view

INFANT
BEHAVIOR
AND
DEVELOPMENT
14,
Individual
Correspondence,
Change in Mother
CATHERINE
S. TAMIS-LEMONDA
143-
I62
I 1991 I
Variation,
Stability,
and
and Toddler
Play
AND
MARC
H. BORNSTEIN
New York University
and
National Institute of Child Health and Human Development
Individual
differences
and developmental
changes
in mother and toddler
nonsymbolic
ond symbolic
play were examined
longitudinally
when toddlers
were
13 and 20
months old. A coding
system was developed
to assess
mothers’
play sophistication
bosed
on extant
coding
systems
used lo assess
toddlers’
play; the application
of
a common
coding
system
lo toddlers’
ond mothers’
ploy permitted
empirical
investigation
of behoviorol
parallels
between
toddler
and mother.
Characteristics
of
mother and toddler
play were regularly
associated
01 the two ages, and aver time
mother
and toddler
play sophistication
changed
in similor
woys.
For example,
between
13 ond 20 months, mothers
and toddlers
both moved toward
higher levels
of ploy, and changes
in one portner’s
play were regulorly
associated
with changes
in the other partner’s
play. Nonetheless,
signikont
changes
in mothers’
and toddlers’
play
remained
ofter
covarying
partner
influences.
These
findings
indicate
that
individuol
differences
ond oge-related
changes
in mother
and toddler
second-yeor
play ore bartly mediated
by matched
partner
ploy ond partly motivated
by processes
independent
of partner
ploy.
ofav
individual
differences
mother-infant
interaction
stability
Over the course of the second year, children are increasingly able to
conceptualize abstract relations between symbols and their external referents (Mandler, 1983). This fundamental change in representational
ability is reflectedin a rangeof newly emergingcompetencies,a noteworthy
one being play (Piaget, 1962). In the first year, play is predominantly
characterizedby sensorimotor manipulation; the goals of children’s play
appear to be to extract information about objects, what objects do, what
perceivable qualities they have, and what immediate effects they can
C. Tamis-LeMonda
was supported
by a New York
University
Fellowship,
a Helbein
Scholarship.
an IRTA
Fellowship
from the National
Institute
of Child
Health
and Human
Development.
and by research
grants (HD20559
and HD20807).
M. Bornstein
was supported
by research
grants (HD20559
and HD20807)
and by a Research
Career
Development
Award
(HD00521)
from the National
Institute
of Child Health
and Human
Development.
We thank
L. Cyphers.
K. Dine. G. Fitzmaurice.
J. Hampson,
J. Tal. and M. Zamor.
Correspondence
and requests
for reprints
should
be sent to Catherine
S. Tamis-LeMonda,
New York
University
Infancy
Studies
Program.
6 Washington
Place, New York.
New York,
10003 or to Marc H. Bornstein,
Section
on Child and Family
Research,
National
Institute
of Child
Health
and Human
Development,
Building
3l-Room
B2Bl5,
9000 RockviIle
Pike. Bethesda.
MD 20892.
143
144
TAMIS-LEMONDA
AND
BORNSTEIN
produce. This form of play is commonly referred to as nonsymbolic,
because children’s actions are tied to the tangible properties of objects,
rather than being representative. In the second year, children’s play actions
take on a “nonliteral”
or “as if’ quality. The goal of play now appears
to be symbolic, to represent experience. Play is increasingly generative,
as children enact activities performed by self, others, and objects in simple
pretense scenarios, pretending to drink from empty teacups, to talk on
toy telephones, and the like. (See Bretherton & Bates, 1984; Fein, 1981;
Garvey, 1977; McCune-Nicolich,
198 1; McCune-Nicolich
& Fenson, 1984;
Power & Parke, 1982; Slade, 1987.)
These two general types of play, nonsymbolic and symbolic, have each
been further analyzed into levels of sophistication.
In nonsymbolic play,
for example, actions tend to first be directed towards a single object (e.g.,
squeezing a foam ball), then they incorporate combinations
of two or
more objects, initially treated inappropriately
and later appropriately (e.g.,
a nesting block might first be juxtaposed with a busy box, only later to
be inserted in its appropriate partner block). In symbolic play, pretense
schemes tend to be applied to self before they are applied to others (e.g.,
pretending to drink from a cup before feeding a doll); single-scheme
pretense appears before multischeme pretense (e.g., pretending to drink
from a cup and later pretending to pour and drink); and pretense with
literal objects precedes pretense with “substitution”
objects (e.g., a telephone will at first represent a telephone, and later a stick might represent
a telephone). (See Belsky & Most, 198 1; Fein, 198 1; Lowe, 1975; Piaget,
1962; Shore, O’Connell,
& Bates, 1984; Ungerer, Zelazo, Kearsley, &
O’Leary, 198 1; Watson & Fisher, 1977.)
