Context effects on children's interpretation and generalization of novel action verbs

MISSING IMAGE

Material Information

Title:
Context effects on children's interpretation and generalization of novel action verbs
Physical Description:
vii, 83 leaves : ; 29 cm.
Language:
English
Creator:
Forbes, James N., 1957-
Publication Date:

Subjects

Subjects / Keywords:
Context (Linguistics)   ( lcsh )
Children -- Language   ( lcsh )
English language -- Verb   ( lcsh )
Psychology thesis Ph. D
Dissertations, Academic -- Psychology -- UF
Genre:
bibliography   ( marcgt )
non-fiction   ( marcgt )

Notes

Thesis:
Thesis (Ph. D.)--University of Florida, 1993.
Bibliography:
Includes bibliographical references (leaves 79-82).
Statement of Responsibility:
by James N. Forbes.
General Note:
Typescript.
General Note:
Vita.

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 001937402
oclc - 30899715
notis - AKB3527
System ID:
AA00002076:00001

Full Text












CONTEXT


EFFECTS ON CHILDREN'S
GENERALIZATION OF NOVEL


INTERPRETATION
ACTION VERBS


AND


JAMES


FORBES


A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY















ACKNOWLEDGMENTS


This


research


was


supported,


part


, by


a National


Institute


of Child


Health


and


Development


predoctoral


traineeship


1T32HD07318


to the


author


awarded


through


Department


of Psychology,


University


Florida.


I would


like


express


sincerest


appreciation


every


child


and


adult


to the


who


participated


directors


, principals


studies.


, and


Thanks


teachers


also


centers


and


schools


: Hidden


Oak


, My


School,


North


Florida


Day


Care


Pride


-N-Joy


Small


World,


and


Terwilliger


their


cooperation.


Marcelle


Beckford,


Scott


Jaffe,


Swati


Patel


and


Susanne


Weinstein


rendered


invaluable


service


with


stimuli


logistical


preparation


details


, data


that


collection


such


, and


a project


myriad


entails


other


owe


Howard


Beck


, Ira


Fischler


, Linda


Lombardino,


and


Scott


Miller


a large


debt


of gratitude


their


generous


help


training


me to become


a researcher


As always


, Jeffrey


Farrar


asked


integral


questions,


answers


to which


shaped


this


research


into


a much


more


substantive


form


than


would


otherwise


have


been


case.

















TABLE


LIST


OF CONTENTS


OF TABLES. ........ .. ... ...


ABSTRACT........... . . . . . . . .

CHAPTERS


CONTEXT EFFECTS ON CHILDREN'S
AND GENERALIZATION OF NOVEL


INTERPRETATION
ACTION VERBS..


*. 0 1


Perspe


ctives


on Lexical


Acquisition


Constrai


n


Lexical A
Restric
Lexical A
Enablin
Importance
Developme
Verb Repres
Verb Learni
Action Even
Initial Exe
Adults' V
Generaliz


ts.
6cqu
tiv
.cqu
g C
of
nt.
ent
ng
ts
mpl
erb
ati


ition


Research


Based


Constraints View....
ition Research Based
straints View.......
rbs in Early Languag


ation.....
Constraint
and Verb L
ar Effects
Interpret
ons ..


on

on


S


e


s . . ..
learning.......
on Children' s
nations and


* t
the

the


..........
* .5**.0t t S

* S
* C C Srtrt~


and


EXPERIMENT


Method...
Subject
Materia
Procedu
Results..
Analysis
Select
Exper
Event E


1... .5C C.. 0 St. .... .


*
* C t t S*** S** C** ***** S at.. ... 1
O O .


s of N
on of
iment
effects


ovel
Four
2. .
on


Action
Events

Initial


Verb
for


Gen
Use


Assumptions


ralizations


. *.......


Y: 4r S-. a -I


!
r
!


13acle


ACKNOWLEDGMENTS .


|













M!ethod.. *** ***.


Materials . . . . .
SubjPects. . . .. ..
Materials ....... .
Procedure. .................
Results. . . . . .
Generalizations...........
Initial Assumptions ........


* S *
. . .
. .

. .

. . .a


* a a a a a *t* *a

. . a a a .



. a . a a a a
* S S S *S* S a a
* a a a a a .
I 00
"'0'''


Initial Experience Effects on Interpretations


of Verb Meaning...... ......
Discussion. ...... .... .. . .


Sg a a a a a0. a
. ......... .000.0. .


4 GENERAL DISCUSSION. . . . . . . ..


APPENDIX. . . . . . . . . . . .




BIREFERENCES.CH. . . . . . . .
BIOGRAPHICAL SKETCH ......,........................,......















LIST


OF TABLES


TABLES


Mean Number of Action
Standard Deviations)
Event Version, as a
for Experiment 1....


Verb Generalizations (and
Summed Across Event, by
Function of Change Type


Mean Proportion of
Change Type, and


Generalizations
Version of Verb


by Verb
Event..


Event,
S. ..


Mean proportion of Action Verb Generalizations
(and Standard Deviations) by Event and Change
Type, Collapsed Across Version...............

Mean Number of Verb Generalizations (and
Standard Deviations) by Age and Initial
Experience Condition as a Function of
Change Type. .... .... ....... ............... ..


PAGE















Abstract


of Dissertation


he University
Requirement f


Presented


Florida


Degree


to the


Partial
f Doctor


Graduate


School


Fulfillment


Philosophy


CONTEXT


EFFECTS


ON CHILDREN'


GENERALIZATION


OF NOVEL


INTERPRETATION


ACTION


AND


VERBS


James


August,


Forbes

1993


Chairman:


Major


Jeffrey


Department


Farrar


: Psychology


Two


studies


examined


how


children


and


adults


learn


and


subsequently


extend


new


verbs


to appropriate


contexts.


Experiment


1 demonstrated


importance


adults


accord


manner


, outcome,


instrument,


and


agent


their


initial


interpretation


of action


verb


meaning.


Adults


were


first


taught


verbs


using


novel


videotaped


action


events.


Next,


subjects


were


asked


whether


these


verbs


labeled


other


action


events


which


only


agent,


instrument,


manner,


outcome


had


been


changed.


Almost


invariably,


adults


considered


outcome


and


manner


changes


more


important


their


initial


interpretation


novel


events.


Two


further


issues


were


of particular


interest


Experiment


The


-first


concerned


whether


children


initial


assumptions









biased


interpretations


are


affected


differences


between


kinds


initial


events


from


which


novel


action


verbs


are


learned.


nearly


extend


viewing


When


identical


novel


three


initial


events,


verbs


events


subjects


When


which


training


were


consisted


least


initial


a single


three


likely


training


event


consisted


component


varied,


subjects


were


most


likely


to extend


novel


verb.


When


initial


training


consisted


viewing


events


which


only


one


component


was


held


constant


while


others


varied


Most

and


, subjects


' made


strikingly


event


component


an intermediate


! combined

varied ac


effects


cording


number


initial


to subjects


extensions.


training


age


Unlike


adults


and


-year


-olds


-year


-olds


' biased


interpretations


changed


of novel


different


action


types


verbs


initial


can


verb


actually


learning


experiences.


Taken


together


, the


studies


show


that


-year


olds


' action


verb


concept


differs


a principled


way


from


that


of 10-year


-olds


and


adults.


Results


are


further


interpreted


terms


of verb


learning


biases


and


developing


action


verb


concepts















CHAPTER


CONTEXT


EFFECTS


ON CHILDREN'


GENERALIZATION


OF NOVEL


INTERPRETATION


ACTION


AND


VERBS


Manifestly,


task


explaining


how


children


learn


words


to this


a daunting

complexity


assume


that


complex


many


children


endeavor.


theories


benefit


A persistent


lexical


from


response


acquisition


a set


language-specific


constraints


without


which,


task


would


impossible


(e.g.,


Markman,


1991;


Nagy


& Gentner,


1990).


In general,


constraint


theorists


tend


to focus


almost


exclusively


on characteristics


language


learners


to the


exclusion


larger


context


which


learning


a language


occurs.


However


, other


researchers


have


shown


that


social-pragmatic


context


is an essential


source


information


which


guides


children


s lexical


acquisition


(Nelson,


1988;


Tomasello


, 1992b).


Thus


adults'


ability


interpret


children


focus


of attention,


shared


experience


of past

language


and


ongoing


input


have


events

been D


or even


proposed


properties

important


determinants


children


s acquis


ition


word


meaning


(Behrend


& Harris,


1991;


Bowerman,


1982;


Nelson,


1985


Tomasello


& Farrar,


1986).


The


premise


research


that


verb


acqui


sition


and











altogether


terms


context


within


which


verbs


are


modeled.


Rather,


a more


complete


explanation


verb


acquisition


and


extension


lies


understanding


how


these


and


other


factors


interact.


interest


from


both


studies


reported


this


dissertation


was


whether


children


and


adults


are


biased


to interpret


some


components


of action


verb


meaning


as more


central


to their


initial


representations.


Of particular


interest


from


second


study


was


how


differences


kinds


initial


exemplars


from


which


verbs


are


learned


affect


children' s


and


adults


initial


interpretation


and


generalization


of novel


action


verbs.


Perspectives


on Lexical


Acquisition


Constraints


Debate


over


precise


nature


of constraints


on word


acquisition


produced


a number


perspectives


about


how


constraints


should


be depi


cted


e.g.


, Clark,


1987


Galli


tel,


Brown


, Carey,


Gelman


& Keil,


1991;


Markman,


1991;


Tomasello,


1992b) .


Part


disagreement


attributable


difficulty


in explaining


specific


mechanisms


which


constraints


operate


, what


exactly


that


they


(Nelson,


1988)


, when


and


how


they


arise


(Merriman


& Bowman,


1989),


press


or even


One


when


way


they


to analyze


diminish


lexical


(Forbes


& Farrar,


constraints


to select











little


explicit


treatment,


restrictive-enabling


dimension.


The


restrictive


-enabling


distinction


nonetheless


crucial


because


determines


many


theoretical


assumptions


concerning


constraints,


which


hypotheses


get


tested,


and


research


methodologies


brought


to bear


on these


questions.


Arguments


restrictive


constraints


tend


to emphasize


indefinite


number


hypotheses


children


could


make


about


meaning


Mervis,

constrai


of novel


words


& Hirsh-Pasek 1

nts restricting


992;

the


9!.


Quine,


Markman,

number c


1960;


1991) .


Golinkoff,


Consequently


)f hypotheses


children


could


entertain


about


meaning


of-novel


words


are


viewed


as essential

constraints


content


for

tend


word

also


knowledge.


learning


to emphasize


Arguments


occur.

innate,

enabling


Restrictive

language-specific


constraints,


other


hand,


tend


to emphasize


strategies


acquired


during


course


of language


learning


that


help


children


learn


ways


.g.,


Keil


which


Bowerma


, 1991


a given

n, 1985;


The


language

Galliste


enabling


organizes

1, Brown,


constraints


semantic


Carey


view


meaning


, Gelman,


stresses


procedural


knowledge


and


strategies


adopted


word


learners


which


allow


them


to pick


semantic


distinctions


specific


to their


language.


Rather


than


a bimodal


a bimodal











and


enabling


elements,


while


others


can


characterized


almost


entirely


one


pole


or the


other.


studies


examine


how the


restrictive-enabling


distinction


applies


to verb


acquisition.


They


so by


first


establishing


whether


children


and


adults


are


biased


favor


some


elements


verb


meaning


over


others.


Biased


verb


interpretations


would


constitute


support


for


either


restrictive


or enabling


constraints


position.


This


because


biases


could


reflect


operation


innate


language


specific


constraints


or the


structure


and


context


verb


whether


learning


biases


situations.


can


Next,


be attenuated


studies


or even


examine


changed


different


and

are


types


young ch

directly


verb


ildren


compared


learning


interpretations

to determine w


experiences.


of novel


whether


Also,


verb


biases


adults'


meaning

change


developmentally.


The


extent


to which


biased


interpretations


can


altered


verb


learning


situation


and


show


no developmental


changes


would


consistent


with


restrictive


biased


constraints


interpretations


explanation.


can


The


altered


extent


verb


to which


learning


situation


and


show


developmental


changes


would


consistent


with


an enabling


constraints


explanation.


At this


point


, a more


thorough


review


research


w











Lexical


Acquisition


Research


Based


on the


Restrictive


Constraints


View


One


example


of where


research


based


on a restricting


constraints


view


leads


can


seen


Markman


(1991)


account


of lexical


acquisition.


