Flexibility in Early Verb Use

Flexibility in Early Verb Use: Evidence from a Multiple-N Diary Study
Author(s): Letitia R. Naigles, Erika Hoff, Donna Vear, Michael Tomasello, Silke Brandt, Sandra R.
Waxman, Jane B. Childers and W. Andrew Collins
Source: Monographs of the Society for Research in Child Development, Vol. 74, No. 2,
Flexibility in Early Verb Use: Evidence from a Multiple-N Diary Study (2009), pp. i-v, vii, 1-144
Published by: Wiley on behalf of the Society for Research in Child Development
Stable URL: http://www.jstor.org/stable/25580860
Accessed: 10-06-2015 02:47 UTC
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at http://www.jstor.org/page/
info/about/policies/terms.jsp
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content
in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship.
For more information about JSTOR, please contact support@jstor.org.
Wiley and Society for Research in Child Development are collaborating with JSTOR to digitize, preserve and extend access to
Monographs of the Society for Research in Child Development.
http://www.jstor.org
This content downloaded from 181.118.153.129 on Wed, 10 Jun 2015 02:47:42 UTC
All use subject to JSTOR Terms and Conditions
FLEXIBILITYIN EARLYVERB USE: EVIDENCE

FROMA MULTIPLE-N DIARY STUDY
Letitia R. Naigles, Erika Hoff, and Donna Vear
WITH COMMENTARY BY
Michael Tomasello,
Silke Brandt,
Sandra
R. Waxman,
and Jane B. Childers
W. Andrew Collins
Series Editor
MONOGRAPHS
OF THESOCIETY
FORRESEARCH
INCHILD
DEVELOPMENT
Serial No. 293, Vol. 74, No. 2, 2009
(^PjWILEY
W
J PL/l%Jv
ELL
Boston,Massachusetts
Oxford,UnitedKingdom
EDITOR
W. ANDREW COLLINS
University
ofMinnesota
MANAGING EDITOR
DETRA
DAVIS
Society for Research
in Child Development
EDITORIALASSISTANT
ANTONELLA
CAIAZZA
Board of Advisory
Editors
Brian K. Barber
University
Michael
University
Glenn
University
Kathleen
University
of Tennessee
P.Maratsos
ofMinnesota
I. Roisman
of Illinois
Thomas
ofMinnesota
Manfred Van Dulmen

Kent
State University
Thomas Weisner
University
of California,
Los Angeles
PhilipDavid Zelazo
University
of Toronto
EDITORIALCONSULTANTS
Emma
Jane Childers
K. Adam
Northwestern
University
Trinity University
Mark Appelbaum
University
of California,
San Diego
Gregory Cook
University
of Wisconsin-Whitewater
University
of Wisconsin-Whitewater
RichardAslin
of Rochester
University
Marian
Bakermans-Kranenburg
Leiden University
John Bates
Indiana
University
Nazli Baydar
Koc University
Theodore
University
Beauchaine
of Washington
Susan
Utrecht
Branje
University
Celia Brownell
University
of Pittsburgh
Melissa M. Burch
Hampshire
College
Susan Campbell
University
Stephanie
University
of Pittsburgh
Carlson
ofMinnesota
Joan Cook
Susan
Crockenberg
of Vermont
University
Pamela
University
Davis-Kean
ofMichigan
Eccles
Jacquelynne
University
ofMichigan
James Elicker
Purdue
University
Michelle M. Englund
University
ofMinnesota
Kurt Fischer
Harvard
Doran
University
French
Illinois Wesleyan
Sarah
University
Friedman
CNA Corporation
Douglas Frye
University
of Pennsylvania
Andrew Fuligni
University
of California,
Los Angeles
Darcia
of Calgary
Nelson
City University
of New
University
Megan
ofMinnesota
University
Susan
University
Robert Pianta
University
University
Mark
Lene Jensen
Ariel Kalil
University
of Chicago
Melissa Koenig
University
ofMinnesota
University
Karl Rosengren
of Illinois, Urbana-Champaign
JudithG. Smetana
Atlantic
University
Eva Lefkowitz
Pennsylvania
Katherine
University
State University
Magnuson
Madison
of Wisconsin,
Ann Masten
University
ofMinnesota
KevinMiller
University
Ginger
Pennsylvania
ofMichigan
Moore
State University
of Rochester
University
Kathy Stansbury
Morehouse
Steve
Brett Laursen
Florida
of Virginia
Roosa
State University
Arizona
of Texas, Austin
Clark University
Santa Barbara
of California,
Aletha Huston
University
of Rochester
YukariOkamoto
Hespos
Vanderbilt
Davis
O'Connor
University
Paul Harris
Harvard
of California,
Thomas
Gunnar
University
York
LisaOakes
ElenaGrigorenko
Yale University
Dame
Katherine
Susan Graham
University
Narvaez
of Notre
University
University
College
Thoma
of Alabama
Michael Tomasello
Max
Planck
Institute
Deborah Vandell
University
of California,
Irvine
RichardWeinberg
University
Hirokazu
New
ofMinnesota
Yoshikawa
York University
Qing Zhou
Arizona
State University
David Moshman
University
of Nebraska
FLEXIBILITYIN EARLYVERB USE:

EVIDENCEFROMA MULTIPLE-N
DIARY STUDY
CONTENTS
ABSTRACT
vii
I. INTRODUCTION
II.PRESENTING
THEDIARY
METHOD
22
A GENERALDESCRIPTION
OF EARLY
III.
VERBGROWTHAND USE
32
IV.PRAGMATIC
AND SEMANTICFLEXIBILITY
INEARLY
VERBUSE
40
VERBUSE
AND GRAMMATICAL
FLEXIBILITY
INEARLY
V PRODUCTIVITY
49
VERBGROWTHAND USEAS A FUNCTION

VI. DIFFERENCESINEARLY
OF
DEVELOPMENTAL
PERIOD,CHILD,AND VERB
68
VII.GENERALDISCUSSION
91
REFERENCES 105
ACKNOWLEDGMENTS 112
COMMENTARY
FLEXIBILITY
INTHESEMANTICS
AND SYNTAX
OF CHILDREN'SEARLY
VERBUSE
Michael Tomasello and Silke Brandt
113
LEARNINGFROM INFANTS'FIRST
VERBS
Sandra
R. Waxman
VERBLEARNERS:
EARLY
CREATIVE
OR NOT?
Jane B. Childers
127
133
CONTRIBUTORS 140
STATEMENT
OF EDITORIAL
POLICY 142
ABSTRACT
are hallmarks
of human
and productivity
language use.
Flexibility
use
to serve a
to
have
the
the
words
know
speakers
Competent
capacity
they
to
new
to
of
refer
communicative
and
varied
functions,
variety
exemplars of
to which words refer, and in new and varied combinations
the categories
with other words. When
and how children achieve
this flexibility?and
when
they
are
truly
productive
language
users?are
central
issues
among
accounts
of language acquisition. The current study tests competing

hy
of
of flexibility
and some kinds of productivity
the achievement
potheses
verbs.
against data on children's first uses of their first-acquired
uses
mothers
recorded
their
children's
first
10
of 34 early
Eight
were
if
those
verbs
of the
within
the
window
verbs,
acquired
produced
were
The
16
children
between
and
months
when
the
20
study.
study began
on when the children started to
(depending
produce verbs), were followed
3 and 12 months,
for between
and produced
between
13 and 31 of the
a
verbs.
These
records
the
basis
for
of the
target
diary
provided
description
pragmatic,
semantic,
and
syntactic
properties
of
early
verb
use.
The data revealed that within this early, initial period of verb use, chil
dren use their verbs both to command and describe,
they use their verbs in
reference
to a
variety
of
appropriate
actions
enacted
by
variety
of actors
and
a variety of affected objects, and they use their verbs in a

variety of
semantic and grammatical
syntactic structures. All 8 children displayed
flexibility before 24 months of age. These findings are more consistent with
a model
of the language-learning
child as an avid generalizer
than as a
user.
conservative
Children's
language
early verb use suggests abilities and
to abstract from experience
inclinations
that may indeed begin in infancy.
with
vii
I. INTRODUCTION
list of words and

in the adult is not a memorized
knowledge
Language
sentences paired with the situations in which they have been experienced.
in their vocabularies
novel set
Adult speakers use the words
flexibly?in
to newly encountered
that
instances of the categories
tings and in reference
to form grammatical
can combine
their words
their words
label. They
sentences
in the study of
that they have never heard. A central question
context
is
the
when
children
achieve
independence,
language acquisition
of form that are the hallmarks of
and productivity
of meaning,
extendability
human
competence.
linguistic
One
to new
is that children
possibility
and
referents
produce
novel
are avid generalizers

sentences
very
who
early
in
extend
the
words
course
of
For example, a child who hears the verb jump in the

language development.
sentence
t jump on the bed would without
Don
further linguistic
imperative
to
use
to
in reference
in
declarative
also
be
able
sentences,
experience
jump
to actions as different
jumping done by other people, and even in reference
or jumping
into the
from the original as jumping up to reach something
use
to
the child would be able
jump in other structures,
pool. Furthermore,
such as I jump, he jumped, and Granny jumped into the pool. An alternative
is that children are initially conservative
learners who
possibility
language
to
close
A
the examples of language use they have already experienced.
stay
child who has heard jump only in Don tjump on the bed would not use jump to
refer to a variety of other sorts of jumping, by a variety of different jumpers,
or
in a range
of
syntactic
constructions.
These
two
positions
define
space
of the child as language learner.

that includes multiple possible descriptions
we present new data on children's first uses of their first
In this monograph,
this hypothesis
verbs and ask what theoretical positions within
space are
consistent with that evidence.
The
view that children
in their very
generalize beyond their experience
that children's
first uses of newly
productions
yields the prediction
a range of meanings
verbs should be flexible,
and
acquired
conveying
a
structures.
in
of
The
view that children are initially
appearing
variety
earliest
1
conservative
that children's first uses
language learners yields the prediction
of newly acquired verbs should be very repetitious
in terms of the events
to
structures
refer
and
the
in
because
they
they participate
they are used
to
contexts
structures
in
in
similar
those
which
and/or
only
they were first
encountered.
Verbs
development
are the focus of this investigation of the process of early language

for two reasons. First, verb acquisition has been less studied
noun
than
the
Thus,
acquisition.
and
rapid-generalizer
conservative
views of lexical development

have been tested more
frequently
on
nouns
the
of
than
data
and
verbs,
against
acquisition
questions about the
noun
are
to
which
and
verb
similar
processes have
learning
learning
degree
& Boroditsky,
&
been raised (e.g., Gentner
Hirsh-Pasek,
2002; Maguire,
learner
2006; Sandofer
Golinkoff,
use
tence
structure
Thus,
alizer
ment
the
is also
studying
and
8c Smith,
of
development
as
Second,
the development
grammar,
of verb
sen
because
accounts,
by many
in relation to the verb (Bloom, 1993).

is essentially defined
the acquisition of verbs provides a test of the rapid-gener
views
conservative-learner
as well
2000).
lexical
with
to
respect
grammatical
develop
development.
first 10 uses of
Eight mothers
kept detailed records of their children's
first verb
their first verbs over a period ranging from 3 to 13months. Those
uses were coded for the pragmatic,
semantic, and grammatical
flexibility
accounts of the child as language
learner
and competing
they evidenced,
were
evaluated
against
those
In
data.
the
remainder
of
this
introductory
to this study
chapter, we present the theoretical and empirical background
the extendability
of
in some detail, first considering
work that addresses
verb meaning,
the flexibility of verb
second considering work that addresses
the theoretical and empirical background
grammar, and third, considering
relations among verb vocabulary growth,
that yields predictions
regarding
and flexibility of verb grammar.
flexibility of verb meaning,
MEANING
LEARNING
VERB
Theories of Early Word Meanings
a new word used in a single utterance and a
first encounters
or
must
somehow figure out both what the word
she
single situation, and he
to
in that situation and what it can refer to in other situations. Current
refers
of how the child makes
accounts of word learning differ in their descriptions
The
child
that word-to-referent
not
to a
just
One
linguistic
single
mapping
referent
and
but
is that the child enters

that guide
understandings
view
in how,
to an
entire
and
when,
the
child
maps
words
category.
task equipped with

the word-learning
and extension
the mapping
processes.
2
INTRODUCTION
Children
for
understand,
that
example,
words
and
refer,
this
understand
words (Golinkoff,
ing guides them to seek a referent for newly encountered
Mervis, Frawley, 8c Parillo, 1995; McShane,
Hirsh-Pasek,
1980). They also
that words refer to kinds of things or events, which allows them
understand
to extend words to new instances of the same kind (Golinkoff et al., 1995;
& Hoff,
Naigles
They
to
competence
syntactic
use
the
word-learning,
the correspondences
of
terpretation
8c Forbes,
2006; Poulin-duBois
the
newly
Children
especially
between
encountered
2006).
For
task.
verb-learning,
to guide
and meaning
syntax
verbs.
also bring
example,
in
a novel
hearing
verb in a structure such as X verbed Y, yields, via syntactic bootstrapping,

the
a
or
that
the
verb
is
causative
verb
hit
rather
than
(like
kicked)
interpretation
a verb like sing or laugh (Gleitman,
8c Swensen,
1990; Naigles
2007).
to the
Other
theories
attribute
less initial linguistic
sophistication
and
child
limited
from
language
learning
predict
generalizations
experi
ence in children's
that
early word use. A child who does not understand
words
refer
may
associate
a new
to what
word
with
the
context
entire
of
its use
(e.g.,
to the refrigerator
door;
"open" applies only
A
child
who
understands
that
refer
words
Dromi,
1982; Mervis,
1987).
only
but not that they refer to kinds will construe meanings
very narrowly. On
some
accounts,
children,
at
one's mother
some
early
does
point,
tend
to
words
map
onto
what
context and to extend words to other things

ismost salient in the concurrent
or events with similar
perceptual
properties
(e.g., "open" may apply to
lateral motions
of doors but not of vertical motions
of doors,
lids, or con
et al., 2006;
8c Golinkoff,
tainers; Hollich,
Hirsh-Pasek,
2000; Maguire
that children eventually use
Smith, 1999, 2000). All theories acknowledge
and syntactic information
in their word learning,
reference,
extendability,
but some models hold that these biases and useable sources of information
are not all available when word
for word mapping
and extension
learning
et
et al., 2000; Maguire
Hollich
et al., 2006). In
al.,
1995;
(Golinkoff
begins
this model,
early verbs (those learned during the 2nd year of life) are likely
to be used in specific contexts or extended
limited ways based on perceptual
to
the
the 3rd year of life can
similarity
original exemplar. Only during
the child use social and linguistic sources of information
for mapping
and
extension
to this Coalition-Emergentist
of verbs. Crucial
is the
model
assertion
that
word-learning
principles
emerge
earlier
for
nouns
than
for
and may not transfer directly from noun learning to verb

learning
et al, 2006).
(Golinkoff et al., 1995; Maguire
These
different models
of word learning make different
predictions
to which children will extend
about the degree
their first verbs to new
contexts of use and to new actions. A model
to the verb
that attributes
verbs
to identify new verbs as verbs

learning child sufficient syntactic knowledge
and sufficient lexical understanding
to map new verbs onto
categories of
actions predicts
that children will use their new verbs in utterances
that
3
serve a range of pragmatic functions and will extend the verbs

they acquire
to actions by a variety of actors or agents. A model
that attributes
less lin
to
the
child
less
extension
of
word
guistic knowledge
1-year-old
predicts
use. The emergentist model
et
et
1995;
al.,
al,
(Golinkoff
2006)
Maguire
verb use followed by later flexible verb use. One
predicts initial conservative
uses of
is by looking at children's multiple
way of testing these predictions
their first verbs. Do children first restrict the use of a new verb to a specific
setting, or do they use the verb in a variety of settings?
Evidence
From Studies of Word
Learning
about the extendability

of newly acquired words have more
Questions
on
nouns
with
been
addressed
data
than verbs. Children
be
frequently
tween 12 and 18 months
do sometimes
underextend
nouns,
restricting
their use to a single context
1987; Harris, Barrett, Jones, 8c
(e.g., Dromi,
children's use of nouns in spontaneous
Brookes,
1988). After 18 months,
not
context
is
bound
(Barrett, Harris, 8c Chasin,
1991), and in word
speech
learning
novel
children
experiments
to more
nouns
one
than
older
exemplar,
than
even
18 months
after
only
routinely
one
extend
presentation
& Maguire,
2004; Markman,
1989).
(Hirsh-Pasek, Golinkoff, Hennon,
children might
restrict
If verb use also shows initial underextension,
newly
to
verbs
acquired
refer
to a
actor,
addressee,
single
action
type,
or
a child might use the verb blow only when

object. For example,
or
to blowing out a candle, but not to refer
to
in
Mom
reference
talking
only
to the dog's
to blowing one's nose. A child might know sit with reference
affected
sitting
but
no
one
else's,
or
eat
only
in reference
to
eating
pizza.
restricted by actor or affected

There
is evidence of early verb meaning
uses
of
verbs have been found to
Children's
earliest
conventional
object.
to their own rather than another's
8c
actions
refer primarily
(Childers
8c
8c
1983;
Tomasello,
2006; Huttenlocher,
Smiley
Charney,
Smiley,
Huttenlocher,
1995; but see Tomasello,
1992). Twenty-month-olds
taught
novel
verbs
via
video
were
unsuccessful
in
understanding
the
extension
of
these verbs to similar actions performed

by new actors (Maguire et al., 2006;
8c Forbes, 2002). Diary studies have revealed other kinds of
Poulin-Dubois
verb use, such as cut used solely when using a knife,
initial idiosyncratic
watch used solely when watching
television, push used solely in a game
to a pacifier
in the pool, and cry used solely in reference
of pushing
1995; Smith 8c Sachs, 1990; Tomasello,
1992).
(Braunwald,
the
verb meanings
Evidence of extendable
2nd year of life also
during
to use the same verb (rock and ride) to
exists. Children have been observed
of both objects and people over the 6-month period
refer to the movements
assess
from 13 to 19 months
1995). Laboratory
(Smiley 8c Huttenlocher,
found
ments of children's understanding
of familiar verbs have consistently
4
INTRODUCTION
of a wide range of English verbs

demonstrate
comprehension
actors
unfamiliar
Hirsh-Pasek,
(Golinkoff,
Cauley, 8cGordon,
performed
by
et al., 1983; Naigles,
1987; Huttenlocher
1997; Naigles & Hoff, 2006), al
of verbs in comprehen
though other studies have found that extendability
sion is limited to children who had reached
either a certain age or
level. Only children over 19 months
and/or with high vocab
vocabulary
that toddlers
familiar verbs to actions by unfamiliar actors (Forbes

ulary counts extended
8c Poulin-Dubois,
of a verb
1997; Naigles 8cHoff, 2006); further extensions
to both new agents and new manners
of action were found only at 26
months
of age (Forbes 8c Poulin-Dubois,
conditions
also
1997). Exposure
to matter.
appear
were
who
Twenty-one-month-olds
two
taught
tran
novel
sitive verbs in a playroom

and
setting with two different actors?themselves
the researchers?were
able to distinguish
those two verbs when presented
on video with a disembodied
hand as the only visible actor (Naigles, Bavin,
et al. (2005) may have been
8c Smith, 2005), and the children
in Naigles
more accepting of yet another actor at test than children
in other studies
because the verbs had been taught with at least two actors.
Questions
to Be Addressed
Regarding Children's Acquisition
of Verb Meanings
The previous literature suggests some early limits on the extendability

of
verb meanings
and some increases in extendability as a function of age or level
of language development
achieved. There are suggestions
in the data from
in
to
Forbes and Poulin-Dubois
(1997) that,
comprehension,
extendability
new
agents
outcomes.
questions
is an
The
earlier
achievement
than
literature
about
leaves
existing
the
extendability
1. How
in the
early
of
process
unanswered,
verbs
as
of
learning
early in the course of learning

emerge?
current
The
study
to new
extendability
they
manners
however,
enter
children's
in
verbs
and
several
basic
lexicons:
how
and/or
general
individual verbs, does extendability

was
to
designed
track
children's
extendability of verbs from the beginning of verb use in the child

and from its first use for individual verbs to address this question.
2. Are
ments
some kinds of extendability

than
others?
tients, affected
(e.g.,
eating
cereal
If children
objects,
and
earlier developmental
use
or different
eating
a carrot;
their
verbs
with
but still appropriate

opening
a door
achieve
different
and
pa
actions
opening
is not closely
jar) itwould indicate that the basis of their meanings
tied to perceptual
et
al., 1995; Goodman,
similarity (Golinkoff
8c Brown, 1998; Maguire et al., 2006).
McDonough,
It is important to acknowledge,
we will not be investigating when
relative to both questions

1 and 2, that
children can extend
their verbs to novel
5
actors,
affected
and
contexts,
objects,
we
Because
actions.
do
not
have
access
we will be unable to
to all the verbs uses the children have experienced,
goes beyond the
specify the extent to which their demonstrated
extendability
in relation to
input given. However,
by noting when flexible use emerges
their age and number of uses of a given verb, we will be able to address
early verb use ismore
children's
whether
in the
verbs
among
spect
than their later verb use.
is the source of individual differences,
3. What
data
restricted
seem
studies
few
comprehension,
The
children.
verbs
children
Differences
all
of
are
and
comprehension
factors
with
as candidate
Across
idiosyncratic.
among
extension?
size
vocabulary
differences
among
more
of verb
scope
verb
and
age
suggest
to individual
verbs
of
re
among
the
experimental
understood
reliably
in more detail by
by all children at a given age. As discussed
a
and
Hoff
such
(2006),
Naigles
cross-study inconsistency is symp
tomatic of a deeper issue with respect to early verb acquisition: the
order
in which
children
acquire
verbs
particular
seems
quite
id
iosyncratic, probably because children's input is idiosyncratic in

the particular verbs it illustrates. Diary studies of production
overcome
this
verbs children
verb
uses
also
development
children.
many
problem
by
avoids
that
a wide
casting
learn. Analyses
assuming
can be found
net
of individual
that
by
to capture
children's
whatever
developing
course
is a single
data
averaged
studying
there
of verb
across
LEARNING
VERB
GRAMMAR
Theories of Early Verb Grammar
verbs first appear in children's speech, they frequently appear in
these
utterances. The nature of the representations
that underlie
multiword
a
with
of
conflict
words
is
first combinations
other
of verbs
among
point
accounts of verb development
and is also a central issue with respect to the
broader question of how children achieve grammar. The view that has been
reflects the
in linguistics for decades
is that word combination
standard
When
of
operation
productive
grammar
that
operates
over
abstract
categories,
or meaning
which
1975, 1981,
(Chomsky,
independent
their verbs because
combine
1995; Crain 8c Lillo-Martin,
1999). Children
of items in the category
they know something about the syntactic properties
in
VERB
subjects, that some verbs
English
require
(e.g., that all verbs
that
many verbs appear with the
require objects while others prohibit them,
to
the verbs in
and
indicate
they recognize
activity)
ongoing
"-ing" suffix
to that category.
their lexicons as belonging
are
of function
6
INTRODUCTION
An alternative view, consistent with the child as a conservative

learner
in production,
who stays close to experience
is that children's combinations
of verbs with other lexical items reflect the operation
of lexically specific
to this view, there is no abstract category VERB to which
rules. According
individual verbs belong
1992, 2000, 2006). A child who has
(Tomasello,
that one verb can appear with a subject, direct object, "-ing"
discovered
this
suffix, and/or prepositional
phrase thus will have no basis for extending
same
to
at
other
verbs
the
in
time
2004;
Pine,
discovery
point
(MacWhinney,
8c Rowland,
Lieven,
1998; Theakston,
Lieven, Pine, & Rowland,
2001).
Most of these latter theorists agree
that children
attain an
eventually
abstract
grammar,
accounts
concerning
these theories
combinations
the
still disagree
they may
although
exact
or
nature
level
of
with
formal
the
abstraction.
generative
Thus,
need to include an account of how children's verbs and verb

are transformed
from lexically based to abstract.
Evidence From the Study of Verb Learning

Evidence for Early Abstract Grammar
The claim has been made from several quarters
that structural regu
are evidence
larities in children's earliest combinations
that children have
abstract grammatical
the
(Borer 8c
categories,
including
category VERB
8c
Fisher
Gleitman
&
Wexler,
1987;
2002;
Gleitman,
1995; Pinker,
Newport,
from comprehension
studies
1984; Valian,
1990). Two kinds of evidence
also
argue
that
young
children
know
more
about
verb
syntax
than
just
the
combinatorial
of individual verbs. First, 21-28-month-old
chil
possibilities
dren who have been taught a novel verb in one syntactic frame show they
can distinguish
that verb from another when both are presented
in a new
et
a
frame
children
know
al.,
Second,
2005).
syntactic
great deal
(Naigles
about sentence frames, independent
of any verb they contain, which also
as properties
of specific verbs.
suggests that frames are not just represented
For example,
children
understand
the
1-year-old
subject-verb-object
sentences when given noun-verb
(SVO) word order of simple English
noun
(NVN) sequences with novel verbs (Gertner, Fisher, 8c Eisengart,
2006). That is, they infer the first noun of an NVN
sequence
(e.g., "The
is the agent of a novel action presented
duck gorps the bunny")
in a video
clip.
Moreover,
2-year-olds
map
novel
verbs
in NVN
sentences
onto
caus
ative actions (in which the duck is doing something

to the bunny), and they
in NNV
sentences
map novel verbs presented
(e.g., "The duck and the
are
onto
noncausative
actions
the duck and
(in which
bunny
gorping")
are
same
as
the
side
action,
bunny
doing
independent
by side)?again,
scenes
to
at
evidenced
the
look
when
sentences
such
with
by
they prefer
7
to them
verbs are presented
1996;
1990).
Naigles,
Naigles,
novel
Further
evidence
that
(Fisher,
1996; Hirsh-Pasek,
represent
2-year-olds
with
meanings
not just with lexical items comes from studies in which children
act out sentences
in which frame meaning
and verb meaning
a sentence
given
as
such
*the
zebra
goes
the
lion,
infer
2-year-olds
&
Golinkoff,
frames
and
are asked
conflict.
to
So,
a causative
to the lion,
from the NVN frame and have the zebra do something
meaning
but given a sentence such as *the zebra brings), children infer a noncausative
of bring) and have the zebra do
meaning
(despite the causative meaning
without
affected
8c Gleitman,
any
something
object (Naigles, Gleitman,
et
al.
also
found
that
and
2128-month-olds
1993). Finally, Naigles
(2005)
could recognize novel verbs in unattested
frames; that is, verbs taught in the
in an intransitive
transitive frame were reliably recognized when presented
(see also Fernandes, Marcus, Di Nubila, & Vouloumanos,
2006).
These
of abstract-level
findings, taken together, support the existence
transitive and intransitive frames in the linguistic knowledge
bases of 1- and
learners.
do
indicate
the range of
not, however,
2-year-old
English
They
or
are
how
how
that
much
toddlers
(i.e.,
flexibility
early)
capable of, nor do
us
muster
of
will
in their own
tell
about
the
level
that
children
they
flexibility
frame
And
is
yet production matters, because production
flexibility
of adult language use in all theories (Chomsky, 1975; H. Clark,
turn to data from
Pinker,
1994; Tomasello,
2000). We therefore
productions.
the hallmark
1996;
children's
of
production
verbs
sentences.
and
Evidence for Early Lexically Specific Grammar

to
contrast
In
the
evidence
from
comprehension
studies
that
young
that allows them to in

possess abstract grammatical
knowledge
of
verbs
the
evidence
from
studies
terpret
production
flexibly,
yields a pic
ture of limited flexibility, which has been taken as support for the view that
is lexically based. Tomasello
(1992) argued for this view on
early grammar
verb uses during her 2nd year
the basis of his diary records of his daughter's
of life. This 1-year-old child tended to use her verbs in only one construc
she might
tion type per verb (i.e., one frame); for example,
say "Mommy
children
break"
Fewer
and
than
"Daddy
one
third
break"
of her
but not
verbs
were
"Break
used
cup" or "Break with

in more
than
two
stick."
construction
in her
the child was uneven
the span of her 2nd year. Likewise,
of
lexical
lexical
and
markers;
subjects
subjects
morphological
production
were produced
for some verbs (take, get) but not for others (put),
consistently
types over
and
some
verbs
received
tense
markers,
others
aspect
markers,
but
very
few
received both (see also Tomasello,

2000). Studies involving more children
have also found
limited productivity
of early verb use (e.g., Matthews,
8c Tomasello,
2005; Pine et al., 1998; Theakston,
Lieven, Theakston,
8
INTRODUCTION
et al, 2001). These

latter findings
Lieven, 8c Tomasello,
2003; Theakston
are not independent,
come
much
of
these
from the
because
data
however,
same Manchester
in
which
children's
12
corpus,
spontaneous
speech during
at
for 6 months,
lunch time and toy play was recorded monthly
beginning
the onset of multiword
speech.
was "counted"
Across all of these studies, productivity
if a child used a
or
frame
marker
with
different
lexical
items
and/or the
given
morphological
same lexical item in different
frames or with different morphological
markers. Using
these criteria, these researchers
have found evidence
for
not
of
the
of
the
marker
but
for
the
present progressive
(-ing),
productivity
SV, or SVO frames (see also
past tense, third person
singular, auxiliaries,
8c Berman,
Armon-Lotem
learn
2003, for similar findings from Hebrew
for
limited
Theakston,
Lieven, Pine, and
arguing
productivity,
uses of go, that
when
children's
(2004) discovered,
analyzing
different forms were used in nonoverlapping
contexts, suggesting different
forms were associated with different meanings
(i.e., The train goes choo-choo
use of one meaning
and Mommy went to work) rather than productive
went
to
and
likes
work/outside
(i.e., Mommy
Johnny
going to the park). The flex
was
that
observed
individual
verbs
seemed
and/or children)
(with
ibility
to imitations of varied uses in the maternal
attributable
input (Theakston
et al., 2001), or emerged piecemeal
to
via one-word
additions or deletions
a
a
want
to
want
from
"I
"I
W"
book")
(e.g.,
previously
produced
phrases
ers). Also
Rowland
8c Tomasello,
Behrens,
(Lieven, 2006; Lieven,
2003). Finally,
Speares,
the children's verb use at Stage
McClure, Pine, and Lieven (2006) compared
I (MLU<2.0)
and Stage II (MLU between 2.0 and 2.5), finding that "old"
structures than
verbs (those used at both stages) appeared
inmore complex
"new" verbs (those first observed at Stage II). Thus,
it appears from these
data
that
children's
ability
to use
verbs
in more
structures
complex
is related
to the length of time they have known the verbs and hence children's
ability
to use a verb in a given frame is not predictable
from their use of other verbs
in that frame, as might be expected
if all verbs are treated as full members
of
the
same
abstract
category
studies
of elicited speech have yielded

similar findings:
Experimental
Children
under 2.5 years who were taught novel verbs for novel actions
verb in an unattested
frame. That
is,
rarely if ever used a just-learned
a
children
in
verb
the
intransitive
Ernie
chams) did not use it
taught
(e.g.,
in the transitive (Ernie is chamming Bert) (Akhtar 8cTomasello,
spontaneously
8c
Brooks
1997;
Tomasello,
1999; Olguin 8cTomasello,
1993; see Tomasello,
a
for
Tomasello
8cAkhtar, 2003).
2000,
summary, also, Naigles, 2002, 2003;
older
and
those
in frames dis
novel
verbs
children,
Only
especially
taught
over
tributed
able
to use
Lieven,
several
the novel
8cTomasello,
sessions
verbs
rather
than
in unattested
massed
frames
in a
single
(Ambridge,
session,
were
Theakston,
2006).
9
In
on
research
sum,
children's
early
use
verb
productive
(i.e.,
before
2.5
that young verb users do not routinely use their verbs in

forms and sentence frames, nor do they routinely use
multiple
grammatical
forms and frames with multiple
verbs. The research that yields
grammatical
these findings
is limited, however, by four major factors:
First, Tomasello's
(1992) verb diary study included only one partici
those findings can be generalized
is un
pant; thus, the extent to which
years) has found
in the current research).

(but will be investigated
Second and more
studies with larger sample sizes have
importantly,
the
children's
that the full
only sampled
speech, with the consequence
use
to
of
verb
is
have
been
in
the speech
flexibility
unlikely
represented
known
available for analysis. Given

that the recorded utterances
comprise just a
it seems unwarranted
very small portion of the speech the child produced,
to
draw
conclusions
about
those
words,
and
morphemes,
con
syntactic
structions that did not appear in the sample because these could have been
used by the child while not being recorded
1996; Naigles,
2002;
(Demuth,
8c Stahl, 2004). Many of the speech samples
Stromswold,
1996; Tomasello
were
also
in context
restricted
(i.e.,
same
the
recording
at each
setting
visit,
2006; Lieven et al., 2003) as well as time, thereby limiting the ob

that could be the topics of conversation.
For ex
and people
actions,
jects,
a
if
child
is
recorded
breakfast
and
after
ample,
consistently
during
Lieven,
breakfast
wanting
(food,
things
toys)
and
words
as,
perhaps,
system continues
production
utterances
her
not
would
would
vocabulary
to develop
be
necessarily
seen,
different times of the day and different

have
the
reported,
verbs
used
for
suggestions
time, the same types of conversations

more
is likely to involve expressions
free play, then the conversation
action
recur, with
and
grows
the
working
at
least,
vary
on
memory/
But very novel
is also
child
settings. As Naigles
by mothers,
of
"). Over
the child adding
her
(Lieven et al., 2003).

unless
. . .
("let's
at
recorded
and Hoff
(2006)
accord
dramatically
ing to setting; it is unlikely that children's verb use would be any different. And
given that mothers do not even always use every word and frame flexibly in
et al., 2003), when they,
their sampled speech (Lieven et al., 2003; Theakston
to
unarguably, have full command of productive grammar, it is unwarranted
attribute children's lack of flexibility to a less-than-full command of the same.
Because
children talk less than their mothers,

they would, in fact, demonstrate
less flexibility in their sampled speech by base rates alone.
Third,
the
"conservative-child"
assumes
argument
that when
children's
verb uses omit a subject or object, such omissions

reflect limited grammat
ical knowledge.
and Tomasello
However,
Matthews,
Lieven, Theakston,
(2006) and Valian, Prasada, and Scarpa (2006) have recently demonstrated
that toddlers' production
of subjects and objects in elicited production
tasks
varies
systematically
toddlers
according
to discourse
use fewer full NPs when
and
the referents
processing
constraints
(e.g.,
of those NPs have just been
10
INTRODUCTION
these are part of

they imitate full NPs less frequently when
and
Allen
Schroder
have
also found that
(2003)
longer target sentences).
toddlers' spontaneous
of
lexical
and
production
subjects
objects is highly
context.
Because
constrained
the
discourse
the
by
linguistic and discourse
context affects whether or not children produce NPs, children's production
of a given verb with only one NP is not good evidence
that the child thinks
that verb has only one argument.
Such utterances may reflect pragmatic
sensitivity on the child's part, rather than limited grammatical
knowledge.
limited
Fourth, the import of the experimental
findings that suggested
is
associated
with
verbs
itself
limited
the
fact
that
grammatical
flexibility
by
most of the types of frame productivity
that have been tested have involved
meaning
changes as well as frame changes. That is, the use of specific verbs
in specific frames is partially governed
by each verb's semantics; for exam
not
causative
verbs
do
frames, motion
ple,
usually appear in intransitive
verbs do not appear with sentence complements,
and so on (Fisher, Gleit
and
labeled,
man, 8cGleitman,
1983, 1990; Levin, 1993). As described
1991; Jackendoff,
inmore detail by Naigles
to use a novel verb in
children's
reluctance
(2002),
an unattested
frame could be traced to their uncertainty
the
about whether
verb's
was
meaning
even
Thus,
when
for
suitable
the frame independently

8cAkhtar, 2003; Naigles,
the
children's
verb
uses
than
rather
frame,
of the verbs already

2003).
show
truly
not
their
knowing
in it (see also Tomasello
used
less
flex
grammatical
system would support, there are two possible

ibility than a fully productive
One
that
is
toddlers'
of simple sentence frames
interpretations:
knowledge
is abstract, but they use a limited range of grammatical
options in their verb
uses because they use their verbs to express only a limited range of mean
On
ings.
this
account,
it is the
limited
of
understanding
verb
meaning
that
the limited grammatical

in
flexibility of verb use. An alternative
is
that
toddlers'
of
sentence
initial
is
frames
terpretation
knowledge
simple
a child might know both
indeed lexically specific. For example,
and
drop
move can be causal, but having heard move but not
drop in the transitive
frame thus far, he only produces move in this frame. On this account,
it is
underlies
limited
syntactic
A Mechanism
Thus
dren
have
knowledge
causes
that
initial
conservative
of Transition From Lexically Specific
far we have been

an
abstract
grammar
considering
that
allows
use.
toAbstract Grammar
the evidence
newly
verb
regarding
acquired
verbs
when
to be
chil
rec
as instances
of a category and freely combined with inflectional

ognized
and
morphemes
syntactic frames. Another question, which is crucial to a full
account of verb development,
is how the child achieves such a grammatical
to
the
the grammatical
system. According
strongest generativist
position,
and
the
could
themselves,
categories
configurations
they
possibly appear in,
11
are innate (Chomsky, 1975, 1995; Crain 8c Lillo-Martin,

1999). At least some
of
this
innateness
be
rendered
if it could be shown
aspects
may
unnecessary
that children discover grammatical
and/or
syn
categories
language-specific
tactic configurations
from their input before combinatorial
speech begins,
and studies of infant learning of language structure make this a plausible
1993; Gerken,
2007; Gomez
2002; Mintz,
2003;
(see Brent,
suggestion
of
such
innateness is also a central tenet of
However,
Naigles, 2002).
rejection
to young children. Ac
that attribute
little syntactic competence
positions
to
the
children
cording
lexically specific grammar view,
begin to combine
verbs with other words before they have achieved grammatical categories and/
or abstract syntactic patterning
2000). Thus,
(MacWhinney, 2004; Tomasello,
a full account consistent with an early lexically
specific grammar requires an
account
of how
children
move
from
to
state
that
an
having
abstract
grammar.
A set of influential suggestions

revolves around the idea that a subset of
are
in
verbs
that
called
verbs,
frequent
input and have general meanings,
verbs,
"light"
serve
as
the
child's
entry
into
abstract
forms
grammatical
and
their meanings
(Chenu 8c Jisa, 2006; E. Clark, 1987, 1996, 2003; Goldberg,
et al., 2004; Goodman
8c Sethuraman,
1999; Goldberg
2006; Ninio,
1999;
but see Campbell 8cTomasello,
This
2001).
proposal that light verbs provide
system comes in several forms, three of which we
entry into the grammatical
summarize
here.
of Goldberg
and her colleagues
(2004) focuses on the
proposal
or
to
of
the
be
frames
constructions
meanings
acquired and relies on both
and
semantic
attributes
of
light verbs. Essentially, Gold
frequency
general
of
children's
the high-frequency
verb go
that
berg hypothesizes
hearings
or
to
the
the
followed by a prepositional
around
store,
go
phrase
locative?go
frame or verb+locative
and
the
table, go outside?(+PP
[VL] construction)
verb put in the + NP PP frame (or V+object+L
[VOL] con
high-frequency
on
to as
the
the
bowl
there?enables
them
bowl
the
table,
put
struction)?put
The
sociate the general meanings

of go and put (inchoative motion
and caused
VL
with
VOL
the
and
frames
motion,
1999; Gold
(Goldberg,
respectively)
et
&
Goodman
8c
Casenhiser,
2006;
2003;
Sethuraman,
al.,
Goldberg
berg
to
of the light verbs are "transferred"
2006). Thus, the children's meanings
of the frames or constructions
that they typically appear in so
the meanings
to be acquired
that the light verbs enable the frames' meanings
(see Case
nhiser 8c Goldberg,
evidence
from grade schoolers'
2005, for supporting
of nonsense
verbs). This theory does not predict,
though, that
acquisition
to
first
in
child
verbs
should
be
the
verbs
appear
speech in a given
light
frame?light
verbs
do
their
work
during
comprehension.
They
are
con
to be the engine of learning about construction meaning,

but not
in
of
the
this
observable
first
children's
consequence
necessarily
learning
(Goldberg et al., 2004). And the theory is silent on how the
productions
sidered
frame
or
construction
forms,
themselves,
are
acquired.
12
INTRODUCTION
second version of the light verb proposal

does address children's
of
that
each
frame is first learned
acquisition
syntactic form, hypothesizing
to other verbs (E. Clark,
with a single light verb, and only later extended
rationale
1987, 2003; Chenu 8c Jisa, 2006; Ninio,
1999). These researchers'
most
for this special role for light verbs is similar to Goldberg's:
verbs
Light
or
encode
the
of
frames
constructions
(i.e., go is the
meanings
specific
closely
motion
the +PP frame, make is the canonical
verb, exemplifying
on
the
transitive frame, and put is the ca
verb, exemplifying
causal/acting
the + NP PP frame). The specific hy
nonical transfer verb, exemplifying
is that light verbs should be the first to appear in a given frame
pothesis
canonical
because
used
the
form-meaning
a PP or L because
with
is so
relation
go
Go
transparent.
encodes
transparently
to be
is
learned
motion,
so
any
ac
involve locatives (go home/to the store). Then

arguments must
companying
children's use of go with a variety of PPs or Ls enables them to abstract a
more general +PP or VL frame, which
they can then transfer for use with
more specific verbs such as run and
(eventually) cry.The idea is that the first
verb used in a given frame is that frame's pathbreaker (Ninio, 1999), which,
facilitates children's formalization
of that
by virtue of its general meaning,
frame and thus their use of other verbs in that frame.
from this light-verbs-as-pathbreakers
Further predictions
hypothesis
include that light verbs should be used in ways both more grammatically
flexible
than heavy verbs learned at the
complex and more grammatically
same time (E. Clark, 1987; Ninio,
et al., 2004). Both of
Theakston
1999;
these predictions
follow from the semantic generality of light verbs: Because
verbs such as go and put can apply to any type of inchoative or caused motion
in these mo
(respectively),
specifics about which entities are participating
tions
are
not
recoverable
always
from
the
context.
For
eat
example,
usually
only involves eatable things, whereas put can involve a wide range of things
in forward horizontal
put, and run usually only involves animates engaged
motion,
whereas
motion
in any
can
go
direction.
involve
Thus,
to be as clear as possible
appear
with
overt
either
arguments
(Grice,
or
animates
to
in response
the
inanimates
engaged
communicative
pressure
1989), go and put should be more
than
run
and
eat.
Furthermore,
in
likely to
semantically
in the adult language

