT WAY
Tomasello & Brooks (1998) Cognitive LinguisticsBrooks and Tomasello (1999) Developmental Psychology
Model: “The boy’s meeking the stick!“
Result: Almost all 2-year-old children say:
“The cow’s meeking the fence”
Akhtar (1999) & Abbot-Smith et al. (2001)Weird Word Order
English-speaking children hear utterances with “weird word order” (familiar and unfamiliar verbs)
– “The cow the horse is meeking/pushing” (SOV)
– They are encouraged to use these same verbs with new characters engaging in these same actions
0
10
20
30
40
50
60
70
80
90
100
2;4 (Abbot-Smith et al)
2;8 (Akhtar) 3;6 (Abbot-Smith et al)
3;9 (Abbot-Smith et al)
4;4 (Akhtar)
Ungrammatical FamiliarVerb Ungrammatical Novel
Grammatical Novel
AKHTAR (1999) & ABBOT-SMITH et al. (2001)
Percentage of mean number of utterances which were mismatches, as a function of condition and age group
COMPREHENSION TESTS BRING NO IMPROVEMENT
Akhtar & Tomasello (1997) Developmental Psychology
Traditional Test:
Model: “Look, tamming! This is called tamming!“
Test: “Show me: the dog‘s tamming the car!“
Result: Very few children at 2;9 act out the utterance in canonical fashion.
Modified Test:
Model: “Look at this fun game!“
Test: “Show me: the pig‘s tamming the horse!“
Result: Very few children at 2;9 act out the utterance in canonical fashion.
PREFERENTIAL LOOKING
Naigles (1990), Fisher (2000)
“Cookie Monster’s glorping Big Bird”
Savage, Lieven, & Tomasello (submitted)Structural Priming
• 3-, 4-, and 6-Year-Old Children
• Pictures that could be described with actives or passives
• Four Conditions (5 Items per Condition):• Active- Passive Prime• Lo-Hi Overlap
Hi Overlap(Active) = “It VERBed it”Lo Overlap (Active) = “A truck VERBxx the car”
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
3-years 4-years 6-years
High ATLow ATHigh PTLow PT
PASSIVE RESPONSESas a function of prime type
Active Primes
Passive Primes
Summary: All Methods IndicateItem-Based Organization
• Spontaneous Speech (+diary)
• Production Experiments (nonce verbs)
• Weird Word Order Studies(nonce verbs)
• Comprehension Experiments (nonce verbs)
• Priming Study (English verbs)
Linguistic Representations
• Token Frequency entrenches concrete part of construction
• Type Frequency creates abstractions
Linguistic representations thus become “stronger” and moreabstract with increasing linguistic experience
PROCESSES OF MAKING AND CONSTRAINING SYNTACTIC
ABSTRACTIONS
Main Processes
• Imitative Learning
• Abstraction: analogy (distributional analysis)
• Constraint: entrenchment, preemption
Cameron-Faulkner, Lieven, & Tomasello (submitted)
What’s .18 Where’s .05What’re .09 Where’re .02What do .05 Where shall .01What did .04What has .03 Who’s .08What about .03 Who did .01What shall .02What can .02 Which one .02What does .02What hppnd .01 Why don’t .01What were .01What kind of .01 How many .01
31 frames =>80% of Wh Qs
13 frames =>65% of Wh Qs
Fragments20%
Questions32%
Imperatives9%
SV(X)18%
Complex6%
Copulas15%
5/20%
9/38% 20/
67%
6/53%
8/77%
4/38%
How Item-Based is Mother’s Speech to Children?Cameron-Faulkner, Lieven, & Tomasello (submitted)
• 51% from 52 frames• 45% start w/ one of 17 words
Input: “Can he eat_ it?”“Shall he eat_ it?”“Does he eat_ it?”“Where will he eat_ it?”