According


to Markman,


lexical


acquisition


very


young


children


made


possible


constraints


restricting


kinds


hypotheses


they


could


example


otherwise


, young


make


about


children


have


meaning


a well


novel


known


words.


tendency


categorize


nonlinguistic


concepts


on the


basis


thematic


relations

& Hemenway


g. ,


banana-monkey)


K., 1984)


Clearly


(Nelson

, young


1977;


children


Tversky


will


I B.,


more


successful


correctly


general


zing


novel


object


labels


such


membership

taxonomi c


ruit" if

based on

relations


they


so on the


perceptual


.g.,


and


banana


basis


functional


-grapes).


category

similarity--


A important


aspect


lexical


acquisition


learning


to extend


words


objects,


events,


and


concepts


similar


to those


with


which


word


was


originally


modeled.


In fact,


Markinan


and


Hutchinson


allows


(1984)


children


claim


that


to override


taxonomicc


their


tendency


assumption"


to categorize


using


thematic


relations


That


children


seek


to extend


a novel


object


word


on the


basis


taxonomic


rather


than











avoid


many


fruitless


hypotheses


about


what


a new


word


might


mean. S

children


object


similarly,


assume


rather


than


Markman' s


that


whole


a novel


some


object


label


specific


constraint


refers


leads


to a whole


attribute


object.


Rather


than


entertaining


an indefinite


number


of potentially


inappropriate


hypotheses


about


meaning


a novel


object


label,


children


s default


assumption


that


label


refers


to the


whole


referent


rather


than


one


properties.


Markman' s


(1991)


insistence


on the


necessity


restricting


whole


constraints


object


such


constraint


taxonomic


motivated


assumption


her


and


acceptance


Quinian


conundrum


depiction


word


learning.


Quine


1960)


depicted


task


learning


a novel


word


as a


logical

logical


induction


problem


problem.

induction


Quine's


give


formulation

n in terms


this

an adult


linguist


a foreign


land


trying


to determine


that


Qavacai


means


rabbit


when


uttered


a native


speaker


as a rabbit


runs


Logically,


rabbit


could


refer


color


animal'


fur,


any


animal


with


four


legs


, long


ears


, the


animal's


position


moment


uttering


gavagal,


and


on indefinitely.


Extrapolating


from


this


extensively


cited


and


influential


adult


example


, Markman


(1991)


claims











There


are


course


many


instances


where


children


would


be misled


a taxonomic


or whole


object


constraint


When,


example


, a child


hears


a novel


term


applied


an object


which


already


knows


name


, a whole


object


constraint


would


apparently


overrestrict


number


hypotheses


to consider.


Ostensibly


, the


immediate


result


would


be that


child


would


fail


to learn


novel


term


or even


become


confused


about


meaning


known


label.


Not


so argues


Markman.


In this


situation


another


constraint,


mutual


exclusivity


ME )


helps


children


override


whole


object


assumption.


Mutual


exclusivity


necessary


children


Markman


do eventually


s scheme


learn


things


names


to account


object


how


parts


properties


, or attributes


Still


at least


children


must


override


circumstances


mutual

is whe


exclusivity

n a novel 1


itself


abel


clearly


Children


part


must


of a hierarchy


override


labels


ME to accept


a given


label


object


"animal"


creatures


known


to them


"dogs


Another


when


labels


from


different


languages


are


used


same


object


Very


early


children


bilingual


English


and


Spanish


must


accept


term


nerro


animals


known


to them


as dogs


they


are


to do


once


they


have


acquired


about











instances


Even


without


these


situations,


ME is


still


an all-or-none


constraint.


Golinkoff


et al


1992


demonstrated


that


there


was


only


a small


statistical


difference


between


children


who


reject


a new


term


they


already


have


a label


an object


and


children


who


accept


new


term


given


same


conditions


Hence


, a more


accurate


depiction


of ME would


as a probabilistic


bias


or perhaps


a default


assumption


that


children


must


regularly


override


Merriman


& Bowman,


1989;


Woodward


& Markman,


press).


Nonetheless,


Markman


(1991)


claims


that


constraints


such


these


are


necessary


children


even


to break


into


language.


They


must


, perforce,


operative


or before


time


infants


begin


use


language


during


first


part


second


year


life.


Indeed


, there


some


evidence


that


are


children


provided


follow


as early


taxonomic


assumptions


18 months


, and


when


that


novel


-year


labels


-olds


respect


point


(Woodward


controversial


& Markman,


Using


press)


perhaps


this


slightly


last


more


demanding


tasks


-year


-olds


, Merriman


and


Bowman


1989)


concluded


that


children


younger


than


30 months


have


mutual


exclusivity


bias.


A further


criticism


of Markman


view


of constraints


lexical


acquisition


that


they


seem


to be


merely











that


allow


children


to override


ME rather


than


just


identify


instances


where


this


appears


occur.


Markman


must


also


explain


why


presence


or absence


of a label


should


lead


children


to generalize


according


to taxonomic


thematic


principles


After


all,


-year


-olds'


initial


assumptions


about


meaning


novel


motion


verbs


are


virtually


unaffected


presence


or absence


of a novel


verb


label


(Forbes


& Farrar


in press).


Forbes


and


Farrar


speculate


that


nonlanguage


-specific


knowledge


such


as causal


concepts


serve


as clues


about


how


novel


motion


verbs


and


events


can


initially


represented


lexicon


organized


in memory


Hence


, the


mere


presence


a verb


label


(with


minimal


further


syntactic


information)


provides


little


additional


useful


information


about


meaning


novel


verb


Without


some


explanation


of why


presence


absence


a novel


object


label


should


affect


children


initial


interpretation


of novel


lexical


items,


we can


not


understand


why


children


observe


ME given


one


circumstances


but


disregard


ME in


other


situations


For


present


purposes


, a complete


theoretical


account


representational


form


taxonomic


assumption,


whole


object


constraint


and


ME, along


with


age


onset


unnecessary


Constraints


may


be innately


specified


analog











(Karmiloff-Smith,


1991).


But


when


and


what


form


constraints


emerge


really


crucial


issue


for


purposes.


seek


only


to review


how


a restrictive


constraints


perspective


frame s


research


on lexical


acquisition


and


compare


that


with


how


an enabling


constraints


perspective


approaches


same


task.


this


latter


perspective


that


now


turn.


Lexical


Research


Based


on the


Enabling


Constraint s


View


Framing


word


learning


as a formal


logical


induction


problem


led


many


adherents


to regard


word


learning


as a


"point


and


name


ostension


paradigm"


(Nelson,


Unfortunately,


this


model


child


induction


problem


is misleading


that


assumes


that


children


must


learn


only


object


labels


and


then


only


situations


entirely


divorced


from


any


supporting


contextual


cues.


Moreover


proponents


restrictive


constraints


tend


view


children,


devoid


only


social


unit


cognitive


analysis


or social


explanations


learning


of how


skills,


language


acquired.


located


Since


entirely


lexical


within


acquisition


child,


has


conveniently


child


been


credited


with


powerful,


innate


language


specific


learning


mechanisms


overcome


wildly


exaggerated


indeterminacy


~cauisiti. on











language


been


reduced


nearly


to the


level


mere


noise.


Enabling


constraints


theorists


, on the


other


hand


, tend


to explain


how


young


children


make


productive


use


three


distinct,


yet


interrelated


sources


information.


One


these


sources


consists


children


s available


conceptualizations


that


they


bring


with


them


any


language


learning


situation.


Many


theorists


have


argued


that


children


learning


s earliest


label


lexical


acquisition


of previously


consists


established


their


concepts


such


those


learned


during


course


familiar


events


Gopnik


& Melt


zoff


, 1986;


Mervis


, 1987;


Nelson


, 1985


Event


familiarity


development,


in part


facilitates


, by


-year


providing


-olds


' language


a conceptual


framework


children


to interpret


novel


lexical


forms


and


functions


Farrar


, Friend


, & Forbes


, in press).


Children


are


remarkably


proficient


at acquiring


lexical


items


modelling


those


used


adults


contexts


that


are


partially,


but


fully


account


d for


their


cognitive


models


(Nelson


, 1992)


Nelson


refers


to this


ability


as the


principle


"Use


Before


Meaning


Children


s skill


apprehending


single


at least


context


part


which


of a word


used


s meaning


was


from


originally


even


termed











make


use


during


course


of early


lexical


acquisition


as well


as other


cognitive


achievements.


Additional


mechanisms


must


be specified


to account


often


slow


and


uncertain


process


of how


children


acquire


conventional


adult


meaning


of novel


words.


such


lexical


instruction


mechanism,


acquisition


offered


and


that

more


another


source


children ex

experienced


information


ploit,


others.


customized


Even


infants


came


pre-equipped


with


innate


constraints


, biases,


and


other


skills


(e.g.,


social


imitation)


to facilitate


their


learning


of object


labels,


this


knowledge


would


naught


without


an environment


to motivate


expression.


And


immediate


environment


with


which


infants


come


into


contact


is often


their


mothers


, who


are


already


well


versed


language


Obviously,


Child


and


eager


mothers


directed


can


speech,


pass


not

fine


on their


tuning,


knowledge.


for their c

and maternal


children.

labeling


practices


located


are


socially


within


distributed


child,


wholly


processes--not


localizable


wholly


within


more


experienced


other.


Thus


word


learning


has


been


characterized


as a social


convergence


process


because


children


converge


on the


adult


usages


through


various


forms


of apprenticeship


(Adams


& Bullock,


1986;


Rogoff,


1990) .











child


dyads


were


given


a picture


book


containing


drawings


from


six


different


categories


animals.


Animal


varied


typicality


(basic


-get


example


, typical


subordinate


-get


example


, atypical


subordinate get


example)


and


context


, presented


with


members


similar


or dissimilar


typicality)


Mothers


were


asked


to play


with


their


children


they


Bullock


normally


found


would


that


using


labels


picture


mothers


book


provided


Adams


were


and


not


simply


given


as a function


their


children


s age


, but


also


typicality


and


context


Apparently


, mothers


were


only


attuned


their


children'


current


level


understanding


function


child's


age)


also


how


their


children


understanding


would


affected


different


example


aspects


, mothers


same


identified


learning


atypical


situation.


members


For


using


subordinate


at above


chance


levels


children


three


groups,


but


only


contexts


where


just


atypical


exemplars


were


presented.


researchers


also


found


global


shift


label


children


provided


following


their


mothers


' labeling


practi


ces


Adams


and


Bullock


interpret


these


data


as evidence


mothers


' accommodating


to their


children


increasing


lexical


development


, to dyads


agreement


about


their


joint


focus


attention


, and











Results


also


showed


that


mothers


initially


carried


bulk


labeling


responsibility.


Children's


beginning


word


production,


and


presumably


their


word


comprehension


well,


was


footing


insufficient


picture


to allow


naming.


participation


Gradually


, children


on an equal


assumed


proportionately


greater


naming


responsibility


as they


began


converge


on the


adult


patterns


of picture


labeling


This


dynami c


pattern


of picture


naming,


Adams


and


Bullock


1986)


argue,


demonstrates


a shift


responsibility


within


picture


naming


task


"indicative


transition


from


interpsychological


to the


intrapsychological


plane,


that


transformation


of a shared


activity


into


internalized


activity"


184).


While


preceding


study


does


demonstrate


how


lexical


acquisition


socially


and


extension


distributed


can


process,


conceptualized


so far,


precise


as a


social


mechanisms


that


actually


drive


process


have


been


discussed.


Specific


lexical


learning


mechanisms


identified


Adams


and


Bullock


(1986


include


hedges


(e.g.,


"A penguin


a funny


kind


of bird.


. 187)


which


serve


to expand


child c

phrases


categories


using


to include


labels


atypical


to highlight


instances.


distinctive


Modifying


perceptual


attributes


.g.,


" A zebra


a horse


with


stripes


were











"[Penguins]


are


birds


that


live


South


Pole


and


they


swim


and


they


catch


fish.


to help


their


children


override


salient,


misleading


perceptual


and


functional


attributes.


strategy


Mervis


mothers


use


(1987)


found


furthering


that


most


their


successful


children's


lexical


development

conjunction


involves


with


introducing


a concrete


an object' s


demonstration


name


perceptual


attributes


and


defining


functions.


To this


short


list


, must


added


what


Trevarthen


and


Hubley,


(1978)


have


termed


"secondary


intersubjectivity.


This


refers


to participation


triadic


interactions


Thou-It)


which


allow


children


to construct


with


more


experienced


about


others


first


a world


place


of shared


(Schaffer,


events


1984) .


converse


importance


of such


interactions


word


learning


is supported


positive


correlation


between


autistic


children


s ability


part'i


cipate


to learn


and


episodes


appropriately


of joint


use


attention


new


and


linguistic


their


forms


ability

Landry


& Loveland,


1986) .


In general


, autistic


children


have


difficulty


understanding


meaning


and


use


language


precisely


because


profound


disturbances


their


social


relationships


(Gordon,


1987).