Parent-infant
interaction is a major setting for the development
and
expression of such social and cognitive skills (e.g., Rogoff, 1990). More
specifically, dyadic experiences that stimulate imaginative play are central
to the development
of self-directed exploration
(Teti, Bond, & Gibbs,
1988). However, the extent to which particular experiences scaffold or
promote learning is thought to be limited by cognitive factors: Children
assimilate what they can understand and often eschew what is too difficult
or long mastered (e.g., Killen & Uigiris, 1981; McCall, Parke, 8c Kavana@, 1977; Watson & Fisher, 1977). Accordingly, optimal contexts for
learning are ones in which mothers and toddlers match in play level and
are mutually sensitive to change in partner activity.
Although there is widespread agreement that maternal activities can
fundamentally guide children’s play and learning (e.g., DeLoache & Plaetzer,
1985; Dunn & Wooding, 1977; Lockman & McHale, 1985; O’Connell &
Bretherton, 1984; Power & Parke, 1982; Rogoff, 1990; Rogoff, Malkin, &
Gilbride, 1984; Rogoff & Wertsch, 1984; Slade, 1987; Tamis-LeMonda
&
Bomstein, 1989; Teti et al., 1988; Vibbert & Bomstein, 1989; Vygotsky,
1962, 1978; Zukow, 1986), little effort has been expended toward detailing
PLAY IN TODDLERHOOD
145
specific actions of mother that might affect child play. Developing unique
codes for maternal play or applying toddler play scales to mothers’ actions
would address this end. The latter would also permit characterization
of
patterns, processes, and determinants
of early play development
as well
as allow direct comparisons of toddler and mother. For the present study,
we developed a coding system for maternal play based on extant coding
systems of toddler play (e.g., Belsky & Most, 1981; Fein, 198 1; Hmcir,
Speller, & West, 1985; McCune-Nicolich,
198 1; Watson, & Fisher, 1977).
Using a common coding system, we then assessed variability in toddlers’
play and variability in mothers’ demonstrating and soliciting specific levels
of play when toddlers were 13 and 20 months old.
In demonstrating,
a mother provides her child with information
about
how to engage in particular activities by modeling the actions herself (e.g.,
mother dials the telephone); in soliciting, a mother places the onus for
play on her toddler by encouraging the child to engage in a play activity
(e.g., mother moves the telephone toward her child suggesting that the
child dial the telephone). Preliminary classifications have shown that these
maternal activities are common during interactive play. At an empirical
level, mothers’ demonstrations
and solicitations of play can be categorized
according to their level of sophistication
using the same coding scheme
applied to children’s play.
In this study, we examined how play in toddler and mother changed
over a 7-month period in the second year, and how changes in toddlers’
play and in mothers’ play informed matched developments
in partner
play at both group and individual
levels. We assessed concurrent associations between toddlers’ play and mothers’ play at 13 and at 20 months
and examined stability of mother and of toddler play between the two
ages. Finally, we asked whether forms of 13-month maternal play predict
forms of 20-month toddler play, and whether forms of 13-month toddler
play predict forms of 20-month mother play. In analyses of play change
and stability, we partialled mothers’ play from toddlers’ play, and toddlers’
play from mothers’ play, in order to assess play development independent
of partner activity. Likewise, in assessing predictive influences, we partialled stability in individual
and concurrent partner influence so that
unique predictions from mother to toddler and from toddler to mother
could be examined.
METHOD
Sample
Forty-five toddlers (22 males, 23 females) and their mothers participated
in two home observations. Subjects were recruited from private pediatric
and obstetric groups in New York City; children had been term at birth
and were healthy through the course of the study. Toddlers averaged 402
146
TAMIS-LEMONDA
AND
BORNSTEIN
days (range = 394-411) at the time of the 1Lmonth observation and 619
days (range = 61l-629) at the time of the 20-month observation. Their
mothers averaged 33.9 years (range = 28-42) and had completed an
averageof 6.3 years of post-high school education (range = O-8). Subjects
were from middle- to upper-socioeconomicstatus households(M = 60
on the Hollingshead Four Factor Index; Gottfiied, 1985).
Procedure of the Home Play Visits
Mothers and toddlers were videorecorded in 15 min of free play during
each of two home visits that took place 7 months apart. Visits were
scheduled at times of the day when toddlers were rested and alert and
when others (e.g.,fathers)were not present.Videotaping beganafter mother
and toddler were acclimated to the experimenter’spresenceand took place
in a room in the home other than the child’s own room.