Child: “He eat_ it”
“Imitation” Leading to Errors
Theakston, Lieven, & Tomasello (submitted)
- number of children (n = 24 in each)
02468
101214161820
Hears -sOnly
Hears Both Hears ฟOnly
says -ssays bothsays ฟ
Schematization
The dog eats the bone.The cat the fish.A bird a ladybug.This one that one.Two pin two dogs.
Analogy
The dog eats the bone.The cat catches the fish.A bird found a ladybug.This one wants that one.Two pin onto two dogs.
Childers & Tomasello (2001) Developmental Psychology
Children at 2:6 hear several hundred transitive utterances over4 days/sessions
• Either familiar or unfamiliar English verbs
• With either nouns only in slots or nouns & pronouns (consistent)
Test is traditional nonce verb learning• child hears nonce verb as intransitive or passive
and must produce in transitive
Number of children in each condition (out of 10) who produced at least one productive utterance with at least one nonce verb during testing
0
1
2
3
4
5
6
7
8
9
FamiliarVerb
UnfamiliarVerb
Control
NounsPronouns
20% = same as inprevious studies
He’s push ing it.He’s killHe’s pullHe’s showHe’s drawHe’s deed
TRANSITIVITY OVERGENERALIZATIONS FROM BOWERMAN
• Mommy, can you stay this open?
• I come closer so it won‘t fall.
• Don‘t giggle me.
• She came it over there.
• I want to stay this rubber band on.
• Eva won‘t stay things where I want them to be.
• You cried her.
• Will you climb me up there?
• „Kannst Du mich hochklettern?“
PROPOSED MECHANISMS FOR HOW
CHILDREN CONSTRAIN CONSTRUCTIONS
• ENTRENCHMENT: Repeated use makes other uses sound unconventionalChild hears “X hit Y“ many times, but never “Y hit“
• PRE-EMPTION: Alternative forms block the extension of a verb to a constructionChild watches as adult tickles sibling.Sibling says: “I can’t stop laughing.”Child now expects sibling to say: “Don’t laugh me”.Sibling says: “Don’t make me laugh.”
• ANALOGIES: Semantic subclasses of verbsChild learns verb for “causing direct motion” (remove)Child assumes it behaves like other verbs of the same type, i.e., as fixed transitive (e.g. bring, take, etc.)
BROOKS, TOMASELLO, LEWIS, & DODSON
Child Development (1999), Entrenchment Study
Fixed Transitivity Verbs
More Entrenched Less EntrenchedHit StrikeTake RemoveCome ArriveDisappear Vanish
Method
Children (3-8 yrs.) see transitive event and then are asked “mismatching” question- “What happened with PATIENT?”
- “What did AGENT do?”
Results
More entrenched verbs overgeneralized less often
Brooks & Tomasello (Language, 1999)Alternatives Study
Nonce verbs
Semantic subclasses- Some verbs from class that alternates- Some verbs of fixed transitivity
Preempting alternatives- Some children hear alternative that could be used to answer mismatching question (e.g. perphrastic causative)- Some children hear no alternatives
Methods
• Children (2.5 - 6.5) learn verb in one construction (e.g., intransitive)• Children are asked “mismatching” Q
Results
Both factors are effective, but only from 4.5 years
Three constraining factors working over developmental time.