(Levin,
general verbs are more
syntactically diverse
8c Hoff-Ginsberg,
their general mean
1993; Naigles
1995, 1998) because
ings allow them to fit into a wider variety of sentence frames (Snedeker 8c
Gleitman,
2004). For example, bring can be used both with and without a PP,
on whether
the speaker chooses to focus on the action or goal of
depending
What's
(e.g.,
bringing
Molly doing? She's bringing thepizza/Where isMolly going
so quickly? She's
bringing thepizza to the library), and go but not run can be used
with a preverbal P (e.g., Up you go! but not Up you runl?for
most speakers).
The current evidence
for light verbs as "pathbreakers"
to an abstract
is
mixed:
Ninio
and
Chenu
and
(1999)
grammar
Jisa (2006) report, based
13
on spontaneous
learning Hebrew,
speech samples from children
English,
and French, that many children first used a given frame with only one verb;
moreover,
(Chenu 8c
they took from 43 days (Ninio, 1999) to several months
to
a
with
verb.
Ninio
that
frame
second
However,
(1999)
Jisa, 2006)
produce
et al. (2004) have also found
and Theakston
were not always light verbs. Ninio
(1999)
finding with the light-verbs-as-pathbreakers
what
notion but broadening
"pathbreaking"
verb for a given frame. In particular, want,
is included as a canonical/general
frames,
does not encode "highly transitive" senses
verbs
that these pathbreaking
this divergent
has reconciled
the
by maintaining
hypothesis
counts as a light or "canonical"
a frequent first verb in SVO
it
transitive verb even though
such as change of state or po
8c Thompson,
8c Hopper,
sition (i.e., Hopper
1980; Thompson
2001).
et al. (2004) also point out, though, that because semantic gen
Theakston
are confounded,
it is not clear which factor
erality and input frequency
to children's
forms.
of grammatical
contributes more heavily
acquisition
no
in
numbers
and
verbs
their
between
found
difference
heavy
They
light
of different
frames, percent of utterances with subjects, nor percent of ut
terances with objects, once input frequency was partialled out. Apart from
the light-verbs
the mixed nature of the evidence, a problem with evaluating
is that all of the relevant data thus far have been
hypothesis
as-pathbreakers
drawn from samples of spontaneous
speech of limited duration and limited
are
tenuous for the sampling reasons dis
such
contextual
range;
findings
cussed
earlier.
Moreover,
almost
all of
the
research
thus
far has
investigated
the behavior of light verbs as a class, even though the light-verbs-as-path

breakers proposal really targets several individual light verbs as pathbreak
ers for specific frames (i.e., go for the VL/+PP
frame and put for the VOL/
+ NP PP frame), and furthermore,
the particular
light verb serving the
for
children.
different
different
could
be
function
pathbreaker
that light verbs are the prod
and Gleitman
Snedeker
(2004) propose
ucts of syntactic acquisition
rather than its engines. They argue that light
that could be inferred from observation
verbs do not have specific meanings
as,
for
example,
the meaning
are more
of
eat could
be
inferred
from
observing
eaters.
on
syntax for their acquisition

dependent
from their
receives support in findings
argument
8c
Gleitman,
Lederer,
Gleitman,
1999),
(Gillette,
Paradigm
to
and
told
conversation
of
video
mother-child
clips
given
a
tone
and
verbs
sounds.
when
is
what
verb
go
put)
(e.g.,
guess
Light
produced
to be accurately
than the
more syntactic
information
guessed
required
One
run
8c
and
verbs
Gleitman,
2004).
throw)
(Snedeker
(e.g.,
"heavy"
verbs
this
is
that
that
from
view
follows
for
children's
light
speech
prediction
should appear less frequently than heavy verbs in verb-only phrases and more
frequently with sentence frames because the light verbs should be acquired
after (and because of) the sentence frames.
Thus,
light verbs
than heavy verbs.
Human Simulation
in which adults are
This
14
INTRODUCTION
to Be Addressed
Questions
users
of
verb
their
children
become
is central
to
and how
when
Understanding
ductive
of Verb Grammar
Regarding Children's Acquisition
lexicons
pro
grammatically
and
when
understanding
users. The
of
foregoing
language
verb
about
literature reveals the following
early
questions
and thus about the flexibility of children's early verb use, that
productivity,
form a focus of the present study:
how
children
adult-like
become
review
the existing
1. How
in
early
course
the
at
ductivity?or
novel
verbs
sitive
versus
verb
1- and
that
demonstrated
of
does
learning
least flexibility?emerge?
intransitive
frames
novel
et
(Gertner
understand
in tran
verbs
al.,
have
studies
can
children
2-year-old
in SVO frames and distinguish
pro
grammatical
Prior
2006;
Naigles,
1998). Still, the existing studies of children of that age have not
in their spontaneous
found good evidence of such productivity
et
et
A big gap in the
Theakston
al.,
1998;
al.,
2003).
(Pine
speech
which
literature,
from
the
fact
that
the
current
no
studies
in spontaneous
ductivity
was
study
have
tracked
in a way
speech
to
designed
several
that
fill,
results
children's
includes
all
pro
uses
of
their very first verbs.

to delineate,
though, how much of the gap in the lit
can
fill.
The productivity
that is the hallmark of
study
one
utterances
human language is the capacity to produce and understand
has never heard before. The most unambiguous
evidence of this capacity in
It is important
erature the current
spontaneous
is the
speech
production
fall that on me"
(Bowerman,
have
heard
an
uous
evidence
by
testing
such
that
children
errors?utterances
1973)?because
utterance.
experimenter-created
such
the child
Comprehension
utterances
understand
children
with
of
can
studies
they
utterances
as
"don't
likely to ever
is not
elicit
unambig
never
heard
have
using
nonsense
the bunny"
words, such as "the duck is gorping
1990). Because
(Naigles,
our study relies on children's
of
verbs in their
spontaneous
productions
conventional
uses,
without
comprehensive
analysis
of
their
input,
we
will
not be able to make claims about this level of productivity. However, many
researchers have looked at spontaneous
speech for evidence of productivity
use
as
of
pro
by taking flexibility
indicating, albeit indirectly, underlying
1989;
Shirai,
1998; Tomasello,
1992). If a child can use a
ductivity (Ingram,
frames or a single frame with multiple
verbs, this
single verb in multiple
a
at
work.
Without
the
child's
system
suggests
productive
knowing
input, it
uses are all input-based
is always possible
that these multiple
rather than
the child demonstrates
the
anew, but the greater the flexibility
generated
less plausible
that account becomes. Using a metric
that the production
of
three to five verbs in a given frame displays some amount of productivity,
15
various
with
have reported productivity

in 1- to 2-year-old
children
such as V-ing, SV, VO, SVO (Ingram, 1981; Pine et al., 1998;
1998; Tomasello,
1992). In the present study we similarly adopt the
researchers
frames
Shirai,
and
three-verbs-per-frame
measures
five-verbs-per-frame
of
productivity
as less unambiguous
but still valuable indicators of the children's ability to
use their frames flexibly and independently
of specific verbs.
In this study we also introduce several additional measures,
which we
of
We
index
the
the
argue
system.
investigate chil
productivity
underlying
we
term gram
dren's ability to use their verbs with multiple
which
frames,
(1992) claimed that Travis's "inability"
in her grammar;
limitations
illustrated
to which our child subjects did use their
the appropriate
the first 10 instances?with
verbs?within
grammatical
to which these
We
and
also
arguments
morphemes.
investigate the degree
latter two types of flexibility are related to each other: Does it automatically
follow that a child who uses eat with both Ss and Os might also use the SVor
matical
to use
flexibility. That is, Tomasello

break with both Ss and Os
therefore, we investigate the extent
VO
frame with multiple
verbs?
2. If children do not initially have adult-like productivity,

process
idence
they
by which
of "pathbreaking"
that
seem
view
that
to
verbs
when
which
they
whether
over,
and
whether
the
earliest
tions
they
show
our
onset
driven,
first
grammatical
grammatical
will
be
which
uses
of
though,
a frame
first
children's
flexibility
Relations Among Different Aspects
of data
we
uses
than
function
not
occur
of
their
for
this
among
with
was
address
all
children,
in
the
driven
all
ques
current
rather
the
whether
verbs
are
and
are
verb
to
degree
More
function
productivity,
addressed
collection
will
differences
in the
speech
and
complexity.
serve
show
unanswered,
be
should
specifically,
ev
is, verbs
with
flexibility
serve
this
grammatical
same
verbs
light
to
verbs
study.
than frame
there
that
in children's
appear
verbs,
still
whether
first
the
Because
dren's
that
suggests
is the
found
speech;
a
The
frame.
given
into
abstract
gram
entry
to
the way
productivity
serve
as the child's
verbs
some
what
have
in children's
verbs
lead
frames
matical
studies
it? Several
achieve
but
light
verbs
demonstrate
light
first uses
of heavy
verbs.
chil
rather
more
of Early Verb Growth and Use
on the nature of early verb

positions
the
predictions
regarding
distinguishable
yield empirically
development
and flexibility of grammar as they
relations between flexibility of meaning
of both types of flexibility of
develop and the relation of the development
to generativist
verb use to growth in the verb lexicon. According
theories,
Some of the contrasting
theoretical
16
INTRODUCTION
the grammar
is formal, abstract, and autonomous
1995, 1998;
(Chomsky,
Borer 8cWexler,
1987). Therefore,
acquisition of a verb lexicon, develop
ment
of semantically
flexible uses of those verbs, and development
of
uses
flexible
of
verbs
those
should
all
be
aside
unrelated,
grammatically
from
the
obvious
command
necessity
be able
would
that
to
children
produce
more
have
who
at
words
their
sentences.
longer
theories predict
there will be relations
Other
lexical size,
among
are
semantic flexibility, and grammatical
There
many arguments
flexibility.
that language development
is the product
of domain-general
learning
et al., 1996;
8c Synder,
mechanisms
1988; Elman,
(Bates, Bretherton,
Saffran 8c Thiessen,
of de
2007), and these predict that all the measures
number of verbs acquired and the semantic and gram
velopment?the
matical flexibility with which
be at least loosely
they are used?should
comes
related. A stronger prediction
relatedness
of
from the view that
to this view lexical devel
grammar
emerges from the lexicon. According
opment
must
precede
grammatical
because
development,
a certain
thresh
old or "critical mass" of vocabulary must be achieved before grammatical

can be abstracted
8c
2006; Marchman
(Childers 8c Tomasello,
patterns
Plunkett
8c
Robinson
8c
Bates, 1994;
Marchman,
1993;
Mervis,
1998) and,
in
some
zational
These
accounts,
grammar
processes
views yield
continues
to
develop
lexicon
by a growing
the prediction
that grammatical
caused
as
a result
of
reorgani
8c Thai,
2006).
(Conboy
should be pos
flexibility
of verbs. The syntactic bootstrap
itively correlated with children's number

also yields the prediction
that the size of the verb lexicon
ping hypothesis
com
should be positively related to the flexibility and overall grammatical
plexity
of
verb
use
because
grammar
is
the
source
of meaning-relevant
to learning verbs. Both the shared mechanism

information
that contributes
and syntactic bootstrapping
hypotheses
predict a positive relation between
semantic and syntactic flexibility of verb use. The shared-mechanism
hy
semantic and syntactic flexibility
pothesis predicts that the relation between
should be across children?children
who show more
semantic flexibility
should also show more grammatical
The
flexibility.
syntactic bootstrapping
hypothesis
additionally
predicts a relation between grammatical
flexibility
and semantic flexibility across verbs, because it is knowledge
of a particular
verb's structural possibilities
that supports its semantically flexible use.
Numerous
studies support the conclusion
that the size of children's
vocabularies
and the grammatical
of their speech are related
complexity
8c
8c Bates,
Goodman,
1999; Caselli, Casadio,
(Bates
1999; Jackson
8c
Maldonado,
Thai, Marchman,
Bates,
1993; Maital,
Guitierrez-Clellen,
& Bates,
Dromi,
2000; Marchman
1994; Marchman,
Sagi, & Bornstein,
8cDale, 2004; Ogura, Yamashita, Murase, & Dale, 1993;
Martinez-Sussman,
8c Mervis,
are less well sup
Robinson
1998). More
specific predictions
The
claim
that
lexical
ahead of
ported.
development
initially proceeds
17
is paced by the
and that grammatical development
grammatical development
on
studies
that
used
theWords
is
based
lexical
largely
preceding
development
form of the MacArthur-Bates
Communicative
and Sentences
Development
Inventory (CDI; Fenson et al, 1994). The CDI includes a checklist of over 600
words plus a section in which pairs of phrases (one closer to the adult form
than the other; e.g., two shoe vs. two shoes) are listed; the parent checks off the
phrase that is closer to the child's current level of speech. These studies have
more words
reported a curvilinear relation in which children who produce
are also reported to produce more complex grammatical forms, with smaller
in vo
associated with early increments
complexity
changes in grammatical
in
with
associated
size
and
grammatical
complexity
larger changes
cabulary
later,
in
increments
equal-sized
vocabulary.
to think that the statistical

a variety of reasons
There
are, however,
in these data do not necessarily
relations observed
imply that grammatical
on lexical development.
measurement
issues
First,
depends
development
et
Robinson
abound. In the studies of spontaneous
al.,
1988;
(Bates
speech
of grammatical
8c Mervis,
1998), MLU was used as the measure
develop
of grammar apart from the size of the
ment, and MLU is not a pure measure
in their study
lexicon. Indeed, as Rollins, Snow, and Willett
(1996) reported
14
of children's
speech, the MLU growth between
sampled spontaneous
and 32 months for a substantial percent of the children was accounted for by
or other grammatical
growth in content words alone, not in morphological
items.
a relation
Thus,
between
size
vocabulary
MLU
and
not
does
neces
de
lexical development
and grammatical
sarily reflect a relation between
it
the
reflect
Rather
greater
may
possibilities
expressive
velopment.
afforded by a larger vocabulary.
Two problems particularly apply to the many studies that have used the
CDI.
the measures
Because
same
report
parental
of
instrument,
vocabulary
they
and
may
grammar
share
the
are
same
taken
from
the
bias.
parental
curvilinear
of the observed
interpretation
critically, the ordering
on the
measures
and grammar
the vocabulary
relation between
depends
Most
assumption
grammar
that
were
the measurement
equally
sensitive
of vocabulary
measures
of
and
change
the measurement
of
across
of
the
range
and Marchman
(2007) have argued that this
development
function an
nonlinear
false. They performed
is demonstrably
assumption
CDI
section of the
mapped
alyses which revealed that, while the vocabulary
are uni
lexicon (i.e., increases in vocabulary
linearly onto the underlying
section
the grammar
in underlying
lexicon),
formly related to increases
are
in
such that early changes
grammar
underrep
nonuniformly,
mapped
resented in the CDI, relative to later ones. The curvilinear relation observed
studies. Dixon
and grammatical
between
lexical development
is, they argue,
development
an artifact of the differences
of lexical and gram
the measures
between
in their sensitivity to early growth.
matical development
18
INTRODUCTION
to
are
the observed
(linear) correlations
or direction
source
of
The
they
causality.
that vocabulary
supports the induction of grammar
knowledge
possibility
and the alternate possibility
that grammatical
supports the in
knowledge
are different
duction of word meaning
mechanisms
that
language learning
size and gram
would both be reflected in a correlation between vocabulary
Finally,
matical
not
that
reveal
neither
Also,
usage.
neous
extent
the
still do
real,
the
the
use have
word
CDI
assessed what
that they use them?nor
use?only
or
argument
structure
bination
and some
sentence
nor
the
children
have
about
that
sponta
the words
children's
words
appear
they
use of
in com
to the degree
that vocab
Thus,
achievement
of grammatical
about vocabulary
is
knowledge
may
knowledge
it is not clear just what
understandings,
and what about grammatical
doing the supporting
ulary
of
analyses
know
they assessed
frames?only
are present.
the
support
inflections
previous
understanding
is being
supported.
Previous
the semantic flexibility
studies have not examined
of chil
dren's verb use, and thus have not captured possible differences
among
children and among verbs in what children know about the meanings
of the
verbs they use. A revised version of the grammar-from-lexicon
hypothesis
suggest that children who know more about the words that they use
might
to knowing more words,
(i.e., are more flexible with them), as opposed
or
to
sooner. Some support for
be
able
achieve
grammar
might
productivity
this revised version comes from an experimental
study that assessed tod
dlers' extendability
of verbs' grammatical
form as well as their extendability
of the verbs across situations (Naigles et al., 2005). As described
earlier, the
21-month-olds
in
erent
which
were
pairings,
on
presented
Moreover,
the
this
study
thus
video,
children
were
had been
were
shown
taught
demonstrating
also
shown
to be
able
to
recognize
in a playroom
one
to
recognize
form
verb-ref
setting, when
of
the
they
extendability.
verb-referent
pairings that had been taught in the transitive frame but tested in the in
transitive frame (e.g., children were taught you're gorping the ball and tested
on the ball is gorping).
it was the children who could distinguish
Strikingly,
in the transitive who were
(i.e., had learned) the verb-referent
pairings
more
them to the intransitive. These
likely to be able to extend
findings
there is some threshold
of semantic
before
suggest
required
learning
could
be
demonstrated.
grammatical
flexibility
to investigate
The current
these issues in more
study was designed
detail, asking whether children who use verbs more flexibly in terms of their
semantics also use them more flexibly
in terms of their grammar. We will
also examine
as a characteristic
the relations among types of flexibility
of
verbs. That is, we ask whether verbs that are used more
terms
in
of
flexibly
their semantics also more
likely to be used more flexibly in terms of their
grammar.
19
THEPRESENT
STUDY
The foregoing
review of the theoretical
and empirical
children's acquisition of verbs makes
it clear that a central and
is when in the course of verb development
children
question
verbs beyond
the functions,
and morphosyntactic
referents,
literature on
unanswered
extend their
structures
in
which
cause
is as yet unanswered
be
they have heard them used. The question
the research methodologies
that have been brought
to bear are lim
ited in several ways. Sampling
naturalistic
interactions
parent/child
(e.g.,
is likely to miss situation
collecting 3 hours of data per week for 4 months)
specific language use. Children use verbs in particular contexts that may not
be recorded,
such as washing in the bathtub, splashing in the pool, sleeping at
and
night,
riding in the shopping cart at the grocery store (Naigles 8c Hoff,
8c Hoff-Ginsberg,
2006; Naigles
1998). Sampling also does not capture the
full range of uses to which children put the verbs they use, thus rendering
about
questions
the
of
scope
and
meanings
of
forms
verb
use
unanswerable.
with one child (e.g., Dromi,

1987; Mervis,
1987;
Diary
Tomasello,
1992) may solve the sampling problems but at the cost of making
to all children from
it impossible to untangle evidence of processes common
uses that are
to
the
child
under
idiosyncratic
single
study.
An ideal data set for addressing
the issue of when and to what degree
children go beyond their input would include a record of all of the verb uses
studies
of many
conducted
children?in
than
that
of
verb uses
dren's
Ingram,
issues
verbs
some
address
these
are
the
used
as
that does
1981; Shirai,
discussed
make
ways are more
variety of different
use
to a
addition
we
If, however,
input.
yet
all
that
of
those
verbs
children's
that
likely to reflect productive
in
only
unanswered
not
of
record
assumption
include
1998)?provided
single
way,
questions
are
verb
used
in a
and extended
then
we
with
a record
can
begin
of chil
to
their input (e.g., E. Clark, 2003;

the record overcomes
the sampling
above.
In this study, we overcome

utility of the existing evidence
earliest uses
turing children's
attributes
of a diverse
positive
1987; Harris et
1993; Dromi,
We
included more
sello, 1992).
of the sampling issues that limit the

to the diary method
for cap
by returning
in the method
of verbs, incorporating
the
set of previous diary studies (e.g., E. Clark,
many
al, 1988; Robinson & Mervis,

1998; Toma
a
than
single child, asking the parent par
ticipants to record their children's uses of only 34 common verbs and only
the first 10 uses of those verbs. In this way, we were able to track early
our parent participants'
changes in verb use without exhausting
capabilities.
We targeted 34 of the most common verbs produced
by children early in
8cGelman,
(Goldin-Meadow,
Seligman,
man & Bates,
& Kruger,
Tomasello
1994; Tomasello,
1992;
next chapter we present
in detail. In the results
the method
1976; March
1992). In the
chapters that
20
INTRODUCTION
we present
the method,
the description
of early verb development
that these data suggest, and we address the questions we have outlined with
nature of children's
respect to the pragmatic,
we
first verb uses. In a concluding
consider
what
the data imply
chapter
with respect to the nature of children's early linguistic understandings.
follow
21
II.PRESENTINGTHEDIARYMETHOD
PARTICIPANTS
The data in the present
study come from diary records kept by eight
uses of 34 common verbs. A
mothers
of their children's first 10 spontaneous
to
total of 18 mothers
but 9 mothers withdrew
initially agreed
participate,
from the study before having collected enough data for analysis (i.e., at least
10 uses each of 10 different verbs). For five of these mothers,
their reason
was that their child (always a boy) was not
for withdrawing
talking at all, and
to bring the child to clinical services for evaluation
and
they planned
treatment.
The
mothers
other
who
withdrew
the
gave
reason
their
that
child was talking so much

that they found they could not record or make
note of every relevant utterance
their child produced,
and so they could not
keep
year
the
diary
from
after
However,
One
accurately.
her
and
child
the
data
from
additional
mother
seven
collected
verbs
were
data
for
over
for a total of 31 verbs.
the diary
completed
collected,
she
lapsed
in her
diary keeping for 6 months. Given this lapse, we could not be certain that
the data from the last 24 verbs really reflected
the child's first 10 uses of
these
verbs,
and
this
child's
data
were
not
analyzed
further.
At the onset of the study, the children (5 girls and 3 boys) ranged in age
from 15 to 19months;
all were European American. Detailed questioning
of
established
the mothers
that none of the children had yet produced
any
verbs in the 20
verbs; this was supported by the absence of any spontaneous
min
the experimenter's
first visit. The
speech sample collected
during
children's
word
this
types during
spontaneously
produced
speech sample
were tabulated; on average,
the children produced
25.12 different words
=
(SD
11.47). At the first visit, the mothers were asked to fill a questionnaire
the child's siblings, the parents' education and occupation,
and
television and reading habits. Three of the children had older
=
= 3
(M 5 years, SD
years); one also had a younger
siblings
sibling
one
All
months
of
but
of the parents
(3
(mothers and fathers) had
age).
= 3.33
=
attended
1.97;
years, SD
college for at least a year (M [mothers]
= 4.0
=
were
M
SD
Their
[fathers]
2.28).
years,
occupations
generally
an
middle
class (the fathers
included an accountant,
several engineers,
concerning
the child's
22
THEDIARY
METHOD
PRESENTING
insurance
executive,
and
landscaper,
the mothers
manager;
computer
All of the
been accountants,
had previously
teachers, and bookkeepers).
The eight children
children were cared for at home by their mothers.
watched an average of 9.83 hr of television per week (SD = 7.54) and were
to the
read to for an average of 11.08hr per week (SD = 7.67), according
maternal
report.
MATERIALS
Each mother was provided with a bound diary with 34 individual pages.
the first
On each page was listed a different verb with 10 rows for recording
10 instances of that verb's use. The 34 verbs were chosen from prior data
et al., 1976;
sets of the children's
verb use (Goldin-Meadow
spontaneous
8c Kruger,
8c Bates,
Marchman
1992; Tomasello
1994; Tomasello,
1992);
these are all words that would be used as verbs in the adult language. They
included 8 light verbs, which have more general meanings
(H. Clark, 1996;
narrower
which
had
Nine
and
26
verbs,
1999),
heavy
meanings.
Goldberg,
of the 34 were obligatorily
transitive, 9 were obligatorily
intransitive, and 16
were
alternating
intransitive
verbs
(i.e.,
frames). They
they
can
in
appear
are listed, by category,
both
transitive
in Table
and
1. In the blank
TABLE 1
List
of
34 Common
Verbs
Transitive
Light
verbs
Intransitive
Alternating
come
bring
give
go
look
put
take
want
Heavy
verbs
hold
clap
bite
like
need
see
cry cut
fall
drop
run eat
sit
JumP
kiss
walk
wave
lay
move
open
pull
push
roll
stop
throw
wash
23
TABLE2
Diary
of Connecticut/Florida
University
Atlantic
Record
University
First Verbs
Child's
Page
ten uses
of first
of:
Study
Birthdate_ PULL
Name_Heather_
For commands,
who or what
is
command
addressed
of
Record
Date
Was
utterance
a command
complete
utterance
or
the
doing
verb action?
description?
1st
10/2
Pull
Command
Mommy
2nd
10/3
Pull
Command
I pulling
Uncle,
pull
Description
Command
Mommy
Heather
3rd
12/5
4th
12/8
5th
12/8
6th
12/8
Pulling
Auntie,
pull
7th
12/8
Auntie,
pull
8th
1/3
9th
1/13
10th
diary,
1/13
space
Mommy,
me
pull
I pull that
I pull
was
this
to?
For descriptions,
who or what
is
or what
Who
the
receiving
verb action?
Sled
Command
Uncle
Sled
Command
Aunt
Sled
Command
Aunt
Sled
Command
Mommy
Cart
Description
Heather
Blanket
Description
Heather
Blanket
to
record
the
Other
comments
Chair
Wagon
Car with
Uncle
provided
is
complete
handle
utterance,
the
date,
or
function
and addressee
of the
(i.e., command
pragmatic
description),
as well as the actor and the affected object (when relevant) of the
utterance,
verb's action. A sample diary page is reproduced
in Table 2.
PROCEDURE
Parents with children between
15 and 19 months of age were found by
of the local newspaper.
searching back issues of the birth announcements
the study were sent, followed by phone calls requesting
Letters describing
volunteers willing to keep detailed diaries of their child's verb development.
Extensive questioning
confirmed
that the child had not produced
any verbs
at that point and that the parent would be in primary contact with the child
at this point because
their child
20 families were excluded
(approximately
had already begun
verbs). All the volunteer
parents were
producing
mothers. The researcher
then visited the family to train the mother on diary
a
to
collect
20-min spontaneous
keeping and
speech sample.
24
THEDIARY
METHOD
PRESENTING
The training session was conducted

the goal of the study was to investigate
so
detail,
that
it was
of
crucial
first. The researcher explained

the child's language development
to make
importance
note
of
every
utterance
that
in
that
a target verb and (b) was not an immediate

of
(a) contained
repetition
then
introduced
the
and
The
researcher
another person's
diary
speech.
a separate list of the 34 target verbs. It was explained
that our
provided
were less well studied than nouns. The
focus was on verbs because
they
between nominal and
the difference
researcher discussed with the mother
a
vs.
in
bite"
"bite this"), described
bite
"take
verb uses of homonyms
(e.g.,
several
of
examples
uses
verb
in one-word
and
multiword
and
utterances,
then
that the study was about verb uses only. The researcher
emphasized
of possible
each column of the diary page, giving examples
discussed
utterances
in the diary. Specifically,
the
and how they were to be entered
were instructed on (a) what constituted
"When
commands
mothers
(e.g.,
to get you or someone
to do?or
else
child
is trying
your
stop
and descriptions
(e.g., "When your child is telling
doing?something")
an addressee
or
an
about
event,
relation"),
(b) what constituted
you
object,
actor
is
directed
is
the verb
the
command
to"),
("Who
("Whom
doing
or
what
is
the
of
the
verb
and
affected
action"),
patient
object ("Who
were
were
not
and
labeled
the
when
these
child, (c) recording
action")
by
the
as
utterance
as
exactly
inflections
including
possible,
such
as
"-ing,"
"-ed," and "-s" if heard, and (d) the use of the right-most column involving
notes (mothers were encouraged
to add notes at all times, but
contextual
when
especially
Then,
the
they
researcher
were
not
emphasized
sure
how
about
or
pragmatic
important
it was
semantic
for
roles).
the mother
to
note
the target verbs until 10 instances had been

every utterance using
told
the
mother
that every utterance meant
that
and
repeatedly
produced
of any target verbs should be
the child's exact and inexact self-repetitions
of
included, as well as new utterances with that verb, but not repetitions
other
people's
speech.
The
researcher
suggested
that
the
mother
put
the diary in an easily accessible place in the house and carry the diary
she and the child went out. The researcher
asked and
along whenever
noted which words the child was currently producing
(in case some verbs
had emerged
since the phone
researcher
then
call; none had). The
described her own role, which would be to phone the family every 2 weeks
until
the child began producing

the target verbs. Once verb production
the
would
researcher
began,
phone the family every week to check how data
collection was proceeding
and to answer any questions
the family might
have. The training session ended with a discussion
of the duration for data
collection; each mother was asked to keep the diary for at least 3 months.
The actual duration of the study varied from child to child, ranging from a
=
to a maximum
minimum
of 3 months
of 13 months
(M 8.625 months,
=
SD
3.62).
25
to play with
After the training session, the researcher asked the mother
so that the child's current
child for about 20 min
level of speech
could be recorded. The researcher
toys for them to
production
brought
own
could
also
with
and
read
their own books
their
with;
toys
they
play
play
her
little book reading was conducted),

(although
doing whatever
they usually
did at that time of day. The researcher
then started the audio recorder and
left the room for 20 min, staying outside or in another part of the house.
to minimize
Audio recording was used instead of video recording
feelings of
and
self-consciousness
concerns
about
privacy.
to each family
After this initial visit, biweekly phone calls were made
until the child began producing
the target verbs. Then,
the family was called
the parents of the diary procedure
weekly, reminding
(especially, to record
and
the
every utterance)
parents might have had in
answering any questions
out
to discuss
the
Mothers
used
conversations
these
diary.
filling
typically
new utterances
and the records
the children's
the
thereof, describing
in detail and receiving confirmation
and/or instruction concern
were
how
recorded.
Mothers
little difficulty with the
ing
they
reported
some
level of detail required by the records; those who did acknowledge
were
ones
soon
to
the
who
leave
the
Table
3
asked
study.
difficulty
displays
each child's age at the onset of the study, duration of the study, and the total
to the 10-instance criterion out of the 34.
number of target verbs produced
were recorded were included in
for
all
instances
which
10
those
verbs
Only
utterances
of
any
the
and
coding
analyses.
MEASURES
of the age at which each verb was
diary records provided measures
and the number of days elapsed from the 1st to the 10th instance of
The
used
TABLE3
and
Age
Age at
Onset
Child
of Verbs
Number
Length
of
Study
by Each
Produced
#
of Verbs
Child
Rate
Produced
(Out of a Possible
34)
1;8Carl
11 months
24
Carrie
1;7
7 months
30 4.42/month
Elaine
1;8
5 months
Heather
1;7
7 months
314.42/month
31
Stacey
19
3.8/month
Mae
13 months
1;5
Ned
3 months
14
4.67/month
1;6
Sam
11 months
31
2.82/month
1.5
SD
12 months
8.625
months
3.62
of Target
Growth
2.18/month
1;7
1;4
Mean
1;6
Verb
2.38/month
20
1.67/month
24.5
7.07
3.21/month
1.15
26
METHOD
THEDIARY
PRESENTING
use. In addition,
the following measures were taken from the diary records
or coded based on the information
the mothers
recorded:
Pragmatic Content
Each verb use was coded as a command or description
(these accounted
the addressee was
and for each command,
for over 98% of all utterances),
1 of Table 2, Heather
in Instance
noted. To illustrate,
says "Pull" as a
to pull a chair; in Instance 3 she says "I pulling" as
to her mother
command
a description
is the
of her own actions with the toy car. The addressee
utterance
to
Instance
in
this
the
child's
is
is
the
whom
directed;
1,
person
4 it is her uncle.
and in Instance
mother,
Semantic Role
The
person
recorded for each verb use were the particular

semantic properties
or objects filling the roles entailed by each verb. For all verbs, this
who
recording
For
transitive
involved
the
verb.
or what
and
as
served
the
actor,
who
verbs,
alternating
agent,
or what
or
experiencer
as
served
of
patient
or theme was also recorded. For example, when a child said "jumping" and
the child in the actor cell, then the child was coded as
her mother
recorded
the actor of the action. With regard to affected objects, an utterance of "Pull"
with "chair" recorded in the affected object cell would be coded with "chair"
as the affected object, whereas an utterance of "pull me" would be coded with
of actor and affected
"Heather" as the affected object. These
assignments
were
object
expressed
to which
multiple
actors
made
or not
regardless
because
the
of
whether
the
of
purpose
the
roles
semantic
coding
was
to capture
were
the
overtly
extent
actors and on
children extend these verbs to actions by multiple
were
on
to determine
mothers
instructed
how
The
carefully
objects.
and
affected
objects
when
these
were
not
overtly
expressed.
Action Referent
Each
instance
of
each
verb's
use
was
coded
as
the
same
or different
from
to. This
previous uses of the verb in terms of the physical action referred
uses
was
on
for
of
of
the
made
34
based
the
28
actor,
verbs,
target
judgment
affected object, and other notes recorded by the mother. Action referent
change was not coded for four internal event verbs (like, look, need, see, want),
and itwas not coded for the verb bring because the specific action referred to
by uses of this verb was difficult to discern from the diaries. In our first pass
of coding the action content of the remaining
verbs, utterance pairs of the
same verb were hypothesized
to refer to different actions if (a) the actors
were of different species (e.g., dog
running vs. child running) or kinds (e.g.,
child coming vs. TV show coming), (b) the affected objects were of different
27
kinds
(e.g.,
vs.
rocks
eating
an
eating
sizes
apple),
(e.g.,
a chair
washing
vs.
washing dishes) or configurations

(e.g., opening a bag vs. opening a plastic
or
to different
the
situations
locations
(c)
container),
pertained
(coming
vs. coming out of the house), directions
downstairs
(taking Legos off vs.
taking a Pop tart out), or activities (taking a nap vs. taking a shower). A total
were
of 93 distinct
action pairs were
then
initially identified. These
to
in
were
11
to
randomized
order
who
asked
presented
undergraduates,
rate them on a scale of 1-7, where
1 indicated the pairs described identical
that the pairs described completely different
physical actions and 7 indicated
were included to anchor the lower
An
actions.
additional
items
20
physical
end of the scale; these included action pairs performed
by similar actors to
similar affected objects in similar situations. The ratings were averaged for
each action pair and only those whose rating averaged 4.0 or higher (n = 76)
were included in the final
analysis (mean rating for these 76 pairs averaged
=
SD
The
list of action pairs and ratings can be
5.63,
0.84).
complete
obtained from the authors.
Frame
Grammatical
The
frame
grammatical
mothers'
records
coded
components
location
of
words,
For example,
the utterance
an
cracker";
"I"
subject
"down."
preposition
marker
"-ing"
gical
there"
includes
the
and
were
not
verb-containing
and
utterance
The
the
locative
"my
clapping"
marker
"no,"
Vocatives
frame
grammatical
and
phrases,
full
"I drop my
a
it encompasses
includes a subject "my" and a
"no
negative
form
"there."
as a
counted
object
"My fall down"
the
The
markers,
subjects, objects, morphological
(SVO) frames.
utterance
me)
utterance.
preposition/prepositional
full SVO frame. The
push
complete
overt
coded from
markers,
negative
includes
target verb use was
of each
children's
included
subject-verb?object
cracker"
the
includes
and
the
the
morpholo
utterance
in
(e.g., Mommy
"Go
Mommy,
component.
Flexibility
In order
spontaneous
in
instances
Changes
to calculate
speech
terms
are defined
flexibility
was
of
its
coded
pragmatics,
below
as
of verb use, each

same
the
semantics,
and examples
or
instance
of a verb
from
changed
and
grammatical
are provided
in Table
in
previous
frame.
4.
Pragmatic Flexibility
types of change were counted as instances of pragmatic
flexibility:
in
function and changes in addressee. A function change was coded
changes
or
to using descriptions,
when a child first switched from using commands
Two
28
THEDIARY
PRESENTING
METHOD
TABLE4
Flexibility
Type
Coding
Instance
Baseline
of Flexibility
Instance
Changed
Pragmatic
Pull
Function
(Command)
=
(Ad Mommy).
Pull
Addressee
pulling
(Description)
=
Uncle,
uncle)
pull (Ad
Semantic
Actor
(A)
Affected object (AO)
=
Eating (A daddy)
=
Daddy eating (AO pizza)
=
Eating (A dog)
=
Jill eating (AO bagel)
Pull
Pull
(AO
chair)
(AO
wagon)
Action
(A
Open
Come
object
Situation
Go
child)
= car
door) Open
(AO
ActorGoing
Affected
(A
car)
=
(AO
Come
(downstairs)
jar)
(out of the house)
Grammatical
Subject Pull I pulling

Object
Locative
versa,
first made
My
cup
my
drop
pulling
Doggie
eating
I pull that
Pull me
object
for
Pull
Morphology
Lexical
subject
vice
me
Pull
My fall
Myfall down
I biteNobite
Preposition
Negation
Lexical
Pull
SVO
Drop
Go Go there
verb.
given
a change
An
addressee
in addressee
with
Daddy
eating
was
change
a given
coded
a child
when
verb.
Semantic Flexibility
Three
types
of
semantic
change
were
coded.
An
actor
change
was
coded
when a child first made a change in actor during the 10 instances of a given
the change from the dog to the father as actor in Table 4
verb. For example,
instantiated this child's first actor change with eat. An affected object change
was noted for the instance with each verb when a child used a different
affected object from the first instance. For example,
the change from the
as
to
4
the
in
affected
Table
instantiated
this child's first
pizza
bagel
object
affected object change with eat. An action change was coded when a child
first referred to an action physically different from his or her initial action
referent
for that verb, according
to the ratings described
earlier.
Grammatical Flexibility
Nine types of changes were coded as instances
all consisted of the addition or subtraction of words
of grammatical
or morphemes
changes;
to/from a
29
attested
frame. Table 4 shows the nine forms followed
by
previously
The
first
six forms (subject, object, SVO, locative, preposition,
examples.
formed the category of syntactic flexibility. Morpho
collectively
negation)
included
any instance in which the child made a change in
logical flexibility
Past tense uses were vanishingly
verb morphology.
rare; therefore, we did
tense
not have to distinguish
forms
from regularized
past
irregular past
tense forms (e.g., "fell" vs. "failed"). In addition to the foregoing measures
in which
the target verbs were used,
of syntactic frames
of flexibility
measures
of the flexibility of verb use with respect to the lexical items filling
the subject and object roles were coded. Lexical subject and lexical object
in the
the child made a change
included any instance in which
changes
a
or
For
term
in
the
lexical
used
change
example,
object position.
subject
from saying "I pull" to "Baby pull" was coded as a change in lexical subject
a change in
use. In contrast, a change from "pull" to "I pull" constituted
syntax.
In sum, a total of 14 kinds of flexibility were coded for in the children's

of pragmatic
first uses of their first verbs. The measures
flexibility were
to which
narrow
to
the
of
the
address
relatively
degree
question
designed
use
uses
were
their
context
who
verbs
bound:
Children
verb
the children's
their mother
and/or only in addressing
(when other
only as commands,
as tied to a
were
these
verbs
understand
addressees
available), may only
were
measures
use.
of
semantic
The
flexibility
designed
specific context of
to
address
the
core
questions
of
extendability:
to what
extent
were
the
children able to use their verbs with a variety of actors and affected objects,
to different actions? Children who use each of
and to refer (appropriately)
to only one
their verbs with only a single actor or affected object, referring
instantiation of that verb, may not have acquired a principle of extendability
of grammatical
The measures
that applies to verb meanings.
flexibility were
core questions of productivity: To what extent were
to
the
address
designed
sentence
able to use their verbs in different
the children
(appropriate)
Children who use each of their
and/or with different morphology?
in only a single frame may not yet have abstract frames that are
or that operate independently
of their verbs. The semantic and
represented
were
measures
used to address the question of
also
flexibility
grammatical
to an adult/abstract
use light verbs as pathbreakers
children
whether
or grammatical
semantic
with
used
verbs
Are
greater
grammar:
light
were also the relevant
measures
latter
These
verbs?
than
heavy
flexibility
ones for the analyses concerning
how semantic and grammatical
flexibility
of verb use might be related during language development.
The coding was first performed
by the third author and then checked in
were resolved by discussion.
its entirety by the first author. Disagreements
For each verb for a given child, the following were calculated for each
in
the change
(a) which of the first 10 uses manifested
type of flexibility:
frames
verbs
30
THEDIARY
PRESENTING
METHOD
usage, (b) how many days elapsed between first use and first different use,
and (c) how many different uses for each category occurred within the 10
in how children use their first verbs can arise both
instances. Variability
from differences
among children and from differences
among verbs. In
order to investigate both sources of variability, all analyses were conducted:
for each
(1) treating children as the random factor and calculating measures
child by averaging across the verbs they produced
and (2) treating verbs as
the
random
factor
and
calculating
measures
for
each
verb
averaging
the children who produced

them (H. Clark, 1973). Means
are
for
both
types of calculations.
analyses
reported
and
across
statistical
31
III.A GENERALDESCRIPTION
VERBGROWTH
OF EARLY
AND USE
In this chapter, we describe the emergence

and early growth of verb use
in the eight typically developing,
children who were the
English-learning
in
this
On
the
children
the first
average,
participants
diary study.
produced
instance of one of the target verbs by 18 months
of age and added verbs
lexicons at a rate of 3.21
(from the list of target verbs) to their productive
verbs per month. On average, just over 1month elapsed from the 1st to the
10th instance of use. Table 5 presents
the data for each child, listing age at
first use of the first verb, mean age of first use of all verbs, number of target
over the course of the
verbs produced
study (i.e., between 3 and 13months;
see Table
3),
mean
and
to reach
days
10
instances.
As Table 5 shows, the children varied considerably

both in age of onset
of days to reach 10 instances of use. Some
of verb use and in number
children
(e.g., Carl) took very little time (e.g., on average, under 4 days) to
produce
took more
instances
once
instances
10
time
were
on
(e.g.,
produced
use
average,
on the
of
a verb
over
same
began;
2 months).
for at
day
other
children
For
least
one
four
verb;
Sam)
(e.g.,
all
children,
for
the
10
other
four children,
the shortest number of days to reach 10 instances ranged
1
from 2 to 18 days (Ned = 2, Carrie = 5, Stacey = 7, Sam = 18). Figure
presents
the
cumulative
children's
target
verb
vocabulary
over
the
course
of
the study. The children can be grouped

into early verb learners (Sam, Ned,
and Stacey) and later verb learners (Elaine, Carl, Mae).
Carrie, Heather,
Most of the early verb learners (except for Sam) were also fast learners, with
over time. Most of the late verb
steep increases in target verb vocabulary
learners (except for Elaine) were also slower learners, with more gradual
increases
in
target
verb
vocabulary.
The
correlation
between
the
age
at
their first of the target verbs and the average age

which children produced
at which they first produced
all the target verbs used during this study was
=
=
<
.001.
.81,
r(n
8)
p
significant,
the analogous
data by verb, also revealing much
Table 6 presents
across all verbs was 649.96
onset
The
of
average
age
variability.
days
32
VERBGROWTHAND USE
OF EARLY
TABLE5
Verb
Overall
Number
Child
Verbs
Carrie
30
of
Elaine
19
31
20
Stacey
Mean
24.63
Age of Onset
Across Verbs
of Days
10 Instances
Number
Reach
to
759.17 3.17
564
744.9320.03
602
674.8924.95
584
49.55
632.97
577 31
Mae
744.93
35.60
511 14
Ned
532.85
22.64
546 31
Sam
600.10
67.48
491
534.94 58.39
559.75
653.1035.23
41.15
6.48
SD
Mean
Mean
Age First Verb

Produced
(Days)
60324
Carl
Heather
by Child
Development,
92.76 21.73
38.47); open was on average the earliest produced whereas give was on
On average, verbs
the latest of these 34 verbs to be produced.
average
reached the 10-instance mark 37.39 days after their first use (SD = 16.71);
wave took the most,
clap took the fewest days to reach 10 instances and
on
wave
are
the
and
least
the
data
reliable,
being used by only
although
clap
two children. Two verbs were produced
all
8
children
(want, wash); on
by
=
5.79
children
(SD
1.59).
average, each verb was produced
by
the verb within the
Across verbs, the number of children who produced
with
the
window of this study correlated
average
age at which
negatively
(SD
35
-|-1
r-S
/J
& 25J**T
\
/ fi
y~r*I
IS
r-//
\ - 7
/
5/
\Z /* //
||
5J
0
r^ I I
J
*
i
'
'
*
;J
Carl//
* / Ned
F^H
Carrie
/
-^Heather
/
Elaine
?Mae
//
I /'
rJ
|
I"^-Stacey
Ii i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i I
Age
Figure
in days
1.?Cumulative
verb
vocabulary
by age.
33
TABLE6
Overall
Number
Verb
of
Verb
Mean
Age of Onset
Across Children
Children
Bite
Rank
Order
7 592.71
44.29
10
26.38
17
34.00
16
57.20
47.14 5
Fall 5 614.20
55.60 8
5741.60
34
40.60
Go 7 589.86
Hold 6663.67
Jump7643.14
Kiss 6661.33
26.00 2
23
23.00
15
61.00
19
50.50
Lay 4 599.25
Like 6 710.83
27.00
33
33.00
7 626.29
12
17.57
Move 3 696.00
30
35.67
Need 6638.33
14
27.67
Open 7 587.57
Pull 5671.20
17.29 1
25
77.60
Push
14.83
6 608.17
48.40
22
6 680.50
30.67
28
26
38.29
6 612.17
29.57
Sit 7616.57
11
13.83
Stop 5 700.00
Take 5661.80
Throw
5 663.00
Run 7672.71
See
to
27
65.86
Cut 7 645.57
Roll
of Days
10 Instances
Drop 5 645.20
Eat 7 604.71
Put
Number
Reach
24
11.50
Cry 7680.57
Look
Mean
32
47.33
Clap 2 668.50
Come 8 616.50
Give
of
Acquisition
3 704.67
Bring
by Verb
Development,
31
25.60
38.80
20
614.20
35.60
Walk7637.43
13
23.71
Want 8651.13
18
43.75
Wash 8 662.13
21
29.88
Wave 2 687.00
29
72.00
that verb first appeared

in the speech of the children who produced
it,
=
= ?
.39, p<.0b.
Thus, verbs that were produced
r(n
34)
early were the
verbs that were produced
by more children within the span of the study.
were
the
verbs
that
21
Also,
(bite,
produced
by at least six of the children
come,
cry, cut,
eat, go,
hold,
jump,
kiss,
like,
look,
need,
open,
push,
10 instances of use in an average

walk, want, wash) reached
= 14.42
the 13 verbs produced
whereas
(SD
days),
by fewer
run,
roll,
sit, see,
of 33.24 days
than six of the
34
OF EARLY
VERBGROWTHAND USE
children
reached
(bring, clap, drop, fall, give, lay, move, pull, put, stop, take, throw, wave)
10 instances of use in an average of 44.07 days (SD = 18.52), a
reliable difference,
Cohen's
one-tailed,
?(32) =1.91,
p<.05,
statistically
In
0.67.
those
sum,
verbs
that
were
the most
widely
acquired
were
also
the earliest acquired.