At the start of the session,mother was asked to remain with her child,
to act in her usual manner, and to do whatever she ordinarily would do
when her child was engagedin play. She was also asked to disregardthe
experimenter’s presenceinsofar as possible. No further instructions were
given. Because the intent of the study was to observe mother-child
interaction in the context of free play, maternal behavior was not constrained in any way, and mothers could interact with toddlers as they
felt appropriate (seeO’Connell & Bretherton, 1984).
A set of toys was placed on the floor in front of mother and toddler
at each observation; toys included a teapot and cover, clown-doll, spoons,
cups and saucers,telephone, book, ball, blocks, nesting set, and vehicle.
These toys were chosen to afford toddlers and mothers opportunity to
exhibit various levels of play ranging from unitary functional acts to
sophisticated pretense (Belsky & Most, 1981; O’Connell & Bretherton,
1984; Ruff & Lawson, 1990; Ungerer et al., 1981).
Scoring
Play was subsequentlycoded from videotapesover successive15-sintervals. The median attention duration acrossall play episodeswas 12.5 s.
Play sessionslasted 15 min; therefore, there were 60 scoring intervals
over the session.Coders noted which of eight statesof play the child and
mother exhibited during each interval. Child or mother could be credited
with more than one level during the interval if different play levels
occurred. The eight play levels were: (1) unitary functional activity, (2)
inappropriate combinatorial activity, (3) appropriate combinatorial activity, (4) transitional play, (5) self-directed pretense,(6) other-directed pretense, (7) sequential pretense,and (8) substitution pretense.Operational
definitions of these play levels and predominant play acts which toddlers
exhibited at each level at the two ages, along with an index of their
frequency, are provided in Table 1.
147
PLAY IN TODDLERHOOD
In coding mother play, each demonstration and solicitation of play was
noted along with its level of sophistication. For example, if mother dialed
the telephone, she was credited with a demonstration at Level 1; if she
pretendedto talk on the telephone,she was credited with a demonstration
at Level 5. Similarly, if a mother moved the telephone toward her child
and suggestedthat her child dial the telephone, she was credited with a
solicitation at Level 1; alternatively, if she suggestedthat her child talk
on the telephone, she was credited with a solicitation at Level 5.
For toddlers, the total number of intervals (out of 60) in which each
play level occurred was calculated. In addition, values for nonsymbolic
Toddler
TABLE I
Ploy levels
Predominant
Play
level
1. Unitory
2.
Definition
functional
4.
13 Months
of an
Throw
effect thot is unique
to o single object
foom
Inappropriate
combinatorial
Inoppropriote
Put boll
Appropriote
juxtaposition
or more
play
of two
6.
7.
Approximate
pretense
but
confirmatory
Self-directed
Clear
pretense
activity
toward
Other-directed
pretense
Clear
Sequential
link two
pretense
pretense
8.
Substitution
pretense
f25%p
Months
telephone
(21%)
in vehicleb
Put boll
in vehicleb
of
without
pretense
Nest
Put telephone
receiver
to ear
(without
vocalization)
Put telephone
(44%)
(57%)
Eat from spoon
cup (39%)
directed
self
pretense
activity
towards
Put lid on teapot
(44%)
blocks
f27%)
obfects
evidence
5.
boll
Dial
juxtaposition
of two
or more obfects
Appropriate
combinotoriol
activity
Transitionof
20
or squeeze
activity
activity
3.
Production
examoles
Kiss or hug
(41%)
directed
other
or more
actions
Dial
speak
145%)
Pretend
activity
involving
one or
more object
doll
Pretend
and
receiver
Pretend
block is
telephone
and talk
into itb
to ear
vocalization)
Eat from spoon
cup (36%)
makes
telephone
into
or
receiver
(without
vehicle
sound
Dial
telephone
speak
(30%)
into
Pretend
or
f48%1
ond
receiver
block
telephone
into itb
is
and
talk
substitutions
D Percentages
in parentheses
reflect
the
frequency
of play acts at that level.
b Because
any inappropriate
combination
exemplifies
that play level, none dominated.
frequency
of the
associated
example
over
the
substitution
for
Level
total
for
Level
2 or any
object
8
148
TAMIS-LEMONDA
AND
BORNSTEIN
play were computed by summing the totals for Play Levels 1 through 4,
and values for symbolic play were cdmputed by summing the totals for
Play Levels 5 through 8. For mothers, totals were calculated for each of
the eight play levels for demonstrations and solicitations separately.As
with toddlers, totals for nonsymbolic and symbolic demonstrations and
for nonsymbolic and symbolic solicitations were calculated by dichotomizing play into Levels 1 through 4 versus Levels 5 through 8. Mothers’
total nonsymbolic play was the sum of nonsymbolic demonstrations and
nonsymbolic solicitations, and total symbolic play wasthe sum of symbolic
demonstrations and symbolic solicitations.
Different coders scored toddler and mother play at each age period.