All


preceding


mechanisms


drive


children











centered i

capitalize


instruction.


on the


Chi


diverse


Idren become

instructional


increasingly

techniques


able


offered.


Instruction


effective


precisely


because


children


active


picture


involvement


labeling,


ongoing


or mother-child


social


activity,


interaction


with


such


toys


other


objects.


A third


source


information


children


use


which


enables


them


acquire


and


correctly


extend


novel


object


words


kind


object


with


which


a novel


lexical


item


modeled.


hypothesized


time


Tomasello


that


would


, Mannle


children


affected


, and


Werdenschlag


acquisition


kind


(1988


target


training


word


they


initially


received


time


one.


To test


this


hypothesis,


authors


used


a training


study


to-assess


children


production


and


comprehension


of novel


object


words


(e.g.


sax


learned


after


having


previously


learned


new


words.


previous


learning


occurred


conditions.


In the


"similar"


condition,


children


were


first


taught


name


a highly s

condition,


similarr


object


children


were


, horn)


initially


In the


taught


"dissimilar"


name


highly


dissimilar


object


hook)


Subsequently,


children'


spontaneous


production


and


comprehension


novel


target


words


sax)


were


assessed.


Overall,


children











This


finding


clearly


shows


that


young


children


acquisition


knowledge


referents


a new


object


of previously


Tomasello


word


learned


et al.


affected


words


1988)


with


point


their


similar


that


similar


-referent


word


had


many


overlapping


features


with


subsequent


target


word.


This,


Tomasello


et al


. claim,


facilitated


children


task


of determining


which


features


target


word


were


relevant


assigning


new


adult


label


Conversely,


dissimilar


-referent


was


different


from


target


word


that


child


task


determining


which


were


relevant


features


assigning


adult


label


was


made


more


difficult.


expanding


language


learning


space


to include


child


context,


enabling


constraints


view


promotes


alternative


depiction


cognitive


mechanisms


language


acquisition.


tricting


constraints


are


replaced


children


others


s ability


, participate


to interact

episodes o


with


f joint


more


experienced


attention


and


imitate


others,


form


concepts


of objects


and


activity


experienced


routine


events


, as


well


as create


and


use


categories


conventional


symbols


(Tomasello,


1992b)


This


say


that


such


theorists


assume


that


infants


are


biologically


prepared


to acquire


language.


On the











preparedness


to acquire


language


been


redefined


accord


with


an alternative


view


language


learning


rather


than


minimized


or denied.


Importance


of Verbs


Early


Lanquaqe


Development


As is


research


evident


based


from


on the


preceding


restrictive


review,


or enabling


most


constraints


view


of early


lexical


acquisition


focuses


almost


exclusively


on concrete


object


nouns.


language


involves


much


more


than


knowledge


about


lists


words.


The


very


glue


that


binds


words


together


and


gives


language


gene rat ivi ty


verbs


Verbs


according


to Braine


1976),


functioning


as conceptual


frames


which


provide


a structure


for -larger


lexical


units.


Braine


argues


that


very


young


children'


understanding


meaning


and


function


verbs


allows


them


to make


crucial


transition


from


one-


to two-word


rules


utterances,


and


thereby


structure


demonstrating


language.


some


However,


knowledge


Tomasello


(1992a


claims


that


before


children


have


any


knowledge


verbs


as a grammatical


category,


their


verb


knowledge


resembles


"conceptual


islands"


around


which


meaning


and


argument


structure


slowly


accrue


and


are


subsequently


generalized.


Only


during


third


year


of life


, and


then


only


"mosaically,


" do


children


begin


to construct











syntax.


Even


very


young


children


rely


on their


knowledge


about


verb


definitions


and


conflation


classes


semantic


elements


that


may


occur


together


a given


language)


to decide


which


verbs


may


correctly


with


what


argument


structures


Pinker,


1989


Given


importance


verb


meaning


and


conflation


classes


, it


surprising


that


so little


empirical


study


on how


children


acquire


this


information


been


undertaken.


Verb


Representation


A serious


impediment


study


verb


acquisition


that


very


little


known


about


how


children,


or adults


that


matter


, represent


verb


meaning.


A reasonable


and


widely


used


strategy


studying


verb


representation


first


observe


how


children


and


adults


use


verbs.


When


children


s usage


does


differ


from


that


of adults


, one


can


assume


that


they


represent


meaning


roughly


same


way


When


children


usage


does


differ


from


that


of adults,


one


can


assume


that


they


represent


meaning


differently


Based


on a diary


study


of his


daughter's


verb


usage


during


her


entire


second


year


life


, Tomasello


1992a)


found


category


verbs


evidence


comprised


The


very


what


characteristic


broad


Tomasello


categories


called


feature


verbs


change


verbs


One


state


this


category











defined


beginning


and


end


points.


The


other


broad


category


comprised


what


Tomasello


called


activity


verbs


These


verbs


were


that


defined


not


as actions


involve


performed


transformations


people


e.g.


or other


climb


animals


watch


Interestingly


, some


activity


verbs


were


what


adults


would


use


as change


state


verbs


example


, T used


paint


primarily


to refer


to the


activity


painting


rather


than


transformation


an unpainted


surface


to a painted


one.


Tomasello


(1992a)


claims


that


s verb


usage


highlights


several


important


characteristic


verbs


that


even


beginning


word


learners


use


to distinguish


them


from


other


lexical


items


One


very


obvious


property


that


differentiates


verbs


from


nonverbs


role


temporal


sequence


Moreover


change


state


and


activity

temporal

verbs em


verbs


were


sequence


phasized


wa

the


differentiated

s highlighted.


result


which


aspect


change


or outcome


of state


an event


whereas


her

took


activity

place.


verbs

For c


emphasized


change


state


manner


verbs


which


, perhaps


event


temporal


sequence


was


especially


well


delineated


because


these


verbs


encode


events


with


definite


beginning


and


end


points


Other


researchers


have


focused


on whether


some


aspects


paint)











encoding


encoding

analysis


manner


change


-motion


.g.,


-of-state


children


stirrincr


, mixing


s spontaneous


speech


than


Based

along


those


on her

with


experimental


evidence


, Bowerman


(1982


found


that


children


overextensions


verb


fill


were


consistent


with


manner


-over-end


Hollander


and


.-state b

Goldberg


ias.


More


(1991)


recently


found


that


, Gropen,

children


Pinker


between


and


six


years


age


interpret


manner


of action


as more


important


meaning


verbs


such


as fill


and


empty


In fact,


state


: Whether


defining


a container


feature


ends


these


filled


verbs


or emptied.


Moreover


, Gropen


et al. found


that


children


misinterpretations


led


to specific


kinds


syntactic


errors


involving


same


verbs.


However,


seemingly


related,


but


completely


novel


verbs


manner


-over


-end


-state


bias


not


always


observed.


For


example


, Behrend


1990


taught


children


and


adults


novel


verbs


using


brief


, unconventional


videotaped


action


events.


After


were


were id

element


training,


examples


.entical


children


new


to the


of meaning


had


were


action


training

been ch


asked


verb.


events


anged.


whether


These


except

Behrend


other


other


that a

found


events


events


single


result


changes


made


children


less


likely


than


manner


changes











made


children


less


likely


than


manner


changes


to generalize


novel


motion


verbs.


One


way


to reconcile


these


differences


suggest


that

while


a manner

a result


bias

bias


occurs

occurs


more

more


often

often


familiar


during


verbs

learning


completely


novel


verbs.


The


studies


reported


this


dissertation


will


contribute


to the


construction


a more


complete


developmental


account


verb


representation.


Verb


Learning


Constraints


Like


verb


representation,


there


little


systematic


understanding


how


children


learn


meaning


of novel


verbs


and


then


generalize


them


to appropriate


situations


Golinkoff,


Jacquet


, and


Hirsh- Pasek


(199


blaze


a path


into


unknown


proposing


that


children'


action


verb


learning


expedited


principles


"category


scope"


action


nameless


verbs--or


category"


"action


(N3C)


scope"


The


and


former


"novel


principle


name,


refers


situations


which


children


though


verbs


labeling


action


events


are


extended


to events


same


type


regardless


superficial


differences


manner


in which


action


action.


In i

N3C


performed


refers


or a change


to situations


agent


which


performing


children


assume


novel


verb


refers


an unnamed


novel


action


rather


than


than











et al.


1992)


presented


34-month-olds


with


two -dimensional


drawings


condition


depicting


children


familiar


were


and


shown


unfamiliar


familiar


actions.


actions,


one


one


just-learned


performed


a different


agent,


and


one


novel


action.


Results


indicated


that


children


generalized


new


action


verb


to the


just


-learned


action


spite


agent

level


change


of chance.


frequencies

In another


significantly

condition ch


higher


ildren


than

were


the

shown


familiar,


one


just-learned


action,


and


one


novel


action.


Results


indicated


that


children


generalized


a novel


verb


a previously


unnamed


novel


action


at significantly


higher


than


Pasek


chance


(1992


levels.


claim


Thus


to have


Golinkoff,


adduced


Jacquet,


empirical


and


Hirsh-


support


N3C


and


AS for


action


verbs.


Several


limitations


above


paradigm


restrict


scope


conclusions


One


that


Golinkoff,


Jacquet,


and


Hirsh


-Pasek


(199


chose


to test


principles


of verb


learning


in young


children


using


two -dimensional


drawings


actions


Novel


name


nameless


category


was


originally


proposed

Mervis,


as a principle

& Hirsh-Pasek,


noun


1992)


learning

In fact,


(see


Golinkoff


Golinkoff


, Mervis,


and


Hirsh-Pasek


generally


assume


(1992)

that


found

novel


that

object


young c

labels


children do

refer to











other


than


just


object


labels.


Action


scope


derives


from


category


scope,


which


another


object


label


learning


principle


identified


Golinkoff


, Mervis,


and


Hirsh-Pasek


1992


Indeed


Golinkoff,


Mervis,


and


Hirsh-Pasek


(1992


found


that


shape


was


primary


criterion


young


children


use


their


generalizations


of object


words.


Thus


another


aim


of Golinkoff,


Jacquet,


and


Hirsh-


Pasek


1992)


was


identify k

Deriving v

to account


.ey

"erb

fo


criteria

learning

r object


children's


principles


label


ifrbm


learning


extension


principles


may


verbs.

proposed


be a perfectly


reasonable


strategy


studying


verb


acquisition.


Unfortunately,


researchers


elected


test


N3C


and


actions


depicted


dimensional


drawings.


Obviously


, drawings


resemble


concrete


object


referents


more


than


dynamic


Jacquet,


and


, transitory,


Hirsh-Pasek


action


1992


events.


used


Because


pictures


Golinkoff


actions,


rather


than


actions


per


, it


is unclear


whether


N3C


and


adequately


characterize


children's


action


verb


learning.


A further


limitation


Golinkoff,


Jacquet,


and


Hirsh-Pasek


1992


study


that


action


verbs


was


incompletely


tested.


Although


researchers


did


test


principle


AS for


agent


changes,


they


did


so for


differences


in the


manner


which


an action


performed.











From


brief


reviews


ve rb


learning


and


representation,


clear


that


children


consider


some


elements


action


verb


meaning


to be


more


important


than


others.


Somewhat


problematically,


same


biases


are


consistently


unknown


found


how


across


biased


different


studies.


interpretations


verb


Also


virtually


meaning


may


change


developmentally.


Nor


much


known


about


whether


verb


learning


biases


can


altered


or even


changed


different


kinds


of verb


learning


experiences.


Action


Events


and


Verb


Learning


The


overall


objective


experiments


reported


this


dissertation


was


to examine


how


children


and


adults


interpret


and


generalize


novel


action


verbs.


All


verbs


used


labeled


events


which


consisted


actors


using


particular


to effect


instruments


an outcome.


to act


For


a given


example


manner


, faufing


on objects


referred


person


using


a fixed


remover


.e.


a specialized


bicycle


wrench)


to lift


plastic


bottles


over


a ramp


and


lay


them


onto


a stool.


According


to Slobin


1985)


, these


events


are


best


characterized


"manipulative


activity


scenes


" which


number


among


events


about


which


children


are


first


motivated


talk


(Tomasello


, 1992b)


Bruner


(1990)


claims


that


this


basic


manipulative


activity


scene











Experiment


1 determined


importance


adults


accord


manner


, outcome


, instrument,


and


agent


their


initial


interpretation


verbs


referring


to manipulative


activity


scenes.


To do


this,


adults


were


first


taught


novel


verbs


using


videotaped


manipulative


activity


events


Next,


subjects


were


asked


whether


these


verbs


labeled


other


manipulative


activity


events


which


only


manner,


outcome,


instrument


, or agent


had


been


changed.


Subjects


should


have


been


comparatively


less


likely


to extend


novel


verbs


to events


which


a component


, interpreted


more


central


meaning


verb,


had


been


changed.