The mean age of first use for light (M = 645.43 days, SD = 96.40) and
=
=
83.19) did not differ, nor did these verb
heavy verbs (M 635.13 days, SD
in mean
number
of days to reach
10 instances
subclasses
differ
(M
=
=
=
=
tran
M
SD
SD
31.82,
18.08;
36.30,
23.74).
Similarly,
[light]
[heavy]
sitive (T), intransitive
(I), and alternating
(A) verbs did not differ in their
=
=
mean
SD= 114.53, M
663.85,
[I] 629.61,
ages of first use (M [T]
=
=
=
SD
90.87, M [A] 626.89, SD
81.33) nor in their mean number of days
to reach 10 instances (M [T] = 32.28, SD = 20.39, M [I] = 31.9, SD = 24.17,
M [A] = 37.19, SD = 24.37). Thus,
the data indicate that the variability on
these
measures
of
the
timing
of
emergence
and
the
rate
of
early
use
was
not
the heavy-light
distinction among verbs or to
systematically
the transitive, intransitive, and alternating distinction
among verbs.
at least one grammatical
All children produced
subject with at least one
=
were
of
the
verb
tokens
verb, but only 32% (SD
22%)
produced with overt
related
subjects.
Overt
direct
to either
objects
were
more
common,
with
produced
an
av
=
at least two
25%). Each child produced
erage of 46% of the tokens (SD
=
tokens of subject-verb-object
with
24%
frames,
(SVO)
(SD
19%) of tran
sitive or alternating
verb tokens in SVO frames. Similar findings were ob
tained when calculations were performed
by verb: 35% of verb tokens were
=
overt
with
20%), 45% of transitive or alternating
produced
subjects (SD
=
verb tokens were produced
with overt objects
(SD
30%), and 24% of
were
verb
tokens
in SVO
frames
transitive/alternating
produced
=
then, the children more
(SD
21%). On average,
frequently
produced
verbs with direct objects than with subjects, t(7) = 1.85, p = .052, one-tailed,
d = 0.59, and transitive and alternating
verbs appeared with direct objects
more
than they appeared with subjects, t(24) = 2.36, p = .013
frequently
d = 0.39. Six of the eight children,
and 18 of the 25 relevant
one-tailed,
verbs, followed this pattern. As expected, uses of verb inflections and prep
ositions were much
less frequent, with inflections included with 6% of chil
=
= 7% and
dren's
verb tokens
calculated
9%; when
(SD
by verb, M
=
SD
included with
11% of their verb tokens
11%) and prepositions
=
=
=
5%; when calculated by verb, M
24%, SD
(SD
25%).
As did the age of onset and rate of production measures
for these verbs,
the measures
dren
of
grammatical
use
revealed
wide
variation
among
the
chil
and
the verbs. The percent

of verb tokens produced
with
among
subjects varied from 74% (Carrie) to 1% (Ned), the percent of transitive or
verb tokens produced
with direct objects varied from 77%
alternating
verb
(Stacey) to 17% (Sam), and the percent of transitive or alternating
35
in SVO frames ranged from 59% (Carrie) to 2% (Ned).

tokens produced
Verbs also varied in the percent of tokens produced with subjects, from 70%
(wave) to 0% (look, stop), transitive and alternating verbs varied in the percent
of tokens produced with direct objects, from 94% (want) to 0% (jump, lay),
in SVO frames ranged from 73% (want)
and the percent of tokens appearing
to 0% (five verbs: jump, lay, stop, push, open). Variation
in morphological
and
was also observed: Only five of the children produced
use
any
preposition
verb tokens with the "-ing" suffix and the percent of such tokens ranged
from 23% (Carrie) to 1% (Ned). Fifteen of the 34 verbs appeared with
"-ing," and the percent of such tokens ranged from 0% (9 verbs) to 37%
included 4 uses of third person singular "-s"
(cry).Other verbal morphology
10 uses of the past tense (three children
children
and
(two
contributing)
the
children
verb tokens with prepositions
of
Six
produced
contributing).
and the percent of such tokens ranged from 17.7% (Sam) to 4.6% (Carrie).
Prepositions were produced with 15 of the 34 verbs, and the percent of such
the children's first verb
tokens ranged from 90% (lay) to 1% (look). Thus,
were
not
structures and typically
utterances
full
SVO
containing
frequently
utterance
the
with
consistent
lacked morphological
inflection,
length lim
nature
of early
with
this
and
the
characteristic
of
itations
age
telegraphic
was
a
in
such
the
chil
the
data
that
also
trend
There
word combinations.
dren who were older when they began using these verbs used them more
in full SVO frames than did the earlier verb users, r(n = 8) = .63,
frequently
=
.09.
7 displays the percent of tokens including subjects, objects, prep

ositions, "-ing," SVO frames, and verb-only utterances used with light and
to the timing of
heavy verbs. In contrast to the null findings with respect
Table
emergence
and
rate
of
production
measures,
these
measures
of
grammat
between
these two categories of
ical frame use revealed several differences
more
in SVO frames
verb. Light verbs appeared
frequently
significantly
less likely to appear
than did heavy verbs, and light verbs were significantly
both verb subclasses included verbs with
in verb-only utterances. However,
few verb-only uses (fewer than 30% of tokens [light verbs in boldface]: bring,
clap, cry,fall, give, hold, lay, like, need, put, take, want, wave), many verb-only
uses (>60% of tokens: bite, go, move, pull, push, see, stop, throw, walk), and
intermediate
verb-only uses (between 30% and 59% of tokens: come, cut, eat,
in the analyses both by chil
kiss,
look,
pull, roll, run, sit, wash). Trends
jump,
dren and by verbs suggest that "-ing" appeared more frequently with heavy
than with light verbs.
Table 8 displays for the transitive, intransitive, and alternating verbs the
"-ing," and SVO
percent of tokens used with subjects, objects, prepositions,
frames. Children used transitive verbs with both objects and SVO frames more
the by-verbs analysis revealed
than alternating verbs. Moreover,
frequendy
both
the
for
of
verb
effects
type
percent of tokens with subjects and
significant
36
OF EARLY
VERBGROWTHAND USE
TABLE7
of
Properties
Grammatical
Light
Item
Grammatical
I. By children
with
Tokens
Heavy
Verb
Use
Heavy
SD
Means
(Proportions)
and
Light
Means
df*
-0.36
SD
Cohen's
subjects
.33
.22
.34
.21
ns
.75
.23
.38
.24
5.31
.001
1.57
with
objects
SVO
.42
.30
.19
.16
2.88
.02
0.96
"-ing"
.03
.05
.08
.11
ns
.09
.10
.07
.07
1.09
ns
.25
.21
.40
.21
5.14
.001
subjects
.37
.24
.34
.19
0.37
32
ns
.73
.17
.38
.28
2.64
23
.01
.43
.23
.23
.26
1.52
23
ns
.04
.08
.08
.12
32
ns
.12
.19
.10
.21
.24
.24
.42
.26
Tokens
with
Tokens
Tokens
with
Tokens
with
Tokens
verb-only
//. By verbs
with
Tokens
Tokens
with
Tokens
with
objects
SVO
Tokens
with
"-ing"
Tokens
with
Tokens
verb-only
Note.?The
df decrease for the % tokens with objects and with SVO
because only the transitive and alternating verbs are included.
-1.46
-1.86
0.17
1.76
0.71
1.32
32
ns
32
.087
0.70
by items
(subject-verb-object)
the percent
of tokens with "-ing": Transitive verbs appeared with subjects more

than
frequently
alternating verbs, and intransitive verbs appeared with "-ing"
more frequendy
than either transitive or alternating verbs.
In sum, this group of typically developing
status
middle
socioeconomic
children
produced
utterances.
ports
to use
began
18 months
some
verbs
in their
spontaneous
speech
at an
average
the following
3-11 months,
(560 days). Within
an average of 25 of the
common
34
verbs
targeted
The
nature
of
generalizations
these
children's
about
early
first
verb
uses
of
development.
their
age
of
the children
in at least 10
first
Contrary
verbs
sup
to
the
picture of idiosyncratic development

suggested by studies that only sampled
children's
speech (see chapter I), these diary data indicate that some verbs
are consistently
earlier than others, and these early-appearing
acquired
verbs are also used by more children.
(We can reject sampling probability as
cause
a
the
of
these two variables because of the
spurious relation between
recorded all verb uses.) Children
unique nature of this data set?mothers
who produced
their first verb earlier tended to be faster learners overall.
at least some of their first uses of first verbs in
The children produced
multiword
utterances, but they varied widely in the percent of verb uses that
were in word combinations.
Consistently,
though, verbs were more likely to
be combined with direct objects than with subjects and combinations
with
prepositions
and
morphology
use
were
rare.
37
for
Negligible
Tokens
P
with
all
types
verb
2.03
22
4.94
.14
.13
.001
.25
.51
SVO
Tokens
with
Note.?The
the
df decrease
intransitive
for
the % tokens
with
Properties
Grammatical
Intransitive,
Transitive,
and
Alternating
Use
Verb
of
TABLE8
versus
objects
Transitive
Intransitive
Alternating
transitive
post hoc
and with
comparison
SVO
yielded
by
(subject-verb-object)
a
significant
2.31
0.23
.13
.23
.10
.19
.18
.09
P
Tokens
with
ns
Tokens
with
"-ing"
.03
.01
.16
.15
.05
2.31
5.99
.08
.006**
2.31
.17
4.50
.26
Tokens
.01**
.20
.36
.18
subjects
with
.49
Tokens
Grammatical
with
Item
Tokens
objects
(Proportions)
2.21
2.24
.06
Tokens
.14
.11
.01
"-ing"
with
.01
.10
ns
Tokens
with
SVO
.35
.28
.12
.11
3.70
7.007
1.08
objects
Tokens
with
subjects
.40
.24
.30
.22
.27
.21
.78
2.21
ns.27
.64
.31
with
SD
Means
.27
.72
.21
.20
SD Means
items
subjects,
%
with
tokens
/(22)
difference,
transitive
hoc
yields
%
d
with
**For
post
alternating
significant
versus
comparison
3.13,
1.32.
tokens
For
jf?
.004,
a"-ing,"
=
SD Fit
.20
df*
because
t(\7)
= ? 2.83,
the
difference,
5.13
p Cohen's
22 5.33
.001
Means
p2.43,
*(23)
comparison,
d=
.02,
.98.
I.
By
children
only
By
//.
verbs
CO
00
.001
OF EARLY
VERBGROWTHAND USE
pattern of combinatorial
class. Light verbs were produced
This
SVO
and
frames,
heavy
verbs
verb use differed

with more objects
were
produced
somewhat
and more
with
more
by verb sub
in
frequently
"-ing"
forms.
that light verbs would serve as "path

However,
contrary to the hypothesis
we
to grammatical
found no evidence
that light verbs
breakers"
structures,
were produced
than
verbs.
Of the 11
earlier in children's
speech
heavy
some were light
before the age of 620 days (on average),
verbs produced
(come, go), but others were not (throw, sit, see, push, open, lay,fall, eat, bite). And
in verb-only utterances,
supporting
light verbs did appear less frequently
that light verbs are the products
Snedeker and Gleitman's
(2004) suggestion
Further analyses of this
rather than the engines of grammatical
acquisition.
V
in
and
will
be considered
VI, when we discuss the
chapters
hypothesis
as
a group and individually.
flexibility of light and heavy verbs
grammatical
verbs appeared with subjects, objects, and SVO frames
verbs did, while
intransitive verbs ap
than alternating
than either transitive or alter
peared with "-ing" forms more frequently
these patterns of differences
among
nating verbs. We further investigate
children and among verbs in chapter VI.
this general picture of verb emergence
and growth in
Having painted
these eight children's use of their first verbs, we turn now to a more detailed
Finally, transitive
more
frequently
description
meanings
of
that
the communicative
they
functions
these
verbs
served
and
the
expressed.
39
IV.PRAGMATIC
INEARLY
VERBUSE
In this chapter, we present analyses that address the

questions of how
extendable
first verb uses are in terms of function and meaning. The model
of the child as a conservative
language learner suggests that children should
restrict the functions of their first verbs to those they have heard in input
and thus, to the degree that input illustrates only some functions children's
verb uses, should show a restricted range of functions for each verb when it
et al, 2006; Tomasello,
(Golinkoff et al, 1995; Maguire
1992).
The conservative-child
model also suggests that children should initially use
some verbs with reference
to only the event that verb was first
paired with.
is first used
The
that
evidence
verb
use
in maternal
to children
speech
is
frequently
tied
to particular
settings (Naigles 8c Hoff, 2006) suggests that a conservative
would be similarly restrictive
learner
in verb production.
Such a
language
conservative
in
only
child
a ball
throw
or
reference
for
might,
example,
"eat"
in reference
only
to themselves
eating
use
to
but
"throw"
eating
not
only
cookies
their
as
but
parents.
a command
not
pizza,
If children's
to
or
initial uses of their early verbs are underextended

in the same fashion
as early noun uses, we should expect to find the
diary records of the chil
dren's initial verb instances replete with identical entries down the 10 rows.
are extended
In contrast,
if verb meanings
and functions
almost from
the beginning,
then we should expect to find the diary records containing
a varied set of entries even within
the first 10 uses of children's
very
verbs.
early
To
test
these
predictions,
we
describe
the
pragmatic,
semantic,
and
action flexibility
that the children demonstrated
in their first 10 uses of
course
the target verbs they produced
the
of this study, and we
during
measures
the
of
for
and
compare
flexibility
light
heavy verbs, and for tran
sitive,
intransitive,
and
alternating
verbs.
40
INEARLY
PRAGAAATIC
VERBUSE
PRAGMATIC
FLEXIBILITY
and Descriptions
Commands
The
pragmatic
of
categories
and
"command"
"description"
together
utterances
in the diary records;
for 98.6 % of the verb-containing
in commands
in descriptions.
and 53.2% were
45.4% of verb uses were
are in line with other findings
These proportions
for this age group (e.g.,
2008). All of the children used at least
Vasilyeva, Waterfall, & Huttenlocher,
account
one
verb
sorts
in both
of
utterances.
On
the
average,
children
produced
with 53% of their verbs (SD = 13%); the

and descriptions
commands
commands only
range was from 33% (Stacey) to 71% (Carl). They produced
with an average of 21% of their verbs (SD = 16%) and descriptions
only with
an average of 25% of their verbs (SD = 14%). Four children used more verbs
than only in descriptions,
three children used more
only in commands
verbs only in descriptions
than only with commands,
and one child used the
same number of verbs only with commands
and only in descriptions
(see
children
tended to use language for different
Figure 2). Thus, different
both
communicative
functions;
however,
all
were
children
able
to use
at
least
some of their verbs in utterances

functions.
serving different
to
in
and others in
Some verbs were more
be
used
commands
likely
a
utterance
in
verb
both
the
of
children
average percent
using
descriptions;
=
were
was
first
10
Some
verbs
within
their
instances
53%
(SD
27).
types
all
the
children
whereas
used in both utterance
of
take),
types by
(bring,
35tl
30- f%
i^i n
.o
Sam
Stacey
Mae
Carl
Carrie
inB?thc&Di
c/2 0 Commands
Heather
Elaine
Ned
Children
Figure
2.?Communicative
functions
of early
verb
uses.
41
others never were
Seven verbs
(fall and drop were used only in descriptions).
and
(bite, bring, kiss, pull, push, take, and walk) were used both as commands
stop was used only in com
descriptions
by 80% of the children or more,
mands by 80% of the children, and like was used only in descriptions
by 80%
seem to follow from the meanings
and typ
of the children. These findings
in ref
ical usages of these verbs: Fall and drop were used most commonly
erence to the actions of inanimate objects, which are not likely to respond
as like is a verb of internal volition,
it is not
well to commands; moreover,
not
to
tell adults to like
be the subject of commands
(children do
likely
were
tied to specific functions the restricted use
things). Thus, where verbs
seems explicable
in terms of the verb's meaning
rather reflecting context
on the part of the child. The majority of verbs were
bound representation
and descriptions
used as both commands
by the majority of children, and
the second function ap
when verbs were used to serve both functions,
use of that verb (M = 4.32,
on
of
the
fourth
instance
average
by
peared
16.91 days (SD = 11.83).
SD = 0.89) and within
Some differences
light and heavy verbs in terms of
emerged between
these pragmatic properties. The children were more likely to use light verbs
than heavy verbs, Ms = 0.39 (0.30) and 0.14 (0.12), re
only as commands
=
=
=
1.09, and less likely to use light
.05, Cohen's d
spectively, ?(7) 3.01, p
=
0.08 (0.13), M (heavy) = 0.30 (0.17),
verbs only as descriptions, M (light)
=
=
= 1.92. Similarly, light verbs tended to be
.05, Cohen's d
3.00, p
t(7)
used
more
than
frequently
heavy
verbs
only
as commands,
Ms
0.33
(0.27)
=
=
and 0.16 (0.25), respectively,
.10, Cohen's d
0.67, and less
/(32)
1.66,p
= 0.32
=
as
M
M
0.10
(0.09),
(heavy)
descriptions,
(light)
only
frequently
=
=
= 0.79.
.10, Cohen's d
1.92, p
(0.31), t(32)
Flexibility (ForCommand
Addressee
When
more
using
than
one
Utterances Only)
to command,
verbs
addressee;
however,
used at least one verb with
all children
there
again,
was
considerable
variability
in the addressee flexibility

Figures 3a
among children
they demonstrated.
and 3b show the mean percent of verbs with which the children used more
in
the first change
instance at which
and the mean
than one addressee
flex
addressee
children
the
took
On
addressee
average,
displayed
place.
=
21%). For those verbs that were used
ibility with 50% of their verbs (SD
so on average before the fifth instance (M = 4.72,
flexibly, the children did
= 10.03
=
on
SD
0.58), and
average within 13.29 days (SD
days) of the first
instance.
Within
the
first
10
instances
of
a verb's
use,
the
number
from 1 to 4 (M= 1.86, SD = 0.66). Most of

dressees
ranged
from at least
were
with more than one addressee
used
(81.25%)
and
six
verbs
however,
(wave, like, jump, cry, clap,
bring) were used
were
on
used with
one addressee
Verbs
child.
each
average
by
of
ad
the verbs
one child;
with only
addressee
42
PRAGMATIC
AND SEAAANTIC
FLEXIBILITY
INEARLY
VERBUSE
:?
100tT^)
90-
80"
11
Sam
llllllli
.
Sam
iiuiiii
Figure
Mae
Stacey
Carl
Carrie Heather
Elaine
Ned
Elaine
Ned
Children
9J(b)
8"
dressee
fl|
Mae
Stacey
Carrie Heather
Children
Percent
of verbs
in early verb use.
3.?(a)
flexibility
Carl
used
with
addressee
flexibility
and
(b) onset
of ad
=
flexibility by 41% of the children (SD 26%). Verbs in commands were used
with a different addressee from the first instance within, on average, 14.47 days
of the first use (SD = 19.96 days) and by the fifth instance of the use of that verb
=
the measures
of addressee
(M 4.36, SD= 1.25). When
flexibility are com
none
for
the
and
of
the
subclasses,
pared
light
heavy
analyses by child yielded
one
differences.
The
verb
difference: Light verbs
significant
analyses by
yielded
were used with significandy more addressees than heavy verbs, M (light) = 2.04
= 0.98.
= 1.58
=
=
.05, Cohen's d
(.55), M (heavy)
(.44), *(30) 2.36, p
In
do
not
sum,
when
seem
to
verbs
uniformly
first
enter
restrict
children's
these
verbs
productive
to a
single
lexicons,
utterance
children
function,
nor
to a single addressee. They use about half of their verbs in both com
mands and descriptions,
and they direct their commands, with about half of
their verbs, to multiple
addressees. One limitation of change of addressee as
a measure
of children's flexibility of verb use is that, over the course of a day,
43
have only a limited number of possible addressees: mother,

and
father. Thus,
children who reach their 10 instances of a
sibling, pet,
be
verb
less
addressee
given
quickly might
likely to demonstrate
flexibility
these children
because
fewer
different
are
addressees
available.
FLEXIBILITY
SEMANTIC
Action Flexibility
4a and 4b show
Figures
to more
reference
in action
change
more
one
than
referent
child
On
the percent
the mean
and
occurred.
than one action with
|
fev?i
Q
for each
action
average,
38% of their verbs
children
(SD
in
of verbs used
at which
instance
the
reference
made
first
to
19%). For those verbs
100-A
(a)
90 - K)
80
70-
I , 60-
mm
^ Sam
Stacey
Mae
Carl
Carrie
Heather
Elaine
Ned
Heather
Elaine
Ned
flexibility
and
Children
io -ri I
9- 0>)
iliililk
Figure
flexibility
Sam
Percent
4.?(a)
in early verb use.
Stacey
Mae
Carl
Carrie
Children
of verbs
used
with
action
(b) onset
of action
44
PRAGAAATIC
INEARLY
VERBUSE
showing
action
children's
flexibility,
first
new
action
use
occurred
av
on
=
=
15.54 days
0.85) and within
erage before the fifth instance (M 4.6, SD
=
first
after
instance.
the
The
of
number
actions
verb
within
(SD
13.50)
per
the first 10 instances ranged from one to four; six of the eight children
to at least three different
referred
actions with at least one verb. Across
=
an
of
of
the
children
33%
verbs,
(SD
20%) referred to more than
average
one action within the first 10 instances of
a verb. Come was used to
producing
refer to more than one action by six of the children; put and open were used
to refer to more than one action by five of the children. The first
change in
=
on average
action reference
occurred
the
fifth
instance
(M 5.1,
by
SD = 1.99) and on average within 27.14 days of the first use (SD = 35.77).
Actor Flexibility
5a and 5b show for each
Figures
reference
change
to more
in actor occurred.
Sam
actor
child
Stacey
Mae
the percent
the mean
and
All children
iluiiii
10
fl*> I
used
Carl
in
of verbs used
at which
instance
the
first
at least 50% of their verbs
Carrie Heather
Elaine
Ned
Carrie Heather Elaine
Ned
in
Children
<a>
te
8 -
Uilllll
Sam
ibility
one
than
Figure
5.?(a)
in early verb
Percent
use.
Stacey
Mae
Carl
Children
of verbs
used
with
actor
flexibility
and
(b) onset
of actor
flex
45
to more
reference
one
actor.
On
average,
children
made
reference
to more
than one actor with 73% of their verbs (SD = 13%). For those verbs showing
on average before
actor flexibility,
the first new actor use occurred
the
fourth instance (M = 3.9, SD = 0.77) and within
15.79 days (SD = 12.06)
after the first instance. The number of actors per verb within the first 10
instances ranged from 1 to 5 (M = 2.68, SD = 1.0). Across verbs, an average
=
of 76% of the children
than one actor within
20%) referred to more
(SD
a verb; eight of the verbs were used with
the first 10 instances of producing
actors by all the children. The first change
in actor reference
different
occurred on average before the fourth instance (M = 3.85, SD = 1.01) and
on average within
16.91 days of the first use (SD = 13.91).
percent of children's verb uses that referred to themselves as actors
from
11% (Stacey) to 70% (Carl). On average, 51% (SD = 18.1%) of
ranged
children's first 10 uses of these early verbs concerned
themselves as actors.
as
uses
actors ranged from
to
of
the
that
referred
self
verbs,
percent
Among
in
0% for the verbs cry, look, move, open, and pull (the latter four primarily
was
to
want.
for
used
exclu
the
verbs
and
100%
like,
commands)
Cry
clap,
to infant siblings or fictional characters. Mean percent of
sively in reference
self as actor uses calculated across verbs was 50.8% (SD = 30%).
The
AffectedObject Flexibility
used
6a
Figures
in reference
and
6b
to more
show
the
than
of
percent
one
affected
object
verbs
alternating
at
instance
the mean
and
transitive
and
their first change in affected object reference.

produced
re
verbs, of course, do not have affected objects.) Children
(Intransitive
affected objects on average with 89% of their verbs
ferred to different
=
(SD
9.8%). For those verbs that were used flexibly, the children referred
to a new affected object on average before the fourth instance (M = 3.33,
11.68 days (SD = 7.93) of the first instance. The
SZ) = 0.79) and within
to within the first 10 instances ranged
number of affected objects referred
of children who made
from 1 to 8 (M = 4.69, SD = 1.77). The proportion
to more
than one affected object averaged
reference
79% across verbs
=
verbs were used with
3l). Twelve of the transitive and alternating
(SD
different affected objects by all children, whereas only two verbs (jump, lay)
The
appeared with only one affected object in every child's productions.
on average before
the
in affected object reference
occurred
first change
fourth instance (M = 3.34, SD = 0.97) and within
12.40 days (SD = 7.61).
The above descriptions
(and compare Figures 5a, 5b and 6a, 6b) sug
gest that the children displayed a higher degree of affected object flexibility
for two of the
this difference
than actor flexibility. Paired t tests confirmed
a significant
tests
t
the
revealed
three flexibility measures:
children,
By
in the percent of verbs showing actor compared with affected
difference
which
the children
46
INEARLY
PRAGMATIC
VERBUSE
iooji(a)
m m
Sam Stacey
Mae
Carl
Carrie
Heather
Elaine
Ned
Carrie
Heather
Elaine
Ned
Children
lllillll
10
iTcw
9
Qt
^ ^
llllilU
Sam Stacey
Figure
Percent
of verbs
6.?(a)
affected
in early verb
object flexibility
Mae
Carl
Children
used
use.
with
affected
object
flexibility
and
(b) onset
of
=
=
=
.0005, Cohen's d
1.39, and a significant
?(7) 6.02, p
object flexibility,
actor was produced
in when
difference
the first different
compared with
=
=
when
the first different
affected object was produced,
.01,
2(7) 3.42, p
=
d
0.73. Comparison
of the number of days to actor and affected object
=
d = 0.40.
t(7)
change showed a trend in the same direction,
1.599,/?<.10,
are
verbs
the t tests
included),
By verbs (only the transitive and alternating
t
revealed significant differences
for the mean
instance of change measure,
=
=
=
2.16, p
.04, d
0.62, and the number of days to first change mea
(22)
=
=
=
of children
sure, 2(22)
2.78, p
.01, d
0.37, but not for the percent
of freedom differ in these
2(24) <1, ns. (The degrees
showing flexibility,
analyses because two verbs displayed no affected object flexibility.)
We next compared
the different verb subclasses in terms of actor and
a significantly higher percent
affected object flexibility. Children produced
47
to different affected objects than that they

of their light verbs with reference
=
=
did for heavy
SD = 0.06, M
0.98,
verbs, M
0.81,
(heavy)
(light)
=
=
=
=
SD
0.20, t(7)
2.29, p
1.15, and light verbs tended
.055, Cohen's d
to be produced with different actors by a higher percent of children than
did heavy verbs, M (light) = 0.87, SD = 0.18, M (heavy) = 0.72, SD = 0.19,
=
= 0.80. There were no differences
=
1.88, p
.068, Cohen's d
?(32)
among
the categories of transitive, intransitive, and alternating
verbs in actor flex
transitive and alternating verbs in affected
ibility and no difference between
object flexibility.
in early
These data reveal a picture of pragmatic and semantic flexibility
verb use. For example,
these children were very likely to use eat in reference
to multiple
eaters. Eating, for them, was not a restricted
foods and multiple
a
to
that
category
single type of situation but one that applied across
applied
foods
different
(at least)
being eaten (which are, of course, eaten in different
are
ice cream is licked, cereal is slurped) and dif
chewed,
ways?cookies
ferent people and animals doing the eating (showing that the children were
to an inclusive rather than
able to generalize across different mouth motions
for eat). This picture suggests that children's first verb
restricted meaning
uses differ from their first noun uses, which can be quite inflexible (Dromi,
1987; Harris et al., 1988).
use
The present findings also differ from other descriptions

of early verb
et al.
to the predictions
in the literature. Contrary
of Huttenlocher
children's
(1983),
their own
someone
very
actions
else's
first
uses
verb
or relations;
actions
as
were
own.
their
not
were
children
just as likely
more
Interestingly,
to
in reference
consistently
to talk about
was
flexibility
ob
that this is be
served with affected objects than with actors; we conjecture
cause there is a greater range of possible affected objects in a child's home
items) than of possible actors (e.g., animates).
(e.g., toys, food, household
more varied affected objects than heavy
were
verbs
with
Light
produced
verbs, which is consistent with the notion that light verbs apply to a wider
than heavy verbs (usually,
range of items (one can bring just about anything)
one eats only food, kisses only animates, and rolls only round things; see also
8c Sethuraman,
Goodman
8c Gleitman,
2006; Snedeker
2004; Valian et al.,
and
these
of
the
functions
semantics of early
Overall,
then,
2006).
analyses
use
verb
are
not
nicative
suggest
to
that
when
tied
particular
to refer
functions,
and relations
they denote
verbs
uses
to a
enter
but
are
variety
with multiple
children's
available
of
actions,
agents
to
productive
serve
and
lexicons
multiple
to refer
to
and affected
they
commu
the
actions
objects.
48
IN
AND GRAMMATICALFLEXIBILITY
V. PRODUCTIVITY
EARLY
VERBUSE
In this chapter, we present analyses that address the question of how

and flexible children are in their initial uses of
productive
grammatically
model predicts
that children
The
conservative-child
verbs.
early-acquired
structures they have heard
will use their verbs only in the morphosyntactic
used with that verb. Each verb might have its own range of structures, but
than the full possible range allowed by the
that range would be narrower
their first verbs as items in
adult grammar. If, in contrast, children represent
an abstract category, with all the morphosyntactic
of occurrence
privileges
allowed by the adult grammar, then in early verb use, each verb should be
each structure should be used
used in a wider range of structures?and
the rapid-generalizer
model
with a wider range of verbs. Thus,
predicts
the conservative-child
verb use, whereas
and productive
early flexible
should
be achieved
and
model
that
only
flexibility
productivity
predicts
or years of input (see Tomasello's
claim
that children
after months
[2000]
with the subject-verb-object
[SVO] frame at 3
only become
productive
of verbs
after
have
heard
thousands
of
222-223],
age [pp.
years
literally
they
used in SVO in their input).
To index the productivity
of children's
syntactic frames, we calculated
verbs used each of the following
for each child the number of different
or particle),
and V-ing. Then,
frames: SV, VO, SVO, V+P
(preposition
we
tabulated
the number
and
Shirai
(1998),
(1981,
1989)
following Ingram
of frames that were used with at least three verbs and the number of frames
used with at least five different
percent
the
verbs. We
also calculated
of all of each child's verb types used
three-verbs-per-frame
(3-V)
and
for each frame the

the age at which
in that frame,
five-verbs-per-frame
(5-V)
levels
of
were
reached, and the number of frames that reached 3-V and

productivity
of age.
5-V productivity
when children were 24 and 30 months
with
To index the grammatical
which
children
used their
flexibility
of different
frames
verbs we counted
the number
(syntactic flexibility),
inflectional endings
different
(morphological
flexibility), different nouns in
49
nouns
in object
and different
(lexical subject flexibility)
subject position
in the
with
which
each
verb
(lexical
object flexibility)
position
appeared
10 uses. A frame was coded as different from previously appearing
recorded
frames if there was an addition or subtraction of a subject, object, locative,
term. Morphological
differences
included the
and/or negation
preposition,
or
addition
subtraction
of
or
tense/aspect
person
on
marker
the
verb.
in the lexical item used in the

involved changes
subject differences
Lexical
involved
in the lexical
differences
subject position.
object
changes
differ from the actor and
item used in the object position. These measures
in the previous
affected object flexibility measures
discussed
chapter be
cause those measures
to produce
did not require the children
the nouns
to
actors
whereas
the
lexical
and
affected
the
objects,
subject and
referring
Lexical
measures
object
depend
on
noun
in actual
changes
use.
INFIRST
VERBUSES
GRAMMATICAL
PRODUCTIVITY
Table 9 displays the mean
across children. The children
in each frame
of verb types produced
SV and VO frames with > 60% of
produced
percent
all the target verbs they used, and the SVO frame with just under half (40%)
these are not frames used with just 1 or 2
of the verbs they used. Thus,
the V+P and V-ing frames with <20% of
verbs. In contrast, they produced
in chapter III, these frames were
the target verbs they used; as described
used less frequently
overall. The children varied in the degree of gram
and Carrie were consistently
matical productivity
they displayed. Heather
the
most
productive
children
across
the
frames;
who
child
was
least
TABLE9
Productivity
of Verb
Use
Measures
Frame
SV
Mean % of V types (SD)
Highest
% of V
types
Highest child (# Vs)

Lowest
% of V
types
Lowest child (# Vs)
61 (28)
93.5
VO
62 (29)
100
SVO
VP
41 (26)
18 (15)
Ned (2)
15 (14)
3571.4
35
Carrie (29) Heather (21) Carrie (15)

14.3
V-ing
Sam (11) Heather (11)
40 0
29.4
Carl (5)
Sam (1)
Carl,
Carl
Elaine
Mean
age of 3-V
21.83(7)
22.28(8)
23.64(6)
22.45(3)
21.11(4)
(n)
productivity
Mean
age of 5-V
22.6(7)
23.1(7)
22.13(4)
23.43(5)
21.35(2)
productivity
(n)
50
PRODUCTIVITY
AND GRAMMATICAL
FLEXIBILITY
INEARLY
VERBUSE
for a given frame varied across four different

children
(Carl,
productive
Elaine, Ned, Sam). Sam was clearly the most variable child, showing the
most productivity
with the V+P frame and the least with the SVO frame.
Table 9 also shows the children's mean age when they achieved
the 3-V
and 5-V levels of productivity
for each frame. All of these mean ages are
than 24 months. The number of children
included is fewer than
younger
seven or eight for the SVO, V+P, and
V-ing frames, indicating that these
frames manifested
less productivity
within the span of the study. The chil
dren also varied in the number of productive
frames (out of five) achieved at
24 and 30 months
(see Table 10). At 24 months, most children had achieved
at least 3-V productivity with the SVand VO frames (Ned left the
study at 20
so we know the least about his
to productivity).
months,
Carrie,
progression
and Stacey achieved 5-V productivity
before 24 months with al
Heather,
most all frames, whereas Carl achieved
less productivity
later, closer to 30
months.
Elaine, Mae, and Sam occupy the middle ground, having achieved
3-V productivity
with four or five frames by 30 months.
Because we do not
have records of the children's uses of their early verbs past the first 10
scores are very
to underestimate
the number
instances, the 30-month
likely
of frames used with multiple
verbs.
Researchers
have proposed
that children's
linguistic relations "go pro
ductive
in free combination
with other
[when] the pattern can participate
lexical items that satisfy its typed variables"
2004,
pp. 188
(Jackendoff,
cannot
be
189). Such "free combination"
usually
unequivocally
supported
from spontaneous
speech data, and we do not claim that we have found it
as operationalized
here. However, productivity
by Ingram (1989) and Shirai
(1998) can be addressed with spontaneous
speech data, and we have found
evidence of productivity
defined
this way in these children's
speech. Well
36 months
before
of age?and
even
before
months
24
of
frequently
TABLE 10
of
Number
At
Ned
Stacey
Frames
24 Months
Name
3-V level
Carl
Carrie
Elaine
Heather
Productive
0
5
55
14 2
5
55
Mae 0 0 4
10
Sam3 3 4
5
45
(Out
of
Five)
At
5-V
level
30 Months
3-V
level
5-V
level
11
5
2
5
4
10
3
4
51
children used a number of syntactic

age?the
thus showing that frames are not necessarily
frames with multiple

to specific
restricted
verbs,
lexical
items.
FLEXIBILITY
GRAMAAATICAL
INFIRST
VERBUSES
All eight children used more than one syntactic frame within the first 10
uses of at least one of their verbs. Figures 7a and 7b display for each child
the mean
instance, across verbs, of the first new frame and the percent of
verbs used with more than one frame. On average, the children used 65.6%
of their verbs (SD = 18%) in at least two different frames within their first 10
= 1
uses; their mean number of different frames per verb was 2.14 (range
this early period did not seem to be one characterized
5). Thus,
by frozen
100T71-1
- (a)
&??
90
-7AI
H ^m
^1 ^1 ^H ^H ^1 ^1 ^H ^H
^1 ^1 ^1 ^1 ^1 ^1 ^1 ^1
$3
Sam
Carl
Mae
Stacey
Carrie Heather
Elaine
Ned
Children
10-,-,
(b)
\j i
lllllii.
Sam
Figure
flexibility
Percent
7.?(a)
in early verb use.
Stacey
of verbs
i
Mae
Carl
Carrie
i
i
i
Ned
Heather Elaine
Children
used
with
syntactic
flexibility
and
(b) onset
of syntactic
52
AND GRAMMATICAL
FLEXIBILITY
INEARLY
VERBUSE
PRODUCTIVITY
form use. For those verbs that showed flexibility, the first new use occurred
on average just after the fourth instance (M = 4.17, SD = 1.47) and within
15.37 days (SD = 12.58) from the first instance. Verbs varied in the percent
of children who used them in different
frames, from 100% (fall, give, take,
wave) to 0% (stop). On average, verbs were used with different frames by 66%
of the children (SD = 23%); the mean number of different frames per verb
was 2.20 (SD = 0.55). For those children who produced
different
frames,
on average before
use occurred
the first different
the fourth
instance
=
17.77 days (SD = 16.96) from the first in
(M 3.81, SD= 1.34), within
stance. Most of these instances of syntactic flexibility
involved the addition
or subtraction of subjects, objects, prepositions,
and locatives. Only 12% of
frame changes
included the use of negation
(range 0-36%). On average,
included negation.
1.8% of utterances
was included in our mea
The addition or subtraction of a preposition
sure of syntactic flexibility; however,
this measure
captured only whether
the preposition
slot was filled. It did not capture whether
the children used
more
than one preposition
with a given verb and/or a given preposition
with more
than one verb. Examination
of the particular prepositions
used
showed that of the six children who used prepositions
with their verbs, five
showed flexibility
in their use of prepositions.
That is, Ned used only one
one
with
verb
but
Mae,
Heather,
Carrie, Stacey, and
(down)
(lay),
preposition
Sam each used at least one verb with multiple
and/or used the
prepositions
same preposition
with multiple
verbs. For example, Mae used put and pull
each with two prepositions
and used in and over each with two different
and down,
verbs, and Heather used put and throw each with two prepositions
to, and away each with two different verbs. Sam was by far the most flexible
use: He used lay, give, pull, push, and put each with at least
in his preposition
two different prepositions
and down and in each with at least two different
verbs.
Consistent
with descriptions
of early combinatorial
speech in English
learners as telegraphic,
these young speakers produced
very little inflec
was
tional morphology,
and, not surprisingly, morphological
flexibility
much
than frame flexibility both across children
less pervasive
(appearing
in approximately
and across verbs (appearing
6% of children's utterances)
in approximately
7% of verb uses). Figures 8a and 8b show the percent of
verbs with which the children used more than one morphological
structure
none of their
and the instance of the first change; two children produced
verbs inmore than one morphological
form. On average, the children used
different morphology
with 15.69% of their verbs (SD = 14.5%). For those
use occurred
on average
verbs that showed flexibility,
the first different
=
before
the fifth instance
26.98 days
(M 4.55, SD= 1.91) and within
=
(SD
21.45) after the first instance. Verbs were used with different mor
=
structures by an average of 18% of the children
(SD
22%); the
phological
53
100-n
90
80
3 ?
||
2|
70
60
so
40
il
1
?
oil
Sam
II
i"
Stacey
Mae
i-iM
Carl
i*
Carrie Heather
i-M?t
Ned
Elaine
Children
10-rj
1
Q
ll.
Figure
8.?(a)
morphological
Sam
Stacey
Mae
Carl
ll
Came
Heather
Elaine
Children
Percent
of verbs used
in early verb use.
with
morphological
flexibility
Ned
and
(b) onset
of
flexibility
from 100% of children