Coders were thus blind to performance of different members of the same
dyad at each ageand to mother and toddler performanceat different ages.
Intercoder reliabilities for toddler play were obtained by having three
independent coders score the same toddler play on 20% of the sample.
Reliability was calculated by dividing the number of agreementsby the
number of agreementsplus disagreements.Agreementon eachof the eight
levels of play for matched intervals averaged97% (range = 96%99%).
Similarly, intercoder reliabilities for maternal activities were obtained by
having three independent coders scoreapproximately 15%of the sample.
Agreement on each of the eight levels of play for matched intervals
averaged96% for demonstrations (range = 92%99%) and 93% for solicitations (range = 88%98%).
Prior to analyses,all univariate data were inspected in box plots and
bivariate relations were examined in scatter plots (Tukey, 1977) in order
to identify extremesthat changedgroup correlations; none were observed.
Inspection of bivariate plots showedthat pairs of mother-toddler activities
were not associatedin any systematic nonlinear way. Neither maternal
status (e.g., age, education level, Hollingshead index) nor toddler age at
testing related systematically to the measured variables; thus, statistical
analyses did not consider these factors. However, mothers of females
more frequently attempted to solicit symbolic play at 13 and 20 months
and also demonstrated symbolic play more frequently at 20 months, ps
< .05. In addition, females exhibited more symbolic play at 20 months,
p < .05. Preliminary analyseswere thus conducted separatelyfor males
and females. Becauseresulting patterns were identical for the two groups,
data presentedhere collapse acrossgender.
RESULTS AND DISCUSSION
Results begin with a description of toddler play at the two agesand agerelated group changein the play of toddlers between 13 and 20 months.
The same format is applied to mothers’ play. Descriptive results and
analyses assessinggroup changesare presentedin Tables 2 and 3 (pp.
149
PLAY IN TODDLERHOOD
150 and 151, respectively).These changesare representedby two F values.
The first F value shows the results of a repeated-measuresanalysis of
variance comparing 13- and 20-month play. The second F value assesses
the same cross-agechange in toddler and mother play using repeatedmeasuresanalysis of covariance where the same level of one partner’s
13-month and 20-month play is covaried from 13-month and 20-month
play in the other. Thus, the first F values reported in Tables 2 and 3 test
overall age-relatedchangesin toddler and in mother play at each level,
and the secondF values test changesin toddler and mother play at each
level which are independent of partner influence. The second approach
to the assessmentof developmental changein play focuseson age-related
individual changein the play of toddlers and mothers. In these analyses,
cross-agechangesin the play of individual toddlers are regressedon crossagechangesin the play of individual mothers, and vice versa. Concurrent
associationsbetweentoddlers’ and mothers’ play at the two agesare next
examined. Unique stability in toddler and mother play between 13 and
20 months is then assessedby covarying partner stability from each
analysis.Finally, unique predictive mother-to-toddlerand toddler-to-mother
associationsare reported.
Developmenta!
20 Months
Changes in Toddler
and Mother
Play Between 13 and
Toddler Play, Toddlers showed strong tendenciesto exhibit particular
play acts at each level at 13 and at 20 months (seeTable 1). At the same
time, Table 2 indicates that children varied greatly in nonsymbolic and
symbolic play at each age. At 13 months, 85% of toddlers’ play was
nonsymbolic versus 15%symbolic; some toddlers never exhibited symbolic
play, whereas others exhibited as much as 51% symbolic play. At 20
months, 71% of toddlers’ play was nonsymbolic versus 29% symbolic,
ranging from as little as 2% to as much as 83% symbolic play.
A repeated-measuresANOVA conducted across the eight levels of
toddler play revealeda significant Age by Play Level interaction, F(7,308)
= 19.63,p < .OO1, indicating that agedifferencesin toddlers’ play depend
on level. This significant effectremained after covarying mothers’ matched
play levels from toddlers’ play, F(7,307) = 7.36,p < .OOl.The significance
of the overall Fs for these interactions warrants exploration of change
occurring at each play level and protects against Type 1 error that may
arise in conducting multiple analyses.Repeated-measuresANOVAs and
ANCOVAs were next conducted separatelyfor each level of toddler play.
As shown by the first F values in Table 2, between 13 and 20 months
children decreasedsignificantly in Level 1 and increased significantly in
Levels 3, 5, 6, 7, and total symbolic play. Thus, across the second year,
toddlers reducedtheir lowest play level while increasinghigher play levels.
150
TAMIS-LEMONDA
Toddler
AND
TABLE 2
Level ot 13 ond
Ploy
13 months
Ploy
Level
1
2
3
4
5
6
7
8
Nonsymbolic
Symbolic
20
M
(Rongel
M
17.6
11-331
IO-141
(O-371
11.4
4.8
6.5
3.2
2.1
1.6
10-251
10-91
IO-81
2.6
3.7
4.3
1.4
0.5
(0-l 51
(O-4)
3.8
0.9
32.0
5.7
” Simple f-difference.