More


specifically


rendered


adults


manner


ess


and


likely


outcome


changes


to generalize


than


should

agent


have

and


instrument


changes


(Behrend,


1990;


Forbes


& Farrar,


press;


Gentner,


1978).


Four


complete


events


used


Experiment


were


selected


use


as experimental


stimuli


Experiment


criterion


selection


was


those


events


that


best


exemplified


agent,


instrument


overall pattern

t, manner, and


of generalizations


outcome


changes.


Initial


Exemolar


Effects


on Children's


and


Adults'


Verb


Interpretations


and


Generalizations


Certainly


very


young


children' s


earliest


productive











1985;


Nelson,


1985;


Snyder,


Bates,


& Bretherton,


1981)


Part


of children


s word


learning


task


to appropriately


extend


words


beyond


exact


context


which


they


were


learned.


developing


important


understanding


ability


words


signals


children


as representative


symbols


classes


similar


objects


, events,


states


, actions


and


so on


(Nelson,


1985).


Very


little


research


specifically


addressed


process


of verb


generalization.


Smith


and


Sachs


1990


looked


relationship


between


event


representation


verb


production,


and


verb


comprehension


different


contexts.


children


under


years


age


, Smith


and


Sachs


suggest


that


their


ability


to represent


events


through


sequences


actions


positively


related


to contextual


flexibility


verb


comprehension.


The


focus


Smith


and


Sachs


study


was


very


young


children


s conceptual


basi


for

that


developing


am aware


verb


comprehension


of has


and


specifically


production


examined


how


No study


children


or adults


' interpretation


and


generalization


of novel


action


verbs


are


affected


differences


kind


initial


exemplars


from


which


they


are


learned


children


must


come


to understand


like


adults


that


meaning


a verb


like


to hammer


tied


to a specific


person,


an exact











time


hammer


user.


Hammering


does


not


always


, alas,


result


driving


a nail


straight


through


wood


to its


head,


nor


even


necessary


use


a hammer


to hammer.


Thus


depending


on children


s initial


exposure


to a hammering


event,


children

Two


The

abou


first

t the


could

issues


have

were


concerned


meaning


a different

of particul


whether

action


interpretation


interest


children' s


verbs


undergo


initial


hammering.


Experiment


assumptions


developmental


change.


Consequently,


children' s


and


adults'


interpretations


of novel


action


verbs


were


directly


compared.


The


importance


subjects


accorded


a particular


element


meaning


was


again


inferred


from


pattern


verb


generalizations.


less


inclined


Children


to extend


and


a newly


adults

learned


would

action


presumably


verb


situations


which


an important


element


meaning


had


been


changed.


Developmental


differences


initial


assumptions


would


consistent


with


enabling


constraints


perspective


focus


on lexical


acquisition


principles


acquired


during


course


word


learning


The


second


issue


was


whether


children' s


and


adults'


biased


interpretations


could


be attenuated


or even


changed


different


types


of verb


experiences.


Subjects


were


first


shown


brief


videotapes


depicting


novel


action


events


in one


events











noise


condition,


only


one


four


event


components


was


held


constant


across


three


training


events.


example,


holding


manner


component


constant


faufing


event


resulted


subjects


viewing


same


manner,


with


a different


different


versions


of ei


agent,


condition,

their the n


instrument,


subj ects


manner,


viewed


instrument,


and


outcome.


three


agent,


different


or outcome


event


, while


other


components


were


held


constant.


For


instance


, varying


manner


component


in the


faufins


event


resulted


subjects


viewing


three


different


manners


which


placed


event


onto


occurs


stool,


(e.g


bottles


bottles


are


are


lifted


launched


and


ramp


and


land


on the


stool,


and


bottles


are


tossed


onto


stool).


Thus


while


manner


component


changed


each


three


training


events,


agent


, instrument,


and


outcome


components


remained


unchanged.


sum,


subjects


same,

one v


constant


version


plus

the


noise,

training


and


different


event


three


conditions


times,


saw

three


versions


training


event


which


one


event


component


remained


constant


while


others


varied,


three


versions


training


event


which


only


one


event


component


varied


across


three


versions.


After


training


trials


three


conditions











whether


was


an example


novel


verb.


This


generalization


trial


was


held


constant


across


three


training


conditions.


To continue


with


faufiner


example,


subjects


three


conditions


saw


another


faufina


event


but


with


a different


manner


any


differences


in verb


generalizations


between


groups


were


due


to differences


between


kinds


of initial


exemplars


used


to model


novel


verbs.


same


condition


was


designed


to show


subjects'


initial


assumptions


about


importance


specific


components


action


verb


meaning


In effect


, they


were


"told"


that


each


components


were


equally


important.


Differences


between


generalizations


verb


components


would


indicate


which


elements


of action


verb


meaning


were


most


central


children


and


adult


s' initial


interpretations


different


condition


was


designed


reduce


subjects


' initial


assumptions


about


importance


very


same


elements


verb


meaning.


In effect


, they


were


"instructed"


that


a particular


component


was


important


to verb


meaning


since


they


saw


three


different


instantiations


The


extent


to which


subjects


ignore


this


information


specific


verb


indicates


learning


biases.


relative

Subjects


strength


their


constant











initial


interpretation


about


importance


of specific


components


action


verb


meaning.


In effect


, subjects


were


informed


that


one


verb


components


was


more


important


than


remaining


components.


final


interest


from


Experiment


was


whether


children


and


adults


differed


total


number


generalizations


they


generalizations


made,


each


and


initial


number


experience


condition.


Forbes


and


Farrar


press


found


that


children


generalized


novel


motion


verbs


more


conservatively


than


adults.


Consequently,


experimental


was


conditions


predicted


would


that


less


children


likely


three


to generalize


novel


action


verbs


than


adults.


Moreover


, subjects


viewing


three


nearly


identical


training


events


should


have


been


least


likely


to generalize


novel


verb


Showing


subjects


three


different


versions


of a novel


action


verb


should


have


made


them


most


likely


to. generalize.


And


conflicting


information


constant


plus


noise


should


have


resulted


subjects


accepting


an intermediate


number


of generalization


trials


as an example


novel


verb.















CHAPTER


EXPERIMENT


Method


snbi


ects


Participants


comprised


young


adults.


Adults


were


college


undergraduates


who


received


course


credit


participating.


Equal


numb


ers


feinal


and


mal


within


each


group


took


part


Only


native


Engli


sh language


speakers


were


tested.


Materials


Videotaped


training


events


depicting


novel


action


events


were


used


to teach


subj


ects


seven


different


novel


verbs


The


events


were


novel


insofar


as seven


different


adults


were


filmed


using


an instrument


to effect


a unique


action


resulting


in an original


outcome


Four


event


components


of principle


interest


included


Agent,


instrument


, manner


of action


, and


outcome


Faufing,


instance

remover


involved


(instrument


bottles


onto


an adult

) to lift


a stool


(agent

and p


using


lace


(outcome).


a fixed


(manner


Thus


, training


cup

action


events


were


so novel


that


they


are


generalizable


to the


*r All ~I Ct AS


A4C 4a











training


event.


Specifically


, the


four


generalization


trials


resembled


training


event


respects


except


a change


outcome.


agent,


A change


manner


of agent


-of-action


consisted


, instrument


a different


, or


person


performing


action.


Changes


manner


were


accomplished


varying


way


in which


person


accomplished


outcome


Instrument


action


event


changes


Outcome


involved


changes


varying


will


tool


involve


used


varying


result


versions


action


every


event.


generalization


Three


trial


slightly


were


different


filmed


total


84 generalization


trials.


All


training


and


generalization


trail


lasted


between


12 and


15 seconds


each.


Training


events


for


novel


verbs


and


generalization


trial


are


outlined


Appendix.


All


these


events


were


video


-recorded


using


a VHS


camera


and


displayed


using


a color


monitor


Procedure


Every


subject


viewed


same


seven


training


events


three


times


each.


Training


events


were


introduced


experimenter


following


way


: "Watch


this


"That


was


xing


" where


x represent


a novel


verb


root.


The


experimenter


showed


each


training


event


twice


more


saying


"Watch


again


" and


"Watch


once


more"


before


onset











Immediately


after


third


repetition


different


training


events,


each


subject


was


shown


one


of each


kind


generalization


trial


agent


, instrument


, manner,


and


outcome)


and


one


consisted


three


slightly


conditions


different


Conditions


versions


generalizations


trials


Thus


each


subject


such


saw


a total


of 49


events


: 21


training


and


28 generalization


events


half


generalization


trials


subjects


were


instructed


to watch


video


and


asked


: "Was


that


xing


, or was


that


something


else


For


other


half


generalization


trial


, the


order


query


was


reversed


: "Was


that


something,


or was


that


xing


Query


type


was


counterbalanced


across


generalization


trials.


After


first


block


training


and


generalization


trials

would


experimenter


six


more


informed


occurrences


subject


same


that


procedure


there

e, but


using


different


events


and


words


Training


and


generalization


events


were


shown


one


of four


random


orders


Subjects


were


tested


individually


and


a complete


session


lasted


about


20 minutes.


Results


Analysis


Novel


Action


Verb


Generalizations


Subjects'


generalizations


were


defined


as the


number











averaged


across


subjects.


The


highest


obtainable


mean


score


was


-seven.


A high


mean


score


a given


change


type


indicates


that


subjects


were


more


likely


to generalize


novel


action


verb


Conversely


, a low


mean


score


for


a given


change


type


indicates


that


subjects


were


less


likely


generalize


novel


action


verb.


One


obj ective


of Experiment


was


to determine


relative


instrument


importance


, and


adults


agent


accord


their


manner


initial


, outcome,


interpretation


verbs


gives


referring


mean


to manipulative


number


activity


of generalizations


scenes.


, summed


Table


across


seven


events


, by


type


change


trial


, agent,


instrument


, manner


, outcome)


and


event


version.


Overall,


outcome


and


manner


changes


most


consistently


rendered


adults


least


likely


to generalize


novel


verbs.


Adults


generalizations


to outcome


changes


were


lower


than


to those


manner


changes


Adults


accepted


agent


changes


more


often


than


any


other


change


type


Adults


' generalizations


instrument


changes


were


intermediate


between


those


agent


and


manner.











Table


Standard


Mean


Number


Deviations


as a Function


Action


Summed


of Change


Verb Generalizations


Across


Type


Event,


Event


Experiment


Version,


Change


Type


Version


Agent


Instrument


Manner


Outcome


Overall


.. 5.67


2.58


1.42


.83)


.03)


.93)


.24)


.03)


. . 6


4.42


.40)


.31)


.87)


.81)


.82)


. . . 6


4.40


.45)


.91)


.01)


.52)


.25)


Total


1.83


.20)


.03)


.93)


.24)


Note


Highest


possible


score


importance


of individual


change


types


adults'


interpretation


novel


action


verb


meaning


was


analyzed


using


Change


Type


: Agent


manner


vs. instrument


outcome)


x 7


(Event


: Blating


vs. faufing


vs. noffing


vs.


oushing


. prebing


vs. smuking


. talking


(Version)


mixed


analysis


variance


with


repeated


measures


on change


and


measures











general nations


individual


change


types


vary


across


verb


events


, but


they


also


varied


across


different


versions


verb


events.


The


second


objective


of Experiment


was


to select


events


, within


and


across


which


pattern


generalizations


did


meaningfully


vary


, for


use


Experiment


Thus


absolute


number


of generalizations


between


event


relatively


versions


unimportant


a specific


What


does


change


matter


type


, is whether


general


zations


among


four


change


types


, within


each


three


different


event


versions,


follow


overall


pattern


of generalizations


shown


Table


For


this


reason,


a useful


way


to examine


significant


three


-way


interaction


portion


versions


Tabl


focusing


interaction


on how


differs


conveniently


change


across


provides


type


three


adults'


event


event


mean


proportion


generalizations


change


type


and


event,


three


different


versions


verb


events.











Tabl


Mean


Proportion


of Generalizations


Verb


Event,


Change


Type,


and


Version


of Verb


Event


Change


Type


Event


Agent


Instrument


Manner


Outcome


Version


Blating


0.25


0.08


.45)


.52)


.45)


.29)


Faufing


0.25


.29)


.52)


.45)


Noffing


0.83


.39)


.52)


.49)


.29)


Rushing


0.42


.39)


.45)


0.08


.51)


Prebing


0.50


.52)


.52)


.51)


.52)


Smuking


0.75


.29)


.45)


.52)


Tulking


0.83


.39)


.49)


.52)


.52)


Version


i- -


n A


A A


r m


A A A











Table


continued.