(wave) to 0% of children (14 verbs).
range extended
The first new use occurred on average before the fifth instance (M = 4.34,
SD = 2.29) and within 22.78 days (SD = 23.66). Almost all morphological
in the
changes involved the addition of "-ing." A few verbs were produced
was
tense
one
and
with
the
third
wave,
past
produced
(drop,
give, go, fall),
were
more
No
"-s"
verbs
child
with
(need).
person singular
produced by any
than
one
overt
morphological
marker.
(ex
subject flexibility was the norm, with seven of the children
more
one
one
at
than
lexical
with
least
of
Ned)
cepting only
using
subject
their verbs. Figures 9a and 9b show for the children
the percent of verbs
with which they used more than one lexical subject and the instance of the
first new lexical subject. On average,
the children used more
than one
different
lexical subject with 30.65% of their verbs (SD ? 27.15). For those
verbs that showed flexibility, the first new use appeared on average before
the sixth instance (M = 5.2, SD = 1.06) and within 20.29 days (SD = 17.79)
Lexical
54
AND GRAMMATICAL
FLEXIBILITY
PRODUCTIVITY
INEARLY
VERBUSE
100-n-1
tS: c
<?
m B
Sam
Stacey
B
Mae
Carl
Carrie Heather
B_
Elaine
Ned
Children
10-rl
fl
fl
?~
3 -B
Sam
Stacey
Mae
fl
Carl
Carrie Heather
Elaine
Ned
Children
Figure
lexical
subject
9.?(a)
Percent
of verbs used
in early verb use.
with
lexical
subject
flexibility
and
(b) onset
of
flexibility
after the first lexical subject use. Verbs also varied in the percent of children
lexical subjects over the first 10 instances, from 100%
different
producing
(wave) to 0 (throw, stop, push, move, lay). On average, verbs were used with
=
lexical subjects by 35% of the children
(SD
26%). For those chil
multiple
new use on average
dren showing this flexibility,
their
first
they produced
=
=
the fifth instance
before
12.74 days
(M 4.77, SD
2.21) and within
=
16.17) after their first lexical subject use.
(SD
All 8 children used more than one lexical object with at least one tran
10a and 10b show for the children
sitive or alternating
verb. Figures
the
percent of verbs with which they used more than one lexical object and the
instance of the first new use. On average,
the children used at least two
lexical objects with 43.38% of their verbs (SD = 26.56%). For those verbs
that showed flexibility, the first different use occurred on average before the
fifth instance (M = 4.13, SD = 1.22) and 20.64 days (SD = 16.22) after the
first lexical object use. Among
the transitive and alternating
verbs (the only
55
"3 3
ioo -n-1
?t 901
Sam
Stacy
Mae
Carl
Carrie
Heather
Elaine
Ned
Children
*
10T1-1
8
ll
*>o
00
M
II
H
M
H
?
I
H
I
H
Sam
Stacy
I
H
I
H
Mae
I
^
H
fll
H
I
I
m
I
H
I
H
IH
I
H
IH
IH
Carl
Carrie Heather
Elaine
Ned
Children
Figure
lexical
object
10.?(a)
flexibility
Percent
in early
of verbs used
verb use.
with
lexical
object
flexibility
and
(b) onset
of
in the
variation
verb categories
that take objects) there was considerable
over
10
the
first
lexical
of
different
children
percent
objects
producing
to
tran
On
from
0%
100%
instances, ranging
average,
(stop, move).
(bring)
lexical objects in the
verbs appeared with multiple
sitive and alternating
=
(SD
28%). For those children showing this
speech of 46% of the children
on average before the fifth instance
new use occurred
the
first
flexibility,
=
=
=
SD
and
19.06
16.09) after the first lexical
(Af 4.74,
2.06)
days (SD
object
use.
11 and 12 present means,

standard deviations,
and statistical
measures
verb
subclass.
for
of
the
flexibility by
grammatical
comparisons
a
of
revealed
that
children
and
verbs
produced
heavy
light
Comparison
with
lexical
verbs
than
verbs
of
heavy
light
higher percent
significantly
of light verbs, and
object flexibility, consistent with the broader meanings
a significantly ear
at
in
first
lexical
their
also
change
object
they
produced
lier instance for light than for heavy verbs. The children tended to produce
a higher percent
lexical subject
of heavy verbs than light verbs with
was
not
this
difference
statistically significant. Similar
although
flexibility,
were produced
with
in
the
item
obtained
analyses: Light verbs
findings
lexical objects by a significantly greater percent of children than
different
Tables
56
11
TABLE
Grammatical
Use
Flexibility
of
for
Light
Versus
Heavy
Verbs
Mean
SD
Mean
SD
30.5021.34
Light
Heavy
Mean
instance
morphology
change5.133.66
5.73
2.43
Only
3children
contributed
^
t
df*
26.58
16.68
30.66
25.17
20.18
13.04
change
days
3.66
to
frame
change
^7Mean
3.82
0.71
instance
3.54
1.39
0.57
frame
change
ns
5.43
Number
of
days
to
frame
change
15.93
11.58
14.58
13.08
0.6
7ns
g
Only
Only
5.10
0.9
instance
32
1.13
1.23
3.53
1.04
Mean
4.08
frame
change
F=
ns
Cohen's
3 5 children
children
Only
children
contrib
15.31
21.35
NJ Note.?*df
different
with
Number
of of days days
Number
70.37
0.74
different
3.38
1.24
0.22
.01
0.36
with
Vs
Proportion
lexical
objects
O
different
subjects
lexical
Proportion
different
with
Vs
frames
0.72
0.14
0.69
0.21
0.76
7ns
morphology
By
I.children
Mean
instance
lexical
object
change
2.98
0.85
4.8
1.47
-9.19
6.001
1.52
>
contrib
contr
vary because not all children,

experi
verbs,
all
0.23
0.26
0.32
2 0.37
1.01
.02
24
2.24
different
0.26
0.32
0.61
Cs
Proportion
with
lexical
objectsobject
to to
lexical
nor
^-1.73
22.72
days
lexical
617.18
Number
16.43
18.25
of
object
to
change
ns
Number
of
to
days
subject
lexical
change
17.0
19.37
11.1
14.99
0.87
27
ns
m
lexical
different
32
0.34
Proportion
0.19
Cs
0.37
with
0.29
subjects
0.23
ns
r*
0.26
0.20
?0.20
different
32
-0.8
0.23
0.13
0.16
Proportion
with
Cs
morphology
ns0.16
0.75
1.31
.02
2.95
18
-2.54
2.02
5.33
Mean
instance
lexical
object
change
m
change
^
2.87
instance
2.13
19
1.26
5.62
4.04
Mean
change
morphology
ns
morphological
27
Mean
^
instance
2.83
lexical
2.01
4.71
4.94
0.24
subject
change
ns
change
Proportion
different
with
Cs
frames
0.76
0.18
0.67
0.23
1.02
32
ns
n
>//.
verbs
By
Ch
- ?
Intransitive
Transitive
Alternating
Flexibility
Transitive,
of
Use
Intransitive,
for
and
Alternating
Verbs
TABLE
12ooGrammatical
df*
p
Mean
SD
Fit
Grammatical
item
Mean
d
Cohen's
frame
instance
3.38
Mean
5.62
2.0
0.85
2.17
3.2
0.46
2,21
change
ns
different
0.78
Proportion
0.75
frames
0.23
0.30
0.60
0.14
Vs
1.35
2,21
with
ns
frame
instance
3.81
1.18
4.26
3.14
2,31
Mean
1.2
3.46
1.01
.04
change
days
frame
14.75
17.2
16.31
0.97
Number
16.14
14.65
2,21
12.84
of
to
change
ns
lexical
2.35
instance
Mean
2.0
5.62
0.87
4.99
5.26
1.01
2,14
subject
change
ns
2.87
3.67
instance
5.65
5.25
Mean
5.0
2.5
2,11
0.05
morphology
change
ns
26.15
36.83
30.42
24.68
Number
19.06
of
days
13.28
to
change
morphological
2,11
0.6
ns14.26
frames
different
0.76
0.77
Proportion
Cs
0.21
0.22
2.02
0.19
0.60
2,33
with
ns
17.17
18.95
Number
13.24
of
25.17
days
lexical
13.75
subject
0.66
to
2,14
change
ns
Note.?*df
different
0.09
0.19
Proportion
with
Vs
0.22
0.26
0.21
0.99
2,21
morphology
nsdifferent
0.29
0.36
0.38
0.28
Vs
Proportion
with
0.29
lexical
0.23
2,21
subjects
ns
lexical
different
0.27
0.32
0.56
Proportion
0.30
2.22
Vs
objects
with
7.06
0.88
5.63
0.8
Mean
lexical
instance
object
3.29
1.87
-2.96
61.63
change
.025
vary
days
14.75
Number
frame
21.13
13.93
19.92
17.37
16.63
0.38
2,31
of
to
change
ns
Mean
instance
morphology
change
4.3
2.99
4.15
1.62
4.62
0.07
2.81
2,18
ns
25.95
22.19
22.15
31.91
22.49
0.11
Number
of
days
19.4
2,18
to
morphological
change
ns
Number
of
days
to
lexical
subject
change
15.28
18.70
15.43
20.66
5.82
7.75
0.71
2,26
ns
lexical
instance
2.61
4.97
Mean
1.94
4.52
2.29
subject
4.8
0.09
change
2,26
ns
because
not all children,
different
0.24
Cs
0.52
0.19
Proportion
with
lexical
0.33
subjects
0.29
0.23
2,33
4.33
.02
different
0.15
4.77
Cs
Proportion
0.14
0.33
0.13
0.10
0.29
.01
2,33
with
morphology
Number
days
lexical
15.77
33.70
12.62
36.12
6-1.62
0.66
object
of
to
.12
change
lexical
different
2.86
Cs
Proportion
0.26
0.30
0.24
0.61
22
with
objects
.01
1.23
lexical
instance
3.41
Mean
1.94
5.83
1.31
18
-3.18
object
.005
change
1.43
verbs,
all
nor
22.55
9.69
14.81
Number
of
18
days
lexical
19.68
object
-1.07
to
change
ns
experienced
%
Cs
with
different
Imorphology:
versus
A:
-2.7,
*(23)
=
.012,
=
pd
=
1.11.
%
different
lexical
Cs
p
with
A:
If(23)
subjects:
2.66,
d
versus
=
.013,
1.10.
Mean
instance
Tframe
change:
versus
A:
=
t(22)
2.67,
.013,
=
pd
=
1.04.
other
All
comparisons
ns.
By
I.
children
//.
verbs
By
C/i
FLEXIBILITY
INEARLY
VERBUSE
AND GRAMMATICAL
PRODUCTIVITY
were heavy verbs, and the change in lexical objects was seen at a significantly
earlier instance for light than for heavy verbs.
of children's use of transitive, intransitive, and alternating
Comparison
lexical objects with
verbs revealed that children tended to produce different
a higher percent of transitive verbs than alternating
verbs (p = .06); they
earlier in
lexical objects at a significantly
also produced
those different
stance for transitive verbs than for alternating
verbs. Analyses
by verb
a greater number
Intransitive
verbs
of
ap
significant findings:
yielded
con
lexical subjects and in multiple morphological
peared with multiple
in more
than did alternating
verbs. Moreover,
children's
speech
transitive verbs appeared with lexical object changes for more children than
lexical
did alternating verbs; transitive verbs also appeared with a different
a
at
in
earlier
instance.
transitive
verbs
appeared
object
significantly
Finally,
new syntactic frames at a significantly earlier instance than did alternating
texts
verbs. Overall,
then, both transitive and intransitive verbs were associated
than were alternating
with more grammatical
verbs, and consis
flexibility
tent with the arguments
each type of verb takes, intransitive verb use was
more characterized
in the subject slot whereas
transitive verb
by flexibility
use was more characterized
in
the
slot.
by flexibility
object
some types of grammatical
To ask whether
flexibility were displayed
more
than others, two one-way repeated measures

ANOVAs
(by children
as a single factor within
with flexibility
and by verbs) were performed,
lexical subject, and
subjects factor with four levels (frame, morphology,
a
lexical object). The AN OVA by children
effect,
yielded
significant
to discover which
F (3,28) = 7.14,/? = .001. Pairwise t tests were performed
the Bonferroni
correction adjusted the re
categories differed
significantly;
= .05.
a
to
level
.0083
with
this correction,
quired significance
Using
only
one contrast was significant:
frame
Syntactic
flexibility was significantly
=
=
than morphological
7.72,
t(7)
greater
jfr .0001, Cohen's
flexibility,
=
a significant
d
AN OVA
3.04. The
also yielded
effect,
by verbs
=
were
contrasts
Five
F(3,120)
23.11,p<.0001.
significant using the same
Bonferroni
correction: by paired t tests, syntactic flexibility was significantly
than morphological
and
lexical
greater
flexibility
subject flexibility,
Cohen's
and
ds>1.25,
?s(33)>7.25,
ps<.00001,
significantly greater than
lexical object flexibility, ?(23) = 4.23,/? = .0003, d = 0.78. Both lexical subject
and lexical object flexibility
flexibility
(by paired t tests) were significantly
=
=
=
than
4.71, p
.00004, rf 0.71,
?(33)
greater
morphological
flexibility,
=
=
and t(23)
d
1.11, respectively.
5.5\,p<.00001,
In
sum,
the
above
analyses
converge
on
the
conclusion
that
the
eight
in this study all used their first verbs in multiple

syntactic envi
within the first 10 uses of those verbs. Many verbs were used with
different
frames?that
is, with subjects, with objects, in SVO frames, with
and
with
verbs were used with different
prepositions
negation. Moreover,
children
ronments
59
and
subjects
as well,
objects,
is, across
That
frames.
the
than
less frequently
although
first
instances
of
verb
in different
more
children
use,
fre
the frame entirely

(via adding or subtracting nouns, Ps,
quently changed
the nouns in the frame. In contrast, much
than changed
and/or negation)
more
at all, let
rarely did the children use their verbs with morphology
children
alone with morphological
these
Thus,
changes.
English-speaking
revealed
the flexibility of their early verb use and, by implication,
the ab
stractness
utterances
needed
of their grammars, more

in the syntax of their verb-containing
is
than in the morphology
of their verb use. Further research
more
to see whether
the same pattern holds in children
learning
rich languages
(e.g., Turkish).
morphologically
The light and heavy verb subclasses did not differ with regard to overall
light verbs were produced with more and
grammatical
flexibility; however,
with
earlier lexical object changes whereas
heavy verbs were produced
more and earlier lexical subject changes. Thus, counter to all versions of a
verb" hypothesis
8c Jisa, 2006; Ninio,
1999),
(e.g., Chenu
"pathbreaking
use and
in
verb
there was no consistent
light
advantage
early grammatical
the light and heavy verbs,
between
flexibility. More detailed comparisons
in
individual
the
variation
children, will be discussed
among
addressing
verb
also
did
VI.
subclasses
The
transitive/intransitive/alternating
chapter
not differ with regard to overall grammatical
transitive
flexibility; however,
in
their
of
verbs exceeded
verbs
alternating
achieving
speed
syntactic flex
amount
of
lexical
and
in
both
and
object flexibility.
achieving
speed
ibility
Intransitive
verbs
exceeded
in amount
verbs
alternating
of morphological
and
lexical subject flexibility.

that the obligatorily
transitive and intransitive
It may seem surprising
flexible"
verbs showed more flexibility
than the "permissibly
alternating
to our definition
verbs. This pattern of results may be partially attributable
of flexibility, which included any change in overt realization of grammatical
that apply equally
such as negation
elements and so also included elements
at
across verbs. More
the chil
of
this
development
early stage
important,
et
and
their
is quite constrained
dren's output
al.,
2006),
(Valian
typically
short
as to whether
unrevealing
(i.e.,
grammatical
from
removed
utterances?especially
tion?were
with
versus
verbs)
alternating
the
direct
context
objects
of
conversa
the
were
(i.e.,
processing
for
omitted
tran
for
nature of direct
It seems likely that the obligatory
sitive verbs) reasons.
to a greater use of objects with those
led
for
transitive
verbs,
then,
objects
in more flexibility. The
verbs than with alternating
verbs, and so resulted
subclasses of verbs reflect
children's different patterns of use with different
the
target
grammar.
alternating
verbs.
(i.e.,
possible
According
lacking
explanations
in children's
flexibility
morphological
minimal
Two
objects)
to
uses
processing
argument
also
exist
of intransitive
demand
structures
for
greater
with
compared
account,
of
the
the
intransitive
more
verbs
60
FLEXIBILITY
AND GRAMMATICAL
INEARLY
PRODUCTIVITY
VERBUSE
resources
to selecting
to devote more of their cognitive
to a se
inflections
for
these
verbs.
and
subjects
morphological
According
mantic account, the fact that the intransitive verbs in our data set tended to
verbs explains
the difference;
the
be activity rather than accomplishment
"
former verbs have already been shown to be used more and earlier with
1980; Vendler,
1972; Wagner,
2002). Further
ing" (Bloom, Lifter, 8cHafitz,
to distinguish
between
these explanations.
research is needed
children
allows
FINDINGS
COMPARISON
WITH PREVIOUS
Tomasello(1992)
(1992, 2000)
findings may seem at odds with Tomasello's
his daughter Travis's first verbs and with the Verb Island
of
diary report
on the basis of those data. However,
these and Toma
Hypothesis
proposed
are
sello's (1992) productivity
consistent:
Both Travis
findings
actually quite
and several of the children in the present study produced
the SV, VO, and
SVO frames with multiple
verbs before the age of 24 months
(see Table 8.4
were
in Tomasello,
were used
in
9
verbs
used
the
SV
33
verbs
1992;
frame,
in the VO frame, and 41 verbs were used in the SVO frame). Tomasello
(1992) states, "Travis did indeed begin using word order as a productive
The
above
syntactic device during the 18-24 month period" (p. 247). The findings that
motivated
the Verb Island Hypothesis
did not have to do with the use of
frames with different verbs but with two other findings
(see also Tomasello,
no
Travis
verbs
with
2000, p. 213): First,
produced
overlapping
privileges of
occurrence
same
time
the
and
almost
half of her
second,
during
period,
one third
verbs were produced
in just one construction
and
than
fewer
type
were produced
in two or more. Thus,
in her
individual verbs behaved
speech
as
"islands"
all different
construction.
of
grammar
and because
The
verb
because
their
privileges
each one was restricted

diary
records
collected
for
of occurrence
to being
the
used
present
were
in only one
study
do
not
the first finding regarding overlap among verbs in their

permit addressing
uses (i.e., because we have only 10 instances for each verb), but
they do
at
were
data
odds
with
Tomasello's
that
individual
verbs
provide
finding
limited in the range of constructions
in which they appear. The children in
the present study used their first verbs both with and without
subjects, with
and without
direct objects, with varied prepositions,

and so on. In fact, they
more
one
almost
of
their
in
verbs
than
70%
produced
syntactic frame by the
10th instance of use. What might account for this discrepancy
between our
and
Tomasello's?
One
is
that
both
Tomasello
and the
findings
possibility
present
study
participants
accurately
differ.
characterize
the
participants
It could be that flexibility
in each
study,
and
the
of early verb use varies among

61
one extreme
of this variability. An
that Travis represented
is that the different
other possibility
findings reflect the different ways in
which flexibility was coded in the two studies. We consider each of these
in the following analyses. First, we reanalyze the data from our
possibilities
to Tomasello's method
child participants
and compare the results
according
with those obtained
Travis.
this exercise
led us to our
from
Performing
second analysis, which involves a more detailed coding of argument use by
children
verb
and
subclass.
One
basis
for Tomasello's
characterization
of Travis's
grammar
as
con
islands comes
from his analysis of the arguments Travis used

taining verb
with her verbs. Tomasello
found that Travis was most likely to use her verbs
or with only a single argument.
utterances
either in single-word
Impor
struc
tantly for the claim that individual verbs did not appear in multiple
as being the same
utterances
treated all single-argument
tures, Tomasello
sort of structure
cannot
be
on the grounds
that children
to have
over
assumed
control
at this point
word
order.
in development
Thus,
"draw
dog"
structure
in grammatical
different
would not be considered
(verb+noun)
the same analysis to the present data,
from "I draw" (noun+verb).
Applying
we coded all 10 instances of our children's verb uses for whether
they ap
or holophrastically,
utterances
in
frames
argument
treating
single
peared
with arguments
in either the subject or object position as equivalent.
The results are shown in Table 13, together with data from Travis. We
of Tomasello
recoded Travis's data from the Appendix
(1992), using only
instances of the 34 target verbs of the present
study. The data are not
because Tomasello
did not consistently
report all to
perfectly comparable
use.
This inconsistency,
kens of single verb
though, could only have the
result
vanced
of
Travis
making
appear
than the children
rather
more,
in the present
than
less,
study. Higher
grammatically
percents
indicate
TABLE 13
Percent
of Verbs
Used
With
One
Argument
or
in
Holophrases
Child
%
79
Carl
Carrie
37
Elaine
84
52
Heather
53
Mae
Ned
86
Sam
90
Stacey
Travis
50
50
62
ad
that
FLEXIBILITY
VERBUSE
INEARLY
PRODUCTIVITY
AND GRAMMATICAL
The findings
verbs are "island like" or unproductive.
again reveal
to
Tomasello's
the
substantial variability: Applying
coding
subjects in the
a
in
which
the
four
of
present
present participants
study yields
picture
less flexibility
than Travis, using over 75% of their verbs in
demonstrate
frames or in single-word
three children
look
utterances;
single argument
or
to
of
their
verbs
in
about
half
similar
Travis, using
very
single-argument
one
more
us
child
looks
flexible
than
Travis,
single-word utterances;
only
or
utterances.
ing only 37% of her verbs in single-argument
single-word
of verb use.
Thus, Travis does not appear to be an outlier on this dimension
to
seven
in
of
Tomasello's
the
children
the present
criteria,
According
eight
to
et al.,
in
the
be
McClure
also
"verb
island"
also
appear
stage (see
study
more
2006).
treating
Beyond
ical structure,
we
would
there
argue
all
aspect
to
erroneously,
as
utterances
single-argument
is another
the
of Tomasello's
same
grammat
that leads,
procedure
that
conclusion
the
are
children
not
pro
in their early verb use. That is,Tomasello's

analysis did not take into
the differing requirements
of the individual verbs. His assessment
of whether
verbs are produced with multiple
and his inference
arguments
not
assume
of limited grammatical
when
do
that the verbs
they
knowledge
not.
In
all require multiple
But
do
intransitive
arguments.
they
particular,
to view a child who uses subjects but
verbs require only subjects; therefore,
ductive
account
not
over,
come
with
objects
of
many
the
or
verbs
go
as
is
More
arguments
inappropriate.
as transitive,
as draw,
are
such
regarded
"missing"
Tomasello
in
transitive and intransitive frames (Le

reality verbs that alternate between
use
such
the
that
of
these
verbs with just one argument
is per
vin, 1993),
can
a
For
"I
draw"
be
fectly appropriate.
example,
fully specified
if the context makes what is drawn obvious, and "draw dog"
description
can be a fully
if the context makes
ob
the addressee
specified command
more
The
vious.
revealing
is to compare
therefore,
analysis,
the
children's
to
argument use for each verb subclass (transitive, intransitive, alternating)
see if the children are using their arguments
for
the
different
differently
subclasses.
14 presents,
Table
ances
by
verb
class,
the
of
percent
that are verb only, verb plus one argument
or
such
arguments
source/goal
word
garding
order),
and
verb
as
plus
"look
two
ball"
or more
(including
and
utter
verb-containing
subject, object,
and
store,"
"go
arguments.
It
disre
is apparent
in these data that the use of holophrases

is high, particularly with intran
sitive and alternating verbs and that the frequency of one-argument
frames
one third of all uses) for all three subclasses.
is similar (approximately
as predicted,
the use of two-argument
frames is highest with the
However,
transitive
and
as
frequent
seemingly
the
verbs;
alternating
one-argument
anomalous
moreover,
frame
for
the
2-NP uses found with
the
two-argument
transitive
verbs.
intransitive
is as
frame
(Note
that
the
verbs arose because

63
TABLE 14
Percent
of Verb
Uses
Produced
in Different
Frame
Child/Subclass
1 Argument3
Holophrase
Frames,
by Verb
Subclass
Type
2 Arguments3
3 Arguments3
Otherb
Carl
Transitive
Intransitive
Alternating
0 40 70
64.3
82.31
0
27.14
5.38
4.62
0
na
na
0
8.6
0
10
64.44
10
30.83
49.2
4.44
na
na
0
11.11
0
0
42.85
0
na
na
0
0
0
Carrie
Transitive
Intransitive
Alternating
0
14.44
21.67
85.5
Elaine
Transitive
Intransitive
Alternating
45
20 35
57.14
51
41 8
Heather
Transitive
Intransitive
Alternating
3.75
24.44
25.38
48.75
43.75
0
47.77
50
22.31
1.25
na
na
2.5
32.2
10
Mae
Transitive
Intransitive
Alternating
35.56
24.44
11.25
12.5
30.77 29.23
11.11
na
na
5.56
6.25
0
10
0
na
na
0
0
12.85
42.2
41.1
6.67
10
50
4.3
31.43
na
14.3
na
11.3
0
na
na
0
10
16.67
12.22
70
32.3
Ned
Transitive
Intransitive
Alternating
65
305
97
30
77.14
Sam
Transitive
Intransitive
Alternating
0 9.3
79.3
Stacey
Transitive
Intransitive
Alternating
2 6632
47.1
35
042.9
11.67
36.67
Travis
Transitive
Intransitive
Alternating
10
64 22
33
25 na
na
14
78 3
na
2.1
2.26
10.31
6.35
43
Mean
Transitive
Intransitive
Alternating
16.89
53.05
50.51
39.4139.19
3.35
33.85
29.04
13.33
required objects, recipients, and locations.

Note.?aSubjects,
vocative uses (e.g., "Mom, come").
bAdverbs, prepositions,
64
FLEXIBILITY
INEARLY
AND GRAMMATICAL
VERBUSE
PRODUCTIVITY
their prepositions,
of uses of look and sit that were missing
e.g., "Grandma
look doggie" and "Dolly sit chair.")
The critical question for the Verb Island Hypothesis
these
is whether
in the number of arguments
data indicate systematic differences
appearing
with subcategories
of verbs that would be predicted on the basis of the adult
grammar. If children's grammars are lexically based, then children have no
of verbs. If, however,
category VERB and, by implication, no subcategories
children's very early verb uses show differences
related to verb subclass, this
of underlying
verb subclasses in children's
implies the existence
linguistic
our
from
data set, we do not know if these go
representations.
(Admittedly,
the input given.) This hypothesis was tested in a 3 (verb type) x 3
beyond
ANOVA, which yielded a sig
(number of arguments)
repeated measures
nificant interaction, F(4,28) = 7.46, p = .0005. Pairwise contrasts using the
Bonferroni
correction
(p< .005) revealed that within the intransitive verbs,
utterances
holophrastic
ment
were
were
utterances, Fs(\,7)>
used holophrastically
more
frequent
than
17, jfrs<.004, Cohen's

with less frequency
either
one-
ds>0.73.
or
two-argu
Transitive
than either
verbs
or
intransitive
ds> 1.35; the transitive verbs also

verbs, Fs(l,7)>
alternating
16,jf?s<.005,
to appear
in significantly more
tended
frames than the
two-argument
=
=
In
d=
the
intransitive
1.72.
sum, under
15.22, p
.006,
verbs, F(l,7)
all
of
the
children?even
Travis?demonstrate
knowl
appropriate
analysis,
in frames privileged
edge of the difference
by transitive and intransitive
verbs from the beginning
of their use of those verbs (see also Goldin
Meadow,
2003,
TheManchester
and McClure
et al., 2006,
for similar findings).
Corpus
The degree of syntactic flexibility observed

in this study is also greater
than that reported
in Lieven and colleagues'
studies of their Manchester
et al., 2001, 2004).
corpus (Lieven et al., 2003; Pine et al., 1998; Theakston
in particular, are at odds. First, the degree of lexical subject
Three findings,
flexibility with intransitive verbs in our data is considerably
higher than that
et
Pine
al.
for
their
25%
for ours;
36-52%
(1998);
reported by
namely,
study,
see Table
et al. (2004; Table 8) reported
12. Second, whereas Theakston
syntactic flexibility on the order of two frames per verb in their children
within, on average, 6 weeks (thus including many more instances of a given
verb than we could), we find comparable
numbers of flexibility within, on
It seems likely that we found more and earlier
average, just over 2 weeks.
grammatical
flexibility because the present data were, to use Lieven et al.'s
term, "denser." That is, every instance produced by our children (within the
first 10 for a given verb) was recorded; moreover,
children's utterances were
across a wide range of situations. As discussed
recorded
in chapter
I,
on
situation-restricted
to
make
relying
sampled
speech
judgments
65
toddlers'
concerning
grammatical
yield underestimates
may
productivity
of
productivity.
Lieven et al. (2003; see also Lieven, 2006) make a further claim that is
also relevant to the present analyses of the diary records. Lieven and col
are not
combinations
leagues have claimed that children's early verb+noun
a
indications of grammatical
because
number
of
these, in
flexibility
large
involved substitution of only one word at a time. For example,
or "Bring cooky" and then "Bring
saying "Doggie go" and then "Kitty go,"
as
not
to
count
Lieven et al. because a change of only
juice" does
flexibility
one word could simply indicate flexibility at the lexical rather than gram
matical
levels. In contrast, in the coding scheme used in the present study,
both of the above examples would have been counted as changes (in lexical
subject or object) and thus counted as indicators of flexibility. We have two
their corpus,
to Lieven
counters
et al.'s interpretation
is practical: Children
of this type of flexibility. The first

in the toddler years are at a stage of
in terms of length; they only produce
if the
1973; Valian et al., 2006). Thus,
counterargument
utterance production
that is limited
two to three words at a time (Bloom,
verb stays the same, only one other word could change over the course of
a time would be the only way
sequential utterances. Changes of one word at
these children could show syntactic flexibility, because of output constraints
not
and
because
one
from
vary
of
that young
finding
matical
than
of
absence
abstract
to
an
the
another
in the number
is a
next
Put
grammar.
are limited
children
utterance
rather
limitations
the
discovery
independent
that can
of
consequence
logical
their
regarding
the
way,
of words
length
gram
knowledge.
comes directly from the diary records. As

second counterargument
in the present
earlier in this chapter, the children
described
study were
more
to
frame
the
demonstrate
(i.e.,
syntactic flex
entirely
change
likely
or
which
included
and/or
nouns,
Ps,
negation)
subtracting
adding
ibility,
lexical subject or
than to change the nouns in the frame (i.e., demonstrate
individual children also demonstrated
lexical object flexibility). Moreover,
went beyond just changing nouns
that
considerable
flexibility
grammatical
in a given frame. The first example of such flexibility comes from one of the
she said "I coming"
"star" children, Heather. Early in her verb production,
to Lieven et al.,
and "Come here" within 30 days (starting at 18 months);
substitution and not real flexibility.
this might be coded as just a one-word
The
she
However,
dropped
"Take
users:
followed
"Drop,"
"I
"I
drop,"
span of time;
23 months
Heather
said
such
Carl
was
were
demonstrations
one
of
our
slower
not
learners,
by (on the same day) "Gavin walk,"
showing
a
a bite,"
yet
he
"I
syntactic
of
property
also
and "I don't
word
multiple
comprise
"take
"I
and
something,"
drop
these
this" all in the same day, clearly
Furthermore,
verb
said
the same
at
Moreover,
changes.
and
also
it" over
take
bite,"
flexibility.
the
only
said "I no
"star"
walk"
like" followed
66
by
FLEXIBILITY
PRODUCTIVITY
AND GRAMMATICAL
INEARLY
VERBUSE
(on the same day) "I like cheese." Both of these couplets involved changes of
more than one word, and thus would count for grammatical
flexibility even
on
et
Lieven
al.'s
We
ations/contexts.
set
data
would
we
above,
argued
grammatical
our
because
corpus
As
criteria.
more
demonstrated
flexibility
includes
argue,
believe
speech
in addition,
a wider
from
the
that
our
that
children
in the Manchester
than those
of
range
situ
scheme
coding
ap
biases the
corpus (and that used by Tomasello)
plied to the Manchester
et al.
That
little grammatical
Lieven
results toward finding
is,
productivity.
claim to have discovered
that children are not productive
in
and Tomasello
is not, however,
of
their use of verbs. This discovery
their
independent
to credit the child with little grammatical
coding decisions. Having decided
on an utterance-by-utterance
basis, it is not surprising that they
knowledge
then find
that the children

overall.
demonstrate
little grammatical
in
knowledge
their speech
CONCLUSIONS
In sum, the data presented
in this chapter support our contention
that
we
the
learners?at
have
in
studied
such
least,
English
eight
in their verb use from its very
detail?demonstrate
grammatical
flexibility
onset. Overwhelmingly,
the first 10 instances of these 34 common
verbs
American
frames. Moreover,
these first 10 uses show sensitivity to
appear in multiple
verb subclass, in that transitive verbs appear less frequently as holophrases
and tend to appear more frequently
than intransitive verbs in frames with
two
use
arguments.
in which
is an
young
"island"
each
These
unto
verb
data
of
descriptions
appearing
do
not
itself.
The
learners
data
children's
in its own
suggest
learners have underlying
language
even very young

their
of
input language.
in which
contradict
data
with
environments.
syntactic
verb
These
as restricted,
are more
are acquiring
an
account
of
grammars
consistent
early
limited
the grammatical
of
acquisition
in which
with
verb
range
an
each
account
patterns
67
VI. DIFFERENCES
INEARLY
VERBGROWTHAND USEAS A
FUNCTION
OF DEVELOPMENTAL
AND VERB
PERIOD,CHILD,
In
V,
chapter
we
presented
analyses
that
some
revealed
grammatical
in the first verb uses of eight children. All chil

and flexibility
productivity
dren used at least one syntactic frame with multiple
verbs; three children
used all five coded frames with multiple
verbs. All the children used at least
some of their first verbs with grammatical
flexibility, and each of these early
was
some children during the win
at
verbs
used
least
acquired
flexibly by
dow of this study. Within
that picture of early and pervasive flexibility, there
was variability among children and among verbs in the degree of
flexibility
It is also possible that there was variability across the duration of
observed.
in the analyses presented
thus far. In
the study that has been unexplored
this
chapter,
we
children,
among
interrelations
provide
examine
and
among
further
differences
among
the
verbs
measured
tests of the theories
among
in more
properties
of
periods
detail,
of
and
we
children's
of verb development
development,
examine
verb
we have
the
use
to
consid
ered.
DOES FLEXIBILITY
OF VERBUSEVARYBYPERIOD
OF VERBDEVELOPMENT?
to the conservative-child
in chapter I,
described
hypothesis
According
are
at
later
is
the
combinatorial
first
pos
grammar
lexically specific; only
in a grammar with an abstract category of
sibilities of verbs represented
VERB
1999, 2007; Tomasello,
2004; Ninio,
2000). The
(MacWhinney,
are
most
used
with
frames
and most
that
verbs
present findings,
multiple
verbs within the first 10 instances of verb use, are
frames used with multiple
It is still possible, however,
that there is a
inconsistent with this position.
such
in the abstractness of children's grammars
progression
developmental
than verbs that
that verbs that are learned first are used with less flexibility
are learned later. To address this proposal, we go back to the data to test the
verb use, suffi
that there is an early stage of nonproductive
hypothesis
68
INEARLY
DIFFERENCES
VERBGROWTH
TABLE 15a
Age
Children's
for
(Months)
Each
Developmental
Period
Child
Carl
Period
1
20-21
Heather
Sam
Mae
16-17
19-20
Elaine
19-20
20-21
Carrie
Stacey
18-19
16-17
2 ends
23.5
22.5
19.5
22.5
23.5
21.5
19.5
3 ends
25.5
24.5
21.5
24.5
25.5
23.5
21.5
4 ends
27.5
26.5
23.5
26.5
5 end
29.5
28.5
25.5
28.5
6 ends
31.5
30.5
7 ends
33.5
32.5
Ned
17-18
20.5
TABLE 15b
Children's
MLU
for
Each
Developmental
Period
Child
Period
Carl
Sam
Mae
111
1
1.05
1.93
2.18
2.68
1.65
1.17
1.71
3.95
2.71
2.06
2.7
4.37
3.08
3.67
2.3
3.8
3.31
2.2
4.04
3.45
Ned
1.98
5.08
2.52
Stacey
2.58
1.15
= mean
Carrie
1.48
Note.?MLU
31
Elaine
2.32
2
4
1.2
Heather
5.72
length of utterances.
brief that it was obscured

and flexible
by the overall productive
of verb use during the long window of this study.
We analyzed each child's verb use across his/her participation
in the
We
child's
of
divided
each
into
span
study.
participation
developmental
as follows: Period 1 included each child's first 6 weeks of
periods,
partic
of the study,
ipation and thus, given the participant-selection
requirements
their very first verbs. Period 2 included each child's next 8 weeks of par
ticipation, Period 3 included the next 8 weeks, Period 4 included the sub
Some children
sequent 8 weeks, and so on to the end of their participation.
ciently
nature
seven periods
(Carl, Mae) participated
(see Table
15a), others
through
five
others (Elaine, Carrie,
(Sam, Heather)
participated
through
periods,
three periods,
and one (Ned) participated
Stacey) participated
through
two
It
to
is
through
periods.
important
point out that because we have
records only for each child's first 10 uses of each verb, the later periods
include data only from later-learned
verbs and not from additional uses of
69
? 70%
- IP
ll
iim mi
iiu
Period
Period
1(7)
2(9)
Period
Figure
early-learned
11.?The
verbs.
Period
^m ' Multiword
?& ^
iH
in
Period
Period
Period
Period
3(10) 4(14)
5(20)
6(22)
7(24)
of Development
development
Thus,
^|
the
asm?'e
frame
(Total Target Verbs)

of syntactic
crucial
flexibility
question
in newly
concerns
learned
whether
verbs:
Carl.
the
first
10 instances of later-learned
verbs consistently display more flexibility
than
the first 10 instances of early-learned
verbs. Tables 15a and 15b present the
for each
children's ages and mean length of utterances
(MLUs), respectively,
in which they participated.
of the periods
The MLUs are estimates, calcu
utterances
in the diaries.
lated using only the verb-containing
recorded
Figures 11 through 18 present each child's timeline individually; Figure
in a single figure. The figures are ordered begin
19 presents all children
for the greatest number of pe
ning with those children who participated
for the fewest
riods, Carl and Mae, and ending with Ned, who participated
11, Carl began with seven verbs in his first
periods. As shown in Figure
period and added new verbs slowly during his next two periods. All of his
verbs in these first three periods were in single-word utterances
and hence
not used grammatically
(gray bars). During his fourth period, he
flexibly
utterances
began to use verbs in multiword
(striped bars); however, both
were
one
verbs
frame (subject-verb
[SV]). It was during his
only used in
fifth period, when he was between 27.5 and 29.5 months of age, that he used
two verbs both in multiword
utterances
and in different
frames within the
first 10 instances (black bars). The frames included SV, verb-object
(VO),
and SVO. Thus, Carl began to show grammatical
flexibility during Period 5.
Mae, shown in Figure 12, displayed a similar although not identical pattern.
Mae began using two verbs in Period
1, then added eight more verbs in
Period 2 but only eight more verbs in Period 3. Unlike Carl, throughout
70
VERBGROWTH
DIFFERENCESINEARLY
1 80%|
1 %-i
? 60%-p
50%~
*
I I
I 40%"
^II i I
I
i^ 30%1
I
I
20%1I 1
, ill
0% I
| 111 I
Period
Period
1(2)
2(10)
Period
Figure
12.?The
these three periods

first
10
instances.
I
I
^
1
P
I, K^
._,
Multiword
I l I
I^H I^1 I^|
I
l
II "^ i y/ft IH
| | |
| I
I22? H
I I
P
I^,
I,
utterances
in
frames
multiple
0Multiword
u~swith
I
lH
I
asin?le
trame
rsingie-word
utterances
'-'
I
III
I^,
I,
Period
Period
Period
Period
3(12) 4(20)
5(27)
6(30)
7(31)
Period
of Development
of syntactic
development
she used
However,
some verbs
as was
the
in newly
flexibility
inmultiword
case
for
Carl,
learned
utterances
each
verb
verbs: Mae.
within
her
in
appeared
one verb appeared with a preposition

(P)
only one frame: During Period 1,
and during periods 2 and 3, verbs appeared only in the SV frame. During
Period 4 (i.e., one developmental
period earlier than Carl), she used three
utter
utterances
and in different
frames of multiword
verbs in multiword
ances within
the first 10 instances. Frames of SV, VO, SVO, and verb
(VP) were all represented. Mae thus began to show grammatical
preposition
of age. Sam,
24.5 and 26.5 months
flexibility
during Period 4, between
this general pattern of (a) restricted use
shown in Figure 13, also displayed
utterances
for their first 10 instances during
of his target verbs inmultiword
use begin
use and multiframe
1-3 and then (b) mostly multiword
Periods
at
in
4.
the
At
Sam
used
his
verbs
Period
his
Period
SV,
SVO,
VO,
4,
ning
and VP frames. The timeline data from these three children (Carl, Mae, and
some support for a weaker version of the conservative-child
Sam) provide
in restricted frames (either single
First
verbs were produced
hypothesis:
word
utterances
or multiword
utterances
using
only
one
frame);
later
verbs
were
in flexible frames.
produced
not all of the children followed this pattern. Heather
However,
(Figure
in their
14) and Carrie
flexibility
(Figure 15) demonstrated
grammatical
verb use starting at Period 1. Both girls produced many verbs even during
in the study (n = 16 and 13, respectively),
their first 6 weeks of participation
utterances
half of these verbs in multiword
and they used approximately
and
in different
frames
within
the first
10 instances
during
Period
1.
71
100% -i
vm
vm
, 90%A |
-%1
? 70%-i
n
fc
| I
ill
j?gv
IB
j%
60%
HI
~
?>50%
II
HI
II
II r?3
Bl
|l|
H 40%
111
|
Wk
*o
H HI
Ml
30%"
S 20%1
?
Period
Period
Period
1(3)
2(9)
3(20)
4(29)
2 70%1 1111
I I
c*
,:]|
gg
g
60%-
-p
3 50%
p
?
!?
I
1":
Figure
5(31)
of syntactic
%Z
I
I
I
I
1
I
I
I
I
I
I
I
I
Period
Period
Period
Period
Period
2(22)
3(29)
4(31)
14.?The
development
in newly
flexibility
of Development
01Single-word
utterances
Period
with
frame
Period
1(16)
Multiword
utterances
a
single
utterances in
multiple frames
of Development
development
Multiword
H
H
r
77/ ,
Zvy Wfk
Period
13.?The
._.
P;%> I^fl
II
II
Period
H^1
I
ll
III
Figure
learned
verbs:
Sam.
I ,_,Multiword
utterances in
multipleframes
P 0 Multiword
r;:?r
???
\
5(31)
of syntactic
flexibility
in newly
learned
verbs:
Heather.
Heather
and Carrie used the SV, VO, SVO, VP, and V-ing frames all during
Period
flexible use continued
1, and grammatically
through each child's
next two periods. Neither
Carrie nor Heather
of
any evidence
provides
restricted verb use in terms of grammar, even for their very first verbs. They
72
VERBGROWTH
DIFFERENCESINEARLY
100%-.
90%-III
^1
^ 80%
o
70% _
^H
^fl
^H
^fl
^H
^1
^1
^1
^1 ^1
^1
^|
utterances
with
B^ 40%- S I I a singleframe
^^|
30%
Period
1(13)
Figure
15.?The
of syntactic
I|
gg
I
-
'd
20%
10% IIII
II-1
pIH
i-1
Period
three
ered
development
point
Carrie.
%60% I multipleframes
[I P
H
Multiword
utterances
with
I ? 1 single frame
I 40%a
30%"
1
I I I?Si?gle-word
utterances
II
///.
rcfl
i-1-1
Period
3(19)
of Development
the ending
conservative-first
i-1
2(15)
Period
also reached
Hi
Period
1(3)
16.?The
III
verbs:
Multiword
utterances in
M
%6
H|
III
? fly
gz
Figure
learned
.-,-
S 50%
o% I
in newly
flexibility
^|
utterances
3(28)
of Development
development
90%I|
m%]
80%| 70% l|l
?
^| |
Period
2(23)
Period
ra Single.word
-p
20%
Period
I-;
utterances
frames
multiple
0 Multiword
~ ^|
^|
^|
-I ^| ^|
II
^0%
50%
fc
?
Multiword
of syntactic
flexibility
of the study more
in newly
quickly
learned
verbs:
Elaine.
than two of the
children.
Elaine, Stacey, and Ned (Figures 16-18, respectively) may be consid

with respect to their speed of reaching grammatical
"intermediate"
73
100%-. _
g 80%- p
jj&f
||
Y/z\?PM 20%
B
Hn <u
rvv>j
io%-1
0%4-B?,-,_
Period
17.?The
i-1
i-1-1
3(20)
of Development
development
Multiword
Period
2(19)
Period
utterances in
H H multiple frames
with
5340%- utterances
III
a
frame
HH
single
t ^|
^H
30%n Single-word
III |
utterances
^1
"
_,-1
Period
1(9)
Figure
g2 60%-
Is
50%-
I Multiword
I^
1I 70%00
^|
^|
of syntactic
in newly
flexibility
learned
verbs:
Stacey.
100%-i
o 90%"
I I
80%"
1=-:-1
I|
c/,
111
H
Is 60%"
||
1
2 70%
H
frames
multiple
- 50%
III
Multiword
utterances in
0
Multiword
utterances
with
40%- IH
I
?HI
|
0% I
*io%-1
,-,-
Period
20%-
a single frame
U I D Single-word
utterances
I w
-1-,-,
Period
2(14)
1(10)
Period
Figure
YM
1_'-'
30%"
18.?The
of Development
development
of syntactic
flexibility
in newly
learned
verbs:
Ned.
frames within the first 10

flexibility of verb use: Each used 1 verb inmultiple
instances during Period 1. Elaine and Stacey both used their one multiframe
this
1 in the SV, VO, and SVO frames. Also during
verb during Period
one
two
frame
in
within
multiword
additional
verbs
used
just
period, Stacey
the first 10 instances (Elaine used none in this way). Leaps in grammatical
or Period 3 (Elaine).
flexibility were then displayed during Period 2 (Stacey)
74
DIFFERENCESINEARLY
VERBGROWTH
100
90-
-i-^--??-II
o
80| 70J
a 60"
c 50-
/
/ A
17
^ ^ ^. /
'
>r
jf i
Sam
Carrie
--*-Heather
~ /1
II I
y40
""*"""Elaine
|
/ / / /^f/l -"H"~~Mae
/
W
\ /
V
/
0 -I?It?i
16-
Carl
//
/' ^Ned
//
*-r*
^ 30 ///
20?
>.
J*
/ \ /
--*?Stacey
18-
i-r?i??
20-
22-
i X24-
19
21
23
25
'?i-1-1-1-1
2628-
30-
32
27
31
33
29
Age
Figure
19.?Change
in multiword
use by age.
than Carl, Mae, or Sam; however, their

they showed earlier flexibility
was
across
as
a number of verbs as was Carrie's
not
wide
flexibility
displayed
or Heather's.
Ned participated
for the fewest periods and, like Carl, Sam,
and Mae, used his first 10 verbs in only restricted single-word or multiword
utterances
1. During
for their first 10 instances during Period
Period 2,
new
two
in
he
four
verbs
and
used
of
them
multiword
though,
produced
utterances and in different frames within the first 10 instances. Ned's frames
Thus,
during Period 2 included SV, VO, and SVO.

to grammatical
that these children
The different pathways
flexibility
can be seen together
in Figure
the
19. Figure
19 displays
demonstrated
utterances
in different
in multiword
cumulative percent of verbs produced
the first 10 instances for each child across periods. As the
frames within
consider
and Ned produced
figure shows, Carrie, Stacy, Elaine, Heather,
utterances
in different
Periods
able multiword
frames during
2 and 3,
whereas Sam, Mae, and Carl reached these levels only during Periods 4-6 of
their
verb
use.
in the
children's
analysis of individual
developmental
changes
with
used
verbs
reveals
which
grammatical
flexibility
they
newly acquired
that the majority of the children, although not all children,
showed gram
even the
matical flexibility
close to the very onset of verb use. Moreover,
child who was slowest to achieve grammatical
Carl,
flexibility,
displayed
"slow" learner,
such flexibility
before the age of 2.5 years; Sam, another
before 24 months
of age.
grammatical
flexibility
actually demonstrated
and
Carrie
demonstrated
Heather,
Ned,
grammatical
flexibility be
Stacy,
This
75
fore 20 months
of age. Both these developmental
patterns and the ages of
are
at odds with those re
the first grammatical
demonstration
flexibility
other
who
claimed
that
there is little gram
have
researchers,
ported by
are
to
matical
children
close
until
2.5
years of age and that
flexibility
children's
verb
2000,
if not exclusively,
is predominantly,
et al., 2006; Tomasello,
2004; McClure
initial use of sentence
frames
specific (e.g., MacWhinney,

p. 213).
TOGRAMMATICAL
FLEXIBILITY?
VERBSPATHBREAKERS
ARE LIGHT
Previous
comparisons
of
the
grammatical
environments
surrounding
III and V)
first verb uses (in chapters
light and heavy verbs in children's
more
with
direct
revealed that light verbs appeared
significantly
frequently
in
and
verbs
and
frames
than
verbs
that
SVO
light
appeared
objects
heavy
earlier and more frequently with different direct objects than
significantly
evidence for the claim
did heavy verbs. These
findings might be considered
of abstract
that light verbs function as facilitators of children's acquisition
sentence
E.
&
frames
Clark, 1987; Gold
(Chenu
Jisa, 2006;
phrase-level
8c
2005b;
1999, 2005a,
Casenhiser,
2006; Ninio,
1999; Goldberg
berg,
the
the
Rice
8c
because
verbs,
Pinker, 1989;
Bode, 1993)
among
target
light
verbs
in
this
study,
did
seem?overall?to
show
more
and
earlier
direct
for light verbs

the finding of greater flexibility
object flexibility. However,
was limited to the direct object slot; light verbs did not show greater gram
matical flexibility
overall
(Table 14). Furthermore,
light verbs may have
more varied direct objects not because of
and
with
more,
earlier,
appeared
of light verbs are general
any syntactic privilege but because the meanings
the direct objects is necessary for communicative
and thus expressing
pur
or
event
is
to
the
child
relation
that
the
talking
poses
specify
sufficiently
et al., 2004).
about (e.g., Snedeker & Gleitman,
2004; Theakston
that these averaged data missed potential path
It is possible, however,
individual
that
functions
light verbs might have served for indi
breaking
vidual children. As listed inTable 1, the light verbs among the target verbs in
this study were fairly diverse, and some of these (go, put, want) have figured
than others
more
of the pathbreaker
in proposals
hypothesis
heavily
2006; Ninio,
1999; Goldberg & Casenhiser,
(Chenu & Jisa, 2006; Goldberg,
such potential functions for each
1999). The following analyses investigated
each light verb that each child produced.
child individually,
considering
we investigated whether
et
individual light
al.
Theakston
(2004),
Following
in a wider range of different frames than verbs with more
verbs participated
the
that were acquired at the same age. We compared
specific meanings
frame flexibility of each light verb with (a) the heavy verb that appeared just
76
DIFFERENCES
INEARLY
VERBGROWTH
TABLE 16
CARL's
Verb
Light/Heavy
for
Comparison
Frame
Flexibility
Age
(Days)
Preceding
Light
Verb(s)
Verb
1st
Following
Use
Verb(s)
Push658
GO680
715
Kiss
Sit 808
COME
of Frames
Number
10th Used in 10
Use
Instances
680
680
715
1 (V)
1 (V) Go
1 (V) Kiss
(Frame)
Examples
Push
Me
sit, No
808
2(SV,
800
Cry
814
800
814
2 (V,SV)
1 (V)
Cry, Baby cry