” Mothers’
matched
ploy level
13 and 20 months,
respectively.
‘p < .05. l * p < .Ol. l **
4.5
14.8
20
Months
months
IRongel
fsll 1,441
I1 -261
IO- 191
23.4”’
0.4
0.2
32.7”’
0.7
5.8’
0.0
18.6”’
0.6
7.0”
1 1.5”’
13.2”’
1.6
7.1”
15.2”’
I .8
(O-4 1 I
IO- 13)
IO-201
(O-261
IO- I81
IO-91
F’lfl ,431
(4-65)
33.3
(4-591
0.4
3.4
to- 191
12.8
t l-361
26.0”’
29.8”’
ot
p
BORNSTEIN
13 and
<
20
months
covoried
from
toddlers’
ploy
level
ot
,001.
The covaried F values show that the increasesin toddler play remained
after considering mothers’ play. By contrast, the decreasein toddlers’
Level 1 play attenuated to nonsignificancewhen mothers’ matched play
was covaried, indicating that this decreasewas mediated by mothers’
decreasedLevel 1 play (see Table 3).
Mother Play. Mothers also varied substantially among themselves.On
average,at 13 months, 65% of mothers’ play was nonsymbolic and 35%
was symbolic. For individual mothers, the latter rangedfrom 9% to 78%.
At 20 months, 56% of mothers’ play was nonsymbolic and 44% was
symbolic, with some mothers never engagingin symbolic play and others
exhibiting 96% symbolic play. At 13 months, an averageof 42% (range
= 12%80%) of maternal activity consisted of play demonstrations as
opposed to 58% solicitations (range = 20%88%); by 20 months, 30%
(range = 5%67%) was demonstrations and 70% was solicitations (range
= 33%-95%).
A repeated-measuresANOVA for mothers’ demonstrations acrossthe
eight levels revealed a significant Age by Play Level interaction, F(7,308)
= 12.09,p < .OOl, indicating that differencesin maternal demonstrations
between the two ages depend on play level. This effect remained after
covarying matched toddler play, F(7,307) = 10.16,p < .OOl. Similarly,
there were significant Age by Play Level interactions for mothers’ solicitations both before and after covarying toddler play, F(7,308) = 35.04
and F(7,307) = 17.41, ps < .OOl, respectively. Repeated-measuresANCOVAs were next conducted to assessthe age effect at each level of play.
151
PLAY IN TODDLERHOOD
Mother
TABLE 3
Level at 13 and
Ploy
13 months
20
20 Months
months
M
IRongel
M
(Rongel
F”l I ,441
PI I ,431
5.3
1.2
4.1
0.0
1.7
2.3
0.8
0.3
(O-20)
IO-51
IO- I a)
10-l)
IO-61
(0-a)
IO-51
(O-41
I .6
0.6
2.9
0.2
1.0
I .5
0.7
0.2
(o-81
IO-51
IO- I 31
(0-2)
IO-71
10-71
10-5)
IO-31
44.6”’
6.0’
5.7
5.0’
4.3’
4.1’
0.1
0.8
19.2”’
5.9’
IO. I l **
4.8’
5.9’
9.1 l *
2.4
0.5
(2-351
5.3
(O-271
56.9”’
61.4”’
fO- I 7)
3.4
IO- 16)
7.3”
14.6”’
15.8
(3-39)
a.7
(I -29)
54.2”’
50.4”’
10.5
0.9
3.0
0.1
4.0
2.3
I .6
0.1
(0-24)
10-9)
IO- I 31
(0-l)
(0-l II
(0-l II
(O-5)
(O-11
3.1
0.8
7.3
0.1
3.8
4.6
1.8
0.5
IO- I 21
(0-l I)
IO- 19)
(O-21
10-13)
IO- 19)
10-l II
IO-61
50.2”’
0.1
32.0”’
0.0
0.3
14.1”’
0.1
4.7’
20.0”’
0.0
7.0’
0.0
0.7
6.4*
0.5
4.4’
14.4
I I -371
II.3
(O-30)
6.0’
6.9’
10.6
(o-38)
6.0’
0.1
Demonstrotion
Level
Nonsymbolic
10.6
Symbolic
5.2
Total
Solicitation
I
2
3
4
5
6
7
a
Nonsymbolic
Symbolic
level
8.0
i I -22)
Totol
22.4
(4-491
21.9
(2-511
0.1
0.4
Totol
Nonsymbolic
Ploy
25.1
(3-57)
16.6
12-391
29.5”’
33.9””
Symbolic
Ploy
13.2
(2-331
14.0
(O-45)
0.4
3.2
q~Simple F-difference.