Change Type


Event


Agent


Instrument


Manner


Outcome


Noffing


1.00


0.00)

1.00


Oushing


(0.00)


0.83


(0.39)

1.00


(0.00)


0.33

(0.49)

0.58


0.25

(0.45)

0.25


(0.52)


(0.45)


Prebing


0.92


0.75


0.83


0.67


(0.29)


(0.45)


(0.39)


(0.49)


Smuking


1.00


1.00


0.50


0.08


(0.00)


(0.00)


(0.52)


(0.29)


Tulking


1.00


0.00)


0.58


(0.52)


0.58

(0.52)


0.25

(0.45)


Version 3


Blating


1.00


0.83


0.08


0.50


(0.00)


(0.39)


(0.29)


(0.52)


Faufing


1.00


0.58


0.58


0.17


(0.00)


(0.52)


(0.52)


(0.39)


Noffing


0.92


0.83


0.83


0.58


(0.29)


(0.39)


(0.39)


(0.52)











Tabi


continued.


Change


Type


Event


Agent


Instrument


Manner


Outcome


Prebing


.29)


.29)


.45)


.00)


Smuking


1.00


0.33


.00)


.39)


.45)


.49)


Tulking


0.25


0.25


.00)


.45)


.52)


.45)


Inspection


of Tabi


reveal


that


majority


generalizations


change


type


and


event


conform


to the


overall


pattern


shown


Tabl


Namely,


adults


generalized


least


often


to outcome


changes


and


most


often


agent


changes.


Furthermore


, their


generalizations


instrument


changes


generally


exceeded


those


manner


changes


Two


obvious


exceptions


to this


overall


pattern


can


seen


mean


proportion


generalizations


manner


changes


within


blatinc


, prebing


, and


tulking


events


versions


two


and


three


of blatina


, along


with


version











generalizations


manner


changes


exceeded


(16)


or equalled


generalizations


to outcome


changes.


A second


exception


overall


pattern


of generalizations


occurred


primarily


with


talking


and


prebing


events.


Across


three


versions


tuikins


and


version


two


or prebing


adults


generalized


to instrument


changes


equally


or less


often


than


to four


manner


changes


and


three


outcome


changes.


Two


further


minor


differences


from


overall


pattern


change


type


event


generalizations


were


noted.


version


three


of oushing,


mean


proportion


generalizations


manner


changes


equaled


that


outcome


changes.


Secondly,


generalizations


agent


and


instrument


changes


were


equal


and


or near


ceiling)


oushinq


version


prebing


(version


and


smukinq


(version


Selection


Four


Events


Use


Experiment


Thus


analysis


indicates


that


generalizations


varied


least


across


three


versions


of faufingq,


noffing,


oushing


, and


smuking.


Generalizations


were


further


analyzed


using


(Event


: Faufing


vs. noffing


vs. oushing


vs.


smuking


vs.


(Change


outcome)


Type:


analysis


Agent


variance


vs.


instrument


with


repeated


. manner


measures


event


and


change


type.


Generalizations


were


collapsed


across


event


version


because


they


did


appear











< .03.


assess


how


generalizations


to each


four


change


types


varied


across


four


different


events


, the


scores


must


first


averaged


across


three


levels


event


version.


Event


effects


on, nitlial


assumptions


Tabi


shows


adult


s mean


proportion


generalizations


change


type


and


event,


collapsed


across


three


different


versions


verb


events.


Proportions


are


given


because


each


individual


change


type


within


a given


verb


event,


adults


could


either


accept


or reject


change


as an example


novel


verb.


An examination


Table


reveals


that


pattern


generalizations


across


four


events


was


virtually


identical.


Mean


generalizations


each


change


type


always


followed


Adults

often


same


generalized


to agent


pattern


least


changes,


within


often


and


each


to outcome


their


four


changes


events


most


generalizations


instrument


changes


exceeded


those


manner


changes











Table


Standar
Across


Mean


proportion


'd Deviations)
Version.


of Action


Event


and


Verb


Change


Generalizations(and


Type


, Collapsed


Change


Type


Agent


Instrument


Manner


Outcome


Event


Faufing


0.41


.17)


.50)


.50)


0.42


Noffing


.28)


.45)


.51)


0.47)


Gushing


.28)


.35)


.50)


0.40)

0.17


Smuking


.17)


.35)


.50)


.38)


Note.


Highest


possible


score


Because


effect


of change


type


did


significantly


vary


across


levels


event,


a posteriori


comparisons


change


types


within


each


verb


event


were


made.


All


following


adjustment


comparisons


to maintain


were


conducted


familywise


using


alpha


Bonferonni


.05.


Within











instrument


compared


to outcome


changes


within


four


events.


that


noffiner.


that


Contrasts


difference


Contrasts


difference


agent


was


and


significant


instrument


was


instrument


only


and


significant


manner


only


changes


faufina


changes


oushin


showed


and


showed


and


smuking.


Finally,


smukina


only


-was


difference


generalizations


between


manner


and


outcome


change


types


significant.


Discussion


Experiment


1 had


principle


aims.


One


these


was


to demonstrate


relative


importance


individual


event


components


adults


' interpretation


of novel


action


verb


meaning.


Based


on their


overall


pattern


generalizations,


adults


almost


invariably


considered


outcome


and


manner


changes


more


essential


than


instrument


and


agent


changes


their


initial


interpretation


novel


events.


With


exceptions,


adult


word


learners


extended


novel


action


verbs


to events


spite


agent


changes.


This


result


partially


supports


Golinkoff


and


coauthors'


(1992)


claim


that


action


verb


acquisition


proceeds


according


to a principle


Action


scope


also


predicts


that


word


learners


will


extend


novel


action


verbs


to events


regardless


of superficial


differences in t


manner


in which


act ions


are


performed.


actions











example


, adults


generalized


least


often


to outcome


changes


version


one


blatinc


, but


generalized


least


often


manner


changes


versions


and


three


same


event.


Evidently


, standing


and


pulling


, away


, then


dropping


near


end


of a table


figured


more


prominently


adult


' interpretation


blatinac


than


either


standing


and


pulling


, toward


, then


dropping


end


table;


end


or b)


a table.


kneeling


and


On the


pushing


other


hand,


, then


adults


dropping


generalized


manner


changes


more


often


than


to outcome


changes


three


versions


as well).


faufin


each


manner


(and


to most


change


other


faufinq


was


events


different.


Two


bottles


were


dragged


into


place


(version


, tossed


underhand


into


place


(version


, or launched


a ramp


into


place


version


This


empirical


evidence


suggests


that


toss


compared


to dragging


bottles


place


relatively


equivalent,


whereas


pulling


a tabl


and


away


from


oneself


compared


pulling


table


and


toward


oneself


not.


Unfortunately,


AS states


only


that


verbs


action


events


will


extended


to events


same


type


regardless


superficial


differences


manner


of action.


Action


scope


could


retained i


a more


complete


theory


of lexical











difference


between


manners


action


for


same


type


event.


Results


from


current


study


also


accord


with


those


of Behrend


(1990)


The


paradigm


current


study


was


very


similar


to that


used


Behrend .


In the


generalization


phase


Behrend


study


, adults


and


children


were


asked


whether


novel


action


verbs


extended


to events


in which


only


result


manner


of action,


or instrument


diff


ered


from


original


training


event


As


current


study


Behrend


found


that


result


changes


had


greatest


effect


action


changes


changes


an intermediate


weakest


effect


effect


both


, and


instrument


studies,


a greater


effect


meant


that


subjects


were


less


likely


to extend


novel


verb.


Behrend


1990


sugg


ested


that


findings


could


explained


terms


"result


verb-


bias


Such


a bias


could


, as Behrend


suggests


, facilitate


verb


learning


restricting


word


learners


hypotheses


about


what


novel


action


verbs


mean.


In the


current


study


, result


changes


did


always


verbs.


render


subjects


In nearly


one


least


willing


third


to generalize


total


number


novel


cases


mean


generalizations


manner


or instrument


changes


were


less


than


or equal


to those


of result


Obviously


, the











verbs.


In other


words


verb


learning


affected


only


principles


and


biases


held


word


learners,


but


also


very


events


from


which


verbs


are


learned.


Certainly


more


than


a principle


AS and


a result


verb


bias


verb


needed


to adequately


generalization.


characterize


Interpreting


then


process


subsequently


generalizing


than


novel


either


action


principle


verbs


of AS


a more


or a result


dynamic


bias


process


suggest.


Superficial


changes


manner


which


an action


takes


place


may


important


in isolation.


What


may


important


determining


whether


a particular


verb


will


extended


to similar


situations


how


such


changes


interact


with


other


components


action


verb


meaning.


Experiment


more


systematically


explores


how


context


in which


initial


training


occurs


affects


generalizations--and


inference,


initial


interpretations


action


verb


meaning.


Of primary


interest


Experiment


was


how


novel


differences


action


verb


between


are


the


learned


kinds

affect


events


children' s


from

and


which

adults'


verb


interpretation


and


generalization.


Therefore,


a second


aim


Experiment


was


to select


four


events


use


Experiment


Experiment


was


important


that


pattern


of generalizations


did











overall


pattern


of generalizations


observed


for


seven


events


combined.


All


four


change


types


affected


generalizations


same


way


for


each


selected


events.


Thus


differences


generalizations


observed


Experiment


are


due


to the


experimental


manipulation


rather


than


to properties


stimuli.
















CHAPTER


EXPERIMENT


Method


Subi


ects


Participants


included


-year


-olds


(mean


= 3;2


, range


10-year-olds


and


young


adults


mean


Preschoolers


10;1

and


, range


= 9;3-10;6)


forth-graders


were


recruited


letter


correspondence


with


parents


Adults


were


college


undergraduates


who


received


course


credit


participating.


Equal


numbers


females


and


mal


within


each


age


group


took


part


Only


native


English


language


speakers


were


tested.


Twenty-


one


-year


-olds


were


dropped


because


of experimental


error


and


three


other


three


-year


olds


were


dropped


because


they


failed


to complete


procedure.

Materials


Generalization


Experiment


: Faufincr,


trials


each


noffinq


four


oushina


and


events

smukincr


used

were


selected


use


Experiment


Since


there


were


three


versions


of each


event,


and


four


general


zation


trials


per


event


, the


total


number


trials


came


to 48


These


trials


1-------------------


A.- I-


---A.-.-.-


I*


-3;7)


~ 3;0


,,L,~


II


I ir r


r











other


change


one


event


had


same


relative


effect


same


combination


a different


event


as well


as for


different


versions


same


event.


follows


that


any


observed


differences


generalizations


obtained


Experiment


are


due


exclusively


to properties


stimuli,


but


rather


to differences


training


conditions.

Procedure


Procedures


paralleled


those


of Experiment


1 except


that


type


training


events


differed


and


were


shown


one


three


conditions


: Same,


different,


and


constant


plus


noise.


In the


same


condition,


subjects


were


shown


three


versions


same


event.


For


example


, manner


training


trials


faufing


consisted


three


virtually


identical


versions


of faufing


which


manner


well


as the


other


three


event


components


did


vary.


Manner


training


trials


were


virtually


, though


complete


identical


because


actor


was


videotaped


faufing


same


manner


three


different


times.


Events


were


filmed


this


way


so that


training


trials


would


simulate


watching


someone


perform


a novel


action


on three


separate


occasions,


where


very


minor


random


variations


action


would


occur.


The


same


procedure


was


followed


ror every


event


with


each











All


other


event


components


were


identical


in the


three


training


faufina


trials.


event


example,


consisted


manner


three


training


different


trials


manners


which


faufinq


occurred.


The


same


agent


did


faufinq,


with


same


instrument


, which


resulted


same


outcome


three


training


trials.


The

elements


third


condition,

previous


constant


two


plus


conditions.


noise,


comprised


Subjects


constant


which


other


plus


a single


three


noise


event


event


condition


component


components


saw


was


varied


three


held


training


constant


slightly.


trials


while


For


example,


manner


training


manner


trials


action


consisted


was


held


three


constant


trials


across


which


three


trials.


The


other


three


components


were


held


constant


across


sequence


only


three


three


trials


trials.


resulted


Thus


a training


a subject


viewing


single


manner


action,


two


different


agents,


two


different


instruments,


and


different


outcomes.


This


same


procedure


was


followed


agent,


instrument,


and


outcome


trials


constant


plus


noise


condition.


each


three


conditions


, all


four


events


were


used


once


as manner


training


trials.


Agent,


instrument,


and


outcome


training


trials


were


presented


same


way.











three


conditions.


The


order


training


event s


was


counterbalanced


across


subjects.


Training


events


were


identified


experimenter


follows


"Watch


this


" and


"That


was


xing


" said


after


event


had


been


viewed


Tomasello


and


Kruger


(1992)


found


that


unlike


nouns,


children


demonstrated


greater


production


and


superior


comprehension


of novel


verbs


learned


non-


ostensive


.e., impending,


completed)


compared


to ostensive


, ongoing)


contexts.