Come
Cut,
Cut
paper,
Mama
cut
Cut,
Give
Cut
paper,
Mama
cut
me,
Sissy
noV)
827
827
Cut 827
827
GIVE828
828
3 (V, VO, SV)

3 (V, VO, SV)
2 (VIO, SVIO)
861
1 (SVO)
Cut 827
827
838
840
3 (V, VO, SV)

2 (V, V locative)
Cut
Hold
861
LOOK
861
1 (SVO)
Like 911
911
928
932
2 (SVO,
2 (SVO,
861
Hold
WANT
932
932
me
give
I hold (flash)light
cut
paper, Mama
Look punch bag
Cut
Cut,
Look,
I hold (flash)light
SV-neg
O)
I don't
3 (SVO, VO,
Me
V)
like,
want
Me
SV+S)
to do
Wash
sit
I like cheese
it, Me
want
it
wash
table, Wash
floor, Wash
=
=
=
=
Verb; S
subject; SV
subject-verb; VO
verb-object;
talized verbs are highlighted
relative to "regular" verbs.
Note.?V
SVO = subject-verb-object.
Capi
before this light verb's first use and (b) the heavy verb that appeared just
after this light verb's first use. We also looked across the range of earliest
verbs used by each child to see which verb(s) first appeared with which
frames.
Given
that
our
data
set
was
verb
centered
rather
than
frame
centered,
though, we are unable to address the related question of whether
the very first uses of a frame appeared with light verbs. The findings are
in Tables
16-23, described below in detail for Carl and summa
presented
rized for the rest of the children.
Carl produced
six light verbs. Each was preceded
in development
by a
more
none
verb
was
Table
and
of
the
verbs
(see
16),
heavy,
specific
light
more flexible in its use than the
verb.
Moreover,
just-preceding
heavy
light
verbs were not consistently
the first ones to be produced
in new construc
at 800 days, was Carl's 13th verb but the
tions. For example,
cry, produced
first verb used with a subject. Cut was the first one used with an object, and
hold was the first one used in the SVO frame. Carl's light verbs were the first
to be used in the verb 4-locative (VL) (inchoative motion)
and VOL (caused
motion)
frames/constructions;
however,
his "first-in-frame"
light verbs were

77
TABLE 17
CARRIE'S
Verb
Light/Heavy
for
Comparison
Frame
Flexibility
Age
(Days)
Number
Preceding
Verb(s)
Light
Verb
Following
1st
10th Used
Use
Use
Push
564
574
Wash
565
583
Verb(s)
of Frames
in 10
Instances
(Frame)
568
582
3 (V, SV, VO)

3 (V, SV, VO)
3 (V, V L, SVL)
569
644
3 (V, SV, SV-ing)
Cry 590
593
601
2 (SV, SV-ing)
2 (V, SV-ing)
GO
Walk
Examples
Push, Arty push, Push mommy
I wash, Wash my hands
Wash,
Go, Go bye-bye, Daddy go work
Walk, Mom walk, Doggie
walking
COME
Run
Open
WANT
Kiss
Hold
LOOK
Pull
BRING
Wave
TAKE
Wave
Roll
GIVE
594
603
585
600
593
626
595
616
604
622
608
625
622
660
628
660
PUT 644
670
See
Need
603
652
679
Need 652
679
659
683
672
691
2 (SV, SV-ing)
3 (V, SV, VO)
1 (SVO)
3 (no V, SV, VO)
3 (SVO, neg, -ing)
2 (V, V locative)
3 (SVO, neg, -ing)
No kiss, Kiss baby, Daddy

No Mom hold it, Daddy
SVing)
I bring my baby
Dad wave, I wave
679
659
681
1 (SVO)
3 (Neg VO,
672
691
NegSVO)
3 (SV, SVIO,
687
709
3 (SVO, SVP, SVing)
My waving
My roll my hands, Wagon
rolling down, all rolling
690
708
3 (VOP, SVOIO,
Give mom
No
SVO,
SVing)
VO)
take it, No my
take my nap,
Daddy take my sippy

Dad wave, I wave at A,
hug, I give my
to A, Give that back
after
V = Verb; S = subject; SV = subject-verb; VO = verb-object;

relative to irregular verbs.
Note.?
at A,
My waving
My need my binka
Need 652
produced
Give
kiss
it
holding
Look, Look my shoe
No pull my shirt, Daddy
pulling it
cookie
No verbs were
open
My see doggie
put shoes
Mommy
My need my binka
My need my binka
1 (SVO)
1 (SVO)
1 (SVO)
1 (SVO)
1 (SVO)
3 (SV, SVIO,
Baby cry, Arty crying

Come, Granny coming
Arty run, My running
teeth, Daddy
Open, Open
I want my daddy
Capi
the ones predicted: Look was his first verb to be used in the VL frame
(lookpunch bag) instead of the predicted go, and give was his first verb used in
frame (give me) instead of the predicted/^.
the V(0)L
Admittedly,
given the
of
of go and his 1st production
time elapsed between Carl's 10th production
not
78
INEARLY
VERBGROWTH
DIFFERENCES
TABLE 18
ELAINE'S
Verb
Light/Heavy
for
Comparison
Frame
Flexibility
Age
(Days)
Preceding
Light
Verb
Verb(s)
1st
Following
Use
Verb(s)
WANT
LOOK
Open
Sit
10th
Use
of Frames
Number
Used
in 10
Instances
(Frame)
See
658
672
Wash
Open
I sit
2 (V, SV)
I see
See,
716
1 (V)
Roll
696
716
696
741
3 (V, VV, SV)

2 (SVO, VO)
Come,
I wash
Roll 695
COME
Examples
I want veggies
Go, I go
Look
602 613 1 (SVO)

GO613 710 2(V, SV)
633 634 1 (V)
651 651 1 (V)
653 656 1 (SV)

Note.?
Come
hands,
roll, My come
Wash hands
Capi
look, it is probably not the case that his VL production with look was his very
first in that frame; however, this caveat does not hold for his 1st production
of a verb in the VOL frame. That is, when Carl produced give me he had not
even one instance of put, bring, or take-, therefore, it is very likely
yet produced
that give me was one of his first, if not his very first, use of the VOL frame.
Carrie, Ned, and Heather each showed a similar pattern to Carl's, in that
(a) none of their light verbs was used more flexibly than concurrently
pro
duced heavy verbs and (b) only some frames were used first with light verbs,
whereas others were used first with heavy verbs (Tables 17, 19, and 21). For
Mae, Sam, and Stacey, some proportion
(one quarter to two thirds) of their
light verbs were used more flexibly than their concurrently
produced heavy
verbs; however, they were no more likely to produce specific frames first with
the predicted
light verbs (e.g., VL with go or VOL with put) (Tables 20, 22,
and 23). Note that the first 10 uses of make were (voluntarily) recorded by
Mae's
and
Stacey's
mothers;
make
has
been
proposed
as
pathbreaker
verb
for the VO frame, yet itwas neither the earliest nor the most flexible verb in
this frame for either child. In fact, only Elaine followed the predicted pattern,
such that both her first uses of grammatical forms and her first flexible uses of
such forms appeared with light verbs (Table 18). These data, then, corrob
orate those of Theakston
et al. (2004), who found no consistent advantage
in
grammatical flexibility for light verbs at this early point in development.
Table 24 presents a summary for seven commonly
frames,
appearing
whether
they were first used by the children with light or with heavy verbs.
for the light verbs here either, as
Clearly, there is no consistent advantage
were
as frequently
at
verbs
the
least
for the S, O, -ing,
heavy
pathbreakers
79
TABLE 19
HEATHER'S
Verb
Light/Heavy
Comparison
for
Frame
Flexibility
Age
(Days)
Number
Preceding
Verb(s)
Light
Verb
Following
10th
Used
of Frames
in 10
Use
Use
Instances
584
629
Cut
585
590
2 (VO, SVO)
1 (V)
Go
585
590
590
593
634
595
629
1 (V P)
1 (V P)
2 (SV-ing, V locative)
4 (V, SV, SVO, SVedO)
Sit down
585
589
708
593
701
4 (V, SVO, VO, SV)

4 (V, V locative,
Push
588
600
SVP, VOP)
2 (V, VO)
Drop
595
629
Wash 593
652
595
624
611
615
618
631
633
653
634
652
Verb(s)
Cut
GO
Sit
Sit
COME
Drop
Kiss
PUT
WANT
Like
Need
LOOK
Eat
Throw
1st
669
(Frame)
Examples
this, I cut this
Sit down
I coming, Come
Drop, I dropped
here
it, I drop
I kiss it,Mommy
kiss, Kiss that
Put here, Put top on, I put on
Push, Push this

4 (V, SV, SVO, SVedO) Drop, I dropped
it, I drop
2 (VO, V)
Wash, Wash it
I want it, I don't want this
2 (SVO, SV+negO)
Like this, I like this
2 (VO, SVO)
2 (VO, SVO)
2 (V, V locative)
3 (V, SV, SVO)
701
5 (SVO, SV, V P,
SVP SVOP)
Me need
this, need this

Look, Look here
eat, Heather
Eat, Heather
eat this
I throw that, throw away
GIVE
687
708
5 (VO, VIO,
VPP, SVPP SVOPP)
I threw down, I throw it away

M give it, Give me, Give to D,
A gave to me
TAKE
701
701
Wave
702
827
2 (VO, SVO)
3 (V bye, V PP VL)
A gave coin to me
Take a bite, Daddy take this
Wave bye, We need to wave
Stop
710
710
1 (V-I-gerund)
to D, I waved M

Note.?
Stop eating
the flowers
Capi
first appear with light verbs;

and P frames. Some frames did consistently
the pathbreaking
verbs that were observed were not consistently
however,
the first one used
the ones predicted. That is, a light verb was consistently
for the VL frame; however, for only two children (Carrie, Stacey) was this, as
was come (one
the pathbreaker
go. For the other four children,
predicted,
or
a
was
used
look (three children). Similarly,
child)
light verb
consistently
first in the VOL frame, but for only one child (Heather)
this was the pre
was give (two), take
the pathbreaker
dicted put. For the other five children,
or
a
stative rather than action verb, was the
(two),
bring (one). Finally, want,
first verb produced
in the SVO frame for six of our children; need, another
80
VERBGROWTH
DIFFERENCES
INEARLY
TABLE20
MAE's
Verb
Light/Heavy
for
Comparison
Frame
Flexibility
Age
(Days)
of Frames
Number
Preceding
Verb(s)
Light
Verb
Following
Verb(s)
Push
COME
Throw
1st
Use
10th Used
in 10
Use
Instances
1 (V) Push
1 (Von)
577
589
616
625
624
630
1 (V)
1 (V)
(Frame)
Examples
Come
on
Throw
654
669
GO 661
684
673
756
1 (V)
2 (V, SV)
Run 731
750
2 (V, VP)
Cut 740
819
757
757
742
776
3(V, SV, VO)

1 (V loc)
3 (V, SV, V-neg)
765
898
5(VO, VOP,
SVOP SVO)
727
733
3 (V, SV, VO)
Open,
Open
Me want
Eat
Fall
LOOK
Cry
Pull
Eat
Push
fall
Fall, Mommy
Run, Run away
Cut, Me cut, Cut it
Look me
Cry, Jill cry, Don't cry
Pull feet, This one pull
off, Me pull it over,
Me pull
Open
WANT
it
Mommy
this
open,
769
838
3 (SVO, VOS-bar,
765
898
5 (VO, VOP, SVOP, SVO)
773
777
791
834
2 (SVO, SV)
1 (SVO)
Me hold her, Me hold

Me bring rock
789
820
3 (VO, SVO, NegVO)
Me need
795
863
3 (VOP, SVOP, SVO)
Drop 802
818
2 (SVO, VO)
812
856
4 (VP, VOP, VOPP,
2 (SVO, S-neg VO)

2 (SVO, S-neg VO)
3 (SVO, SVOIO, SVOPP)
Pull
Hold
BRING
Need
TAKE
PUT
Like
833
910
Like 833
910
840
928
GIVE
VO)
cheese, Want
Panda watch me, I want that
Pull feet, This one pull off,
Me pull it over, Me pull it
that, No need
that plate
Take that off, Me take
my pie out, I take it
SVPP)
Drop my Teddy, Me drop

my Bandaid
Put on, Put me over,
Put me up Jill's table
I like that, I don't
I like that, I don't
like pears
like pears
ring to Jill,
give Jill a toy
I made, Me made tower
Me wash it, wash my hands
Me give my
Me
MAKE
Wash
851
880
855
868
3 (VO, SVO, past)

2 (SVO, VO)

Note.?
Capi
desire verb, was the SVO pathbreaker

for the seventh child, and cut or hold
for the eighth.
The finding that the particular light verb used first in a frame varied across
children
is contrary
to
two
crucial
parts
of
the
light-verbs-as-pathbreakers
81
TABLE21
NED's
Verb
Light/Heavy
for
Comparison
Frame
Flexibility
Age
(Days)
of Frames
Number
Preceding
Light
Verb
Verb(s)
1st
Following
10th
in 10
Used
Use
Use
Instances
Roll
519
549
COME
526
549
2(V, Ving)
2 (V, V P)
526
580
549
Lay
Like
556
611
Verb(s)
Bite
593
580
Examples
Roll, Rolling
Come, Come
1 (V)
Bite
2 (V, V P)
2 (V, VO)
568
WANT
578
Need
(Frame)
Lay, lay down,

Like, Like it
Want some
1 (VO)
2 (VO, SVO)
615
a cookie,
a pen
Need
needs
Note.?V
highlighted
=
=
Verb; S
subject; VO
verb-object;
on
lay back
Pop
verbs are
Capitalized
verb's trans
targets the pathbreaking
hypothesis. One part of the hypothesis
to
to
is
enable
children
conceive
of or to
which
parency of meaning,
proposed
8c Sethuraman,
That
abstract the frames (Goldberg, 1999; Goodman
is,
2006).
children learn how go encodes inchoative motion, or put encodes caused mo
tion,
or make
involves
caused
action,
and
then
deduce
the
from
verb's
meaning
that go is to be followed by locatives, make is to be followed by object names, and

put is to be followed by object names and locatives. From these "semantic
frames VL (or VPP), VO, and VOL (or
frames," they abstract the grammatical
and
all
be considered equally representative
VOPP). Bring, take,give,
put might
of the caused motion
construction VOL, such that the variation here is not
a
look could not possibly be considered
particularly problematic. However,
transparent
look
involves
representative
a more
abstract
of
the
sense
inchoative
motion
with
of motion,
construction
someone's
VL,
gaze
because
being
di
to
toward a ground rather than an object or person actually moving
Mae's
and
Sam's
&
ward a ground
1985;
Gleitman,
Levin,
1993).
(Landau
innovative use of the VL frame first with look seems inexplicable in terms of an
argument based on semantically transparent underpinnings.
rected
verbs is
variability with the SVO frame pathbreaking
If se
for
the
hypothesis:
similarly problematic
of
children's
ab
is
the
action
of
transitive
mantic
transparency
instigator
as Hopper
and
straction of the SVO frame, then want hardly qualifies;
(1980) and others have discussed at length, verbs such as make or
Thompson
The
push,
which
children's
capture
caused
changes
of
state
or
position,
are
better
candidates.
et al, 2004) have already grappled

Ninio (1999) and others (e.g., Theakston
so early in the SVO frame
with the inconsistency of verbs like want appearing
82
DIFFERENCES
INEARLY
VERBGROWTH
TABLE22
SAM's
Light/Heavy
Verb
for
Comparison
Frame
Flexibility
Age
(Days)
Number
Preceding
Verb(s)
Light
Verb
1st
Following
10th
Used
of Frames
in 10
Use
Use
495
566
1 (V)
GO 531
567
2(SV, V)
1 (V)
Verb(s)
Open
Cut
Cut
LOOK
Eat
Eat
COME
Hold
Drop
WANT
540
583
540
583
553
592
569
660
Instances
(Frame)
Open
Go, I go
Cut
1 (V) Cut
1 (VPP)
1 (V)
1 (V)
Examples
Look
at me
Eat
569
660
576
602
571
607
602
783
4 (V, SV, VO, VPP)
632
722
3 (SVO, VO,
Eat
1 (V)
1 (V)
Come
Hold
SVS-bar)
M drop, Drop
Drop in tub
I want
I want
Roll
644
711
602
783
634
663
644
711
Need 646
674
664
739
Drop
TAKE
Roll
BRING
Lay
PUT
Run
cup, Want
to go
2(V, VO)
4 (V, SV, VO, VPP)
Roll ball, Roll

M drop, Drop
Drop in tub
1 (V)
2 (V, VO)
2 (VO, SV)
Take
4 (VO, VOP,
Bring
diaper,
that, Bring my
plate over, No bring,
673
705
3 (V, VO, negVO)
679
739
3 (SV, SVP, SVPP)
694
732
697
750
2 (V P, VOP)
5 (V, Ving, SVPP,
SV, VPP)
Note.?
cup,
Roll ball, Roll

I need, Need this
negV, Ving
Like
diaper,
Bringing
Like those, M like,
Don't like you
Teddy
lay, Lay down,
Lay down with me
Put away, Put that back
I run in water
Running,
Capi
but not encoding highly transitive actions, and

they have proposed different
resolutions. Ninio
that an even more abstract
(1999) suggested
("prototyp
the subject's/agent's
relation is more
ical") notion of transitivity, in which
facilitates children's abstraction of the SVO frame, and Ned's use
prominent,
of need fits this notion. However, Carl's use of cut does not (cut
clearly changes
the patient and ismore similar to the
and Thompson
version); thus,
Hopper
there is still individual variation unaccounted
for.
83
TABLE23
STACEY's
Verb
Light/Heavy
Age
for
Comparison
Light
Verb
Verb(s)
1st
Following
Use
Verb(s)
10th
Use
of Frames
in 10
Used
Instances
(Frame)
529
1 (V) Sit
491
498
495
624
2(WhSV,
V)
5 (V, Ving,
503
533
VingO)
1 (V)
512
611
2(VO,
516
611
503
533
2 (Ving, SVing)
1 (V)
COME
518
600
TAKE
520
595
Sit 491
GO
Eat
Open
Run
Open
Hold
LOOK
Jump
Nanny go? Go
Eat pizza, I'm eating, I eat it
Where
Open
I want juice, Want
some more
SVO)
Running,
Open
Come
3 (V, V P, SV P)
2 (VO, VOP)
2 (VO, SVO)
540
593
572
621
559
584
2 (V, V PP)
4 (Ving, SVing,
581
SVed, V)
2 (SVing, VO)
555
Wash
Examples
SVO,
SVing,
WANT
Flexibility
(Days)
Number
Preceding
Frame
PUT
572
598
MAKE
598
675
1 (VOP)
3 (SVO,
(5th)
VO,
Hold
in, You come out

it, Take it off
that duck, I hold this
Look,
Look
Take
in mirror
Kitty jumping,
Kitty jumped
Mommy's
washing,
Wash my hands
Put it on, Put phone
Daddy make it,
making messes
-ing)

Note.?
Girl running
back
Capi
The
second
part of the
hypothesis,
between
the hypothesis
resolution
of the inconsistency
potential
frames is highly
and the data, is that children's abstraction of grammatical
on
with
which
the
particular
light verbs appear
frequency
dependent
and
in
variety
abstraction
or
put
encodes
in
frames
of
the
caused
that go is frequently
followed by object
frequently
names
and
locatives.
Then,
That
is,
an
alternative
learn how go encodes

or
motion,
observe
object
input.
is that after children
make
followed
names,
over
involves
caused
action,
by locatives,
and put is frequently
hearings
motion,
they
next
(or want) is
followed by
make
numerous
of
process
inchoative
of
these
verbs
with a wide variety of locatives and object names in the relevant positions
in the input, the children
abstract
the VL
(or VPP), VO, and VOL
not rely on the meaning
does
latter
This
frames.
(or VOPP)
procedure
but does rely on the frequent
the abstraction
of the verb to enable
(with varied lexical items) of a single verb in a given frame
appearance
84
DIFFERENCES
INEARLY
VERBGROWTH
TABLE24
First
of Grammatical
Uses
Forms
by Verb
for
Subclass
Each
Child
Child
Grammatical
Form
SVCarrie,
Elaine*
Heather,
VO
Heather*,
Sam*,
Carrie
VP
Sam,
VL Carl
Mae
VOL Carl
Mae
Sam,
First Use With
Verb
Stacey
Elaine*
SVO Carrie*,
"-ing"
a Light
First Use With
Heather*,
Elaine*,
Mae*
Carl,
Carrie,
Mae,
Ned
Carl,
Carrie,
Mae,
Ned,
Sam,
Stacey
Ned
Carl,
Verb
Heavy
Stacey*
Heather
Carrie,
(come)
Stacey, Ned
(look), Carrie
(come)
(go), Heather
(look), Sam (look), Stacey
(go)
(give), Carrie
(give), Heather
(put)
(take), Sam
(bring), Stacey
Heather,
Mae,
Ned
(lay)
(take)
Note.?*Child's
first use was with the verb want.
V = Verb; S = subject; SV = subject-verb; VO = verb?object;
et al., 2003; Chenu

8c Jisa, 2006; Goldberg,
(e.g., Casenhiser
8c Sethuraman,
Goodman
8c Tomasello,
2006; Kidd,
Lieven,
et al., 2004).
Theakston
1999;
2006;
This frequency-based
does indeed account for the predom
procedure
inance of want as the children's pathbreaker
for the SVO frame; however,
it encounters
in accounting
difficulties
for the occurrences
of look or
come
as Mae's,
and
Heather's,
Sam's
pathbreakers
for
the VL
frame,
because
look and come are much

less frequent
verbs than go in children's
input
and output (Naigles 8cHoff-Ginsberg,
the highly differential
1998). Moreover,
of put, bring, take, and give (put appears
2.5-3
frequencies
approximately
times more frequently
than the other verbs; Goldberg,
8c
1999; Naigles
in
of
and
children's
make
1998)
corpora
Hoff-Ginsberg,
caregivers'
speech
this process
less viable as an explanation
of children's acquisition
of the
so
VOL frame/construction:
is
in
the
VOL
construction, why do
Ifput
frequent
Sam, Stacey, Carrie, Mae, and Carl all use other verbs with their first VOL
frame?
In other
words,
the
semantic
transparency-based
version
of
the
light
hypothesis may account

but it does not account
for the children's first uses of

for their first uses of the VL
version of this hypothesis may ac
verbs-as-pathbreakers
the VOL construction,
and SVO frames. The frequency-based
count for the children's first uses of the SVO frame, but it has difficulties
for their first uses of the VL and VOL frames. Our data, then, do
accounting
not support the
as a process
pathbreaking-verbs
hypothesis
by which chil
dren
learn
to abstract
sentence
frames.
85
HOWAREDIFFERENT
MEASURES
OF VERBDEVELOPMENT
AND USERELATED?
to generativist
that underlies
theories, the abstract grammar
is
the
that
children's
autonomous,
language productivity
yielding
prediction
use
to the
of
flexible
verb
should
be
unrelated
development
grammatical
use
or
a
to
of
flexible
verb
the
of
verb
development
growth
semantically
lexicon (e.g., Chomsky,
includ
1981). Other theories of verb development,
results from domain-general
ing theories that grammatical
development
According
theories
processes,
learning
that
grammar
from
emerges
the
lexicon,
and
the syntactic bootstrapping

hypothesis,
predict that the acquisition of a verb
of semantically flexible verb use, and the devel
lexicon, the development
flexible verb use should all be positively
related
opment of grammatically
across children. The syntactic bootstrapping
pre
hypothesis
additionally
are
dicts that the grammatical
with
which
different
verbs
used
flexibility
should be related to the semantic flexibility with which those verbs are used
because
semantic understanding
of a verb is based on syntactic analysis of
verb's
that
environments.
One way to investigate these hypotheses

with the current data might be
to compare the degree and onset of flexibility demonstrated
by the children
in the different domains. Although
of
it is not a given that our measures
are
so
and
semantic,
(i.e.,
syntactic flexibility
directly comparable
pragmatic,
as a change in syntactic frame?),
is a change in actor of the same magnitude
we
flexibility
syntactic
of
number
t tests
pairwise
performed
using
between
four
number
instances,
of
days,
affected
actor,
addressee,
indices
of flexibility
of
percent
(percent
and
verbs)
object,
and
of children,
the
Bonferroni
to set alpha at .0083. Of all 16 comparisons,

only 1 reached
significance: Children were significantly more flexible with their affected
=
= .0018. Im
object use than with their syntactic frame use, ?(7) 4.83, p
correction
none
portantly,
viding
earlier
A
of
the
no evidence
than
better
time-related
reached
comparisons
for the notion
pro
significance,
in one domain
that flexibility
emerges
in another.
way
to address
these
proposals,
we
feel,
is to examine
intercor
and use. To reduce the num

relations among properties of verb development
on
we
of descriptive
focused
selected measures
ber of correlations calculated,
and
theoretical
interest,
and,
where
possible,
we
created
composite
scales.
OF CHILDREN'S
VERB
AMONGMEASURES
INTERCORRELATIONS
AND USE
DEVELOPMENT
measures
were
Composite
use
to
verbs and
dren started
created to index the age at which the chil

the semantic and grammatical
flexibility of
86
DIFFERENCES
INEARLY
VERBGROWTH
TABLE25
Intercorrelations
Among
of Children's
Measures
for
all
use
composite
4. Number
composite
flexibility
of verbs at 21 months
5. Number
of frames
6. Number
of productive
frames
*p<.05
5.
.042
.527+
Use
-.456
.749**
.624*
.537+
?
? ?
at 21 months
(one-tailed),
4.
-
3. Grammatical
Note.?+^?<.10
and
3.
2.
-.670*
?
begins
flexibility
Development
(n
= 8
Analyses)
1.
1. Age verb
2. Semantic
Verb
6.
-.189
.110
.586+
.222
.796**
.119
.839**
.183
?
at 24 months
(one-tailed),
**p<.01
(one-tailed).
their first 10 uses of early-acquired

verbs. The scales were created as follows:
was
use
at
verb
which
the
average of two measures?the
Age
age of first
begins
verb use and the mean age of first use of all the target verbs the child used in
a for this scale was .75. Semantic
the window of this study. Cronbach's
flex
use
was
measures:
verb
the
of
three
the percent of verbs used
average
ibility of
with different actors, the percent of verbs used with different affected ob
jects, and the percent of verbs used to refer to different actions (Cronbach's
oc= .89). Grammatical flexibility
of verb use was indexed by the average of four
measures:
of
verbs
used
with different frames (excluding vocatives),
percent
of
verbs
used
with
different
percent
percent of verbs used with
morphology,
different
lexical subjects, and percent of verbs used with different
lexical
=
of
.87). We also included two snap-shot measures
objects (Cronbach's a
at 21 months,
children's verb lexicons and verb grammars
which was the
oldest age at which we have data for all participants:
the number of target
verbs in the child's vocabulary and the number of frames (out of 5: SV, VO,
SVO,
VP,
productive
frame
V-ing)
in the
child's
utterances.
Last,
frames each child had at 24 months,
criterion.
For
each
correlation,
we
expected
we
counted
employing
that
the
of
number
the 5-verbs-per
if the measures
were
related they would be positively related, except that age would yield a neg
ative correlation
because a younger
of verb use is a
age at the beginning
of
indicator
verb
all
tests were
therefore,
positive
development;
significance
one-tailed. The correlation matrix
is presented
in Table 25.
The correlations
of children's verb use in Table 25
among measures
that children who began verb use at a younger
age showed more
in their first verb uses, and they also tended to have
at 21 months
larger verb vocabularies
compared with children who began
verb use at an older age. The semantic flexibility with which children used
their first verbs was also significantly related to the grammatical
flexibility of
their verb use, to the size of their verb vocabularies
at 21 months,
and
to
the
of
number
at
frames
had
months.
24
(marginally)
productive
they
reveal
87
TABLE26
intercorrelations
among
verbs'
of use
properties
of Verbs
in early
1.
1. Age
onset
of
child
speech
2.
4.
3.
.122
-.277+
children
3. Percent
using
children
4. Percent
using
children
5. Grammatical
Note.?+p<A0
using
flexibility
(one-tailed),
to refer
with
with
different
different
of use
*p<.05
to different
actions
actors
affected
.100
5.
-.036
(28) (34) (25)
2. Percent
(number
in Analysis)
.373+
.336*
(34)
.294+
(28) (19)
(28)
.349*
.020
?
(25)
(34)
.307+
objects
(25)
?
composite
(one-tailed).
at 21 months
verb vocabularies
showed a marginally
with
correlation
overall
grammatical
significant positive
flexibility of their
first verb uses and a significant positive
correlation
with the number
of
at
frames
had
months.
The
24
they
grammatical
flexibility of
productive
children's verb use also was positively related to children's number of pro
that flexibility of use is, in fact, a
ductive frames at 24 months,
suggesting
The
good
size of children's
index
of
grammatical
productivity.
AND USE
OF VERBS'DEVELOPMENT
AMONGMEASURES
INTERCORRELATIONS
of verbs, as they are
To analyze the intercorrelations
among properties
of grammatical flexibility
first used in children's speech, a composite measure
was created by averaging
four measures:
the percent of children using the
one
more
more
inflec
with
than one morphological
than
verb with
frame,
one
more
more
one
lexical
and
with
than
lexical subject,
than
tion, with
=
indices of semantic flexibility did
.79). The multiple
object (Cronbach's a
not form an internally consistent scale?as Table 26 shows, the three indices
of semantic flexibility were not significantly related to each other. Thus, the
were entered
the per
into analysis individually:
three measures
following
cent of children using the verb to refer to more than one action, the percent
than one actor, and the percent of
of children using the verb with more
children using the verb with more than one affected object. The final mea
sure of how verbs appear in children's speech was the average age at which
the
the verb first appeared. As was the case for the by child correlations,
predictions
were
that positive
indicators would
be positively
related,
and all
88
DIFFERENCESINEARLY
VERBGROWTH
tests were one-tailed. The correlation matrix

in
is presented
significance
Table 26.
The correlations
of how verbs were used within these
among measures
first instances of children's verb production
show that later-appearing
verbs
an
to
to
with
of
verbs
older
tended
be
used
children
fewer
onset)
(i.e.,
age
by
refer to varied actions and were used by significantly more children with
grammatical
flexibility There were also trends toward the number of chil
dren using a verb to refer tomultiple
actions being positively associated with
the number of children using that verb to refer to different affected objects
and negatively associated with the number of children showing grammatical
flexibility of use. Verbs used by more children with different actors were also
used by more
The
children
correlations
verbs were
with grammatical
with
age
of
onset
flexibility
indicate
that,
on
average,
early
used with
grammatical
appearing
flexibility by fewer children
than later-appearing
verbs. Such correlations might appear to be consistent
with hypotheses
that grammar
later, and thus perhaps as a con
emerges
Bates
of
lexical
8cGoodman,
1999). However,
sequence,
development
(e.g.,
these correlations were only significant in the by-verb analysis; across chil
to their
dren, the age at which children began to use verbs was unrelated
the analyses
of individual
grammatical
flexibility.
composite
Similarly,
of
earlier
in
described
this chapter,
growth patterns by period
development,
indicated that some of the children used their very first verbs with gram
matical flexibility. This appears to be a case in which averaged data (in this
case average properties
of verbs rather than of individuals) present a pic
ture that is true for some individuals but not for others (see Molenaar,
2008,
for discussion).
CONCLUSIONS
In this chapter, we looked at the grammatical
flexibility of early verb use
in more detail, asking whether
the overall grammatical
flexibility we ob
served might obscure a limited, very early period of less flexible verb use
and whether
the lack of average differences
between
light and heavy verbs
obscure a pathbreaker
role played by some light verbs for some
might
children. We found that a small number of children did manifest
conser
vative before
flexible verb use, but the majority of children demonstrated

to the beginning,
close
if not at the beginning,
of verb use. We also
flexibility
a light verb was the first to appear with a
found that sometimes
particular
in syntactic
syntactic frame, but that sometimes
heavy verbs appeared
frames before light verbs. These findings
that early
support the conclusion
flexible verb use is widespread
and not limited to a late stage of develop
89
ment
or a particular
out that
kind of verb. These
findings also pointed
not
of
do
reflect
all
average developmental
necessarily
descriptions
paths
individual developmental
paths.
We also asked whether different aspects of early verb use were related.
Direct comparisons
of the degree and onset of grammatical,
semantic, and
indication
of
differences
these
little
between
pragmatic
flexibility
yielded
domains;
the
however,
were
correlations
more
informative.
Across
children,
grammatical
flexibility and semantic flexibility of verb use were related to
each other; similarly, verbs showing actor flexibility with more children also
children. These
showed grammatical
flexibility with more
findings are
lexical
about the interrelation
between
consistent with many hypotheses
reflect
the
of a
The
correlations
effect
and grammatical
may
development:
cor
common
common underlying
of
and
the
effects
and/or
ability
input,
effects of syntactic under
relations may also indicate mutually
supportive
standings
on
and
semantics
versa.
vice
The
that
hypothesis
grammar
a particular
threshold
size received only
emerges
a
the
correlation
between
weak
very
significant
support?in
marginally
and their overall
number of the target verbs children used at 21 months
we found that those verbs
flexibility of verb usage. Although
grammatical
were
at younger
with
used
that appeared
ages
flexibility by
grammatical
once
fewer
do not
the lexicon
on
children,
average,
interpret
who
children
flexible
in their verb uses

children
were
than
this as reflecting
because
those
reaches
who
verb
began
began
those
use
than children
to
use
verbs
verbs
earlier
were
later
less
later. We
showed
we
ages,
of verb
period
not
who began
earlier
at
acquired
a pregrammatical
use
grammatically
did find that
greater
semantic
verbs may be used with less grammatical

flexibility. Thus, earlier-learned
seman
but
earlier
children are likely to be more
verb-learning
flexibility,
in
the
of the
discussion
these
flexible.
We
consider
again
findings
tically
theoretical
implications of these data.
90
VII.GENERALDISCUSSION
is children's first uses of their first verbs.

the study of verb development
has bur
of early verb comprehension
geoned,
including experimental
investigations
and verb learning,
studies of production
using samples of spontaneous
a
one
and
of
child's
speech,
early verb use (e.g., Fisher & Gleit
diary study
8cGolinkoff,
man, 2002; Hirsh-Pasek
2006; Lieven, 2006; Tomasello,
1992;
The topic of this monograph
In the past decade and a half,
8cMerriman,
Tomasello
1995). There have not, however, been studies of
in the very earliest stage of verb production
several children
that would
a
more
use
can
a study
of
verb
than
provide
generalizable
early
description
of a single child. This absence of detailed documentation
of the beginning
of
verb use was a significant gap in the data on verb development
because
there
are
disputes
the
regarding
nature
of
early
language
representation
and, indeed, the nature of the language acquisition process addressable

by
was
data on how children use their very first verbs. The present
study
to fill this gap.
designed
of any verbs, eight mothers
Starting before their children's production
kept records of their children's first 10 uses of 34 target verbs that the
literature suggested
should be among children's first verbs. For
previous
each instance of verb use, the mothers
in which
recorded
the full utterance
and also noted whether
the verb appeared
the utterance
the
containing
verb
served
the
function
of
a command
or
who
description,
or
was
what
the verb action, and who or what was the object of the verb
performing
action. These diary records were coded to provide measures
of the prag
uses. Our
and
of
children's
first
verb
matic, semantic,
grammatical flexibility
discussion
of these data and of their theoretical
is
implications
organized
around the following questions:
1. To
what
flexible
2. To
what
flexible
extent
are
in pragmatics
extent
are
children's
early
verb
uses
conservative
versus
verb
uses
conservative
versus
and semantics?
children's
early
in grammar?
91
3. To what extent is early flexibility particularly

some (i.e., light, or
verbs?
pathbreaking)
are the flexibility
4. To what extent
are
5. What
the
theoretical
of meaning
of
implications
flexibility
the
characteristic
of only
and form related?

observed
of children's
pragmatic,
first verb uses?
ARECHILDREN'S
EARLY
TOWHAT EXTENT
VERBUSESCONSERVATIVE
VERSUS
FLEXIBLE
INPRAGMATICS
AND SEMANTICS?
Pragmatic Flexibility
data of this study suggest that children's early verb use is prag
at least some of their first
flexible. All of the children produced
matically
verbs in both commands
and descriptions,
and the majority
of the verbs
were used as both commands and descriptions,
same
the
child
(and these
by
two functions captured >98% of the childrens recorded utterances). Thus,
to be completely
context bound in
neither any child nor any verb appeared
The
use
and,
ation
in mental
inference,
by
verbs
among
terances,
not
variation
by
such
no
commands;
can
verb
some
that
be
There
representation.
were
verbs
were
verbs
used
used
in
only
in
descriptive
commands.
the
verbs'
verbs
Furthermore,
addressees,
that
suggesting
the
as
used
children
were
commands
had
ut
The
inherent
and the suitability of these meanings

for the command
meanings
For all of the target
is not a sensible command).
(e.g., "Need"
are plausible
which both commands
and descriptions
functions,
within
the first 10
both commands
and descriptions
produced
recorded.
vari
however,
in
only
to differences
attributed
was,
function
verbs for
children
instances
with
used
of
representations
multiple
verb mean
we
actors. Although
ing that were sufficiently abstract to apply to multiple
did not explicitly assess the situations in which children used their verbs, the
fact that the verbs with the highest number of addressees were the ones with
more general meanings
(i.e., the light verbs come, go, look, put) suggests that
the children were applying the verbs across situations. This flexibility of use
across
situations,
combined
with
the
verbs'
use
to
multiple
addressees,
an account
ar
of early verb use as context bound. This discovery

gues against
use
across addressees
of the flexibility of verb
and, probably, situations was
the researcher
employed. Unlike
possible only because of the diary method
in a single setting, the mother keeping a
recording a sample of conversation
con
settings with multiple
diary observes children's verb uses in multiple
of
the
the
addressees
versational partners and potential addressees.
Indeed,
utterances
children's
in television
shows,
included
and
parents,
siblings,
other
relatives,
characters
pets.
92
GENERALDISCUSSION
Semantic Flexibility
from the be
data from this study also reveal semantic flexibility
use of a verb
as
is
defined
the
where
verb
semantic
use,
flexibility
ginning of
actors
and on
different
to refer to multiple
enacted
actions
by
appropriate
child's
10
of
the
first
instances
within
the
affected
different
produc
objects
in
tion of that verb. (The actors and objects need not have been expressed
or
use
her
that
his
indicates
child's
the point is that the
the utterance;
that
of the verb is not restricted to something
of the meaning
representation
a
to
a single actor does or something
that is done
single object.) Six of the
one
to
of their verbs in reference
in this study used at least
eight children
one
more
more than one action, and
third of the verbs produced
than
(38%)
were used to refer to multiple
actions. A total of seven verbs (come, cut, go,
to multiple
in reference
actions by at
open, put, take, wash) were produced
The
least half of the children.

in this study used at least some of their verbs with
All eight children
actors within the first 10 uses of those verbs, and most of the target
multiple
actors
verbs produced
(73%) were used with multiple
by these children
the children's very first use of approx
within the first 10 uses. Moreover,
was in reference
to an actor other than them
imately half of their verbs
was limited, the limitations
to
actor extendability
seemed
selves. Where
the child.
limitations within
rather than linguistic
reflect verb meaning
have
Children
and
may
like were
reflect
access
little
used
to others'
primarily
to use
occasions
restricted
wants,
needs,
in self-reference.
particular
and
Other
verbs
so want,
likes,
restricted
(e.g.,
roll
need,
actor uses
is used
only
there is a ball and thus ball is the only actor to appear with
(Naigles 8cHoff, 2006).
than actor flexibility:
Affected
object flexibility was even greater
when
at
some
children
used
affected
object, and 90% of the transitive
refer
to
an
action
least
on
more
verbs
in reference
than
one
to actions
and alternating
affected
object.
with
more
than
roll)
All
one
verbs were used