I’ Toddler-matched
ploy
and 20 months, respectively.
l p
< .05. l * p < .Ol.
level
l **
ot I3
p
<
ond
20
months
covoried
from
mother
ploy
level
at
13
.OOl.
These analyses showed that mothers’ demonstrations decreasedsignificantly for all but Levels 4, 7, and 8, as did their nonsymbolic demonstrations, symbolic demonstrations, total demonstrations, solicitations at
Level 1, nonsymbolic solicitations, and total nonsymbolic play. In contrast,
solicitations at Levels 3, 6, and 8, and total symbolic solicitations all
increased significantly. Decreasesin mothers’ play remained after par-
152
TAMIS-LEMONDA
AND
BORNSTEIN
tialling toddlers’ matched play. However, the increasein symbolic solicitations attentuated to nonsignificanceafter partialling toddler play, suggestingthat this increasewas mediated by toddlers’ tendency toward more
frequent symbolic play.
Changes in Toddlers’ Spontaneous
Versus Prompted
Symbolic Play. As
noted above, toddlers shifted from lower to higher levels of play between
13 and 20 months, and in particular symbolic play increasedsignificantly.
We further examined the nature of advancesin children’s symbolic play
by considering the context in which that play occurred. A toddler might
exhibit pretenseprompted by a mother’s demonstration or solicitationmother talks on the toy telephone,then her toddler does-or the toddler
might spontaneously vocalize into the receiver in the absenceof any
maternal prompt. In order to compare the two scenarios,each toddler
symbolic play act was restored as“spontaneous” or “prompted.” Prompted
play acts were defined as those that followed a maternal demonstration
or solicitation of the same symbolic activity within the same coding
interval or in the prior coding interval without an intervening child play
act. Spontaneous play acts were those in which the child exhibited a
symbolic activity which was neither demonstratednor solicited by mother.
We then tested whether children’s symbolic play became distanced from
proximal social support from the start to the end of the secondyear by
assessingage changesin children’s relative distributions of spontaneous
versus prompted symbolic play.
Coding reliability for the spontaneous/prompteddistinction was calculated by having two coders independently score approximately 15% of
each age sample. Reliability averaged88% at 13 months and 84% at 20
months.
Age changesin children’s spontaneousversus prompted symbolic play
were tested in a 2 (Age) by 2 (Symbolic Play Type) repeated-measures
ANOVA. Results indicated a main effect for Age, F( 1,44) = 23.00, p <
.OOl, replicating the finding of increased symbolic play between 13 and
20 months. There was a main effect for Play Type, F(1,44) = 52.00, p
< .OOl, with spontaneoussymbolic play greaterthan prompted symbolic
play (74% vs. 26%, respectively). Finally, a significant Age by Play Type
interaction, F( 1,44) = 14.30, p < .OOl, indicated that toddlers engaged
in more spontaneousversus prompted symbolic play at the older than
the younger age.Thus, over this ‘I-month period, children’s symbolic play
not only became more frequent, but also grew more independent of
maternal prompts. It appears that symbolic play, initially “mother-generated,” becomes more “child-generated” over the secondyear.
Change in the Play of Individual Toddlers and Mothers Between 13 and
20 Months. Our next approachto assessingchangein play was to examine
PLAY IN TODDLERHOOD
153
individual toddlers and mothers. Changes in group means for Paflicular
levels of play or for summary statistics on play do not necessarily inform
about changes that are occurring in the play of individual
members of
the dyad. For example, a particular level of play might show no significant
change over time when group means are compared even though individuals
differ substantially among themselves on how much they change in their
play.
Are changes in the play of individual
toddlers between 13 and 20
months associated with changes in their mothers’ play over the same time
period? That is, are mothers and toddlers matched in their play development over time? To overview our findings, over age, the play of individual
mothers and toddlers changed in parallel; thus, mothers who exhibited
increased play at particular levels of play had toddlers who also increased
play at those same levels, and mothers who decreased play had toddlers
who decreased, and vice versa. The associations between mothers’ and
children’s changes in play for the eight levels of play and for nonsymbolic
and symbolic play are represented by the raw score regression coefficients
(P-weights) and F values presented in Table 4 (p. 154). Column 1 presents
the P-weights obtained from a simple regression of changes in toddlers’
play on changes in mothers’ play at each play level. Thus, for a given
level of play, as change in mothers’ play increased by 1 unit, change in
toddlers’ play increased on average by the corresponding
P-weight. Similarly, column 2 presents the P-weights for changes in mothers’ play
regressed on changes in toddlers’ play. The F values in column 3 test the
significance of the standardized slopes of the lines regressing changes in
partners’ play on one another.