Generalization


trials


immediately


followed


third


training


trial


three


conditions.


The


experimenter


introduced


half


generalization


trials


saying:


"Was


that


xing


, or was


that


something


else?"


remaining


generalization


trials


order


query


was


reversed:


"Was

was


that

shown


something


one


else


or was


generalization


trial


hat

per


xing?"


Each

three


subject

training


events


a total


four


generalization


trials.


The


same


four


generalization


trials


were


used


three


conditions.


Generalization


trials


consisted


of a different


agent


, manner,


instrument,


or outcome


version


training


event.


For


example


, manner


training


trials


were


followed


a manner


generalization


trial,


instrument


training


trials


an instrument


generalization


trial,


and











Preschoolers


were


screened


to establish


whether


were


able


a video


to label


monitor


simple


and


familiar


then


and


whether


unfamiliar


subsequent


actions


events


seen


were


same


or something


different


e.g


., uncoupling


camcorder


from


a tripod


followed


panning


a cameraless


tripod


or uncoupling


camcorder


again


All


preschoolers


performed


prescreening


flawlessly


Results


General


zations


Generalizations


were


defined


number


times


subjects


accepted


a general


zation


trial


as an example


novel


verb


Subjects


could


either


accept


or reject


change


type


trial


as an example


novel


verb.


Thus


range


scores


went


from


zero


one


for


any


given


change


type


within


a single


verb


event


S.Table


4 gives


mean


frequency


of generalizations


individual


change


types,


condition


type


and


indicates


age.


that


A high


subjects


mean


were


frequency


more


a given


likely


change


to generalize


novel


action


verb.


Conversely


, a low


mean


frequency


a given


change


type


indicates


that


subjects


were


less


likely


to general


novel


action


verb.











Tabl


Mean


Deviation
Function


Number


by Age
Change


and


of Verb


Initial


Generalizations


Experience


Type


Condition


Standard


as a


Change


Type


Age


Agent


Instrument


Manner


Outcome


Overall


Same


.48)


.85

.37)


.51)


0.20


.51)


.41)


.37)


.05

.22)


.28)

.41

.23)


Adult


0.95


0.20


.22)


.50)


.31)


.41)


.19)


Total


.36)


.50)


.41)


.13

.34)


Constant


+ Noise


0.46


.38)


.52)


0.65


.47)


Adult


.48)


.52)


.48)


.52

.31)

.46


.35

.48)


0.75


0.60


.41)


.44)


.51)


V


.47)


.241


land











Tabl


continued.


Change


Type


Age


Agent


Instrument


Manner


Outcome


Overall


Different


0.71


.43)


.43)


.49)


.66

.25)


.46)


0.25


0.63


.47)


.48)


.44)


.21)


Adult


0.75


.00)


.36)


.47)


.51)


Total


0.78


0.44


.29)


.42)


.50)


.50)


Initial


Assumptions


The


focus


of Experiment


was


on whether


some


elements


action


verb


meaning


are


more


central


to children


s and


adults


' initial


interpretations


The


importance


subj


ects


acc


orded


specific


change


type


to their


initial


assumptions


of verb


meaning


was


erred


from


pattern


generalizations.











. manner


vs. instrument


vs.


outcome)


mixed


analysis


variance


with


repeated


measures


on component


As predicted,


three


main


effects


were


significant


The


kind


training


events


from


which


subjects


learned


novel


action


verbs


significantly


affected


how


likely


they


were


generalize


El2


17.48


.0001.


A look


at Table


shows


that


, overall


, subjects


different


condition


generalized

generalized


most

least


often;

often


those


Table


the s

4 also


ame


condition


shows


that


generalizations


subjects


constant


plus


noise


condition


were


generally


intermediate


between


those


same


and


different


conditions.


age


of subjects


significantly


affected


how


likely


they


would


generalize


151)


An inspection


data


summariz


Tabi


shows


that


children


generally


made


fewer


generalizations


generalizations


than


made


adults


-year


However


-olds


, the


was


mean


number


invariably


lower


than


those


made


adults


or even


10-year


-olds


Subjects'


general

affected


zations


kind


novel


of chang


verbs

e type


were

F(3


also


significantly


50.99


.0001


The


data


Table


show


that


outcome


and


manner


changes


generally


rendered


subjects


least


likely


to accept


change


trial


as an example


of a novel


verb.











A second


young


objective


children'


Experiment


initial


was


interpretations


to find


action


whether


verbs


differed


systematically


from


those


adults.


further


interest


was


whether


children


s and


adult


s biased


assumptions


of novel


action


verb


meaning


could


altered


different


Table


types


suggests


verb

that


learning


experiences


generalizations


for


Reinspection


instrument


manner,


and


outcome


change


types


varied


across


initial


training


conditions


-year


-olds


a way


that


adults


and


-year


-olds'


generalizations


did


not.


Therefore,


reasonable


strategy


analyzing


three-way


interaction


compare


how


-way


interaction


of change


type


and


initial


experience


varies


within


each


three


different


groups.


More


formally


, initial


experience


significantly


affected


-year


-olds


' extensions


novel


verbs


across


level


change


type,


S111i


= 2.46


.04.


Tests


simple


effects


were


next


conducted


within


each


four


different


.003;


change


and


types


outcome


instrument


changes


, 37)


initial


experience


significantly


affected


-year


-olds'


generalizations


Initial


experience


had


no significant


effect


on 3


-year-olds


' generalizations


agent


or manner


, 37)


'(2


Or manner











Of particular


interest


is precisely


how


initial


experience


affected


-year


-olds


generalizations


within


agent,


instrument,


manner


, and


outcome


change


types


Analysis

showed t


Compared


condition


of cell


hat


means


initial


to the


were


same


within


experience


the

had


condition,


significantly


instrument


change


predicted


-year


more


-olds


likely


type


effects


different


to extend


novel


verb


.0001


No further


significant


differences


between


cell


means


were


observed


instrument


changes


Contrasts


of cell


means


within


outcome


change


type


also


showed


initial


experience


had


predicted


effect


Here,


-year


-olds


generalized


significantly


more


often


diff


erent


condition


than


same


constant


plus


noise


condition:


e 1,


= 11.63,


.002;


.01.


Rather


conspicuously


, contrasts


of cell

initial


means


within


experience


manner


did


have


change


type


predicted


showed


that


effects


year


-olds.


Nor


were


any


significant


differences


observed


contrasts


cell


means


within


agent


change


type


sum,


year


initial


-olds


experience


generalizations


had

only


predicted

selected


effects

change t


on 3-


ypes


Instrument


and


outcome.


Visual


inspection


of Tabl


suggests


that


, 25)


J











analysis

analyses


effects


were u

showed


on 10


ndertaken


that


-year


to confirm


initial


-olds'


this


experience


or adults


observation.


had


These


no significant


extensions


novel


verbs


across


marginal


means


level


can


change


interpreted


type.


Consequently


unambiguous


-year-


olds


and


adults


because


effects


initial


experience


significantly


vary


across


four


different


change


types.


A priori


contrasts


initial


experience


marginal


means


-year


-olds


showed


that


children


were


significantly


more


likely


compared


Adul-ts


to extend


to the


also


same


generalized


novel


condition,


verb


different


F(1,


significantly


.005


more


often


different


compared


to the


same


condition,


F(1,


= 27.09


.0001


No further


significant


differences


between


levels


of initial


experience


were


observed


-year


-olds


or adults


A priori


comparisons


of change


type


marginal


means


revealed


mostly


similariti


between


generalizations


made


-year


-olds


and


adults


Both


10-year


-olds


and


adults


generalized


significantly


less


often


outcome


than


agent;


, 57)


= 43


respectively;


< .0001


or instrument


and


, 57)


changes;


= 26


, 57)


.0001


11.63











= 15.12,


.0001


and


, 57)


= 22.93,


.0001.


Where


-year


-olds


and


adults


did


very


subtly


differ


was


how


likely


they


were


to generalize


for


manner


compared


instrument


and


outcome


changes


Ten


-year


-olds


made


significantly


fewer


outcome


than


manner


generalizations


.12,


.003


Adults


made


fewer


outcome


than


manner


generalizations,


but


difference


was


significant


manner


compared


to instrument


changes


adults


generalized


significantly


less


often


to the


former


.005


I7


Ten-year


-olds


also


made


fewer


manner


than


instrument


generalizations,


but


difference


was


significant.


These


as well


as all


previous


planned


comparisons


were


made


using


Bonferonni


adjustment


keep


familywise


0.05.


Thus


analysis


shows


that


generalizations


year


-olds


manner


changes


diff


ered


from


those


-year


olds


and


adul t


across


initial


training


conditions


The


specific


sources


significant


three


-way


interaction


were


as follows


For


manner


changes


-year


-olds'


general


zations


did


significantly


differ


between


same


and


different


conditions


whereas


-year


-olds


' and


adults'


did.


Additionally,


preceding


analysis


and


an examination











*type


initial


experience.


On the


other


hand


for


year


-olds


and


adults


comparative


importance


change


types


did


vary


within


three


levels


initial


experience.


This


apparent


change


type


age


interaction


can


more


closely


examined


reanalyz


significant


three-


way


interaction


a slightly


different


way


two-way


interaction


involving


change


type


and


age


analyzed


within


each


level


initial


experience


Change


type


and


did


significantly


interact


within


constant


plus


noise


condition.


Therefore,


main


effects


initial


experience


, change


type


, and


age


can


interpreted


unambiguously.


within


different


condition


, change


type


and


age


did


significantly


interact,


, 1535


Tests


for


simple


interactions


showed


that


within


different


condition,


generalizations


-year


-old


differed


significantly


from


those


-year


-olds,


.005;


and


adults


F(3,


A posteriori


contrasts


revealed


only


one


significant


comparison.


Three-


year


-olds


' general


significantly


more


often


agent


compared


manner


changes


different


condition,


.008.


Of greater


interest


was


that


unlike


, 96)











different


condition.


However


, the


difference


was


significant,


, 13)


.45,


.055


Within


same


condition,


change


type


and


age


significantly


interacted,


, 150)


.005


Tests


simple


general


effects


zations


showed


-year


that


-olds


within


differed


same


condition,


significantly


from


those


-year


-olds,


.01; and


I3


adults


.002


A posteriori


contrasts


revealed


significant


comparisons


Three


-year


-olds'


gen


eralized


significantly


more


often


agent


compared


to instrument


and


outcome


changes


same


condition


: F(1,


= 14.00


.003;


, 12)


14.00


.003.


Of greater


interest


was


that


unlike


adults


and


-year


-olds


-year


-olds


' manner


generalizations


exceeded


instrument


generalizations


same


condition.


More


concisely


-year


-olds


' outcome


generalizations


exceeded


their


manner


generalizations


diff


erent


condition.


same


condition,


-year


-olds


' manner


generalizations


exceeded


their


outcome


general zations


Similarly


, in


same


condition,


3-year


-olds


made


fewer


instrument


than


manner


generalizations


different


condition,


-year


-olds


instrument


generalizations


exceeded


their


manner


generalizations


This


pattern


, 93)











change


within


levels


of initial


experience.


These


results


show


that


-year


-olds


' biased


interpretation


can


changed


type


initial


verb


learning


experience


they


receive.


On the


other


hand


, adults'


and


older


children


biased


interpretations


action


verb


meaning


are


relatively


unaffected


type


initial


verb


learning


experience


they


receive


Discussion


In accord


with


Experiment


, individual


event


components


again


had


predicted


effects


on subjects


generalizations.


Overall,


outcome


changes


had


greatest


effect.


Manner


changes


had


next


greatest


effect,


followed


instrument,


and


agent


changes


Given


this


pattern


of generalizations,


it would-


appear


that


outcome


component


initial


figured


assumptions


most


prominently


about


children


meaning


novel


s and


action


adults


verbs


Previous


research


shown


that


some


elements


meaning


are


more


central


than


others


to the


initial


representation


verbs


similar


to those


used


current


study


(Behrend,


1990;


Forbes


& Farrar


in press


However,


unlike


previous


research,


initial


experience


variable


of Exp


eriment


2 provides


unique


data


relative


strength


children


s and


adults


biased











different


condition


was


specifically


designed


to attenuate


same


biases.


The


constant


plus


noise


condition


was


designed


to have


an intermediate


effect


on subjects'


interpretations.


Overall


, these


three


conditions


produced


hypothesized


effects


In particular


, when


initial


training


consisted


viewing


three


nearly


identical


events


same


condition),


subjects


were


least


likely


extend


novel


verb.