There
were
to
more
affected objects per verb in children's first 10 instances of use than there
were actors per verb; the time elapsed between first verb use with a different
affected object was less than the time between first verb use and first use
with a new actor, and all children showed a higher degree of affected object
than actor flexibility. Thus, most of the time, children did not
flexibility
relations to a single patient or theme;
restrict their talk about two-argument
to up to eight different patients or
in
used
their
verbs
relation
instead, they
in the early verb-learning
themes within 10 uses. These are the first findings
in produc
such a high level of semantic flexibility
literature to document
tion. They are consistent with the evidence of early flexibility from studies of
or novel
verb comprehension
unconventional
affected
using
objects
&
It
is
that
the
between
Hoff, 2006).
asymmetry
(e.g., Naigles
possible
93
actor and affected object flexibility

in the underlying
reflects a difference
actor
of
the
and
affected
roles
such
that children are
representations
object
more
a verb
to extend
reluctant
to a new
actor
to a new
than
or
patient
we
think a practical explanation

is more
However,
likely: In the
a
there
exists
wider
of
affected
child's world,
variety
possible
objects than
actors. That is, actors in children's worlds are mostly animate beings and
these consist of family and friends. In contrast, affected objects are primarily
theme.
inanimate objects, and these include all kinds of food, toys, household
items,
and so on. The set of possible affected objects is just bigger than the set of
actors.
possible
actor flexibility
and affected object flexibility
patterned
Importantly,
measure
were
in the composite
of semantic
combined
together?they
in the analyses of the intercorrelations
among verb use measures;
flexibility
was
both
and
their
zero-order
correlation
positive
significant,
=
=
=
that children who demonstrated
.831, jb .011. The
S)
r(n
finding
more
more
also demonstrated
and
and earlier affected object flexibility
earlier actor flexibility
resources
other.
the
Moreover,
that there was no tradeoff
suggests
to one
allocated
area
in fewer
resulted
children
were
who
early
in which
resources
verb
cognitive
for
available
learners
the
were
overall
that semantic
semantic flexibility,
the ones showing more
indicating
a
a
at
not
of
later
is
verb
stage.
property
developmental
learning
flexibility
actor
in
first
verb
children's
of
The
present
finding
flexibility
also
uses
is
at
preferred
odds
with
Huttenlocher
to use
verbs
in reference
et
al.'s
to themselves;
(1983)
that
report
the
however,
toddlers
discrepancy
et al.
inmethods
in terms of the difference
used. Huttenlocher
a
thus
in
and
mother-child
interaction
only
single setting,
dyadic
is explicable
sampled
two
possible
actors
were
available.
The
diary
method,
in contrast,
was
able
in which other
the day and across different
settings
instances
of actor flex
that
available.
The
fact
be
many
participants may
are
or
not
our
set
who
in
involved
data
pets,
usually included
siblings
ibility
to
credence
this
in recording
adds
further
sessions,
methodological
expla
actor flexibility with at least one
nation. In fact, each child demonstrated
the present
the action of his/her pet. Furthermore,
verb by describing
are
actor
of
with
the
of
consistent
comprehen
findings
flexibility
findings
sion studies, which show that toddlers can extend familiar and newly taught
et al.,
verbs to new actors and/or agents (Naigles 8c Hoff, 2006; Naigles
to tap uses
2005;
across
8c Forbes,
Poulin-Dubois
However,
the
children's
2006).
of
extension
new
verbs
to new
actors
was
not
On average, children had a verb in their lexicon for half a

instantaneous.
it to a new actor. Thus, it is
month and used it three times before extending
and Forbes's
that Poulin-Dubois
still possible
(see also
(2006) finding
not
match
et
do
that
20-month-olds
al., 2006)
just
consistently
Maguire
learned
verbs
to their
actions
when
these
are
performed
by
new
actors
94
may
GENERALDISCUSSION
an initial reluctance
children have (within the 10 minutes
a
to
extend
novel
verb's action to a new actor. And
usually provided)
et
contrast
between
al.'s (2005) findings
that children
the
Naigles
perhaps
and other researchers'
that
do extend verb meanings
findings
they do not
as
that
the
verbs
with
different
actors,
suggests
Naigles et al. did
presenting
indeed
during
reflect
training,
make
children
helps
their
own
In
extensions.
the
real
world, most children would experience most common verbs with multiple
actors well before
it seems that if such
the onset of speech; therefore,
actor
is
the
crucial
factor
for
then most children
experience
extendability
to new actions
would have had this experience
and be able to generalize
one
were
actor.
in the present
So
with
the
why
presented
1-year-olds
only
more
in
than
in com
flexible
others
have
demonstrated
study
production
are
A
is that 1-year-olds
prehension?
likely possibility
simply less efficient
learners, such that (in the absence of experiencing
actors) they
multiple
need more time with the teaching stimulus (more than the 18 or so seconds
in experimental
time after
usually provided
settings), or possibly more
to
consolidate
the representation
before extending
the verb to an
teaching,
action
performed
by
a novel
actor.
Context-bound
use,
then,
may
be more
or time with particular verbs rather

function of the amount of experience
a
course
than
of verb learning.
property of the overall
et al.'s (1995) proposal
These findings also cast doubt on Golinkoff
that
the principle
of extendability
that children apply to new nouns must be
relearned for verbs. The speed with which our eight children demonstrated
action, actor, and affected object flexibility
(within 10 to 16 days of the first
use of that verb) suggests that whatever
principle was acquired for noun
to use with verbs. Stated inmore
learning was easily and rapidly transferred
general terms, the picture of early flexible verb use suggested by the present
data contrasts with those descriptions
in the literature that paint early word
use,
attested,
in
as
general,
underextended
context
bound.
usage
may
While
be more
instances
of
characteristic
are
underextension
of nouns,
whose
acquisition begins earlier in English, than of verbs. Our data show that early
verb use is not typically context bound.
TOWHAT EXTENT
ARECHILDREN'S
EARLY
VERBUSESCONSERVATIVE
VERSUS
FLEXIBLE
INGRAMMAR?
The data from this study show clearly that within their first 10 instances
a verb in spontaneous
of producing
speech, children demonstrate
flexibility
in the syntactic environments
in which
those verbs appear. The children
used multiple
frames with two thirds of their verbs, and these first verbs
were used in different frames, on average,
by two thirds of the children. On
children's
first
in
average,
change
syntactic frame took place within half a
95
month
of the first verb use; thus, there was no extended
period of frozen
form use. On average, only 12% of such frame changes included the use of
in
1.8% of utterances
(range 0-36% across children; on average,
negation
most
of
frame
observed
the
cluded negation markers);
therefore,
flexibility
structure of the verb. Mor
in the argument
involved children's flexibility
reflecting just the children's uses of verbal suffixes,
flexibility,
phological
at a much
lower rate. Some children did not use any inflections
occurred
with
"-ing")
their verbs
and
examples
(most of the morphological

of
past
tense
or
forms
third-person
that were
singular
use
used were
were
rare.
close to a
addition or subtraction of such a morpheme?took
Changes?the
to be observed after the first use.
month
in verb use was the extent to which verbs
of flexibility
Another measure
were used with different
slot or in the subject
lexical items in the preposition
or object positions.
in filling all these
Children
demonstrated
flexibility
In
the children
lexical
items.
with
slots
particular,
multiple
grammatical
in which a verb was used with at least two
showed lexical subject flexibility,
one third of their verbs; just over one
different subjects, with approximately
with
used
verbs
the
third of
children
subjects. Lexical object flex
multiple
was
half of the children and by
for
under
verbs
demonstrated
just
ibility
by
was seen in
half
of
their
verbs.
children for just under
flexibility
Preposition
at
these
all.
five of the six children who produced
Tellingly,
prepositions
measures
of
cohered with the other
indicators of lexical flexibility
gram
matical
flexibility
our
in forming
composite
grammatical
flexibility
score.
in syntactic flexibility
individual differences
analyses
Finally,
course
of syntactic flexibil
the
developmental
(Figures 10-18) investigated
to
fewer than
in
the
conservative-child
child.
each
hypothesis,
Contrary
ity
initial conservative
half of our children (n = 3) demonstrated
verb-only use
4-5 months).
for
verbs
after
followed by later syntactic flexibility
(i.e.,
using
use of verbs with arguments
demonstrated
five
children
More
tellingly,
for two of these children,
from the start of verb acquisition;
syntactic flex
was evident within the first 6 weeks of verb use.
verbs
with
ibility
multiple
in the use of one verb appeared during
For the others, syntactic flexibility
the first period and more verbs came to be used with flexibility during the
almost always before the age of 2 years.
8 weeks?and
subsequent
of grammatical
In addition to these measures
flexibility, we also applied
measures
that have been
of productivity
to the present data corpus-based
We
found that
literature
in
the
Shirai,
1989;
1998).
(Ingram,
employed
a frame must be used with five
more
criterion?that
the
stringent
using
of productive
number
verbs to be considered
different
productive?the
frames children had at 24 months was positively correlated with the overall
is
measure
of their verb use. This correlation
of grammatical
flexibility
is an indicator of productivity.
that flexibility
consistent with the argument
3-Verb or 5-Verb
of the children who achieved
We found
that many
of
96
GENERALDISCUSSION
with a given frame (e.g., using VO with three or five different

productivity
so
did
before the age of 24 months;
three children even achieved
verbs)
such productivity with almost all of the relevant frames before 24 months of
age. Across the entire corpus of first uses of first verbs, we found that the
children used the SV and VO frames with > 60% of their target verbs, the
SVO frame with almost half of their target verbs, and the VP and V-ing
frames with fewer, but still multiple
verbs. Thus, well before 36 months
of
even
of
before
months
children
24
age?and
frequently
age?the
provided
evidence
that their frames were not necessarily
restricted to specific lexical
items.
The measures
of productivity
are
measures
indirect
only
study
and flexibility employed

in the present
as we were un
of hallmark productivity,
able to assess how any of the children's utterances
from their
diverged
the
children
had
heard
these
verbs
used
Indeed,
input.
likely
flexibly by the
adults around
them. The picture of early verb use that these measures
is in stark contrast to the description
of early verb use that
reveal, however,
has been the basis for claims that young children are not productive
verb
users.
use
With
across
a
of different
variety
frames
others
and
measures,
assessing
some
children's
assessing
frame
children's
use
across
verb
verbs,
these eight children have demonstrated

that their early verb use was flexible
rather than restricted
and thus more
consistent with an account of chil
dren's early verb use as reflecting abstract underlying
than with
knowledge
a description
of early verb use as strict imitations of input. The abstract
need not be innate, as the traditional generativist position would
knowledge
hold. Rather, children's flexible verb usage may reflect children's
learning
that flexibility
is general characteristic
of verb use from the evidence
of
flexible uses of particular verbs in their input. Put another way, children's
flexible verb uses may reflect learning from input, but it is learning that is
more abstract than learning the
uses of individual verbs modeled
multiple
in the input.
TOWHAT EXTENT
ISEARLY
FLEXIBILITY
PARTICULARLY
CHARACTERISTIC
OF ONLY
SOME (i.e., LIGHT,
OR PATHBREAKING)
VERBS?
The
picture of early grammatical flexibility of verb use suggested by the

was not limited to only some verbs. There was no evidence of a
data
present
role for light verbs in early grammatical
special pathbreaking
development.
were
verbs
neither
the earliest verbs produced,
nor were
Light
they, as a
flex
class, the verbs with the earliest or greatest syntactic or morphological
Of
11
the
verbs
before
the
of
620
age
ibility.
produced
days (on average),
some were light (come, go) but others were not (throw, sit, see,
push, open, lay,
when the children are considered
fall, eat, bite). Moreover,
individually, only
97
an advantage for light verbs, such that

one child consistently demonstrated
her first use of each frame and first instances of grammatical
flexibility were
shown for light verbs. Only sporadic evidence of a light verb advantage was
found for three other children, and for the other four children none of their
than heavy verbs that
light verbs showed more
grammatical
flexibility
at the same age. When
the frames were considered
emerged
individually,
the
SVO
first with a single verb, and this
frame
only
consistently
appeared
verb (want for six of the eight children) was hardly the most semantically
first varied widely
transparent. For all other frames, the verbs that appeared
across children and were neither consistently
transparent nor
semantically
our findings
consistently highly frequent in the input. Taken together, then,
not
do
1999;
support the proposals
(e.g., Chenu & Jisa, 2006; Goldberg,
of grammatical
Ninio,
1999) that light verbs are the primary promoters
development.
in our data that the children

There were, however,
several indications
did use light verbs differently
than heavy verbs. Light verbs appeared more
in descriptions
in commands
and less frequently
than did heavy
frequently
addressees more than
verbs. Moreover,
light verbs were used with multiple
heavy
verbs;
heavy
verbs,
also used with more
they were
and
there
actors
with
was
a trend
in the
different
same
affected
a
for
direction
than
objects
use
greater
of
light than with heavy verbs. Finally,

light verbs
more
in
SVO
frames
than
with
direct
and
objects
frequently
appeared
more
were
with
earlier
and
lexical
and
used
verbs
did,
object
they
heavy
than heavy verbs. All of these findings derive in a straightforward
flexibility
of light and heavy verbs. The
in the meanings
fashion from differences
can be attributed
to the behavior
association of light verbs with commands
different
of
three
monly
in
verbs
as
used
included more
not
be
come,
particular,
commands
in
as commands.
used
In
contrast,
com
to be
likely
the
subclass
heavy
states (like, need, cry), which
to internal
the more
Furthermore,
seem
look, which
as well.
input
verbs referring
and
go,
verbs
general
could
in the
light
situations and so can be used with

verb subclass can apply to more different
more different
That
and
addressees.
affected
actors,
is, anyone
objects,
can go or come and anything can be brought or taken, whereas only eatables
can
eaten
be
and
animates
only
can
eat.
And
because
heavy
refer
verbs
to specific events in the world, their direct objects can

directly and obviously
common
of
between
because
be omitted
speaker and hearer
ground
more
to
with
be
retained
forms
need
these
whereas
general verbs. The
to have
some
at
the
addressee
need
of
of
the
children seemed aware,
level,
some
affected
al., 2006).
of
common
objects
made
If such objects
ground,
then,
more
explicit
than
are required more

because
of base
others
with
rates,
(see
also
et
Matthews
light verbs by the dictates

more
different
NPs
should
of light verbs, appearing

the grammatical
be used. Thus,
advantage
more with objects and in SVO frames, is directly traceable to the semantic/
98
GENERALDISCUSSION
of those verbs in usage, not to their unique map

of specific frames. This is consistent with Snedeker
that light verbs, because of their appli
and Gleitman's
(2004) suggestion
a
to
of
such
wide
situations, must be the products rather than
cability
variety
the engines of early grammatical
acquisition.
pragmatic
ping onto
requirements
the semantics
ARETHEFLEXIBILITY
OF MEANINGANDOF FORMRELATED?
TOWHAT EXTENT
Semantic flexibility and grammatical flexibility of verb use were related,
of verbs. The children who
both as properties
of children and as properties
in their first verb uses (indexed by a
showed more
also showed more grammatical
composite of multiple measures)
flexibility
a
the target verbs in this study,
indexed
(also
score). Among
composite
by
those that were used with grammatical flexibility by more children were also
actors by more children. Thus,
these findings
used with multiple
extend
accounts
theoretical
of the relations between
early verb syntax and verb
semantics in a couple of ways. The findings
suggest that it is not just know
a
certain number of verbs (Marchman 8c Bates, 1994) but also having
ing
that underlie
the understandings
semantic flexibility of verb use that sup
ports the acquisition of verb grammar
(Naigles et al., 2005). The findings
also suggest that the relationship
between
semantic flexibility and syntactic
is reciprocal. In particular, the finding that semantic flexibility and
flexibility
of verbs, in addition to being
syntactic flexibility were related as properties
related as properties
of children,
supports the syntactic bootstrapping
hy
pothesis:
elaborate
Verbs used with more

semantic
grammatical
are verbs with more
flexibility
as well.
representations
The
semantic and grammatical
of verb use did, however,
flexibility
differ in their relations to the age at which the child began to use verbs and
to when the verb appeared
in the children's
speech. Children who began
verb
use
at a younger
age
showed,
on
average,
greater
semantic
flexibility
in
their verb use. Children may become early verb learners, then, if they have
the ability to extract verb referents
from their observed
early acquired
contexts.
In
contrast,
these
earlier
verb
learners
did
not
show
greater?or
in their early verb uses, nor did they achieve

lesser?grammatical
flexibility
a greater?or
of productive
lesser?number
frames by the age of 24
use
to
months
with
children
who
verbs later. Earlier verb
began
compared
in their ability to extend their
learners, then, may be specifically precocious
verb meanings in multiple ways, but they are neither advanced nor behind
other children in their ability to use those verbs inmultiple ways in terms of
the picture is a bit different: Those verbs
grammar. Among verbs, however,
that appeared
earlier in the children's speech were used with grammatical
(in their first 10 uses) by fewer children than were later-appearing
flexibility
99
of chil
showed no relation to the proportion
verbs, but age of appearance
dren who used a verb with semantic flexibility. Thus,
the data present a
use
use
to
in
which
children who begin
verbs earlier
their first verbs
picture
with a similar degree of grammatical
flexibility as children who begin later,
but verbs that are used earlier than other verbs are used with grammatical
flexibility by fewer children. Less grammatical
flexibility, then, is a property
of
earlier-learned
verbs
but
not
of
children.
earlier-verb-learning
It
is im
verb (open) was used

portant to point out, though, even the earliest-learned
we
with grammatical
almost
of
the
children.
30%
Overall,
by
flexibility
found
no
children,
evidence
that
there
exist
or
verbs,
early-learned
early-starting
that are consistently inflexible.
THEORETICAL
IMPLICATIONS
OF THE
AND
OBSERVED
PRAGMATIC,
SEMANTIC,
FLEXIBILITY
OF CHILDREN'S
FIRST
GRAAAAAATICAL
VERBUSES
The data from this diary study indicate that 1-year-old children use
in multiple
their newly acquired verbs flexibly,
situations, with multiple
actions, actors, affected objects, and paths or locations. These
findings are at
are ini
in which children's verb meanings
odds with models of acquisition
et
et
The
Golinkoff
conservative
1995;
al.,
al.,
2006).
Maguire
(e.g.,
tially
present data show that 1 -year-old children also use their verbs in different
sentence
frames,
with
different
subjects,
objects,
and
and
prepositions,
all within the first 10 instances of

(somewhat) with different morphology,
use. Such flexible use suggests as well that children younger
than 2 years do
the restricted use that inspired the Verb Island Hy
not, in fact, manifest
pothesis.
These
findings
are
more
consistent
with
the
generativist
view
of
8c Fisher, 2002)
child language acquisition
1981; Gleitman
(e.g., Chomsky,
that children are conservative
than with the positions
language users and
do not have abstract syntax (e.g., Goldberg,
that 2-year-olds
1999; Lieven,
2006; Tomasello,
2000).
verb use was
in the present data, fully flexible and productive
However,
not evident in all children from the moment
verb use began. On average,
the percent of verbs that children used flexibly varied from 16% (mor
to 30% (lexical subjects) to 38% (actions) to 46% (lexical objects) to
phology)
to 66% (syntax) to 73% (actors) to 90% (affected objects).
50% (addressees)
on the
some verbs in the same way (i.e., conservatively)
Every child produced
measures
10
their
first
for
all
of
lexical, and/or grammatical
pragmatic,
instances.
Even
when
verb
uses
were
flexible
across
the
first
10
instances,
they
were rarely flexible by the second instance. Thus, children were swift to show
all kinds of
some kinds of flexibility and productivity but did not demonstrate
some
children
and
Moreover,
ap
instantaneously.
productivity
flexibility
more flexible and productive
than others.
peared to be swifter and/or
100
GENERALDISCUSSION
There
are
two
and
possible,
not
sources
exclusive,
mutually
of
this
in the flexibility with which

children use their first verbs. A
variability
had full grammatical
would
that
the
children
argue
knowledge
generativist
and con
but that a variety of personality,
from the beginning
pragmatic,
of that underlying
textual factors limited children's expression
grammatical
more interested than others in talk
knowledge. That is, some children are
than other children
(Nelson,
ing about new things, and to new people,
some
8c
verbs afford more flexible
1973; Reznick, Corley,
Robinson,
1997);
use than others and sometimes
the occasions for the use of a verb may be
with this position,
limited. Consistent
studies, where such
comprehension
and
semantic-,
pragmatic-,
influences
personality-based
should
not
apply,
of most
that most toddlers demonstrate
significant generalization
et
et al.,
items
Gertner
introduced
lexical
al., 2006; Naigles
(e.g.,
newly
Future
8c
Poulin-Dubois
2005;
Forbes, 2006).
comprehension
investigations
could further solidify these findings by extending
the semantic and gram
matical properties
studied
(e.g., to novel affected objects, and to [in the
relevant languages] PPs and case markers).
indicate
in which
the present data, though,
is a model
not
is
from
Some
the
grammatical
knowledge
fully present
beginning.
are
use
achieved
in
before
verb
any
grammatical
understandings
produc
un
in the flexibility
tion resulting
that is observed,
but some grammatical
after
The
clear
frame
differences
derstandings
develop
begins.
production
in these data suggest that the correct description
and individual differences
as verb learners will contain
of children
elements
of both the rapid
measure
in
and
conservative-child
different
for differ
accounts,
generalizer
ent children, different verbs, and different frames (see Maratsos,
2007, for a
in a different domain). When all the supportive factors are
similar argument
in place (e.g., the verbs have been used flexibly
in the input, the child is
new
situation
the
is
and
feeling talkative,
interesting but not too new and
more
and
the
frame
is
children are more
interesting,
possibly
transparent),
to
be
flexible
and
likely
swiftly
productive?and
possibly even show gen
Also
consistent
to novel
eralization
with
instances,
such
although
latter
are
demonstrations
un
in spontaneous
the supportive
factors
likely to be manifest
speech. When
are not in place, children may be more
to
some
be
and
conservative,
likely
more
some
factors
be
available
for
verbs
than
for
others.
may
supportive
in their speech processing
Differences
abilities may also
among children
in the onset or rate of flexibility, productivity,
lead to variation
and gen
consonant with other recent findings
that early perceptual
fac
eralization,
tors predict later language measures
8c
Perfors,
Marchman,
2006;
(Fernald,
8c Pruitt,
Padden, Nelson,
8c
Dow,
2006).
Jusczyk,
Kuhl, Conboy,
Jusczyk,
Thus,
categories
one
position
are learned
consistent
from
with
our
2005; Newman,
data
is that
input, but the process
Bernstein
abstract
of learning
Ratner,
grammatical
begins
early,
101
at all or before production

of the relevant
lin
speech production
or
see
As
frame
construction.
summarized
Gerken
also
(2007;
guistic
by
have demonstrated
2002; Saffran 8c Thiessen,
2007), researchers
Naigles,
that considerable
about
the
patterns?at
knowledge
varying levels of ab
a given language are accrued during the first 12-18 months
stractness?of
before
of life. If children have indeed been actively processing

their input language
it is
their first verbs, then perhaps
for over a year by the time they produce
not surprising
that they should show the level of productivity
and flexibility
view from a
that our data reveal (see also Snyder, 2007, for a concurring
more generativist
if
awaits
framework).
Similarly,
grammatical development
it is a threshold that
the achievement
of a threshold of lexical development,
children achieve very early.
The present findings have some further implications for accounts of the
The observed
that underlie
process or processes
language development.
common
the grammatical
flexibility and semantic flexibil
and
25
26, chapter VI) suggest there may be some
(Tables
common experiences
that support both
and/or
learning processes
semantic
and grammatical
between
intercorrelations
ity of verb use
ticular,
suggest
that
syntactic
development.
properties
The
of
verbs
correlations
reveal
their
by verb,
in par
semantic
pos
sibilities, and vice versa. For example,

learning quickly that open can be used
affected objects might enable children to quickly understand
with multiple
that it should also appear with grammatical
objects; learning quickly that
run
does
not
involve
affected
objects
effectively
limits
its grammatical
uses
them from those of open. Another way to think about this

and distinguishes
to be learned quickly also
is that whatever
input enables verbs' semantics
is
relevant for meaning
information
enables their grammatical
aspects?the
is consistent with the
also relevant for grammar. The input interpretation
and Hoff-Ginsberg
(1998), who found that verbs used
findings of Naigles
are the ones used more frequently
with more syntactic diversity by mothers
10 weeks later.
and with more syntactic diversity by children
for future re
the
relevant
that
The present
question
findings suggest
even?can
be flexible and
search, then, is not whether children?toddlers,
lexical semantic meanings.
with syntactic frames and extend
productive
can.
in
is
how
The
issue
early
development, and how quickly
really,
Clearly, they
a
new
this
and
what
factors influence
do
this
do
verb,
they
upon meeting
process.
LIMITATIONS
This study is, of course, limited by the fact that we have investigated
of only a small number
the development
(34) of verbs and have tracked
for only their first 10 uses. In particular, our exclusion of
their development
102
GENERALDISCUSSION
about the gram

make and do from the targeted verbs limits our conclusions
as the
of light verbs, as these have been proposed
contributions
transitive
frame
for
the
1999;
Ninio,
1999). How
(Goldberg,
pathbreakers
matical
ever,
our
verbs
were
chosen
because
data
previous
indicated
that
these
34 were
among the earliest produced by 1-year-old English learners (e.g., Naigles 8c

1998); therefore, we do not anticipate that many other verbs
Hoff-Ginsberg,
would elicit earlier use and/or flexibility. Moreover, when the first uses of make
were examined for two of the children, whose mothers recorded them at their
own initiative (Mae and Stacey; see Tables 20 and 23, respectively),
there was
for either one.
that make functioned as a pathbreaker
is also limited by the fact
The generalizability
of the present findings
that we have only included eight children, although an n of 8 is large for the
in this study. Because
individual data analytic approach
the
employed
were
current method
for
different
both
children
who
reasons,
excluded,
talkative and those who talked very little (see chapter II), the
extremely
toddlers at large is probably
range of behavior of American English-learning
wider than seen here. It is likely that both more and less swiftly flexible
children exist. Our conclusions might also be limited by the fact that we
relied on the parents to note each of the first 10 instances of these 34 verbs,
no evidence
we have reason to believe

that these
However,
including
self-repetitions.
we
out
to
mothers
carried
this
task.
talked
each
First,
parent every
eight
just
week, once their child had begun producing
verbs, giving them weekly
reminders of how to report the relevant aspects of their children's verb use
to the diary pages. Mothers who were not able to do this were quite open to
telling us. Only the children whose mothers were able to carry out the task
in
the manner
prescribed
are
in
included
the
study.
Moreover,
if our
doc
umented
child variations reflected only maternal
variation
in keeping
the
or
then
children
should
in
how
slow
differ,
diary,
consistently,
speedy
they
were
to
That
is,
reach
if some
10
instances,
mothers
or
were
in how
flexible
consistently
they
more
were
on
conscientious
all measures.
or
more
attuned to their child's speech, then some children should show precocity
on all measures
whereas
other children
should be consistently
slow. In
all 10 instances of at least
contrast, though, all of our children produced
some verbs in a very short
period of time and all also took a very long time to
reach 10 instances of other verbs (chapter III).
Finally, it is possible that, because they were listening for their children's
use of these verbs, the mothers
might have been more
likely to use those
verbs, themselves. More frequent use by mothers,
though, would only have
led to earlier use by the children, not to more or less flexible or productive
use. That is, there is no reason to expect that more
frequent use by mothers
would have also meant more diverse use by mothers?in
fact, we would
that
the
of
folk
theories
held
argue
language "training"
by most middle
class mothers
would
lead them to less diverse uses (i.e., repetitions
of the
103
same verb phrases). Thus,

it seems quite unlikely
of diary records altered the development
under
that the mothers'
keeping
study.
CONCLUSIONS
The overall picture of the functions, meanings,
and grammar of verbs
one of flexible use. Children
is
when they first appear in children's
speech
use their verbs in utterances
that serve multiple
communicative
functions,
actors and affected objects, and in multiple
with multiple
syntactic frames.
in the literature of early verb use as
This picture contrasts with descriptions
for these differing
restricted and context bound. Part of the explanation
use may be in the unique nature of the database
of
verb
descriptions
early
the completeness
of sam
for the present analyses. Our database provided
our database included several
pling that is characteristic of diary studies, but
individual children can have
children. Thus, our data reveal that whereas
restricted uses of individual verbs, restricted use is not a stage of language
through which all children pass. Like children's overexten
development
sions (e.g., calling strange men Daddy; Rescorla,
1980) and over regular
Rosen, & Xu,
Hollander,
(e.g., goed) (Marcus, Pinker, Ullman,
verb use
restricted
and
of
context-bound
instances
1992),
grammatically
are
more
than
salient
be
may
they
frequent.
izations
The
present
findings
argue
that
children
are
not
as
conservative
lan
guage learners as some theories of language acquisition would have them

or in structure. At least for verb learning, they do not
be, either inmeaning
start small, with verb-specific
but instead begin in the middle,
grammars,
with some flexibility and productivity manifested
by the time they begin to
from the very
use verbs at all. Children quickly match their input, observing
in the target
distinctions
of language use the subcategorization
beginning
uses
of newly acquired verbs children show
their first 10
language. Within
include categories of verbs, which are differentiated
that their grammars
by
occurrence. They do not appear to have a single
of
privileges
grammatical
seem to
category, VERB, in which all verbs are treated equally, nor do they
To
reveal these
have 34 separate and isolated verb-specific
representations.
a
new
we
to
method?a
had
diary that targeted only verbs
develop
findings,
to
record only their child's first 10 instances of use.
and that asked mothers
the diary tractable for
the aspects of language use, we made
By restricting
more
children could be studied. By using the diary
data collectors such that
in which
we
were
to
collect uses in almost all of the contexts
able
method,
we
were
able to
talk. And by asking for 10 successive
children
instances,
demonstrate
that?and
track
how?children's
very
first
uses
changed.
104
REFERENCES
with word
children's
M. (1997). Young
N., & Tomasello,
productivity
Psychology, 33, 952-965.
morphology.
Developmental
M. (2000). The
of words
and word
social nature
Akhtar, N., & Tomasello,
& L. Bloom
K. Hirsh-Pasek,
Golinkoff,
(Eds.), Becoming a word learner: A
Press.
New York: Oxford
University
(pp. 115-135).
acquisition
structure
in
H.
Preferred
S. E. M., & Schroder,
Allen,
(2003).
argument
E.
L.
In
W.
&
data.
DuBois,
W.J.
Ashby
spontaneous
J.
Kumpf,
speech
Akhtar,
argument
structure: Grammar
as architecture for function
(pp. 301-338).
order
and verb
In R.
learning.
debate on lexical
early
Inuktitut
(Eds.), Preferred
Amsterdam:
John
Benjamins.
M. (2006). The distributed
A. L., Lieven,
E. V. M., 8c Tomasello,
B., Theakston,
Ambridge,
of an abstract
construction.
effect for children's
Cognitive
syntactic
acquisition
learning
21, 174-193.
Development,
R. A. (2003). The
of grammar:
Armon-Lotem,
S., 8c Berman,
emergence
Early verbs and
30, 845-877.
of Child Language,
beyond. Journal
for clues to word
reference.
D. A. (1993).
the speaker
Infant's
ability to consult
20, 395-418.
of Child Language,
Journal
and maternal
Barrett, M., Harris, M., 8c Chasin, J. (1991). Early lexical development
speech:
uses of words.
initial and subsequent
A comparison
of children's
Journal
of Child Lan
Baldwin,
18, 21-40.
I., 8c Snyder, L. (1988). From first words togrammar: Individual differences
E., Bretherton,
Press.
and dissociable mechanisms. New York: Cambridge
University
In B.
of grammar
from the lexicon.
the emergence
Bates,
E., 8c Goodman,
J. (1999). On
New Jersey: Erlbaum.
(Ed.), The emergence of language
(pp. 29-80).
MacWhinney
guage,
Bates,
Mouton.
One word at a time: The use of single word utterances. The Hague:
UK: Cambridge
Uni
(1993). Language
development from two to three. Cambridge,
versity Press.
of verbs and the development
of verb
L., Lifter, K., 8c Hafitz,
Bloom,
J. (1980). Semantics
in child language.
inflection
386-412.
56,
Language,
K. (1987). The maturation
of syntax. In T. Roeper
8c E. Williams
Borer, H., 8cWexler,
(Eds.),
Parameter
Reidel.
Dordrecht:
123-187).
setting (pp.
Bloom,
L.
Bloom,
L.
(1973).
M.
Structural
Bowerman,
(1973).
relationships
In T. E. Moore
semantic?
(Ed.), Cognitive
in children's
and
utterances:
Syntactic
or
the acquisition
development
of language
Press.
New York: Academic
197-214).
in the acquisition
of early verbs: Evidence
from diary data
S. (1995). Differences
In M. Tomasello
8cW. Merriman
from sisters.
(Eds.), Beyond names for things: Young
children's acquisition of verbs (pp. 81-111).
Hillsdale,
NJ: Erlbaum.
(pp.
Braunwald,
to lexicon:
grammar
Unsupervised
243-262.
19,
Linguistics,
Computational
M. (2001). The
of dative
A., 8c Tomasello,
acquisition
Campbell,
253-267.
22,
cholinguistics,
Brent,
Caselli,
M.
R.
(1993).
From
learning
of
constructions.
lexical
syntax.
Applied
Psy
M. C, Casadio,
of the transition
from first words
P., 8c Bates, E. (1999). A comparison
to grammar
in English
and Italian. Journal
Child
69-111.
26,
Language,
of
105
a phrasal
A. E. (2005). Fast mapping
between
form and mean
D., & Goldberg,
Science, 8, 500-508.
ing. Developmental
in early
constructions
and semantic
F., & Jisa, H. (2006). Caused motion
Chenu,
generality
In E. Clark & B. Kelly
of French.
in acquisition
(Eds.), Constructions
(pp. 233
acquisition
Casenhiser,
CA: CSLI Publications.

261). Stanford,
M. (2006). Are nouns
easier to acquire
Childers,
J., & Tomasello,
children
Pasek & R. Golinkoff
(Eds.), Action meets word: How
New York: Oxford
Press.
University
Chomsky,
N.
(1975).
N.
(1981).
In K. Hirsch
than verbs?
learn verbs
(pp. 311-335).
on
Books.
Reflections
language. New York:Pantheon
Lectures on government
and binding. Holland:
Foris Publications.
on
A constraint
The
of contrast:
principle
language
acquisition.
(Ed.), Mechanisms
of language
acquisition
(pp. 1-33). Hillsdale,
Chomsky,
E. V. (1987).
Clark,
B. MacWhinney
In
NJ:
Erlbaum.
E. V.
Clark,
The
(1993).
E. V.
Clark,
(2003).
in acquisition.
lexicon
First
language
Cambridge,
acquisition.
UK:
Cambridge,
Press.
University
Cambridge
UK:
University
Cambridge
Press.
Press.
UK: Cambridge
(1996). Using language. Cambridge,
University
A
statistics
H. H. (1973). The
of
fallacy:
language
language-as-fixed-effect
critique
and Verbal Behavior,
research. Journal
12, 335-359.
of Verbal Learning
psychological
the lexicon and grammer:
Cross-sectional
B. T, & Thai, D. J. (2006). Ties between
Conboy,
H.
Clark,
Clark,
toddlers.
studies of bilingual
longitudinal
D. (1999). An introduction
S., 8c Lillo-Martin,
UK: Blackwell
Publishers.
Oxford,
and
Crain,
in
Child development,
77, 712-735.
to
and
theory
language acquisition.
linguistic
K. (1996). Collecting
spontaneous
production
S. Cairns
(Eds.), Methods for assessing children's
Press.
MIT
Demuth,
H.
data.
In D. McDaniel,
syntax
(pp.
3-22).
C. McKee
Cambridge,
8c
MA:
in
V. A. (2007). Grammar
and the lexicon: Developmental
ordering
J., & Marchman,
Child
190-212.
78,
Development,
language
acquisition.
Press.
UK: Cambridge
E. (1987). Early lexical development. Cambridge,
Dromi,
University
K.
M. H., Karmiloff-Smith,
E. A., Johson,
A., Parisi, D., & Plunkett,
Elman,
J. L., Bates,
on
MA:
innateness: A connectionist perspective
(1996). Rethinking
Cambridge,
development.
Dixon,
MIT
Press.
in early
S. (1994). Variability
Fenson,
L., Dale, P., Reznick,
J., Bates, E., Thai, D., & Pethick,
in Child Development,
communicative
of the Society for Research
Monographs
development.
59(5, Serial No. 242).
Fernald,
V. (2006). Picking up speed in understanding:

A., & Marchman,
Speech
across the second year. Developmental
and
Psy
growth
vocabulary
efficiency
98-116.
A., Perfors,
processing
chology, 42,
K. J., Marcus,
Fernandes,
to syntax and back
G. E, Di Nubila,
again: Argument
J. A., & Vouloumanos,

structure
in the third
A.
year
(2006). From semantics

of life. Cognition,
100,
B10-B20.
Fisher,
C.
(1996).
Structural
limits
of sentences.
on
verb mapping:
The
Cognitive Psychology,
interpretations
L. (1991). On
H., 8c Gleitman,
C, Gleitman,
zation frames. Cognitive Psychology,
23, 331-392.
L. R. (2002).
8c Gleitman,
Fisher, C,
Language
acquisition.
R. (Vol. Ed.), Steven's handbook of experimental
8c Gallistel,
Fisher,
role
31, 41-81.
the semantic
in children's
of analogy
content
In H.
of
subcategori
Pashler
(Series Ed.)
Vol. 3. Learning,
8c Sons.
and emotion (3rd ed., pp. 445-496).
Hoboken,
motivation,
NJ: John Wiley
in young
D.
8c
Poulin-Dubois,
N.,
Forbes,
(1997).
change
Representational
J.
verb meaning.
of familiar
children's
24,
of Child Language,
Journal
understanding
psychology:
389-406.
106
REFERENCES
Gentner,
M.
L. (2002).
D., 8c Boroditsky,
Individuation,
Bowerman
& S. Levinson
(Eds.), Language
and early word

In
relativity,
learning.
acquisition and conceptual development
(pp.
Press.
University
UK: Cambridge
215-256).
Cambridge,
L. (2007). Acquiring
structure.
Gerken,
linguistic
handbook of language development
(pp. 173-190).
In E. Hoff
Oxford,
words
Learning
Y, Fisher, C, &: Eisengart,

Gertner,
J. (2006).
in early sentence
of word order
comprehension.
H., Gleitman,
L., 8c Lederer,
Gillette,
J., Gleitman,
73, 135-176.
ulary learning.
Cognition,
A.
8c M.
UK:
Shatz
Publishing.
and rules: Abstract
knowledge
Science, 17, 684-691.

Psychological
of vocab
simulations
(1999). Human
sources of verb meanings.

L. (1990). The
structural
Language
E. L. (1995). The
L. R., 8c Newport,
of language
invention
Gleitman,
on the
ronmental
and biological
influences
of language.
acquisition
to
M. Liberman
invitation
(Eds.),^4n
cognitive science: Vol. 1. Language
MA: MIT
Press.
Cambridge,
Gleitman,
A. (1999). The
Goldberg,
In B. MacWhinney
Blackwell
(Eds.),
Blackwell
of the semantics
of argument
emergence
(Ed.), The emergence
of language
(pp.
1, 3-55.
Acquisition,
Envi
by children:
In L. R. Gleitman
(2nd ed., pp.
structure
197-211).
8c
1-24).
constructions.
Mahwah,
NJ:
Erlbaum.
D. (2006). Learning
A., 8c Casenhiser,
argument
Goldberg,
Clark & B. Kelly
in acquisition
(pp.
Publications.
structure
constructions.
185-204).
Stanford,
In E.
CA: CSLI
D. M., 8c Sethuraman,
A. E., Casenhiser,
N. (2004). Learning
structure
argument
15, 289-316.
generalizations.
Cognitive Linguistics,
S. (2003). The resilience of language: What gesture creation in deaf children can tell
Goldin-Meadow,
us about how all children learn
Press.
language. New York: Psychology
Goldberg,
M. E., 8c Gelman,
R. (1976). Language
in the two-year-old.
S., Seligman,
4, 189-202.
Cognition,
L. (1987). The eyes have it: Lexical
and
R., Hirsh-Pasek,
Golinkoff,
K., Cauley, K., 8c Gordon,
in a new paradigm.
14, 23-45.
syntactic
comprehension
Journal
of Child Language,
R. M., Hirsh-Pasek,
C. B., Frawley, W,
8c Parillo, M. (1995). Lexical
Golinkoff,
K., Mervis,
can be extended
to
In
of
verbs.
the
M.
Tomasello
& W. Merriman
principles
acquisition
Goldin-Meadow,
(Eds.), Beyond names for things: Young children's acquisition of verbs (pp. 185-222).
Erlbaum Associates.
NJ: Lawrence
R. L. (2002). Variability
structure.
and detection
of invariant
Gomez,
Psychological
431-436.
Hillsdale,
Science,
13,
context
role of semantic
and
L., 8c Brown, N. B. (1998). The
J. C, McDonough,
in the acquisition
of novel nouns.
Child Development,
69, 1330-1344.
memory
8c Sethuraman,
N. (2006).
Interactions
in the development
of constructions
Goodman,
J. C,
In E. Clark & B. Kelly
and the acquisition
of word meanings.
in
Goodman,
Grice,
Stanford:
acquisition
(pp. 263-281).
H. P. (Ed.) (1989). Further
notes
CSLI Publications.
on
Studies in the way of words
logic and conversation.
MA:
Harvard
Press.
from Syntax and
41-57).
(pp.
Cambridge,
University
(Reprinted
semantics: Vol. 9. Pragmatics,
1978, New York: Academic
pp. 113-128,
by P. Cole, Ed.,
Press).
S. (1988). Linguistic
M., Barrett,
M., Jones,
D., & Brooks,
input and
15, 77-94.
meaning.
Journal
of Child Language,
R. (Eds.),
K. & Golinkoff,
children
Hirsh-Pasek,
(2006). Action meets word: How
Oxford
UK: Oxford
Press.
University
Harris,
early
R. M., Hennon,
E. A., & Maguire,
K., Golinkoff,
Hirsh-Pasek,
(2004). Hybrid
M.J.
at the frontier
of developmental
The emergentist
coalition model
psychology:
as a case in point.
a lexicon
In G. Hall & S. Waxman
(Eds.), Weaving
learning
204).
Cambridge,
MA: MIT
word
learn verbs.
theories
of word
(pp.
173
Press.
107
use of syntactic frames

L. (1996). Young
children's
Hirsh-Pasek,
K., Golinkoff,
R., 8c Naigles,
to derive meaning.
In K. Hirsh-Pasek
8c R. Golinkoff
(Eds.), The origins of grammar:
Evidence
MA: MIT
123-158).
from
early language
comprehension
(pp.
Cambridge,
Press.
G. J., Hirsh-Pasek,
R. M. (2000). Breaking
the language
barrier: An
K., & Golinkoff,
for the origins of word
coalition model
emergentist
learning. Monographs
of the Society for
Research
65(3, Serial No.262).
P. J., & Thompson,
in grammar
S. A. (1980). Transitivity
and discourse.
Hopper,
Language,
56,251-299.
R. (1983). The emergence
of action categories
in the
Huttenlocher,
J., Smiley, P., 8c Charney,
Hollich,
from verb meaning.

child: Evidence
Review, 90, 72-93.
Psychological
of grammatical
In R. Stark (Ed.) Language
Ingram, D. (1981). Early patterns
development.
New York: Elsevier North-Holland.
behavior in infancy and early childhood (pp. 327-352).
Press.
UK: Cambridge
Ingram, D. (1989). First language acquisition. Cambridge,
University
R. (1983). Semantics and cognition. Cambridge,
MA: MIT Press.
Jackendoff,
Jackendoff,
R.
(1990).
Semantic
Jackendoff,
Oxford
R.
(2004).
Foundations
Cambridge,
Brain,
of language:
MA: MIT
Press.
evolution.
grammar,
meaning,
New
York:
Press.
University
D., Thai,
Jackson-Maldonado,
lexical
Early
structures.
V. (1993).
D., Marchman,
V, Bates, E., 8c Guitierrez-Clellen,
in Spanish-speaking
infants and toddlers. Journal
of Child
development
20, 523-549.
Language,
M.
the role
E., Lieven,
E., 8c Tomasello,
Kidd,
(2006). Examining
children's
and processing
of sentential
complements.
acquisition
of
lexical
frequency
Cognitive
in
Development,
21, 93-107.
and
P., Conboy,
D., Nelson,
T, & Pruitt, J. (2005). Early speech perception
B., Padden,
for
the
"critical
later language
Language
Learning
Implications
period".
development:
and Development,
1, 237-264.
and experience: Evidence from the blind child.
L. R. (1985). Language
B., 8c Gleitman,
Landau,
MA: Harvard
Press.
University
Cambridge,
Kuhl,
The
B. (1993). English verb classes and alternations: A preliminary
investigation. Chicago:
of Chicago
Press.
University
utterances.
In E. Clark 8c B. Kelly
E. (2006). Producing
multiword
Lieven,
(Eds.), Construc
CA: CSLI Publications.
tions in acquisition
Stanford,
(pp. 83-110).
A
M. (2003). Early syntactic
H., Speares,
E., Behrens,
Lieven,
creativity:
J., 8c Tomasello,
Child
333-367.
30,
Journal
of
Language,
usage-based
approach.
Levin,
B.
MacWhinney,
(2004).
A multiple
solution
process
883-914.
Child
31,
Language,
of
Journal
acquisition.
M. J., Hirsh-Pasek,
Maguire,
verb
learning:
Putting
Action meets word: How
to the
logical
problem
of
language
R. M.
of word
& Golinkoff,
(2006). A unified
theory
R. Golinkoff
context.
In K. Hirsh-Pasek
in
8c
(Eds.),
acquisition
New York: Oxford
children learn verbs (pp. 364-391).
University
K.,
Press.
S., Dromi,
Maital,
development
tions. Journal
Maratsos,
M.
(2007).
E.,
Sagi, A.,
8c Bornstein,
M.
inventory:
Language
specific
27, 43-67.
of Child Language,
Commentary.
Monographs
H.
(2000).
properties
The
and
Hebrew
cross-linguistic
communicative
generaliza
72,121-126.
A
in lexical and morphological
V, 8c Bates, E. (1994). Continuity
Marchman,
development:
test of the critical mass hypothesis.
21, 339-366.
of Child Language,
Journal
nature
V A., Martinez-Sussman,
C, & Dale, P. S. (2004). The
Marchman,
language-specific
learners. Developmental
from bilingual
Evidence
of grammatical
language
development:
Science,
7, 212.
108
REFERENCES
F. (1992).
T. J., 8c Xu,
G. E, Pinker,
M., Rosen,
M., Hollander,
S., Ullman,
in Child
in language
of the society for Research
acquisition. Monographs
Overregularization
57(4, Serial No. 228).
Development,
in children: Problems of induction. Cambridge,
E. (1989). Categorization
and naming
Markman,
MA: MIT Press.
Marcus,
M. (2005). The
in
role of frequency
A., 8c Tomasello,
order. Cognitive Development,
20, 121-136.
M. (2006). The
effect of perceptual
A., 8c Tomasello,
use of
on young
children's
Ap
referring
expressions.
E., Theakston,
Matthews,
D., Lieven,
of English word
the acquisition
E., Theakston,
Matthews,
D., Lieven,
and prior discourse
availability
27, 403-422.
plied Psycholinguistics,
E. V. M. (2006).
of chil
the abstractness
K., Pine, J. M., 8c Lieven,
McClure,
Investigating
structure. Journal
of argument
dren's early knowledge
33, 693-720.
of Child Language,
Press.
to talk. Cambridge,
UK: Cambridge
McShane,
University
J. (1980). Learning
In U.
and early lexical development.
object
categories
in
and intellectual factors
Ecological
conceptual development:
Press.
New York: Cambridge
University
(pp. 201-233).
categorization
as a cue for
in child directed
T
frames
Mintz,
(2003).
categories
grammatical
Frequent
90, 91-117.
speech. Cognition,
of
theorems:
P. C. M. (2008). On the implications
of the classical ergodic
Molenaar,
Analysis
C. B. (1987). Child-basic
Mervis,
Neisser
(Ed.), Concepts and
variation.
has to focus on intra-individual
processes
Developmental
developmental
60-69.
50,
chobiology,
use syntax to learn verb meanings.Journal
L. (1990). Children
of Child Language,
Naigles,
357-374.
L.
Naigles,
verb learners,
too? In E. Clark
one-year-olds
English-speaking
research
the
annual
child
28th
Stanford,
199-212).
forum
(pp.
language
of
and Information.
for the Study of Language
(1997).
Proceedings
The Center
Are
in the use
L. (1998). Developmental
Naigles,
changes
L. Lipsitt, & H. Hayne
Rovee-Collier,
London:
Ablex.
298-318).
L., Bavin, E., 8c Smith, M. (2005). Toddlers

sentences.
Science, 8, 424-431.
Developmental
L. R. (1993).
L., Gleitman,
H., 8c Gleitman,
Naigles,
In E. Dromi
from syntactic
evidence.
ponents
Naigles,
(pp.
opmental perspective
L., 8c Hoff-Ginsberg,
Naigles,
of syntactic bootstrapping.
L., & Hoff-Ginsberg,
Naigles,
Effects
of
Language,
L. R.
Naigles,
language.
L. R.
Naigles,
input frequency
25, 95-120.
(2002).
Form
86,
Cognition,
(2003). Paradox
of
structure
(Eds.), Advances
17,
(Ed.),
CA:
In C.
learning.
research (Vol. 12, pp.
in verb
in infancy
verbs
recognize
Psy
in novel
situations
and
com
acquire word meaning
and cognition: A devel
(Ed.), Language
Norwood,
104-140).
NJ: Ablex.
E. (1995).
for the plausibility
Input to verb learning: Evidence
Developmental
are some verbs
E. (1998). Why
learned before
other verbs?
and
structure
on
is easy, meaning
157-199.
Children
children's
is hard:
early
Resolving
verb
a
use. Journal
of Child
in early
paradox
child
to Tomasello
lost? No, paradox
found! Reply
8cAkhtar
(2003).
88, 325-329.
Cognition,
L. R., 8c Hoff,
E. (2006). Verbs at the very beginning:
between
Parallels
compre
Naigles,
In K. Hirsh-Pasek
and input.
& R. Golinkoff
hension
children
L. R.,
Naigles,
M.
Shatz
learn verbs
New York: Oxford

Press.
(pp. 336-363).
University
L. D. (2007). Syntactic
8c Swensen,
for word
supports
learning.
New
(Eds.), The handbook of language development
(pp. 212-231).
In E. Hoff
York:
&
Black
well.
K. (1973). Structure
and
Nelson,
Research
strategy
38(1-2,
to talk. Monographs
in learning
No.
149).
of the Society for
Serial
109
Newman,
R.,
Infants'
later
Bernstein
early
to
ability
language
N.,
Ratner,
Jusczyk,
A. M., Jusczyk,
conversational
P. W.,
the
segment
A retrospective
development:
analysis.
8c Dow,
speech
K. A.
(2006).
signal
Developmental
predicts
42,
Psychology,
643-655.
A.
in syntactic development
and the question
Child
26, 619-653.
of
Language,
typical transitivity. Journal
in syntactic development.
the role of semantic
Ninio, A. (2005a). Testing
similarity
Child Language,
32, 35-61.
Ninio,
(1999).
verbs
Pathbreaking
of proto
Journal
A. (2005b). Accelerated
semantic
Indirect
similarity:
learning without
objects.
nitive Linguistics,
16, 531-556.
A. (2007). Language
and the learning curve: A new theory of syntactic development.
Ninio,
York: Oxford
Press.
University
Ninio,
of
Cog
New
from the Japanese

Yamashita,
T, 8c Dale, P. S. (1993). Some findings
Y, Murase,
the
communicative
inventories.
Memoirs
early
of
development
Faculty
of Education,
Shimane University,
27, 26-38.
(In English).
M. (1993). Twenty-five
month
do not have a gram
old children
R., 8c Tomasello,
Olguin,
T,
Ogura,
of verb. Cognitive Development,

8, 245-272.
category
different
C. (1998). Comparing
of the development
models
E., 8c Rowland,
J., Lieven,
of the English
verb category.
36, 807-830.
Linguistics,
MA: Harvard
S. (1984). Language
Pinker,
Cambridge,
learnability and language development.
Press.
University
matical
Pine,
S. (1989). Learnability
Pinker,
MA: MIT Press.
Pinker, S. (1994).
410.
How
could
and
The acquisition
cognition:
a child use verb
syntax
of argument
to learn verb
structure.
semantics?
Cambridge,
Lingua,
92,377
to system building: Acquiring

V (1993). From rote learning
verb
K., 8cMarchman,
nets. Cognition,
in children
and connectionist
48, 21-69.
morphology
to intentions-in-action
in learning
attention
D., 8c Forbes, J. N.
Poulin,
(2002). Toddlers'
novel action words. Developmental
Plunkett,
model
and action: A two-tiered
D., 8c Forbes, J. N. (2006). Word,
intentions,
In K. Hirsh-Pasek
8c R. Golinkoff
word
learning.
Press.
New York: Oxford
children learn verbs (pp. 262-285).
University
L. A.
in early
Rescorla,
(1980). Overextension
development.
Journal
of Child
language
Poulin-Dubois,
of action
7, 321-335.
Language,
in
twin study of intelligence
Reznick,
J. (1997). A longitudinal
J. S., Corley, R., & Robinson,
the second
62(1, Serial
year. Monographs
No. 249).
in the verb lexicons of children with specific
L., 8c Bode, J. (1993). GAPS
language
First Language,
13, 113-131.
impairment.
Mod
C. B. (1998). Disentangling
B. E, 8c Mervis,
Robinson,
early language
development:
an extension
of case-study
using
methodology.
acquisition
eling lexical and grammatical
Rice,
M.