As Table 4 indicates, changes in mother and toddler play at Levels 1,
2, 3, 6, and 7, total nonsymbolic
play, and total symbolic play were
associated with parallel changes in partner play at those same levels. In
general, the F values testing the regression slopes were moderate to strong
in magnitude. Moreover, the positive P-weights indicate that individual
mothers and toddlers covaried systematically at specific levels.
Figures 1A and 1B (p. 155) further illustrate these trends using symbolic
play as an example. Figure 1A plots the mean number of symbolic acts
at 13 and 20 months for toddlers of mothers who increased versus stayed
the same or decreased in symbolic play. Figure 1B plots the mean number
of symbolic acts at 13 and 20 months for mothers of toddlers exhibiting
increased symbolic play versus those who stayed the same or decreased.
As shown in Figure lA, children of mothers who stayed the same or
decreased in symbolic play showed a marginally significant 5 1% increase
in symbolic play over this 7-month period, F of the slope difference from
zero = 3.54, p = .O7. Children of mothers who increased in symbolic
play showed a highly significant 202% increase in symbolic play, F of the
slope = 31.53, p < .OOl. A 2 (Group) by 2 (Age) repeated-measures
154
TAMIS-LEMONDA
Associations
Between
Chonges
in Mothers’
AND
TABLE 4
Ploy ond
20
Ploy
level
BORNSTEIN
Changes
in Toddlers’
Ploy
From
13 to
Months
I+
p’,
F,Il,441
I
2
3
.44
.43
.67
.47
.29
I I .3”
6.1’
4
5
6
-.Ol
.20
.45
.33
.oo
12.2”’
0.0
7
8
.50
-.09
.20
.40
.32
- -.06
I .9
9.5”
8.2”
.42
.23
4.5’
.47
15.7”’
Nonsymbolic
Symbolic
ploy
ploy
I’ P-weight
months.
mo;tt;;eight
l
.57
0.3
of
toddler
ploy
change
regressed
on mother
ploy
change
between
13 ond
20
of
moth er ploy
change
regressed
on toddler
ploy
change
between
13 ond
20
Test of the
p < .05.
l
regression
between
* p < .Ol. l ** p
mother
<
and
toddler
ploy
changes.
,001.
ANOVA testing differences in symbolic play between the two groups
showed a significant Group by Age interaction, F( 1,43) = 9.99, p -C .Ol.
Thus, the two groups did not differ significantly in their initial levels of
symbolic play, F at 13 months < 1.00, but by 20 months the difference
between the two groups was significant, F at 20 months = 10.71, p <
.Ol.
Similarly, as depicted by Figure lB, mothers of toddlers who stayed
the same or decreasedin symbolic play showeda significant 38% decrease
in their symbolic play, F of the slope difference from zero = 8.94, p <
.Ol, whereasmothers of toddlers who increasedin symbolic play showed
a significant 27% increasein symbolic play, F of the slope = 6.24, p <
.05. A 2 (Group) by 2 (Age) repeated-measuresANOVA showed a significant Group by Age interaction, F( 1,43) = 13.27,p < .OOl. Again, the
two groups of mothers did not differ significantly in 13-month symbolic
play, F at 13 months < 1.00,but by 20 months the group differencewas
significant, F at 20 months = 5.23, p < .05.
In summary, between 13 and 20 months, toddlers becamemore active
participants in play and grew into more sophisticatedplayers as indicated
by a specific increasein symbolic play and a consistent movement toward
higher play levels. These changesin children’s play remained after partialling mothers’ play. Moreover, children showed an increasedtendency
to exhibit spontaneoussymbolic play; that is, symbolic play appearedto
depend less and less on mothers’ prompts. As these increaseswere occurring in toddlers’ play, mothers’ play tapered: Specifically, demonstra-
155
PLAY IN TODDLERHOOD
A: TODDLERS
SYMBOLIC
20,
PLAY
I
15 -
10 -
5r
I
0
13 Months
B: MOTHERS’
2Or
SYMBOLIC
I
13 Months
Figure
1.
of mothers
I
AGE
20 Months
PLAY
AGE
20 Months
A: Toddlers’
symbolic
ploy OS o function of change
in mothers’
who increased
I” symbolic
ploy fn = 23) versus
toddlers
remoined
unchanged
In toddlers’
symbolic
mothers
of toddlers
symbolic
ploy: toddlers
of mothers
who either
or decreased
In = 22). 6: Mothers’
symbolic
ploy OS o function
ploy: mothers
of toddlers
who increased
in symbolic
ploy fn =
who either remoined
unchanged
or decreased
fn = 13).
of change
32) versus
tions reduced, and, by contrast, solicitations did not change. Paralleling
toddlers’ play, mothers dropped lower levels from their repertoires, and
their predominant play shifted toward higher levels of sophistication. For
individual dyads, as mothers increasedor decreasedparticular play levels
156
TAMIS-LEMONDA
over time, their toddlers
play, and vice versa.