2 3 4 5 6
Entrenchment
PreemptionVerb
Subclasses
Growing abstractness ofthe transitive construction
GiggleChortleLaugh
Many overgeneralizations b/c not entrenched
No overgeneralizationsb/c Verb Islands
Low overgeneralzations b/cpreemtion and verb subclassesin addition to entrenchment
CHILDREN’S PRODUCTION:Usage-Based Syntactic
Operations
LIEVEN, BEHRENS, SPEARES, & TOMASELLO(in press)
General
• One English-Speaking girl, 24 months• One hour tape, 5 days per week, 6 weeks• Maternal diary 7 days per week, 6 weeks [focus on “new” utterances]
This study
• Last hour-long tape = 527 utterances (295 multi-word)• Search for precursors on 6 weeks previous
Total Utterances = 537Novel = 21%
Multi-Word = 295Novel = 37%
One Operation = 74%Subst. = 84%Add = 15%Drop = 1%
Multiple Op = 26%
OVERALL ANALYSIS
Where’s the bus
Where’s the mega-blocksWhere’s the babyWhere’s the plate
Where’s the _____
50x +
Bertie busBertie @ bus
bus
Put a bit more here
Put cup herePut some more herePut it herePut some more herePut it herePut the apple herePut some here (3x)Put it here (10x)
Put ____ here
a bit more
A bit more hereA bit more milkA bit more (2x)Want a bit more
Mommy’s toast not ready
___ not ready
Mommy’s ___
Dinner’s not readyIt’s not readyIs@ not ready (2x)
Not ready (5x)
50x+
My toast
2x
COMPLEXCONSTRUCTIONS
DEVELOPMENT OF S-COMPLEMENTSDiessel & Tomasello, Cognitive Linguistics (2001)
Subjects: Adam, Eve, Sarah, Naomi, Peter, Nina - 1 to 5 yearsComplex Ss: 2807 tokens
Examples from Sarah: Examples from Nina:I think he’s gone See that monkey cryingI think it’s in here See Becca sleepingI think my daddy took it See that goI think I saw one See my hands are washedit’s a crazy bone, I think See he bites meI think dis is de bowl See him lie down
% Subjects in Complex Ss
1-P 2-P 3-P Lex ImpGuess 100 -- -- -- --Bet 100 -- -- -- --Mean 52 48 -- -- --Know 36 55 05 04 --Think 85 13 02 -- --Wish 97 -- -- 03 --Hope 88 12 -- -- --
See 07 01 01 -- 91Look -- -- -- -- 100Watch -- -- 11 -- 89Remember 6 6 -- -- 88
- Virtually no complementizers
- Virtually no non-present tenses
- Virtually no modals or negations
Diessel & Tomasello, Cognitive Linguistics (2001)
Development of Matrix Clauses
Age Sarah Adam
>2;11 I think [2;11] (2)
3;0-3;11 I think [3;1] (26)
(Do) you think [3;7] (2)
I think (7)
Do you think [3;3] (4)
Does he think [3;3] (3)
You don’t think [3;5] (1)
What do you think [3;5] (1)
I don’t think [3;8] (2)
4;0-5;0 I think (42)
Do you think (3)
I thought [4;1] (7)
I’m thinking [4;2] (1)
They think [4;3] (1)
What do you think [4;4] (1)
I don’t think [4;8] (2)
I’ll think [4;10] (1)
I think (99)
Do you think (5)
I don’t think (2)
Why do you think (2)
What do you think (1)
One think [4;6] (1)
Paul think [4;10] (1)
Diessel & Tomasello, Cognitive Linguistics (2001)
DEVELOPMENT OF RELATIVE CLAUSESDiessel & Tomasello, Cognitive Linguistics (2000)
- Subjects: 4 CHILDES children from 1;9 to 5;1
- Total of 324 relative clauses
Here’s the toy that goes around.
That’s the sugar that fell out.
There’s the ball I bought
This’s the bird that sings.
That’s the one that goes moo.
Here’s the boy that ran into the water.
Earliest All
NP ONLY:
“The girl that came with us” .05 .19
PRESENTATIONALS
“This is the car that turns around” .75 .47
OBLIQUES
“I’m going to the zoo that has snakes” 0 .06
OBJECT
“She has a bathtub that goes with it” .20* .26
SUBJECT
“The one that not finished is up there” 0 .01
* 50% of these = “Look at all the chairs Peter’s got”
Diessel & Tomasello, Cognitive Linguistics (2000)
CONCLUSIONS
• Imitative Learning
• Schematization & Analogy
• Entrenchment & Preemption
• Distributional Analysis
• Usage-Based Syntactic Operations