Developmental
P. R., Snow, C, 8cWillett,
Rollins,
J. (1996). Predictors
First Language,
16, 243-259.
developments.
E. D. (2007). Domain-general
Saffran, J. R., & Thiessen,
Shatz
(Eds.),
Blackwell
handbook
of
language
of MLU:
learning
development
Semantic
and morphological
capacities.
68-86).
(pp.
In E. Hoff
& M.
Oxford,
UK:
Blackwell
Publishing.
C. M., 8c Smith,
Sandhofer,
different
frequencies,
Shirai,
(1998).
disposition?
L. B.
kinds
The
emergence
First Language,
nouns and verbs in the input: Differential

(2000). Counting
of learning? Journal
of child language, 27, 561-585.
in Japanese:
Universal
of tense-aspect
pre
morphology
18, 281-309.
no
REFERENCES
Smiley, P., 8c Huttenlocher,

and
events, objects,
J. (1995).
persons.
and
Conceptual
development
In M. Tomasello
8cW. Merriman
for
early words
(Eds.), Beyond names for
Erlbaum.
NJ: Lawrence
the child's
things: Young children's acquisition of verbs (pp. 21-61). Hillsdale,

8c Sachs, J. (1990). Cognition
and the verb lexicon
in early
C,
11, 409-424.
Applied Psycholiguistics,
lexical
Smith,
L.
Smith,
Children's
(1999).
ized
noun
learning: How
In B. MacWhinney
mechanisms.
learning
80). Hillsdale,
Smith, L. (2000).
development.
make
processes
learning
(Ed.), The emergence of language
general
special
(pp. 29
NJ: Erlbaum.
crane.
how to learn words: An associative
In R. Golinkoff,
Learning
8c L. Bloom
(Eds.), Becoming a word learner: A debate on lexical acquisition
York: Oxford
Press.
University
Hirsh-Pasek,
New
51-80).
K.
(pp.
L. R. (2004). Why
it is hard to label our concepts.
In G. Hall 8c S.
Snedeker,
J., 8c Gleitman,
a lexicon
Waxman
MA: MIT
Press.
(Eds.), Weaving
(pp. 257-293).
Cambridge,
Press.
(2007). Child language: The parametric approach. Oxford, UK: Oxford University
Snyder, W
K.
In D. McDaniel,
children's
Stromswold,
(1996).
spontaneous
Analyzing
speech.
8c H. S. Cairns
(Eds.), Methods
MA: MIT Press.
Cambridge,
Theakston,
A., Lieven,
E., Pine, J., 8c Rowland,
itations in the acquisition
of verb argument
Child Language,
28, 127-152.
C. McKee,
A.
Theakston,
generality,
61-99.
L.,
children's
assessing
syntax
(pp.
23-54).
C. (2001). The
role of performance
lim
structure: An alternative
account.
Journal
of
E. V. M.,
C. F. (2004).
Semantic
Pine, J. M., & Rowland,
and the acquisition
of syntax. Journal
31,
of Child Language,
Lieven,
input
for
frequency
A. L., Lieven,
M. (2003). The
Theakston,
E., 8cTomasello,
of third person
in English.
verbs
singular
Journal
Research, 46, 863-877.
role of the input

of Speech,
Language,
S. A., 8c Hopper,
Thompson,
P.J.
Evidence
from conversation.
structure,
and argument
clause
(2001). Transitivity,
In J. Bybee 8c R Hopper
in the acquisition
and Hearing
structure:
(Eds.), Frequency and the emergence of

the Netherlands:
linguistic structure (pp. 27-60). Amsterdam,
John Benjamins.
M. (1992). First verbs: A case study of early grammatical
UK:
Tomasello,
development. Cambridge,
Press.
Cambridge
University
M.
children
have adult syntactic
Tomasello,
(2000). Do young
74,
competence?
Cognition,
209-253.
Tomasello,
M.,
8cAkhtar,
N.
(2003). What
paradox?
to Naigles
A response
(2002).
Cognition,
88,317-323.
on actions:
verbs in ostensive
Tomasello,
M., & Kruger, A. (1992). Joint attention
Acquiring
and non-ostensive
contexts.
Child
19, 311-333.
of
Journal
Language,
W. (1995). Beyond names for
Tomasello,
M., & Merriman,
things: Young children's acquisition of
verbs. Hillsdale,
NJ: Erlbaum.
M., 8c Stahl, D. (2004). Sampling
31,
enough? Journal
of Child Language,
Valian, V. (1990). Null
subjects: A problem
35, 105-122.
quisition.
Cognition,
Tomasello,
Valian,
V, Prasada,
S., 8c Scarpa, J.
children's
imitation
of sentences.
M., Waterfall,
H.,
Vasilyeva,
across
and differences
Z. (1972).
Vendler,
Press.
Wagner,
L.
Language,
(2002).
Res
spontaneous
How
speech:
much
is
for
parameter-setting
models
of
ac
language
on young
Effects
object predictability:
Child
247-269.
33,
Journal
of
Language,
8c Huttenlocher,
of syntax: Commonalities
J. (2008). Emergence
children.
Science, 11, 84-97.
Developmental
cogitans: An
Aspectual
28, 661-682.
children's
101-121.
(2006).
Direct
essay in rational psychology.
influences
on
early
tense
Ithaca,
NY: Cornell
comprehension.
Journal
University
of Child
Ill
ACKNOWLEDGMENTS
in part by the Large Research Grant Pro

This research was supported
of Connecticut.
gram of the University
We thank Cynthia Fisher, James Magnuson,
and our reviewers for their
constructive
comments
on
previous
versions
of
the
manuscript.
We
are
to Danielle
also
Popp, Roberta Golinkoff, Aylin Kuntay, Marilyn Shatz,

grateful
Lars Bergman,
Yuriko Oshima-Takane,
and Whitney Tabor for helpful dis
the work. This research would not have been possible
cussions regarding
without
the dedication
of the parents who kept these diaries of their chil
Eliane Ramos,
dren's first verbs. Special thanks go to Jennifer Coughlan,
for their assistance during data collection and manu
and Ashley Maltempo
script
preparation.
112
COMMENTARY
EARLY
FLEXIBILITY
AND SYNTAX
OF CHILDREN'S
VERBUSE
INTHESEMANTICS
Tomasello and Silke Brandf
Michael
are where the

Objects are boring; they just sit there. Events and actions
that probably 90% of all of
That is why it is so puzzling
action is?literally.
the work
done
on
children's
early
word
on children's
the current monograph
on
focuses
learning
acquisition
and
nouns,
why
is so welcome.
of verbs
OF VERBS
INPRAISE
Linguistic
almost
communication
some
event,
or
action,
concerns
always
or informatives
states of affairs. Declaratives

state
of
affairs,
and
events,
or
imperatives
or
actions,
invite the listener
to attend
directives
to
enjoin
to bring about a desired action or state of

the listener to do something
affairs. Thus, even when young children are using object labels as single
word
whole
from
utterances,
there
is almost
the
always
point
some
of view
of
underlying
the
communicative
event
or
action
as a
intention
at
issue.
When
to attend to it or
the infant exclaims "Airplane!" she is exhorting her mother
to notice its presence,
and when the infant requests
'Juice!" she is rousing
into action to satisfy her desire. One could argue that the ap
her mother
is something
along the lines of "Look at
propriate gloss of such utterances
the airplane!"
(or "The airplane is therel") and "Get me some juice!" The
action or state of affairs intended, and its corresponding
verb, is implicit; the
utterance
is what has been called a holophrase.
The one potential
is naming objects. But naming objects is
exception
a
kind
of
actually
metalinguistic
speech act. It is not using language but
rather
"mentioning"
it, mostly
teaching
it.Western,
middle-class
parents
do
this with some regularity with their children, and their children
learn the
names and then show off by using them in return. But in many other
or demonstrative
cultures the pedagogical
naming of objects is a very rare
act
of
in the acquisition of language.
and
little
role
type
speech
plays very
113
AND SILKE
MICHAELTOMASELLO
BRANDT
And even in Western,

culture parents only rarely explicitly
middle-class
teach words other than object labels. With verbs and other types of words,
children must in almost all cases learn them within the ongoing flow of social
and
interaction
discourse?on
their
so
own,
to
on
speak?based
un
their
of the intentional actions of others in the social interaction. For

derstanding
children are able to infer the referent action of a novel verb
young
example,
even when they hear it only as a directive
to act or as an anticipation
of an
even
never
see
at
if
action
the
referent
action
they
impending
actually
on the adult's accompanying
all?based
behavior and context
(Tomasello,
in the acquisition
is so interesting.
1992, 1995, 2001). This social-cognitive

complexity
cess is another reason why the acquisition of verbs
a
From
are
as
referents,
are
of
point
more
nouns
sponding
Even
semantic
much
also
complex
children's
or
causative
not
(e.g.,
their
are
nouns
vs.
for
for other
concrete
corre
kinds
in whether
have),
they
of
objects).
learn and use vary in whether
that children
give
used
mostly
verbs
corresponding
and
their
objects
than
for the moment
nouns
early
the very earliest verbs
and
conceptually
aside
(setting
events
view,
pro
they
results
designate
(e.g., clean the table vs. wipe the table), in whether

they designate
of motion or not (e.g., roll vs. move), in whether
they are defined by a
they are about spe
particular objects or not (e.g., hammer vs. hit), in whether
or more abstract changes of state (e.g., lick, involving
cific bodily movements
or not
manner
movements
specific
of
on what
depending
different
vs.
events
tongue,
from
the
vs.
events
completed
cut,
to a
that may
of actions
aspects
vs.
lawn
the
is cut,
different
very
to mention
involving
finger)?not
be designated
grammatically
vs.
events
impending
events
past
actions
all of
the
(ongoing
vs.
future
events, etc.). Indeed, Gentner

(1982) famously stressed that inmany cases it is
to
to
know how
difficult
"package" the ongoing flow of expe
conceptually
into
rience
discrete
Further
one or more
as
events,
to
referred
by
typically designated
participants,
event
not in any way "given" by the phenomenal
can be used so as to highlight different participants
use,
e.g.,
John
broke
the vase,
at all.
verbs,
is the fact that events
in this direction
two
with
participants
and verbs always involve

linguistically by nouns and
itself (indeed, some verbs
on different occasions of
vs. The
indicated,
vase
broke,
with only one indicated). And so, in a sense, events incorporate objects but
not the reverse. And this is the final way in which verbs are especially
in the study of child language acquisition. As
important and interesting
from all theoretical backgrounds
stressed by researchers
(e.g., Pinker, 1989;
the participants
Tomasello,
1992), verbs have as part of their very meaning
if there is
famous
Blake's
involved
observation,
(paraphrasing William
a
must
of verbs
the
be
there
dancer). This makes
acquisition
dancing
a
already
verbs
that
step
involves
form
the
on
the
way
the kind
backbone
to
grammatical
of verb-argument
of mature
sentence
competence,
as
the
(event-participant)
learning
of
structures
structure.
114
COMMENTARY
is so
It is thus for all of these reasons that the current monograph
in the field. Verbs are arguably the most complex and important
and they have been sorely
type of word in early language development,
in
8c
8c
the
Hirsh-Pasek
Golinkoff,
2006; Tomasello
(see
papers
neglected
some
notable
for
Merriman,
1995,
exceptions).
welcome
METHOD
INPRAISE
OF THEDIARY
in other areas of developmental
Researchers
sometimes
psychology
over
use
of
of
the
in
the
diaries
parent
puzzle
study
But there are very good reasons for the use of this method,
and the current
monograph
child pairs.
involves
the innovative
extension
of using multiple
parent
is that children talk too much. A child is awake and

basic problem
around
but the normal sampling regime in
talking
lOhr/day, 7 days/week,
a
of one of those 70 hr/week.
research
is
maximum
This is a sample of 1-2% of what the child says. The reason researchers do
not sample more
is because
it takes from 10 to 20 hr to transcribe
1hr of
sam
conversation.
this
this
1-2%
But, despite
taped
good practical reason,
The
ple is inherently limited, potentially biased in all the ways that small samples
may be biased in all scientific inquiry, and it very likely misses completely
most
And so, for example,
it has been quan
low-frequency
phenomena.
a
demonstrated
that
with
small
the age at which
titatively
sample estimating
or
the child acquires low-frequency
structure
items
is highly un
linguistic
reliable (Rowland 8c Fletcher, 2006; Tomasello
8c Stahl, 2004).
Diaries
have their own limitations
of course. Most obviously,
they
one
involve
external validity?and
par
prototypically
only
child?limiting
ents are not trained scientific observers. But with the collection of
multiple
diaries
study?the
of
comparable
first
problem
children?again,
is to some
an
degree
innovation
overcome.
of
And,
the
in
current
terms
of
the limitations of parents as observers,

it has been established by researchers
the
MacArthur
Communicative
(a paper
using
Development
Inventory
assessment
of children's vocabulary
filled out by parents)
that
and-pencil
are
at
overt
reliable
their
child's
behaviors
parents
actually highly
recording
true if parents are directed what to
(Fenson et al., 1994). This is especially
look for. And the current study did an especially good job of this
by focusing
parents' attention on a smallish number of specific verbs that their children
might be using.
to diaries
that is often not
Finally, there is one further advantage
stressed. Because
the parent
is recording
(of some
basically everything
one
can
focal
conclude
with
reasonable
designated
phenomena),
certainty
115
MICHAELTOMASELLO
AND SILKE
BRANDT
the child does not do something, which can never be concluded with
small- or even medium-sized
samples. In the study of language acquisition,
to which
this is often important since one of the key issues is the degree
children use their early language flexibly and productively,
and to deter
mine
that we must know not only what they have done but also what they
is con
might have done but did not. And indeed, the current monograph
and productivity,
and so the
cerned precisely with the issues of flexibility
when
diary method
is especially
appropriate.
FLEXIBILITY
INEARLY
VERBUSE
data of the current monograph
provide us with by far the most
in
of how English-speaking
the
literature
picture
complete,
quantitative
the limitations of the diary
children learn and use their early verbs. Within
focused parents on a subset
method
(including the fact that the researchers
we
now
at our disposal a vast and
have
of 34 preselected
target verbs),1
The
array
important
of
new
facts.
Flexibility in Verb Use: Pragmatics and Semantics

Parents
of
each
kept the diary until they had recorded
targeted
verb,
that
is,
10 utterances
containing
10 actual
the
instances of use
verb
in question.
Across all of the verbs used by all of the children, the average time it took for
the 10 instances to be produced
Prag
by the child was a bit over 1month.
as commands or
either
of
their
verbs
used
about
half
the
children
matically,
as descriptions
only; the other half were used in both functions. Foreshad
an
issue that will come up later with regard to syntax, we do not know
owing
in which of these instances children heard their parents using the verb in
and so we do not know if children are simply
both pragmatic
functions,
use
or
the
adult
doing something more creative in using these
following
there are no ex
with
verbs. And because
functions
particular
pragmatic
we
in
this
the literature addressing
simply do not know.
question,
periments
current
found that
the
From the point of view of semantics,
monograph
the children used their newly acquired verbs relatively flexibly from rel
was not a foregone conclusion. Although
research has
atively early. This fact
shown that children use their object labels in fairly flexible ways from early
on (Harris, Barrett, Jones, & Brookes,
1988), a number of different nat
of young children using
have provided
uralistic observations
examples
more
context
in
bound
verbs and other types of words
ways for some time
never
But
been clear whether
it
has
1988; Tomasello,
1992).
(e.g., Gopnik,
or
are
reflect some
whether
those
they might
examples
unrepresentative
more systematic difference with object labels.
116
COMMENTARY
study, children did not confine their use of verbs to a

defined
action type; they used almost 40% of their verbs to
single, narrowly
refer to different
action types (e.g., a person coming vs. a TV show com
if "different action
this number would undoubtedly
be higher
ing)?and
In the current
had
type"
been
more
defined
new
These
loosely.
uses
on
occurred,
average,
after about four uses of the verb in question. And children quite often used
their verbs for different actors and patients of these actions as well (whether
these were
for something
around
three
lexically realized or not). Thus,
to
events
actors
of
their
children
referred
different
verbs,
quarters
involving
on
different
events
actors
question.
or
Although
other diary
to novel
patients
are
about
different
involving
and/or
usage
and
occasions,
being
studies
referents
we
same
the
patients
occurred
after
cannot
be sure
more
creative
in
around
In both
Bowerman,
uses
three
if the children
their
semantic
in the past have noted many

(e.g.,
were
proportion
of the actions.
to refer
used
to
cases,
the new
of
verb
the
are following
in all
extensions
cases of creative
in
adult
cases,
extensions
1982).
use much
of their early language in recurrent social

seem fairly high, and
situations and routines, these percentages
they would
seem to refute any proposal
that children are inflexible in their early use of
verbs to refer to events and actions in the world. One might ask how this
flexibility compares with children's early use of object labels, but the answer
is that we do not know because we do not have quite such detailed data in
this case. There are some diary studies (especially Dromi,
1987) that have
focused on the acquisition of object labels in enough detail to provide some
relevant data for comparison,
but since the two types of word are so differ
ent semantically,
it is difficult to think how such a comparison
could be done
Given
that children
quantitatively.
In
any
case,
the
general
conclusion
would
seem
to be
that
children use both their early nouns and their early verbs
English-speaking
in reasonably flexible ways to refer to all kinds of objects and events in the
world.
This having been said, it is unclear that?despite

the literature review
in the current monograph?anyone
any
espousing
theory of word learning
would predict that children in the second half of their 2nd year of life would
still be using any kinds of words in highly context bound ways. Virtually all
of the proposals for early context bound use have focused on the first half of
the 2nd year of life or even earlier.
Flexibility in Verb Use: Syntax

The most
contentious
issue in the monograph
is children's
flexibility
and productivity
in using their early verbs syntactically. We deal first with
the issue of flexibility, which,
in the current context, means children using
their verbs in syntactically diverse contexts.
In the following section, we will
117
MICHAELTOMASELLO
AND SILKEBRANDT
deal with the issue of productivity,

in the current context, means
which,
children using their verbs in syntactic contexts
that go beyond
those in
which they have heard them being used. Productivity
is an especially con
tentious issue as it involves the degree to which young children's language is
underlain
by abstract, lexically general,
linguistic categories and construc
tions (rules). But the two issues are tied together; hence, let us first sort them
a bit.
out
Several authors,
have proposed
that children do
including ourselves,
not have innate and abstract grammatical
that
categories
apply equally
across all words as part of their beginning
basic
linguistic competence?the
Chomskian
proposal.
revolves
around
petence
Instead,
specific
of
much
or
verbs
their
other
early
items
grammatical
as
such
com
pronouns
& Baldwin,
1997; Tomasello,
1992, 2000, 2003). Thus,
with
constructions?some
of which have
begin
lexically specific
been
called verb
island constructions
because
revolve
around
they
verbs?with
In
abstractions.
this
local,
only
lexically
specific
theory,
if the young child says "Doggie kiss me," the item-based construction
in
to such
volved might be something
like kisser KISS kissee, without reference
less subjects and objects.
things as agents and patients, much
verb-general
With this item-based construction,
the child could learn a new object label
use it as either "kisser" or "kissee"
and immediately
without
productively,
on
an
others
do
based
of
all
the
items
involved.
this,
hearing
understanding
Pine,
(Lieven,
children
But
what
she
could
not
do
is to
learn
a new
verb,
say
hug,
and
immediately
an analogy to kiss and use it with a

"hugger" and a "hugee" without
is a structural
having heard others do it. Each lexically specific construction
make
island.
comes from both naturalistic

The evidence
for this proposal
observa
tions and experiments.
Many studies of natural samples of children's early
language show restricted use of words and lexically specific schemas (e.g.,
Lieven, Behrens,
2003; Lieven et al., 1997), and this
Speares, & Tomasello,
restricted use has been taken to reflect a relative lack of lexically general
that can be used totally productively.
But most
categories and constructions
to
current
the
data isTomasello's
(1992) original diary study. In
comparable
that study there were two basic findings relevant to issues of verb flexibility
and productivity.
The first was that the majority of this diary child's early verbs were used
in fairly simple sentence frames, often only one frame per verb, and differ
ent verbs most often had different frames?suggesting
that they were each
an
and
island
graph
studies
not
addresses,
are
very
and, as noted
generally
for both. The
criteria
half full, depending
flexible.
very
These
are
the
data
by the authors,
comparable,
once
one
that
the
current
mono
the data of the two diary

adopts
the
same
coding
outcome
is the story of a glass either half empty or
on one's perspective.
Here
is a quick summary of the
118
COMMENTARY
(based on the diaries
relevant data from the current monograph

children that were fully analyzed).
Of the few verbs the children used at 18 months,

in more
used
children's
were
verbs
of
55%
months,
sentence
one
than
sentence
the
children's
the entire period

verbs
Across
more
with
the
16% of
entire
their
or
present
the
of diary collection,
one
than
point
of
view.
children
used
as one
used
intransitive
verbs
word
the perspective
the eight
one
children
morphological
object.
used
only
form
(e.g.,
over
course
the
or with
utterances
word
of
the
lexically ex
subject and object
one
only
with
five
four
children
The
authors
the
children
three
the
claims of lexical
ground general
pretty flexible.
But the alternative
the
course
of
the
(53%
lexically expressed
one
to five
current
at 24 months
frames,
not
have
child
such
single
had one
of age,
frame
such
that
or not
depends
and
on one's
as
monograph?taking
specificity
they
frame,
frames.
this syntactic flexibility

of
over
used
sentence
did
or more,
verbs
had
children
the subject
utterances).
of
of
used
that
used with
as one
used
one considers
Whether
collection,
and
subject,
expressed).
the
three
expressed
diary
than
used with both
study, 30% were
From
that
verbs
(56% were
argument
lexically
children used 31% of
tense).
past
transitive
pressed
were
24
one
than
than one lexically expressed
of
period
with more
verbs
study, 35% were
Of
the
At
frame.
in more
used
of
44%
sentence
were
verbs
43% of their verbs with more
Of
20 months,
one
than
15% were
only
frame.
Across
their
in more
used
At
frame.
of the eight
back
in early grammars?think
it is
of these kinds of data by proponents

interpretation
of lexical specificity has typically used as background
the expectation
that
children's language is structured by a set of abstract syntactic categories and
constructions
that apply across all relevant
lexical items equally. In this
one
use
one third of their transitive
wonders
children
about
context,
why
and intransitive verbs with all of the syntactically required arguments,
but
the other two thirds without. And why do children at around 2 years of age
still use almost half of their verbs in only one sentence frame? And why
are
there
perspective,
so
few
sentence
although
frames
children
used
across
multiple
show some flexibility,
verbs?
it does
not
From
this
look even
119
MICHAELTOMASELLO
AND SILKE
BRANDT
like all of their utterances
remotely
neutral,
lexically
syntactic
derive
and
categories,
from a set of preexisting,
abstract,
rules.
this disagreement
about flexibility
is really the issue of
In
the
verb
island
children
may be as flexible as
productivity.
hypothesis,
use
in
like
their
of
verbs
in
you
particular
syntactic frames and morpho
as
logical paradigms?so
long
they have heard all of these uses in the lan
around
With
them.
each verb, they learn various ways of talking from
guage
the discourse
interactions
is, for each verb
they have with adults?that
is cut
(1992) diary, a paradigm
separately. And so, in Tomasello's
example
Underlying
versus
draw.
frame,
whereas
similar
was
Cut
draw
used
(learned
was
semantics)
by
used
Tomasello's
at about
in many
child
same
the
different
one
in
basic
only
time
and
having
ones.
The
general
sentence
somewhat
explana
to many
tion is that, for whatever
reason, the child either heard or attended
different syntactic uses of draw, but not so for cut. Thus, flexibility of use for a
particular verb is purely a function of the language the child has heard and
to with that verb, along with her motivation
attended
for speaking about
these kinds of events.
And so flexibility with particular verbs does not signal lexically general
across all verbs, unless one has evidence
that the flexibility
is
productivity
a
to
due
child's creative generalization
and not simply to a reproduction
of
adult flexibility with each particular verb individually. Children will become
across verbs
in all kinds of ways at some point?including
productive
and
constructions?it
is just
that
initially
they
are
not.
This
is
important
If children begin lexically based and become

it speaks to process.
because
more productive
only gradually, then the likelihood is that the process is not
one
of
constructing
general
from
activating
ever
cognitive
mature
innate
more
syntactic
general
processes
language
and
categories
syntactic
(e.g.,
categories
statistical
rules,
but
and
rules
learning,
rather
on
analogy)
the
one
of
basis
of
and
input
users.
INEARLY
PRODUCTIVITY
VERBUSE
in the
is thus the key theoretical
issue. Indeed,
Syntactic productivity
on
was
we
the
entire
focus
first draft of this monograph
reviewed)
(which
reason
from
theoretical
for
this
derives
The
Chomsky's
mainly
productivity.
children go be
If from the beginning
poverty of the stimulus argument.
ways, this
yond the language they have heard in creative and productive
and
schemas
have
abstract
that
(or
suggests
categories
they
preexisting
that the
is the argument
their early language. This
rules) that generate
authors tried to make in the first draft. But the reviewers simply noted that
without any data about what language the children had heard, one can talk
about flexibility but not about productivity. One cannot say that children are
120
COMMENTARY
going beyond what they have heard if one does not know what they have
is indeed the issue these authors wish to
heard. That syntactic productivity
near
in
the end of the monograph:
is
evident
their
quote
push
These
are more
findings
guage acquisition
the positions
olds do not
sello, 2000).
that
have
consistent
(e.g., Chomsky,
children
abstract
are
syntax
with
the
view
generativist
of
child
lan
1981; Gleitman & Fisher, 2002) than with
conservative
(e.g.,
Goldberg,
language
1998;
users
Lieven,
and
that
2-year
2006;
Toma
(p. 107)2
as well as in
in the original verb island formulation,
But, to reiterate,
are
other item-based
children
approaches,
syntactically productive?but
only in limited
Thus, as alluded
teaches a child
diately stick it
terms
predicate
within
their item-based
constructions.
ways constrained
to above, from as early as they are combining words, if one
a novel noun in an experiment
(e.g., the wug), they imme
into the syntactic slot of many of their existing verbs and
straightaway,
saying such productive
things as "Wug gone"
never
and "Find wug"?even
before heard wug com
though they have
bined with any other word (Tomasello, Akhtar, 8c Rekau,
1997). Early in
most sentence frames are structured by verbs (like gone and
development
slots freely and pro
find), each of which fills new words into its argument
children are productive with their language in some ways
ductively. Hence,
from near the beginning
of multiword
speech.
It is just that transfer of structure across verb island constructions
does
not happen as the result of the pregiven
and abstract categories of formal
as a result of both what the child hears and her
linguistics but develops
across initially very low-scope schemas. This is
generalizations
developing
illustrated by another analysis performed
in Tomasello's
(1992) diary study.
The
the child learned one of her verbs in a full
logic was this. Suppose
for example, saying many things of the form x FINDy.
transitive construction,
If this construction was fully abstract and verb general, one would expect to
see her using other verbs in the same frame pretty soon after, that is,
given the
so.
to
not
one
But
do
that
is
what
the
data
If
showed.
looked at
opportunity
the development
of the sentence frames used with a particular verb across
time, the child added grammatical
complexity only gradually and incremen
never happened
for
each
verb
It
almost
that a verb that was at
tally
separately.
one point used in a simple sentence frame all of a sudden was used in a more
complex one, for example, based on some kind of transfer from another verb
that had recently been used in this more
frame. The summary
complex
statement was as follows: By far the best predictor of this child's use of a
given
verb on a given day was not her use of other verbs on that same day, but
rather her use of that same verb on immediately preceding days. Hence,
the
evidence here was that the overall developmental
initial
pattern suggested
for each verb.
syntactic independence
121
MICHAELTOMASELLO
AND SILKEBRANDT
on individual children
in the
comparable
analysis was performed
study. The analyses of the individual children (displayed in Figures
In terms of the number of sentence
11-19) basically showed the following.
frames used with each verb (flexibility), across time two children showed U
two children
showed a several
(Heather, Elaine);
shaped development
month
followed by a relatively rapid increase in
period with no flexibility
the
four children showed relatively quick
and
other
Sam);
(Carl,
flexibility
on.
This
kind
of
analysis does not look for any kind of
flexibility
early
No
current
across
transfer
sentence
frames.
recent
In another
study,
on
based
samples
of children's
(2006) claims to find
speech (not on a thorough diary), Ninio
lies the rub. Even in this case one cannot claim
such transfer. But herein
if one does not know what language
anything about transfer or productivity
the children have heard, or how their other cognitive abilities are devel
it could be that once a child's parents hear her using tran
oping. Hence,
sentences
sitive
more
sentences
with
one
often,
or more
verb,
this
saliently,
them
encourages
with
other
verbs.
to
use
Or
transitive
perhaps
the
the indi
this time, making
is expanding
child's working memory
during
of more complex verb island constructions
vidual acquisition
possible for
the first time.
di
The fundamental
point is that one cannot investigate productivity
have
the language the child has heard. Researchers
rectly without knowing
it of special importance when children overgeneralize
therefore considered
and
say
tion
that
things
this
like
sentence
"She
giggled
must
be
me"
generated
1982)?on
(Bowerman,
by
productive
the
schema,
assump
as the
have heard such a sentence with that verb in their ev

errors usually do not occur
(these productive
eryday linguistic environment
before age 3 years; Pinker, 1989). The other approach
is, of course, exper
in
iments. We and others have performed
literally dozens of experiments
see
a
we
in
various ways and
novel verb (e.g., tamming)
which
teach children
like the transitive or intransitive
if they then use it in existing constructions
to do so.
or passive?when
and motivation
opportunities
given multiple
as the authors of the monograph
is that children
The outcome,
note,
with newly
than 2.5-3 years of age show very little productivity
younger
child would
never
learned novel verbs?even

though older children show much productivity
with these same verbs learned in these same ways (see Tomasello,
2000,
2003, for reviews).
focus their critique only on the
The authors of the current monograph
we
and others have also done similar exper
but
experiments,
production
un
iments tapping the child's comprehension
(e.g., to see if children can
derstand who is doing what to whom in "The dog is tamming the cat") (e.g.,
8c Tomasello,
8c Tomasello,
Akhtar
2009; Dittmar,
1997; Chan, Lieven,
8c Tomasello,
Abbot-Smith,
Lieven,
2008a), as well as structural priming
using familiar English verbs (e.g., Savage, Lieven, Theakston,
experiments
122
COMMENTARY
8cTomasello,
across
ing
verb
the
in the
constructions
of
existence
we are testing for productivity
In all of these paradigms
2003).
island
abstract,
of
absence
relevant
input?suggest
in all cases
schemas?but
verb-general
the
before about 2.5-3 years of age.

finding is of very little syntactic productivity
The current authors focus instead on a recent preferential
looking study
with English-speaking
children by Gertner,
Fisher, and Eisengart
(2006),
in children under 2 years of age.
showing syntactic productivity
supposedly
of how to interpret
the results of a looking
aside the question
Leaving
in terms of the nature of underlying
that
discrimination
representations,
a
the
before
children
entered
the
used
experiment
study
training technique
that taught them key aspects of the construction
(i.e., it used the same actors
nouns
in both training and testing in the same semantic
and associated
When
children were tested without
such an initial
roles).
German-speaking
not
at this young
did
show
any syntactic productivity
training period,
they
so it seems
And
8c
Abbot-Smith,
Lieven,
Tomasello,
2008b).
age (Dittmar,
very
to us
clear,
at
that
least,
the
data
experimental
children across
and production,
in young
syntactic productivity
in both comprehension
ologies,
miliar verbs. One can always criticize
valid, but in this case there are many
show
several
a consistent
of
lack
different
method
both novel and fa
using
as not being ecologically
experiments
different methodologies
involved.
The main point is that there is no evidence
in the current monograph
for syntactic productivity
in 1- and 2-year-old
children?in
the sense of
across
generalizing
to
verbs
create
verb-general
construc
syntactic
we do not know what

tions?because
language
they have heard being
used around them, and there is no other indirect evidence either. Children
are flexible from the beginning,
as they learn from adults multiple
frames
for a given verb. And they are productive
in the limited sense that they can
for one another
in the syntactic slots of each
freely substitute participants
thus creating schemas such as x
verb-specific
syntactic frame individually,
on
But
based
Tomasello's
(1992) longitudinal analysis failing to find
FIND)).
transfer
children
across
verbs,
remain
on
and
conservative
the
extensive
with
their
syntactic
is that the grammatical

time, our own conclusion
children's
is not
competence
early linguistic
but rather
is concrete
and item-based?leading
productivity.
Perhaps
dren's
early
one reason
syntax
is that
for the slow buildup

to make
literature
experimental
generalizations
constructions
in which
some
for
organization
underlying
abstract and rule-based,
to highly
constrained
of productivity
across
such
in young
things
chil
as verb
island constructions
children must make complex analogies between com
structures
plex
(e.g., aligning agents with agents and patients with patients
in utterances with different
transitive verbs). The age at which they seem to
do this most readily in nonlinguistic
domains
corresponds
fairly well with
the age at which they seem to be doing itwith their language (see Gentner
8c
123
AND SILKEBRANDT
MICHAELTOMASELLO
Markman,
1997). And it is perhaps worth noting as well that psycholin
research
has shown that even adults' representations
of argument
guistic
structures are still fairly closely tied to specific lexical items (e.g., Trueswell
8cTanenhaus,
1994).
CONCLUSION
current monograph
is extremely
valuable and advances
the field
in
the
level
of
detail
and
thor
ways.
By exploiting
multiple
significantly
across
of
method?and
then
it
the
multiple
oughness
diary
employing
The
children?the
were
authors
to
able
collect
data
on
children's
early
acqui
assessments
sition of verbs that enabled all kinds of quantitative
simply not
now know new and important facts
in
We
other
kinds
of
studies.
possible
that
about children's
semantic flexibility with verbs in early development
that are
will help us to better formulate
theories of lexical development
inclusive of all word types. And we now know new and important facts about
children's syntactic flexibility with verbs that will help us to formulate better
come to generate
structure
and grammatically
theories of how children
we
in
critical
And
have
been
their early multiword
although
productions.
this commentary
about the claims of certain
in the current data set?we
actually believe
as
experiments?nevertheless,
flexibility
children's
dren
the
kinds of syntactic productivity

that that is impossible without
authors
out
point
in
several
places,
of
and so the investigation
is a precondition
for productivity,
a
is very important first step in figuring out how chil
flexibility
construct
ultimately
their
abstract
syntactic
constructions.
NOTES
of
for the exclusion
of verbs to focus on is well justified,
selection
except
early verbs, make, do, and have.
important
on p. 109 we get the somewhat
with
consistent
"One position
2. Although
contradictory:
are learned from the input, but the process
of
our data is that abstract grammatical
categories
or before
at
of
the
relevant
all
before
production
early,
speech production
learning
begins
frame or construction."
linguistic
1. The
three
authors'
fairly
References
Akhtar,
N.,
8cTomasello,
morphology.
M.
Bowerman,
late speech
146). New
M.
(1997).
children's
productivity
Developmental
to language
to talk worse: Clues
(1982). Starting
errors.
In S. Strauss 8c R. Stavy (Eds.), U-shaped
York: Academic
Young
with word
order
and word
acquisition
behavioral
from
children's
growth
(pp.
Press.
124
101?
COMMENTARY
M. (2009). Children's
of the agent-patient
A., Lieven,
E., & Tomasello,
understanding
Canton
between
in the transitive
construction:
relations
comparisons
Cross-linguistic
and English.
ese, German
20, 267-300.
Cognitive Linguistics,
com
M. (2008a). German
children's
K., Lieven,
E., 8c Tomasello,
Dittmar, M., Abbot-Smith,
Chan,
of word
prehension
1152-1167.
79(4),
order
K.,
M., Abbot-Smith,
Dittmar,
dren's
early syntactic
23, 48-66.
E.
Dromi,
Fenson,
case marking
and
Lieven,
competence:
in causative
Early lexical development. Cambridge,

P. S., Reznick,
J. S., Bates, E., Thai,
communicative
Monographs
development.
(1987).
L., Dale,
early
M.
E., 8c Tomasello,
A preferential
looking
sentences.
(2008b).
Child Development,
chil
German
Young
Science,
study. Developmental
UK: Cambridge
Press.
University
in
S. (1994). Variability
D., & Pethick,
in Child Devel
of the Society of Research
opment, 59(5).
versus natural
nouns are learned before verbs: Linguistic
D. (1982). Why
Gentner,
relativity
In S. A. Kuczaj
Vol. 2. Language,
(Ed.), Language
thought and
development:
partitioning.
Erlbaum.
culture (pp. 301-334).
Hillsdale,
NJ: Lawrence
A. B. (1997). Structure
in analogy
and similarity. Amer
Gentner,
D., & Markman,
mapping
ican Psychologist,
52, 45-56.
words and rules: Abstract
8c Eisengart,
Gertner,
Y, Fisher, C,
J. (2006). Learning
knowledge
in early sentence
of word order
Science, 17(8), 684-691.
Psychological
comprehension.
A. (1988). Three
Social words,
words
and
types of early word:
Gopnik,
cognitive-relational
names
to cognitive
First Language,
and their relation
8, 49-70.
development.
S. (1988). Linguistic
M., Barrett, M., Jones,
D., 8c Brookes,
input and early word
15, 77-94.
meaning.
of Child Language,
Journal
R. M. (Eds.). (2006). Action meets word: How children learn verbs.
K., 8c Golinkoff,
Hirsh-Pasek,
New York: Oxford
Press.
University
Harris,
usage-based
Lieven,
E., Pine,
Ninio,
M.
H., Speares,
(2003). Early syntactic
J., 8c Tomasello,
creativity: A
333-367.
30(2),
of Child Language,
approach.
Journal
and early grammatical
G. (1997). Lexically
based
J., 8c Baldwin,
learning
187-219.
24(1),
of Child Language,
Journal
and the learning curve: A new theory of syntactic development. Oxford,
Language
E., Behrens,
Lieven,
development.
A. (2006).
UK: Oxford
University
S. (1989). Learnability
Pinker,
MA: MIT Press.
Press.
and cognition:
The acquisition
of argument
structure.
Cambridge,
on estimates
C. E, 8c Fletcher,
S. L. (2006). The
effect of sampling
of lexical
and error rates. Journal
33(4), 859-877.
of Child Language,
specificity
M. (2003). Testing
the abstractness
of
A., 8c Tomasello,
E., Theakston,
Savage, C, Lieven,
con
children's
Lexical
and structural
of syntactic
linguistic
representations:
priming
Rowland,
structions
Tomasello,
M.
Cambridge
M.
Tomasello,
Merriman
in young
(1992).
University
(1995).
Cambridge,
York:
contexts
In M. Tomasello
for early verb learning.
& W.
names for
children's
verbs
acquisition
of
Beyond
things: Young
(pp. 115?
Erlbaum.
NJ: Lawrence
Pragmatic
(2000).
Do
young
M. (2001). Perceiving
Tomasello,
M. Bowerman
8c S. Levinson
132-158).
M.
Tomasello,
New
(Eds.),
146). Hillsdale,
M.
Tomasello,
209-253.
Science, 6(5), 557-567.