Outcome


and


manner


changes


tended


figure


more


prominently


10-year


-olds


' and


adults


' initial


interpretation


verb


meaning.


outcome


and


instrument


changes


tended


to figure


more


prominently


in 3


-year


-olds'


initial

training


interpretation


consisted


verb


of viewing


meaning


three


When


events


the initial

which a single


event


component


varied


different


condition)


subjects


were


most


likely


to extend


novel


verb


Outcome


and


manner


changes


again


tended


to be


more


central


to 10


year


-olds


' and


adult


' interpretation


verb


meaning


For


-year


-olds


, however


, manner


changes


became


most


central


component


novel


their


verbs


initial


When


assumptions


initial


about


training


meaning


consisted


viewing


events


which


only


one


component


was


held


constant


while


others


varied


constant


plus


noise),











These


data


show


that


children' s


and


adults'


preferences


specific


event


components


can


be enhanced


or diminished


simply


making


them


more


or less


prominent


within


events


from


which


verbs


are


initially


learned.


Even


more


strikingly,


data


show


that


unlike


adults


and


year-olds,


3-year-olds'


biased


interpretations


novel


action


verbs


can


actually


be changed


different


types


initial


verb


learning


experiences.


Apparently


, verb


learners


are


much


less


procrustean


than


a restrictive


constraints


perspective


implies.


Verb


learners


of all


ages,


and


especially


3-year-olds,


just


blindly


impose


same


static


of one-size-fits


preferences


and


interpretations


on action


events


types.


Employing


such


a strategy,


one


could


never


hope


to acquire


changing


senses


differences


among


same


closely


verb,


related


alone


verbs.


subtle


as Nelson


(1992


puts


A person's
information


particular


word


can


use,


interpret


came


to the


lexicon
about


word


may
the


forms


be perfectly
and the child


what


expectations


contain


much


complexities


but


no given


predicted


must


is meant


language


about


with


word


relevant


of possible
instance o


outside


somehow


that


highly


meaning


use


context


uses


the
context


situation


A child


constrained


would


who


at a loss.


The


most


interesting


finding


from


Experiment


was


that











three


levels


initial


training


a way


that


did


not


happen


10-year- olds


and


adults.


One


interpretation


this


three-way


interaction


that


3-year-olds


learn


action


verbs


differently


than


adults


or even


10-year-olds.


When


same


adults


condition)


and

which


10-year-olds

elements of


are


acti4


"asked"

on verb


meaning


the

are


more


central


their


initial


interpretations


, they


respond


generalizing


least


often


to outcome


and


manner


changes.


When


informed


different


condition)


that


element


verb


meaning


can


vary


without


changing


meaning


verb,


they


generalize


more


often


change


types-


-but


relative


importance


of each


verb


component


does


vary


their


initial


interpretations.


When


youngest


subjects


are


"asked"


which


elements


of action


verb


meaning


are


more


central


their


initial


interpretations


, they


respond


general


zing


least


often


outcome


and


instrument


changes.


When


informed


that


element


of verb


meaning


can


vary


without


changing


meaning


verb,


they


generalize


accordingly--except


manner


change


In other


words


, the


outcome


and


instrument


bias


-year-


olds


showed


same


condition


completely


disappeared


different


condition.


Here,


manner


changes


had


greatest


effect


on 3-year-olds


s initial











year-olds,

assumptions


initial

about


experience


had


importance


no effect


manner


on their


in the


initial


meaning


novel


verbs.


for


outcome


and


instrument


components,


initial


experience


significantly


affected


year-olds


initial


assumptions.


The


evidence


argues


against


3-year-olds


merely


making


idiosyncratic


interpretations


of novel


action


verbs.


Consider


, agent


changes


example.


For


subjects,


same


and


different


initial


experience


conditions


had


effect


on their


initial


assumptions


about


importance


agent


olds'


meaning


and


3-year-olds'


the

made


novel


verbs.


virtually


Adults'


same


, 10-year-


number


generalizations


both


conditions.


According


to the


principle


precisely


what


word


learners


are


predicted


to do:


Extend


novel


action


words


to events


spite


changes


agent


(Golinkoff,


Jacquet,


& Hirsh-


Pasek


, 1992)


instrument


and


outcome


changes


too,


initial


experience


similarly


affected


generalizations


subjects.


Adults,


10-year-olds


, and


-year-olds


made


fewest


generalizat ions


same


condition,


most


different


condition,


and


an intermediate


number


constant


plus


noise


condition.


Only


for


manner


changes,


along


with


instrument


and


outcome


changes


different


condition















CHAPTER


GENERAL


DISCUSSION


Together


, results


from


both


experiments


provide


empirical


data


on the


process


of developing


action


verb


concepts.


Specifically


, the


studies, support


two


conclusions.


The


first


that


3-year-olds'


initial


action


concepts


differ


a principled


way


from


those


of adults


even


10-year-olds.


Support


year-olds'


first


manner


conclusion


generalizations


found


across


mainly


levels


initial


experience


and


their


outcome


and


instrument


generalizations


different


condition.


For


-year-


olds,


three


initial


experience


conditions


had


same


effect


on their


extensions


novel


verbs


manner


changes.


Moreover,


relative


importance


-year-olds


accorded

changes


outcome


was


and


complete


instrument

v reversed


changes


compared

different


manner


condition


from


what


had


been


same


condition.


On the


other


hand


, for


adults


and


-year


-olds


, the


same


and


different


conditions


consistently


had


predicted


effects


on their


extensions


novel


verbs


for


four


change


types.


A t-


1


,I ,











condition

central


They


to action


also

verb


considered

meaning t


outcome


han


and


instrument


manner

t and


more

agent


Most


notably,


individual


unlike


change


types


year

did


-olds

not


change


relative

within


importance


levels


initial


experience


with


either


group


older


subjects


The


data


show


that


youngest


children


are


simply


ess


likely


than


adults


to extend


novel


verbs


to events


which


a single


element


of meaning


been


changed.


If 3


year


-olds


were


just


more


conservative


than


adults,


one


would


expect


that

four


absolute


change


types


number

would


generalizations


be systematically


lowest


each

for


-year


-olds


compared


this


to adults,


did


-year


always


-olds


did


occur.


extend


For


example


novel


verb


ess


often


outcome


changes


same


condition


and


respectively


However


-year


-olds


also


made


more


manner


generalizations


= 0.36


same


condition


than


either


-year


-olds


.20)


or adults


.10)


The


same


condition


initial


experience


was


specifically


designed


to reduce


likelihood


general


zations


. So if


-year


-olds


were


plainly


more


conservative


that


adults


, the


evidence


would


have


been


found


same


condition


each


four


change


types


better


explanation


data


that


-year


-olds


weighted











Why


about


were


youngest


importance


children's


manner


changes


initial


nearly


assumptions


unaffected


different


types


initial


experience?


Perhaps


3-year-


olds


failed


to notice


manner


of action


as much


as they


noticed


other


three


event


components.


Consequently


their


interpretation


novel


verbs


would


have


favored


different


semantic


elements


than


adults'


or 10-year-olds'


Compared


other


three


components


tested


present


studies,


manner


component


was


more


dynamic


and


transitory.


Agent


and


instrument


components


were


available


perceptual


analysis


before,


during,


as well


as after


events


were


completed.


Moreover,


events


culminated


particular


available


ceased


outcome


brief


occur.


which


continued


inspection


This


to be


after


to suggest


perceptually


manner


that


of action


manner


component


somehow


more


complex


than


other


components.


Subjects


ages


probably


had


no difficulty


perceiving


manner


action.


But


even


when


shown,


different


condition,


that


manner


of motion


could'


change


without


affecting


meaning


novel


verb,


year-olds


made


same


number


generalizations


same


condition.


If a failure


to notice


manner


component


entirely











across


levels


initial


experience.


this


did


occur.


more


likely


explanation


why


importance

initial ex


manner


perience


that


-year

they


-olds

are


was


unaffected


moderately


biased


assume


that


manner


an important


element


of action


verb


meaning


Furthermore,


this


moderate


manner


bias


relatively


impervious


to the


inconsistencies


initial


action


verb


experiences.


The

appears


outcome


to be


bias


at odds


observed


with


present


previous


research


studies

showing


manner


-over


-end-


state


bias


very


similar


verbs


(Bowerman,


1982


Gentner


, 1978;


Gropen,


Pinker


, Hollander


, & Goldberg,


1991)


Although


younger


children


this


outcome


bias


was


eliminated


different


types


initial


experience.


Unlike


present


experiments


or Behrend


1990)


, previous


research


showing


a manner


bias


used


familiar


, rather


than


completely


novel


verbs


this


reason


, even


youngest


children


with


3-year


kinds


-olds)


events


probably


that


had


verbs


some


such


previous


as stirrinQ


experience


and


fill


label


So rather


than


initial


assumptions


of novel


action


verbs


, previous


researchers


more


likely


found


intermediate


assumptions


an action


event


becomes


familiar


, the


manner


action


may


become


more


central











condition.


Three


repetitions


same


manner


of action


was


sufficient


to make


-year


-olds


and


especially


adults


more


reluctant


to accept


a diff


erent


manner


event


occur


When


subjects


were


shown


three


different


manners


which


event


could


occur


different


condition)


adults


and


-year


-olds


rightly


assumed


that


a particular


manner


action


was


relatively


unimportant


to the


verb


meaning


According


this


reasoning


, further


training


trials


year


same


-olds


and


different


generalizations


conditions


manner


changes


could


make


resemble


those


of adults


and


-year


-olds.


more


serious


criticism


of previous


research


showing


robust


manner


bias


concerns


line


drawings


Gropen


et al


1991


used


to depict


verbs


being


studied.


In effect


subj ects


were


asked


to recall


since


familiar


events


were


used)


or imagine


appropriate


manner


action


accomplish


event


outcome.


The


one


element


of semantic


meaning


that


using


line


drawings


can


represent


manner


action.


To best


repre


sent


manner


action


dynamic c


events


such


those


used


current


studi


are


best


A reasonable


strategy


resolving


discrepancy


between


previous


research


and


present


studies


would











Grope

than


et al.


static


(1991)


pictures


Using


dynamic


familiar


novel


events


may


events

well


rather

result


an end


state


rather


than


manner


bias


Whether


children


would


still


make


syntactic


errors


unclear


given


that


appropriate


known.


adult


Nonetheless


usage


novel


would


be quite


words


would


interesting


see


pref


erred


syntactic


constructions


children


generate


with


completely


novel


verbs.


second


conclusion


clearly


supported


current


studi


that


verb


earning


more


like


sides


equation


than


restrictive


constraints


perspective


implies


events


How


, like


children


novel


are


actions


equipped


, constitutes


to interpret


one


structured


side.


current


studies


indirectly


examined


children


s general


cognitive


skills


specific


such


faufincr


representation


well


as more


and


extension


generalized


action


concepts


action


verb


Clearly


these


and


other


skills


such


ability


acquire


symbols


and


integrate


them


into


interrelated


structures


are


essential


verb


learning


and


lexical


acquis


ition


general


children


s available


conceptualizations


and


cognitive


skills


but


one


side


lexical


acquisition


equation.


Very


little


known


about


how


children


use


different


sources











which


constitutes


another


side


equation.


This


expanded


view


lexical


acquisition


, encompassing


both


word


learner


and


word


learning


situation,


precisely


what


enabling


constraints


position


advocates


The


overall


point


that


verb


acquisition


inexplicable


solely


terms


what


children


already


know


about


verb


meaning.


Indeed


, results


from


present


studies


show


that


adults


and


older


children


whose


interpretations


unaffected


of novel


different


verb

types


meaning


learning


relatively

experiences


Young


children


s preferred


interpretations


verb


meaning


appear


to be


much


more


tractable


to experience


This


suggests

meaning


that

arise


children's


more


biased


as a consequence


interpretations


of experience


verb

than


innately


acquisition


examined


specified


, language


Information


present


fro


specific

m sources


studies


constraints


such


provide


on lexical


those


children


with


clues


about


conventions


verb


use















APPENDIX


Outline


of all


training


and


generalization


events


Blating


: Place


racquet


ball


on top


a table


Person


whacks


table


with


fists


to make


ball


bounce


and


down


and


hits.


ball


with


articles


of folded


clothing


to make


roll


across


table


and


bounce


a wall.


Manner


change


: (a)


Person


kneel


and


pushes


from


underneath


table;


Person


tilts


table


to and


fro;


Person


grabs


back


table


, lifts


and


drop.


Instrument


change


Person


whacks


table


with


monitor


hanger;


Person


whacks


tabl


with


heavy


stick;


Person


whacks


tabi


with


a book


Outcome


change


Object


remains


stuck


piece


poster


table;


tape)


place


Object


on table


attached


Object


to a six


inch


fall


tether


side


so that


does


move


very


far.


Agent


change:


Three


different


people


perform


original


event


Noffing:


a tall


, thin


box


on-a table


Person


inserts


a fire


extinguisher


, drags


box


backwards


along


table


, removes


fire


extinguisher


, rights


box,


and


then











Manner


change


Rolls


box;


Picks


box


and


carries


Upright


box


pushed


and


made


to revolve


continuously.


Instrument


change


: (a)


Camera


tripod;


holder;


Stick


inserted


through


box


to which


a bicycle


tire


liner


attached


to form


a yolk.


Outcome


change


Fire


extinguisher


left


behind


abruptly


pulling


box


from


underneath;


Fire


extinguisher


placed


chair


near


side


table;


Fire


extinguisher


remains


covered


box.


Agent


change:


Three


different


people


perform


original


event.


Prebing:


Press


, using


a the


edge


of a lamp


base


to cut


shaped


piece


of playdough


into


quarters.


Manner


change:


Use


a chopping


motion;


Use


sawing


motion;


Use


a twi


sting


and


pulling


motion.


Instrument


change


Bicycle


brake


cable;


Clear


, round


baking


dish;


ece


cardboard


Outcome


change:


one


quarter


and


then


quarter


into


thirds;


Flatten


playdough;


Pile


quarters


on top


of each


other.


Agent


change


: Three


different


people


perform


original


event.











Manner


change


: (a)


Tip


box


over


onto


toy


then


remove


Lift


box


over


and


pull


it out


from


underneath


sugar;


Lift


box


over


toy


and


pour.


Instrument


change


Paper


sack;


Piece


construction


paper


form


a snow


cone


with


a hole


bottom;


Glass


juice


container.


Outcome


change:


Very


small


amount


salt


comes


box;


Sugar


completely


misses


toy;


Sugar


comes


but


there


is no toy


cover.


Agent


: Three


different


people


perform


original


event.


Faufing


: With


a bicycle


fixed


remover


kind


wrench),


pick


small


filled


plastic


bottles


and


place


them


(without


using


attached


ramp)


on top


a small


stool.


Manner


change:


Underhand


toss


bottles


into


place ;


Drag


bottles


into


place


a ramp;


Launch


bottles


ramp


into


place.


Instrument


change


Use


hedge


clippers;


Bicycle


brake


cable;


Large


Outcome


change:


Place


bottles


upright;


Place

Drop


bottles

bottle i


half


nto


on stool


sandbox


and


under


allow


and


to fall

front o


to ground;


f stool.


Agent


change


:Three


different


people


perform











-Tulking:


Loop


two


bicycle


inner


tubes


around


cinder


blocks


and


drag


them


backward


along


a brick


fence


(toward


person).


Manner


change:


Pick


and


carry


(walking


backward)


cinder


blocks


Pick


blocks


, pivot


like


crane


and


pose


blocks


on fence;


Fling


cinder


block


several


times


(without


letting


inner


tubes)


Instrument


change


Use


a wooden


frame;


a crow


bar;


Throw


a sack


over


brick


and


drag


Outcome


fence;


change


block


: (a)


Both


drops


blocks


side


fall


fence;


end


Blocks


stacked


end


on end.


Agent


change


: Three


different


people


perform


original


event.


Smuking:


Stick


three


soda


cans


horizontal


line


large


piece


of gummed


cardboard


affixed


to wall.


Manner


change


Bring


cardboard


down


cans


Stick


three


cans


once;


Bring


cardboard


halfway


can


and


can


halfway


to cardboard.


Instrument


change


Monkey


puppet;


Magazine


and


stick;


Screwdriver


and


stick


Outcome


change:


Stick


cans


and


replace


on floor


one


at a time;


Place


can


on cardboard,


but


cans


fall;















REFERENCES


, A. K., &
use: Social
categorical
Barrett (Ed
Progress in
197). New


Bullock, D. (1986). Apprenticeship
convergence processes in learning
ly related nouns. In S. A. Kuczaj &
s.), The development of word meaning
cognitive development research (pp.
York: Springer-Verlag.


in word

M. D.

155-


Barrett, M. (1985). Early semantic
early word usage. In S. Kuczaj
The development of word meaning
Verlag.


representations and
,& M. Barrett (Eds.),
. New York: Springer-


Behrend, D. A. (1990). The development of verb concepts:
Children's use of verbs to label familiar and novel
events. Child Development, 61, 681-696.


Behrend, D. A.,
syntactic
mapping of
meeting of
Development


& Harris,
bootstrapp
verb mean
the Socie
t, Seattle


L. L. (1991, April). In
ing: Verb endings and the
ings. Paper presented at
ty for Research in Child
, WA.


search
initial
the


Bowerman, M. (198
and syntacti
Gleitman (Ed
art (pp. 320
Press.


2). Reorganizational processes in lexical
c development. In E. Wanner and L. R.
s.), Language acquisition: The state of the
-346). New York: Cambridge University


Bowerman, M. (1985). What shapes children's grammars?
D. I. Slobin (Ed.), The crosslinguistic study of
language acquisition (Vol 2, pp. 1257-1319).
Hillsdale, NJ: Erlbaum.


Brain, M. (1976)
Monographs c
Development,


Children' s
the Society
41 (1).


first
for


word combinations.
Research in Child


Bruner, J. (1990). Acts


Meaning.


Cambridge:


Harvard


Adams











Carey, S. & Bartlett, E. (1978).
word. Papers and reports on
(Department of Linguistics,
17-29.


Clark, E. (1987)
on language
Mechanisms
Hillsdale,


Acquiring a single new
child language development
Stanford University), 15,


The principle of contrast: A constraint
acquisition. In B. MacWhinney (Ed.),
language acquisition (pp. 1-33).
: Erlbaum.


Farrar, M. J., Friend, M. J.,
Event knowledge and early
Journal of Child Language


& Forbes,
language


N. (in press).
quisition.


Forbes, J. N.,
initial as
verbs: Bia


:Farrar, M. J. (in press). Children's
umptions about the meaning of novel motion
ed and conservative? Cognitive Development.


Gallistel, C
Keil, F
study o
Gelman
biology
Erlbaum


R.
(1
co
Eds
and


, Brown,
991). L
gnitive
.), The
cogniti


A. L.,
lessons
develop
epigene
on (pp.


Carey, S.,
from animal
ment. In S.
sis of mind:
3-36). Hil


Gelman,
learning
Carey
Essays
lsdale,


Gentner
of


, D.
verb


1978) .
meaning.


On relational meaning: The acquisition
Child Development, 49, 988-998.


Golinkoff, R. M., Jacquet, R. C.,
Lexical principles underlie
Unpublished manuscript.


Hirsh-Pasek,
learning of


K. (1992).
verbs.


Golinkoff, R. M.,
Early object
principles.


Mervis, C.
principles
Manuscript


B., & Hirsh-Pasek, K. (1992).
: The case for lexical
submitted for publication.


Gopnik, A., & Meltzoff, A.
semantic and cognitive
stage: The specificity
57, 1040-1053.


(1986). Relations between
development in the one-word
hypothesis. Child Development,


Gropen, J., Pinker,
(1991). Syntax
locativee verbs.
151.


S., Hollander, M.,
and semantics in
Journal of Child


& Goldberg, R.
the acquisition
Language, 18,


of
115-











Landry,
and
and


S., & Loveland
developmental
Developmental


, K. (1986). Joint attention
language delay. Journal of
Disorders, 16, 335-349.


in autism
Autism


Mandler, J. M.
Conceptual


in press).
primitives.


How to build a
Psvcholocrical


baby: I
Review.


Markman, E. M.
mutual exc
on word me
(Eds.), Pe
Interrelat
University


(1991). The whole obj
lusivity assumptions a
anings. In J. P. Byrn
rsDectives on language
ions in development.
Press.


ect, taxonomic, and
s initial constraints
es and S. A. Gelman
and cognition:
Cambridge: Cambridge


Markman, E. M., &
sensitivity
vs. thematic
27.


Hutchinson, J. E. (1984). Children's
to constraints on word meaning: Taxonomic
relations. Cognitive Psychology, 16, 1-


Merriman, W. E., & Bowman, L. L. (1989). The mutual
exclusivity bias in children's word learning.
Monographs of the Society for Research in Child
Development, 54, Serial No. 220.


Mervis, C. (1987). Child-basic
lexical development. In U
conceptual development (pp
Cambridge University Press


object categories and early
. Neisser (Ed.), Concepts and
. 201-233). New York:


Nagy,


W., &
lexical
5, 169-


Gentner, D.
categories
201.


(1990). Semantic constraints on
Language and Cognitive Processes,


Nelson,


(1985)


Making sense.


New


York:


Academic


Press.


Nelson, K. (1988
DeveloDment


gi


Constraints
221-246.


on word


learning?


Cognitive


Nelson, K. (1992, April). Contexted
acquisition of shared meaning.
British Child Language Seminar,


relevance
Paper pre
Glasgow,


and the
ented at
cotland.


Nelson, K.
acqui
107.


E. (1977). Facilitating children
sition. Developmental Psychology,


S s
13,


yntax
(2),


101-











Rispoli, M. (1991). The
categories. Journal


mosaic acquisition
of Child Lanouace


of grammatical
, 18, 517-552.


Rogoff, B.
Oxford


1990). Apprenticeship
University Press.


thinking.


Oxford:


Schaffer, E
world.


R. (1984). The
London: Academic


child' s
Press.


entry


into


a social


Slobin, D. I. (1985).
acquisition (Vol.


The
2).


crosslincruistic
Hillsdale, NJ:


study of
Erlbaum.


lanquare


Smith, C. A., & Sachs,
lexicon in early
Psycholinguistics


J. (1990). Cognition and the
lexical development. Applied
, 11, 409-424.


verb


Tomasello, M. (1992a)
University Press


First


Verbs.


Cambridge:


Cambridge


Tomasello, M. (1992b). The 'social bases
*acquisition. Social Development, 1,


language
-87.


Tomasello,
early


M., & Farrar, M.
language. Child


J. (1986)
Development,


Joint
57,


attention
1454-1463.


and


Tomasello, M
actions
context,


, & Kruger,
Acquiring
Journal


A. C. (1992). Joint
verbs in ostensive and
of Child Language, 19,


attention on
non-ostensive
313-333.


Tomasello, M., Mannle, S.
effect of previously
acquisition of words
Child Language, 15,


, & Werdenschlag, L. (
learned words on the
for similar referents
505-515.


1988). Th
child's
Journal


Trevarthen, C., & Hubley
intersubj activity.
and symbol: The eme
New York: Academic


. p. (
In A.
rgence
Press.


Tversky, B., & Hemenway, K.
categories. Journal c
General, 113, 169-193.


1978). Secondary
Lock (Ed.), Action, gesture.
of language (pp. 183-229).


1984). Objects, parts,
Experimental Psvcholoqv


and


Woodward, A. L., & Markman, E. M. (in press). Constraints
on learning as default assumptions: Comments on


re

of















BIOGRAPHICAL


SKETCH


James

Oregon. I


. Forbes


n May


was


1981


born


20 November


Mr. Forbes


received


1957


an A.A


Eugene,

. from


University


of Maryland,


European


Divi


sion,


Iraklion,


Crete


and


then


attended


University


of Caen


, France


until


1985


He graduate


ed from


University


Washington


magna


cum


laude


with


a B.A.


French


and


a B.S


.in


psychology


May


, 1988


The


University


of Florida


conferred


a Master


Science


degree


psychology


on Mr. Forbes


May


of 1992














I certify that
opinion it conforms
presentation and is
a dissertation for


I have read this study and that in my
to acceptable standards of scholarly
fully adequate, in scope and quality,
the degree of Doctor of Philosophy.


M. Jsr te
Associate


Farrar,
Professor


Chair
Psychology


I certify that
opinion it conforms
presentation and is
a dissertation for


I have read this study and that in my
to acceptable standards of scholarly
fully adequate, in scope and quality,
the degree of Doctor of Philosophy.


Howard W. Beck
Assistant Professor
Engineering


I certify that
opinion it conforms
presentation and is
a dissertation for


Agricultural


I have read this study and that in my
to acceptable standards of scholarly
fully adequate, in scope and quality,
the degree of Doctor of Filosophy.


Ira S. Fischler
Professor of Psychology


I certify that
opinion it conforms
presentation and is
a dissertation for


I have read this study and that in my
to acceptable standards of scholarly
fully adequate, in scope and quality,
the degree of Doctor of Philosophy.


Scott A. Miller
Professor of Psychology


3YIC-


L^ /3PY~
















This
of the Dep
Arts and S
accepted a
degree of


dissertati
artment of
sciences an
s partial
Doctor of


on was submitted t
Psychology in the
d to the Graduate
fulfillment of the
Philosophy.


o the Graduate Faculty
College of Liberal
School and was
requirements for a


August,


1993


Dean,


Graduate


School
















































UNIVERSITY OF FLORIDA
3 121112 08556 70881
3 1262 08556 7088