Developmental
verbs: A case study of early grammatical
development.
Press.
children.
First
children
intentions
have
adult
and
syntactic
learning
words
competence?
in the second
acquisition and conceptual

Press.
University
(Eds.), Language
UK: Cambridge
Cambridge,
a
(2003). Constructing
language. A usage-based
MA: Harvard
Press.
University
Cognition,
74(3),
year of life. In
development
theory of language
(pp.
acquisition.
125
MICHAELTOAAASELLO
AND SILKEBRANDT
L. (1997). Differential
in
K., 8c Rekau,
productivity
and verbs. Journal
24, 373-387.
of Child Language,
W.
Tomasello,
M., 8c Merriman,
(Eds.).
(1995). Beyond names for things: Young children's ac
Erlbaum.
quisition of verbs. Hillsdale,
NJ: Lawrence
children's
is
Tomasello,
M., 8c Stahl, D. (2004). Sampling
spontaneous
speech: How much
101-121.
31(1),
enough?/oMrna/
of Child Language,
Tomasello,
young
M., Akhtar,
N., Dodson,
use of nouns
children's
a lexicalist
M. K. (1994). Toward
8c Tanenhaus,
framework
of constraint
In C. Clifton
8c K. Rayner
resolution.
(Eds.),
Jr., L. Frazier,
syntactic
ambiguity
on sentence
Erlbaum.
Hillsdale,
Perspectives
processing
NJ: Lawrence
(pp. 155-179).
Trues well,
based
J. O,
126
COMMENTARY
FIRST
FROMINFANTS'
VERBS
LEARNING
Sandra
R.Waxman
the centuries, people have been fascinated with infants' first words.
of parents of young children,
is not a special characteristic
fascination
or psycholinguistics.
Instead, this fascination
developmental
psychologists,
Across
This
is
widespread,
discussions
and
of
topics
infants'
as
far
first
ranging
can
words
as
innate
serve
as
entry
knowledge,
to heated
points
the nature
of in
of national character.
telligence, and the development
we can trace
to the writings of the Greek historian Herodotus,
Thanks
an
to the time of Psammetichus,
the fascination with infants' first words
to
BO
in
the
who
According
reigned
seventh-century
Egyptian pharoah
were
to
the
belief
that
the
the
held
Psammetichus
Egyptians
legend,
firmly
most
in the world, but this was disputed
ancient people
hotly by the
were
To
in
the
settle this
that
fact
who
they
originals.
argued
Phrygians,
to
their
claim
the
(and
rightful place), Psamme
dispute
Egyptian people
a passionate
interest in infants' first words, a passion that
tichus developed
the origin of human
stemmed from a desire to discover
language and that
on language development
in
led him to conduct the first known experiment
to bring two newborn
children. Apparently,
he somehow managed
infants
to a shepherd,
living alone among his flock of sheep. The protocol for this
was simple and clear: itwas the shepherd's
responsibility
proto-experiment
to feed and care for the infants, to make sure that they heard absolutely no
human
language, and to wait patiently and listen carefully for the infants'
was equally clear: he reasoned
first words. The hypothesis
that in the ab
sence of any linguistic input, the first word uttered
these
infants would
by
reveal which language was the origin of all human languages. As it turned
that the first word uttered by the children was
out, the shepherd reported
and excitedly with their arms out
"becos," a word they uttered repeatedly
stretched. When Psammetichus
learned that "becos" was the Phrygian word
for bread, he accepted for the first time that Phrygian,
and not Egyptian,
was the original language of humankind.
127
SANDRA R.WAXMAN
In
but
the
stems
21st
from
century,
a different
our
interest
source.
in
We
infants'
are
first
no
longer
words
remains
strong,
with
consumed
dis
is the lingua franca of humankind,

but instead
covering which
language
with what infants' first words can reveal about the nature of the human
mind and how it is shaped by experience.
In their monograph,
Naigles, Hoff, and Vear (2009) provide an out
a careful analysis of infants' first words,
how
of
and
standing example
their first verbs, can inform current
theories and debates
in
especially
on
to
from
16
20
Focusing
eight infants, ranging
at the start of the investigation,
months
traces each infant's
the monograph
first uses of their earliest acquired verbs. The diary records, compiled with
the apparently
tireless support of the infants' mothers,
offer a rich depiction
of the pragmatic,
and
of
infants' early verb
semantic,
syntactic properties
contexts
in
Infants'
and
the
which
productions.
productions,
they occur,
then serve as an empirical base against which competing
about
hypotheses
the flexibility and productivity
of infants' early-acquired
verbs can be tested.
RICHLY
DESCRIPTIVE
fills an important niche. In essence,
it represents
the
This monograph
It provides an impor
first of what we might call a "focused diary design."
tant counterpoint
to more
that the
traditional diary studies, illustrating
diary
of
one
any
child,
we need
erything
this
monograph
in middleraised
no
matter
how
were
to
drawn
from
cannot
comprehensive,
to know. After all, although
the eight children
a rather
homogenous
families
by majority
upper-middle-class
tell
ev
us
in
included
(e.g.,
population
culture,
stay-at
even within
this population
and although
home mothers)
they represent
a
to language acquisition
of
the midrange
continuum
approaches
along
too much or too little were excluded),
their di
(e.g., those who produced
a
of
distinct
aries nonetheless
range
vary considerably,
revealing
approaches
to early verb learning. Importantly,
then, this richly descriptive monograph
serves
as
resource
to which
we
can
turn
to
later
to
pursue
a host
of
questions.
DIARY
DATA
GUIDEDBYAND APPLIED
TOTHEORY
is its strong commitment
Another outstanding
feature of this monograph
to theory. The data were guided by, and then applied to, theory. In addition,
and
the monograph
offers insight into the tight coupling between method
on
on
one
the
other.
and
and
the
hand,
interpretation
theory
coding
128
COMMENTARY
TENDENCIES
AND INDIVIDUAL
DIFFERENCES
CENTRAL
developmental
analyses of these eight diaries not only uncover
within
individual
the group as a whole, but also identify potential
patterns
the importance of bearing inmind that not all
differences. This underscores
in the same way. An
children go about the process of language production
means
of
and
offers
evidence
of the central ten
group patterns
analysis
to discern different
the surface, it is often possible
dencies, but beneath
to these individual
individual styles (Bloom, 1973; Nelson,
1973). Attending
differences
is instructive, and in the case of this monograph,
it permits the
It
authors to describe different developmental
also
underscores
that
paths.
a
children
of
task
the
with
may approach
although
set of linguistic competences,
universal
these alone do not determine
the
The
child's path of production.

This path is also shaped by linguistic character
istics of the input (e.g., Greek vs. English),
the amount of input (rich vs.
and
of
the
individual
child (cautious vs. intrepid)
sparse),
personalities
&:
8c
2005; Gleitman
Trueswell,
(Gleitman, Cassidy, Papafragou,
Nappa,
Fisher, 2005; Huttenlocher,
man 8c Lidz, 2006).
Vasilyeva,
Cymmerman,
& Levine,
2002; Wax
THE
OF "ABSTRACTNESS"
MATTERS
AND "FLEXIBILITY"
to focus on the first 10 uses was motivated
by the goal of
the
breadth
of
children's
first
of
verb use and
identifying
representations
verb meaning.
This is a key question, because
there is currently consider
over the breadth or abstractness
able controversy
of infants' early verb
an
and
For
if
infant
the verb
meanings
applications.
example,
produces
The
wave
decision
in the
context
of
flag-waving
event,
what
can
we
say
about
her
rep
resentation? Does she construe the verb narrowly, applying it to

flag-waving
events only? Does she construe
it too abstractly, applying
it to any and all
events that involve moving
one's hand while grasping an object? The au
thors of the monograph
this question
the
approached
directly, bringing
to bear on key theoretical perspectives
toddlers' productions
(Fisher, 1996;
& Golinkoff,
Hirsh-Pasek
2006; Lidz 8c Gleitman,
2004; Lieven, Behrens,
&
2003; Naigles,
2002; Tomasello,
2000; Waxman
Speares, 8c Tomasello,
Lidz,
2006).
In addition
to recording each infant's first 10 uses of a series of verbs,

and researchers
the mothers
also worked
together to record the range of
utterances
in which the newly produced
verbs appeared
and the range of
were
to which
situations
on the con
to
The
decision
focus
they
applied.
ditions under which the verb was produced
is also motivated
by theory.
129
SANDRA R.WAXMAN
At issue here is the question of flexibility of verb use. Of course, although

it
utterances
is not a simple matter to operationalize
with
flexibility, especially
as short as those produced
the authors were
by these very young children,
careful to guide their decisions by theory.
Another
strength of the analysis is the authors' efforts to tease apart
several (potentially) distinct elements of flexibility
(e.g., pragmatic flexibil
But
in this
semantic
there is a wrinkle
ity,
flexibility, grammatical
flexibility).
a
can
occur
in
of
frames
which
verb
because
kinds
the
analysis
grammatical
are not unrelated
to their semantics
1990;
(Fisher, 1996, 2002; Gleitman,
to bear in
8c Gleitman,
Lidz, Gleitman,
2004). As a result, it is important
mind that measures
of semantic and grammatical
flexibility are, of necessity,
not independent.
Although
focusing on flexibility and abstractness of early verb use is an
advance,
important
at
same
the
it raises
time,
several
thorny
of
questions
its
own. Chief
among them, does the child's flexibility of verb use, as measured

It might,
of the underlying
grammar?
really reveal the abstractness
here,
but
at
this
and
point,
based
on
the
current
data,
we
cannot
rule
strong
are merely
that in using verbs flexibly,
children
hypothesis:
to
the
which
have
been
(i.e., the various
exposed
mirroring
input
they
frames in which they have heard these very verbs in the ambient language).
the input lan
After all, if children have indeed been actively processing
a
more
first verbs,
their
for
than
then
the
time
guage
year,
by
they produce
As a
in
of
these
verbs
different
frames.
would
have
each
heard
many
they
even
uses.
to
in
first
To
them
be
10
flexible
their
result, we might
expect
alternative
resolve
this important matter,
itwill be important
to search for and analyze
errors.
children's
infants'
What this suggests is that although carefully cataloguing
is an important step, itmay not take us far enough. The first
to resolve some of the finely
may prove to be too blunt an instrument
Still, this first step certainly
questions of early language acquisition.
uses
far
enough
to
see where
we
should
step
first 10
10 uses
honed
takes us
next.
ENDEAVORS
OPENINGTHEDOORTONEW RESEARCH
opens the door to several creative lines of ad
Certainly, this monograph
the importance of analyzing the
it highlights
ditional research. For example,
even
to
before
children,
young
they
produce their first verbs, and pro
input
vides guideposts for the kinds of input analyses that are most likely to bear fruit.
It also sets the stage for a careful analysis of how the view we obtain from diary
in general relates to the view that we obtain
studies and language production
from
more
standard
experimental
work
based
on
language
comprehension.
130
COMMENTARY
the monograph
makes
clear the importance
of a more
uses
10
the
first
these
of
comprehensive
analysis,
looking beyond
early
verbs. How can we best characterize
the trajectory of these verbs?
produced
Do children exhibit a steep increase, or explosion,
in flexibility or abstract
a
characterized
reliance on the early
ness, or is development
by deepening
to
frames?
it
will
be
pursue focused diary studies
acquired
Finally,
important
of infants acquiring
other
than
languages
English, focusing especially on
in
that
differ
ways.
languages
theory-relevant
In addition,
The
strongest
for
signature
a "classic-in-the
is not
making"
whether
the
work answers all of our questions, but whether

it offers to carry us forward
in our inquiries. There
is no doubt that when considering
the contribution
to their monograph,
of Naigles, Hoff, and Vear, the answer is "yes." Thanks
we can move beyond asking whether young children are
capable of learning
verbs.
Clearly
they
are. We
of representing
capable
fashion.
Apparently
can
also
language
can.
they
move
beyond
we
Instead,
whether
asking
in an abstract
and grammar
now
can
move
they
are
or flexible
to
forward
pin
that best support the acquisition of abstract knowledge

point the conditions
to identify how early in development
and the use of flexible representations,
this abstract knowledge
becomes evident, and to discover how quickly it is
a new verb.
when young children encounter
apprehended
ACKNOWLEDGMENT
Preparation
of the commentary
was
by NIH
supported
HD030410.
References
L. (1973).
Bloom,
Mouton.
Fisher,
C.
(1996).
interpretations
Fisher, C. (2002).
One word at a time: The use of single-word

Structural
utterances
before syntax. The
Hague:
on verb
The
role of analogy
in children's
mapping:
Cognitive Psychology, 31, 41-81.
limits on verb mapping:
The
role of abstract
structure
in 2.5
of novel verbs.
Science, 5, 56-65.
Developmental
limits
of sentences.
Structural
year-olds'
interpretations
L. R., 8c Fisher, C. (2005). Universal
Gleitman,
to Chomsky
(Ed.), The Cambridge Companion
Press.
University
aspects
(pp.
of word
123-142).
learning.
Cambridge,
In J.
McGilvray
UK:
Cambridge
L. (1990). The
sources of verb
structural
meanings.
Language Acquisition: A Journal
1(1), 3-5 5.
of Developmental
Linguistics,
L. R., Cassidy,
K., Papafragou,
Gleitman,
A., Nappa,
R., 8c Trueswell,
(2005). Hard
J. T
words. Journal
and Development,
1, 23-64.
of Language
Learning
K. 8c Golinkoff,
R. (Eds.). (2006). Action meets word: How children learn verbs. New
Hirsh-Pasek,
York: Oxford
Press.
University
Gleitman,
Huttenlocher,
J., Vasilyeva,
child syntax. Cognitive
M.,
E., 8c Levine,
Cymmerman,
45, 337-374.
S. (2002).
Language
input
and
Psychology,
131
SANDRA R.WAXAAAN
Lidz, J., Gleitman,

the mental
L. (2004). Kidz in the

H., 8c Gleitman,
In D. G. Hall & S. R. Waxman
lexicon.
'hood:
Syntactic
(Eds.), Weaving
and
bootstrapping
a lexicon
(pp. 603
MA: MIT
Press.
636). Cambridge,
L. (2004). Argument
structure
and the child's contribution
J., 8c Gleitman,
Trends in Cognitive Science, 8, 157-161.
learning.
M.
E., Behrens,
H., Speares,
Lieven,
(2003). Early syntactic
J., & Tomasello,
30, 333-367.
approach.
of Child Language,
usage-based
Journal
Lidz,
to
language
creativity:
a
L. R. (2002).
Form
is easy, meaning
is hard: Resolving
in early child
paradox
86, 157-199.
Cognition,
language.
L. R., Hoff,
in early verb use: Evidence
from
E., 8c Vear, D. (2009).
Naigles,
Flexibility
a
multiple-Af
study. Monographs
diary
Serial No. 293).
74(1-111,
to talk. Monographs
K. (1973).
in learning
Structure
and strategy
Nelson,
of the Society for
Research
Serial No.
38(1-2,
149).
Naigles,
Tomasello,
M.
(2000).
Do
young
children
have
adult
syntactic
competence?
Cognition,
74,
209-253.
In D. Kuhn
S. R., 8c Lidz, J. L. (2006). Early word
Waxman,
learning.
Handbook
child
6th
2,
ed.,
Hoboken,
(Vol.
299-335).
pp.
of
psychology
8c R. Siegler
NJ: Wiley.
132
(Eds.),
COMMENTARY
OR NOT?
EARLY
CREATIVE
VERBLEARNERS:
Jane B.Childers
describes a longitudinal
study of eight children's first
an analysis of the variety of words used in conjunction
and the patterns
with 34 targeted verbs, the variety of utterances produced,
uses
in
10
the first
of these verbs. These data are
of developmental
change
a
most
important because
diary studies have included very few children at
on
not
time and have
focused
the beginnings
of verb learning. Thus,
these
of an early stage of verb learning that
results advance our understanding
This monograph
verb uses including
has received
The main
the
verb
young
kinds
learner
one. Verb
servative
debate
relatively little attention.

discussed by the authors
controversy
or
the
other
of studies
there
However,
as
a creative
tend
researchers
and,
of
conducted
are
dangers
user
course,
and
to
these
to view
concerns
or
language
the nature
as
a more
the child on one
differing
the ways
both
of
world
in which
views
data
or
overestimating
are
of
con
side of this
influence
the
interpreted.
underestimating
If we assume more
children's
spontaneous
knowledge.
creativity by the
child than we should, we may begin down a path of experimentation
that
will ultimately be less fruitful than it would be, while ifwe ignore creativity
that is present, we will miss revealing a capacity of the human mind that is
the debate is likely
profound. How to resolve this issue is unclear; however,
to rage
for
some
time.
of how to view children's productions

in this study
Putting questions
an important contribution
aside for a moment,
of this study is that it sheds
that is relatively unexplored.
light on a stage of verb development
Although
a new focus of research is directed at understanding
event understanding
in
8c
much
of
what
is
known
about
verb
Lakusta, 2009),
infancy (e.g., Wagner
concerns
learning
children's
comprehension
and/or
production
after
24 months
Lieven,
Pine, 8c Rowland,
(e.g., Fisher, 2002; Theakston,
a
of
the
better
initial
2004). Thus,
stages of verb learning is
understanding
to produce
It is difficult to get children younger
needed.
than 24 months
133
JANEB. CHILDERS
verb utterances
reliable
record)
parent
scarce.
been
speech
On
the
other
can
a researcher
is that
in the laboratory, and so sampling (or having their

in everyday contexts provides data that to date have
a
hand,
control
great
how
of
advantage
many
experimental
a child
has
exposures
in which that particular verb has

verb, the specific sentences
In one study in which these
and the timing of those sentences.
about a week of daily
children
needed
controlled,
2.5-year-old
a verb to produce
it (Childers 8cTomasello,
2002). A limitation
studies
to a
given
been
heard,
factors were
to
exposure
of laboratory
abilities
children's productive
is that they probably underestimate
the child has to produce
the new verb at the right time in an
studies
because
these kinds of studies (e.g., Abbot-Smith,

setting. However,
experimental
Akhtar
8c Tomasello,
&
Tomasello,
2004;
Lieven,
1997) are important for
to
be
that is, their
about
children's
productive,
ability
answering
questions
a verb they have not heard in a particular sentence type to a
to
extend
ability
new
sentence
type.
ARETHESE
CHILDREN'S
PRODUCTIONS?
HOW CREATIVE
Flexibility in Verb Uses
An
important
in verb
question
dividual
verb's
children
are initially context
in
the
particular
early
meaning
situational
in
learning
concerns
A
development.
bound
contexts
the scope of an in
view
prevailing
in their verb uses, producing

and
to
failing
is
a new
extending
that
verbs only
verb
1983); recent exper

very widely
Smiley, & Charney,
(e.g., Huttenlocher,
this view (Forbes & Poulin-Dubois,
imental evidence has partially supported
8c Brandone,
Golinkoff,
Hirsh-Pasek,
1997; Maguire,
2008). This type of
idio
would be very useful to children given the somewhat
conservativism
an individual verb as well as the less
nature
of
of
the
meaning
syncratic
structure of the verb category compared with the hierarchical
organized
in this study included
structure of concrete nouns. Some of the analyses
semantic flexibility, or the use of a verb with different nouns or in conjunc
tion with different actions. These analyses are important because they show
that young children are more flexible in the use of different affected object
than in their use of different agents and that both of these types of
across different
are more common
actions (al
than is flexibility
flexibility
to code from these kinds of records).
uses
was
across
difficult
actions
though
levels of flexibility,
that there are different
these results demonstrate
Thus,
names
even
what
cause
it is still unclear
in the semantic domain, which is important. However,
be
is like for these early verb representations
the scope of meaning
the
analyses
likely affected
rest
on
by a myriad
children's
of factors
productions
(e.g., objects
of
the new
verbs,
which
available, motivation).
134
are
In
COMMENTARY
addition, even though new objects and agents were produced with the verbs
(and these were the most common ways in which children were semantically
it is possible
that children were still using the verb in particular
flexible),
contexts
situational
(and, thus, were
bound), while mixing
situationally
which particular
entities
in that context were named. That
is, it is still
somewhat unclear from these data how widely these new verbs were used
across
situational
or how
contexts,
"portable"
these
verbs
were.
In my
view,
this is the kind of contextual boundedness

in early verb
others are reporting
et
we
are exam
In
Huttenlocher
al, 1983).
my laboratory,
learning (e.g.,
an
how
children
achieve
of
level
(or
ining
might
appropriate
"portability"
new verb by
of a particular
the range of
extension)
actively comparing
situations
(Gentner, 1983, 1989) in which that verb is heard (Childers, in
Childers
8c Paik, 2009).
press;
Because all of the semantic flexibility exhibited
in these children's ut
terances could have stemmed from uses by the adult, what that means
in
terms of children's verb representations is unclear. If children are
verbs
using
to refer to different agents and to more
than one action, but these are all
initiated by the adult, could it be that a more
imitative process
superficial
underlies
this flexibility? That

this semantic
is, if the process underlying
is
could
their
verb
remain
imitative,
flexibility
representations
fairly small,
seems possible. Studies
restricted, and context bound? This interpretation
of verb use in everyday contexts are important because
can
they
potentially
reveal the breadth and depth of a young child's developing
representation
some
of a particular
verb. This study provides
helpful data relevant for
are like, but a detailed
inferences
about
what
those
making
representations
a
of
situations
in
which
verb is produced
the
(and perhaps whether a
coding
was
could
verb
have
been
to reveal
but
not) is needed
particular
produced
more
fully
the
Obviously,
point
in
scope
of
children
development.
children's
early
verb
representations.
are likely to be productive

Adults
are
productive
verb
in their verb uses at some

users,
creating
new
verbs
(e.g., "texting") and extending

existing patterns of use to those new words
centers on how early children
are
("I'll text you later"). The debate
creative
in
that
the
(i.e.,
ways
productive
spontaneously
go beyond
input)
because later productivity
allows for the possibility
that children could learn
to be productive, while early
that support
suggests mechanisms
productivity
that creativity could be innate. These
data do not bear directly on this
the input, we are unable to discern how these
question because, without
are tied to specific utterances
children's flexible productions
they may have
heard from an adult. That is, the question of
productivity
requires an eval
uation of the strength of the link from parent
speech to child speech.
this limitation does not mean
that these data are not important.
However,
The question of whether
children can go beyond the input in the early
of
verb
will
be
difficult to test. It will be difficult to record all
stages
learning
135
JANEB. CHILDERS
it with the child's

verb and then compare
input for a particular
new productions
to
not
heard.
Other argu
of
that
verb
reveal
production
ments of creativity in productions
have included a discussion of errors made
that would not fit the input and of errors that are possible but are never
of verb errors in older
made. For example, Bowerman's
(1982) description
children
is important because, by these ages, her children were producing
of the
errors
an adult English
that did not fit anything
say. In
speaker would
to
number
and
in
which
the
studies
addition,
exposure
types of
laboratory
sentences
that are sen
in which a novel verb is heard, with methodologies
it possible to control the input and address
sitive to toddler abilities, make
that we currently have very little of
Given
the question
of productivity.
a
either type of data (diary or laboratory) for this stage of verb acquisition,
it
of
demonstrates
the
kinds
is
that
of
the
contribution
present study
major
in the early stages of verb learning, even
children can produce
flexibility
links to specific utterances
created by
have
direct
this
flexibility may
though
an
adult.
Flexibility in Utterances
is not
though flexibility
examined
that these researchers
Even
evaluated
for
them
dren
were
most
of
the
as well.
flexibility
verbs
producing
data
reported
the same as productivity,

children
the utterances
These
in more
here
than
concern
one-
that
show
analyses
one
it is important
and
produced
of
utterance.
type
and
two-word
these
chil
However,
and
utterances,
are scaffolded by an underlying

these utterances
thus it is unclear whether
and
Martin Braine's
structured grammar. Lois Bloom's
(1971) work
(1993)
the
demonstrate
same
as
tensions
reflected
in
this
current
work.
For
ex
to
are too few words in the utterances
ample, Braine pointed out that there
or
not.
does
the child has a grammatical
be sure of whether
Thus,
category
some of the differences
the present
between
(1992)
study and Tomasello
utter
a rich (L. Bloom-like)
of
reflect the tension between
interpretation
ances
as
seen
here
and
a more
conservative
analysis
of utterances
as used
by
in
and Tomasello.
(And, although we usually prize conservatism
own
as
has
its
conservative
limitations.)
science,
noted, being overly
to command
and to de
In addition,
children were using utterances
Braine
scribe,
but,
at
the
same
time,
there
are
many
more
communicative
functions
as
This could be viewed as pragmatic
that could be expressed.
flexibility
as
a
set
or
of
be
restricted
it
viewed
could
and
argue,
colleagues
Naigles
functions. These are just the type of utterances parents are
communicative
of
and descriptions
likely to use frequently with young children: commands
to use these forms could be evi
events. Thus,
the tendency for children
dence of a link between parent and child speech in this study.
136
COMMENTARY
Are
Individual Differences
Important?
is that it provides
Another major contribution made by this monograph
as
a sense of individual variation in verb learning in this sample. However,
not
is
of
the
and
this
note,
range
colleagues
fully representative
Naigles
or even of the range of parents who initially
actual range of variation,
to participate
in this labor-intensive
agreed
study. Instead, a surprising
of families initially agreed but found
verbs. In addition,
several families
producing
number
drop
sented
out
their
because
here
children
a
represents
were
of
midrange
too
that their children were not

began the study but had to
verbal.
children
Thus,
who
were
the
neither
repre
range
too
verbal
nor not verbal enough at the time of the study, and the interpretation
of the
in mind. Per
results should be made with these participant
characteristics
in this study is even more
in development
haps the variability
surprising
then, given the fact that children at each end of variation were not included.
The
careful analysis by these researchers

information
provides
important
in
children
take
verb
about different
which
is impor
trajectories
learning,
tant for researchers,
and
alike
for
parents,
speech therapists
understanding
unfolds.
ways in which normal language development
Are There Specific Verbs That Provide a Particular Benefit in Verb Learning?
in this monograph
addresses questions
Finally, the evidence presented
about whether
the acquisition of some verbs ismore useful for breaking into
the
verb
refer
Thus,
lexicon.
to concrete
researchers
An
nouns
advantage
objects
can
have
can be ordered
examine
the
over
verbs
is that
hierarchically
acquisition
of nouns
the
nouns
that
within
a category.
at different
levels
of
a hierarchy and have found an

early tendency for children to learn nouns
or subordinate
to the superordinate
for entities at the basic level as opposed
levels (e.g., Rosch, 1978). Verbs in a language do not appear to have a clear
structure. There
basic level or a single hierarchical
is a small set of verbs that
is fairly general
As
and
note, there are
verbs").
("light
Naigles
colleagues
researchers who have proposed
that this set is especially
important in verb
can
these
the
child
because
verbs
discover
patterns of
learning
general
help
it has been unclear whether
syntax that hold for groups of verbs. However,
these verbs actually serve this function of "pathbreaking,"
and there is no
evidence
that they do so here. Even if this small set of very general
(light)
data to the child, this is obviously not the same
instead of thinking of verb
learning. Thus,
as
a
in a hierarchy, children
dimension
learning
proceeding
along
particular
to
more
be
both
and
appear
learning
general
specific verbs at the same time,
as the study reported
in this monograph
In ad
demonstrates
effectively.
verbs does provide important

as the "basic level" in noun
seem
to
dition,
verbs
share
sets of syntactic
group
into
frames,
clusters
instead
of
meaning,
of being
or
clusters
hierarchically
of verbs
that
structured
137
JANEB. CHILDERS
across
the verb category, and thus it is unclear whether general extension

across different verbs would actually be useful. Children may be
strategies
in verb learning, given
the major
better
served by being conservative
in the
in the meaning
of individual verbs and the differences
differences
of
range
syntactic
frames
that
verbs
particular
allow.
CONCLUSION
The
dren,
study reported
focuses
across
semantic
on
early
and
here
verb
is important
uses,
syntactic
and
contexts.
it includes
because
reveals
information
Future
studies
several
about
are
chil
uses
those
needed
to
clarify
be
and the relationship
the scope of children's early verb representations
tween parental
this
and
children's
However,
study
productions.
input
to the field by revealing more about an
makes an important contribution
this study
Thus,
early stage of verb learning that is poorly understood.
event
an
and
of
infants'
link
studies
between
processing
important
provides
a link that is important for un
studies of verb learning after 24 months,
the entirety of verb development.
derstanding
References
M. (2004).
E. V. M., & Tomasello,
K., Lieven,
Abbot-Smith,
The role of frequency,
construction:
duce the transitive
Science, 7, 48-55.
Developmental
syntactic distribution.
Training
semantic
2-6-year-olds
similarity
and
to pro
shared
with word order and verb

M. (1997). Young
children's
N., 8cTomasello,
productivity
Developmental
morphology.
UK: Cambridge
Uni
L. (1993). Language
Bloom,
development from two to three. Cambridge,
versity Press.
Akhtar,
M.
Bowerman,
L. Gleitman
In
in lexical and syntactic development.
(1982). Reorganizational
processes
art (pp. 320-346).
& E. Wanner
(Eds.), Language
acquisition: The state of the
Press.
UK: Cambridge
University
Cambridge,
In D. I.
of grammars.
of the internalization
Braine, M. D. S. (1971). On two types of models
Press.
New
York:
Academic
The
Slobin
153-186).
(Ed.),
(pp.
ontogenesis of grammar
new
two
events helps
extend
to multiple
Childers,
1/2-year-olds
J. B. (in press). Attention
verbs.
First Language.
and
8c Paik, J. H.
(2009). Koreanto learn novel predicate
information
Childers,
J. B.,
situational
English-speaking
terms. Journal
children
use
of Child Language,
cross
36,
201-224.
M. (2002). Two-year-olds
learn novel nouns,
verbs, and con
Childers,
J. B., 8c Tomasello,
or
actions from massed
ventional
Psychology, 38, 967
Developmental
spaced exposures.
978.
in 2.5
structure
The
role of abstract
limits on verb mapping:
Fisher, C. (2002). Structural
year-olds'
interpretations
of novel
verbs. Developmental
Science,
5, 55-64.
138
COMMENTARY
Forbes,
J. N.,
D. (1997). Representational
in young
children's
change
verb meaning.
24, 389-406.
Journal
of Child Language,
A theoretical
framework
for analogy. Cognitive Sci
Structure-mapping:
8c Poulin-Dubois,
of familiar
understanding
D. (1983).
Gentner,
ence, 7, 155-170.
D. (1989). The mechanisms
(Eds.), Similarity and analogical
sity Press.
Gentner,
of analogical
reasoning
P., 8c Charney,
Huttenlocher,
J., Smiley,
child: Evidence
from verb meanings.
M. J., Hirsh-Pasek,
K., Golinkoff,
Maguire,
facilitate
relation:
Fewer
examples
R.
(pp.
learning.
199-241).
In S. Vosniadou
New
York:
8c A. Ortony
Univer
Cambridge
of action
in the
Emergence
categories
72-93.
Review,
90,
Psychological
on the
A. C. (2008). Focusing
R. M., 8c Brandone,
De
initial verb learning
and extension.
children's
(1983).
velopmental Science, 11, 628-634.

In E. Rosch & B. B. Lloyd
E. (1978). Principles
of categorization.
(Eds.), Cognition and
Erlbaum Associates.
Hillsdale,
NJ: Lawrence
categorization
(pp. 27-48).
A. L., Lieven,
E. V. M., Pine, J. M., 8c Rowland,
C. F. (2004). Semantic
Theakston,
generality,
of syntax. Journal
and the acquisition
31, 61-99.
input frequency
of Child Language,
Rosch,
M. (1992). First verbs: A case study of early grammatical

Press.
Cambridge
University
to navigate
8c
L.
L.,
Lakusta,
(2009).
Wagner,
Using
language
Science, 4, 177-184.
Psychological
Tomasello,
development.
Cambridge,
the infant mind.
Perspectives
UK:
on
139
CONTRIBUTORS
R. Naigles
of Pennsylvania)
is Professor of
(Ph.D. 1988, University
at
Her
the
of
Connecticut.
research
the
Psychology
University
investigates
in children with autism and compares
of language acquisition
processes
across languages and cultures.
language development
Letitia
Erika Hoff
of Michigan)
is Professor of Psychology
(Ph.D. 1981, University
at Florida Atlantic University. Her research investigates
the role of input in
to
the
of
and
relation
memory
early language development
phonological
in
lexical
and
children.
early
monolingual
bilingual
development
Donna Vear (M.A. 2002, University
of Connecticut)
at NRI Community
Rhode
Services, Woonsocket,
Michael
is the Director
Tomasello
Department
tionary Anthropology
in Germany. His major
Psychology
social
social
cognition,
and
comparative,
cultural
learning,
is a children's
Island.
of the Max Planck

of Developmental
research
interests
communication
and
Institute
for Evolu
and Comparative
are in processes
of
from
aspects
perspectives?especially
clinician
developmental,
to
related
lan
theoretical
focus is on processes
of
involves mainly human chil
research
current
guage and its acquisition. His

shared intentionality. His empirical
dren from 1 to 4 years of age and great
is a postdoctoral
apes.
at the Max
Planck Institute for

Dr. Brandt's research interests are in language
Evolutionary Anthropology.
and processing
of complex
in German and English, acquisition
acquisition
learning mechanisms.
syntax, language and theory of mind, and general
Silke Brandt
researcher
at
in the Department
of Psychology
is a professor
on
between
the relation
Northwestern
University. Her research is focused
across development
and across
and conceptual
development
language
on
and
of biological
and
the acquisition
reasoning
knowledge
languages,
Sandra
R. Waxman
140
CONTRIBUTORS
across
cultures. She is a recent recipient of a Guggenheim

Fellowship,
Ann L. Brown Award for Excellence
in Developmental
Research
form
of Illinois, and a James McKeen Cattell Award.
University
the
the
of Texas-Austin)
is an Associate
(Ph.D., 1998, University
Jane B. Childers
at Trinity University,
Professor of Psychology
San Antonio. The main focus
of her research
is to examine
children's early verb learning, both in lab
studies
oratory
research
son
and
examines
processes,
may
contexts
in naturalistic
how domain-general
contribute
to
the
across
cultures.
mechanisms,
acquisition
of
Her
including
current
compari
verbs.
141
STATEMENT
OF EDITORIAL
POLICY
The Monographs
series aims to publish major reports of developmental
new findings and uses these to foster a
research that generate authoritative
or integration of findings on some
fresh perspective
conceptually
significant
issue.
Submissions
from
research
programmatic
are welcomed;
projects
these
or group-authored
may
reports of findings from a
or
a
from
sequence of experiments
single large-scale investigation
centering
on a particular question. Multiauthored
sets of independent
studies that
center on the same underlying
a critical
question may also be appropriate;
consist
of individually
in
requirement
issues and
such
and
substantial,
well-integrated.
research
multidisciplinary
including
contexts
advance
instances
that the contribution
evidence
are
of
from
particular
knowledge
on
is that
the
various
from
arising
Manuscripts
on significant
reporting
cultural,
racial,
interest.
Because
the
topics
aim
but
also
be unique,
or
and those
questions
ethnic,
of
common
address
interdisciplinary
developmental
diverse
specialized
authors
the set as a whole
the
or
national,
series
to
is not
enhance
other
only
cross
to
or subfields,
fertilization
the links between
the specific
among disciplines
issues under study and larger questions relating to developmental
processes
should emerge clearly for both general readers and specialists on the topic. In
short, irrespective of how itmay be framed, work that contributes
significant
data or extends developmental
thinking will be considered.
to be members
Potential authors are not required
of the Society for
or affiliated with the academic discipline of
Research
to submit a manuscript
for consideration
psychology
by the Monographs.
of the work in extending
significance
developmental
theory and in
new
information
is
crucial
the
consideration.
contributing
empirical
a minimum
should
of 80 manuscript
Submissions
contain
pages
The
of
is
tables
and
150-175
references).
upper
pages
(including
boundary
more flexible, but authors should try to keep within this limit. If color artwork
is submitted, and the authors believe color art is necessary to the presentation
of their work, the submissions letter should indicate that one or more authors
or their institutions are prepared
to pay the substantial costs associated with
The
142
color art reproduction.

Online
Monographs
Please
to the SRCD
submit manuscripts
electronically
at
and Review
Site (MONOSubmit)
contact the Monographs
office with any
Submissions
Please
www.srcd.org/monosubmit.
at
questions
The
letter,
monographs@srcd.org.
warrant
manuscript.
coauthors,
reviews
author
corresponding
that
all
for any manuscript
coauthors
are
in agreement
must,
with
in the submission
the
content
of
the
The corresponding
author also is responsible
for informing all
in a timely manner, of manuscript
editorial
submission,
decisions,
received,
and
any
revisions
recommended.
Before
publication,
the
author must warrant in the submissions

letter that the study
corresponding
was conducted according to the ethical guidelines of the
in Child Development.
Potential authors who may be unsure whether
the manuscript
they are
an
are
to
would
make
submission
invited
draft an
planning
appropriate
outline of what they propose and send it to the editor for assessment. This
as well as a more detailed description
of all editorial policies,
mechanism,
and format requirements,
evaluation processes,
is given in the "Guidelines
for the Preparation
of Publication Submissions," which can be found at the
or by contacting
SRCD website by clicking on Monographs,
the editor, W.
Andrew Collins, Institute of Child Development,
of
51
Minnesota,
University
E. River Road, Minneapolis,
MN 55455-0345;
e-mail: wcollins@umn.edu.
Note
to NIH Grantees
to NIH mandate,
will post the
Society through Wiley-Blackwell
version
of
to
Contributions
authored
NIH
accepted
by
grantholders
PubMed Central upon acceptance.
This accepted
version will be made
see
after publication.
For further information,
publicly available 12 months
Pursuant
www.wiley.com/go/nihmandate.
143
CURRENT
in Early Verb
Flexibility
Marital
Conflict
and Children's
293,
Behavior:
Externalizing
Diary
Multiple-Af
Vear (SERIAL NO.
and Donna
Erika Hoff,
From
Evidence
Use:
R.
Study?Letitia
Naigles,
2009)
Between
Interactions
and
Parasympathetic
Sympathetic Nervous System Activity?Mona El-Sheikh, Chrystyna D. Kouros, Stephen

Erath, E. Mark Cummings, Peggy Keller, and Lori Staton (SERIAL NO. 292, 2009)
The
of Early
Effects
and Relationship
Social-Emotional
St. Petersburg-USA
Orphanage Children?The
on
Experience
the Development
Orphanage
of Young
Team
Research
(SERIALNO. 291,2008)
Mother-Adolescent
Understanding
Conflict
Discussions:
and Across-Time
Concurrent
Prediction
From Youths' Dispositions and Parenting?Nancy Eisenberg, Claire Hofer, Tracy L. Spinrad,
Elizabeth T Gershoff Carlos Valiente, Sandra Losoya, Qing Zhou, Amanda Cumberland,
(SERIAL NO. 290, 2008)
Jeffrey Liew, Mark Reiser, and Elizabeth Maxon
and Gary Lupyan
The
in Infancy:
Concepts
Object
Developing
and Environmental
Genetic
An Associative
(SERIAL NO.
Learning
H.
Perspective?David
Rdkison
289, 2008)
of Learning
Origins
Years?Yulia Kovas, Claire M. A. Haworth,
Abilities
in the Early
and Disabilities
S. Dale,
Philip
School
and Robert Plomin
(SERIALNO. 288, 2007)

The
of Two
Preservation
Children's
Kagan,
Adolescence?Jerome
and Sara Towsley (SERIAL NO.
A Mechanism
Questions:
Into
Infant Temperaments
Vali Kahn,
Snidman,
Nancy
for Cognitive
287,
2007)
M.
Development?Michelle
Chouinard
(SERIALNO. 286, 2007)

Best
in Quantitative
Practices
R. Burchinal,
Foundations
for
Methods
and Kristen L. Bub

An
for Self-Awareness:
Developmentalists?Kathleen
(SERIAL NO.
Exploration
285,
Autism?R.
Through
Parental
Time,
A History
of the First 25 Years of the Black
T.
1973-1997?Diana
Slaughter-Defoe,
(SERIAL NO.
Harrison-Hale
(SERIAL NO.
282,
283,
Control,
K.
Psychological
Support,
and Method?Brian
Culture,
Peter
284,
2006)
in
Caucus
Aline M.
O.
Garrett,
Algea
2006)
Across
Control:
Relevance
Assessing
E. Stolz,
A. Olsen
and
Joseph
and Behavioral
Heidi
Barber,
2005)
Accounts
and Moral
Narrative
and Hurting
Others:
Children's
A. Brehl,
and
Conflicts?Cecilia
Interpersonal
Beverly
Wainryb,
of Their
Judgments
Sonia Matwin
Hurt
Being
Own
Margaret
Hobson,
(SERIAL NO.
Gayathri Chidambi, Anthony Lee, and Jessica Meyer

Too:
Our Children
Child Development,
McCartney,
2006)
(SERIALNO. 281,2005)
Childhood
Sexual
Assault
Victims:
After Testifying
Outcomes
Long-Term
in Criminal
Court?
Jodi A. Quas, Gail S. Goodman, Simona Ghetti, Kristen W. Alexander, Robin Edelstein,
Allison D. Redlkh, Ingrid M. Cordon, and David P. H.Jones
(SERIAL NO. 280, 2005)
The
of Social
Emergence
Malinda
Carpenter
of Physical
Trajectories
in Three
Cognition
(SERIAL NO.
Young
279,
Correlates, and Outcomes?NICHD
and
2005)
From Toddlerhood
Aggression
Tomasello
Chimpanzees?Michael
to Middle
Childhood:
Predictors,
Early Child Care Research Network
(SERIALNO. 278, 2004)

Constraints
Folksociological
on Conceptual
Knowledge
Carey (SERIAL NO.

Origins
Heron
A Case
Development:
in Madagascar?i&ta
277,
Body
Knowledge?
Virginia
Conversations
About
Susan A. Gelman, Marrianne

Development
Ulrich Muller,
of Executive
Douglas
and
Susan
and Michelle
Slaughter
in Collaboration with Linda Jenkins and Elizabeth Tilse (SERIAL NO.
Mother-Child
The
of Folkbiological
and
Solomon,
2004)
of Human
and Early Development
Study of the Acquisition

E. A.
Astuti,
Gregg
Gender:
Understanding
the Acquisition
Taylor, and Simone Nguyen

Function
in Early
Childhood?
Frye, and Stuart Marcovitch
of Essentialist
(SERIAL NO.
Philip
David
(SERIAL NO.
276, 2004)
Beliefs?
275, 2004)
Zelazo,
274,
2003)

Flexibility in Early Verb Use

Încărcat de

Informații document

Descriere originală:

Drepturi de autor

Formate disponibile

Partajați acest document

Partajați sau inserați document

Opțiuni de partajare

Vi se pare util acest document?

Este necorespunzător acest conținut?

Drepturi de autor:

Formate disponibile

Flexibility in Early Verb Use

Încărcat de

Drepturi de autor:

Formate disponibile

Flexibility in Early Verb Use: Evidence from a Multiple-N Diary Study

FLEXIBILITYIN EARLYVERB USE: EVIDENCE

Letitia R. Naigles, Erika Hoff, and Donna Vear

and Jane B. Childers

Manfred Van Dulmen

FLEXIBILITYIN EARLYVERB USE:

VERBGROWTHAND USEAS A FUNCTION

of language acquisition. The current study tests competing

a variety of affected objects, and they use their verbs in a

list of words and

are avid generalizers

For example, a child who hears the verb jump in the

of the child as language learner.

view that children

are the focus of this investigation of the process of early language

views of lexical development

in relation to the verb (Bloom, 1993).

is that the child enters

task equipped with

verb in a structure such as X verbed Y, yields, via syntactic bootstrapping,

context and to extend words to other things

and may not transfer directly from noun learning to verb

to identify new verbs as verbs

serve a range of pragmatic functions and will extend the verbs

From Studies of Word

about the extendability

a child might use the verb blow only when

restricted by actor or affected

these verbs to similar actions performed

of a wide range of English verbs

familiar verbs to actions by unfamiliar actors (Forbes

sitive verbs in a playroom

Regarding Children's Acquisition

The previous literature suggests some early limits on the extendability

early in the course of learning

individual verbs, does extendability

extendability of verbs from the beginning of verb use in the child

some kinds of extendability

but still appropriate

relative to both questions

than their later verb use.

is the source of individual differences,

iosyncratic, probably because children's input is idiosyncratic in

An alternative view, consistent with the child as a conservative

need to include an account of how children's verbs and verb

Evidence From the Study of Verb Learning

ative actions (in which the duck is doing something

Evidence for Early Lexically Specific Grammar

that allows them to in

cup" or "Break with

received both (see also Tomasello,

et al, 2001). These

of elicited speech have yielded

that young verb users do not routinely use their verbs in

in the current research).

available for analysis. Given

2006; Lieven et al., 2003) as well as time, thereby limiting the ob

different times of the day and different

time, the same types of conversations

is likely to involve expressions

free play, then the conversation

But very novel

the child adding