AND
BORNSTEIN
demonstrated
matched
changes in their
own
Concurrent Associations Between Toddler and Mother Play
Concurrent associations between toddler and mother play at 13 and 20
months are presented in Table 5. As shown, variation in toddler play
was associated with variation in mother play at both 13 and 20 months;
moreover, these associations tended to be highly specific. At 13 months,
mother nonsymbolic play related positively to toddler nonsymbolic play,
but not to symbolic play, and mother symbolic play related positively to
toddler symbolic play, but not to nonsymbolic
play. The same general
pattern held at 20 months. Mother nonsymbolic play related positively
to toddler nonsymbolic play and negatively to toddler symbolic play, and
mother symbolic play covaried positively with toddler symbolic play and
negatively with toddler nonsymbolic
play. The negative associations between nonsymbolic
and symbolic play of mothers and toddlers at 20
months suggest that over the 7-month period dyads became increasingly
matched in their play tendencies, particularly when considering mothers’
nonsymbolic
and symbolic solicitations. Thus, by 20 months, mothers
with toddlers who often engaged in symbolic play tended not to solicit
nonsymbolic play, and mothers with toddlers who often engaged in nonsymbolic play tended not to solicit symbolic play.
More differentiated
analyses showed that mothers’ nonsymbolic demonstrations related to toddler nonsymbolic play, but not to toddler symbolic play at 13 months, and did not relate to either form of toddler 20month play. By contrast, mothers’ symbolic demonstrations did not relate
to 13-month play but covaried positively with symbolic play and negatively
with nonsymbolic play at 20 months.
TABLE
Concurrent
Relations
Between
Mother
5
ond
Toddler
Toddler
Ploy
at
13 ond
13 Months
Mother
Ploy
.42”
.l I
1
.37”
-.I
Symbolic
Demonstrotions
Nonsymbolic
Solicitations
Nonsymbolic
.39”
-.06
Symbolic
<
-.02
.27’
-.13
Symbolic
‘p
20
Symbolic
Nonsymbolic
Nonsymbolic
.05.
*’ p <
.Ol.
l **
p
<
.OOl
20
.
Months
Ploy
Nonsymbolic
.42”
-.32’
.18
Months
Symbolic
-.27’
.52”’
.03
-.27’
-.I0
.48”’
.16
.53”’
.43”
-.28’
-.29’
.44”
PLAY IN TODDLERHOOD
157
Mothers’ solicitations showed a somewhat different pattern of covariation. At 13 months, mothers’ nonsymbolic solicitations were associated
with toddler nonsymbolic
play but not with toddler symbolic play; by
contrast, mothers’ symbolic solicitations related to toddler symbolic play
but not to toddler nonsymbolic play. At 20 months, nonsymbolic
solicitations covaried positively with nonsymbolic
play but negatively with
symbolic play, and symbolic solicitations were positively associated with
symbolic play but negatively with nonsymbolic play. The pattern of these
associations might partly reflect the differences between demonstrations
and solicitations and partly mothers’ sensitivity to their children’s capabilities. Mothers might only solicit actions they believe their toddlers
can perform. Accordingly, mothers who solicit symbolic play more often
at 13 months do so because their toddlers are capable of engaging in
symbolic play, and mothers who solicit nonsymbolic
play at 20 months
do so because their toddlers are functioning mostly at lower levels of play
sophistication.
To explore these concurrent
correlations
further, we focused on
mother-toddler
symbolic play associations, again distinguishing
toddler
spontaneous versus prompted play. Associations between mothers’ symbolic play and toddlers’ symbolic play could reflect the fact that toddlers
engage in symbolic play only in response to their mothers’ symbolic acts,
or toddlers could spontaneously exhibit symbolic acts. To distinguish these
possibilities, we compared associations between mothers’ symbolic play
and toddlers’ prompted symbolic play with associations between mothers’
symbolic play and toddlers’ spontaneous symbolic play. Mother-toddler
associations were essentially the same for spontaneous and prompted
forms at both ages, rs(43) at 13 months = .30 and .33, ps -K .05, and
rs(43) at 20 months = .38 and SO, ys < .Ol, respectively. Thus, the
concurrent relations between mothers’ and toddlers’ symbolic play could
not be explained by toddler immediate imitation
only. Toddlers with
mothers who are more symbolic engage more in symbolic play spontaneously as well as following maternal prompts.
To summarize, concurrent mother-toddler
play styles are specifically
associated, but patterns of association differ somewhat for demonstrations
versus solicitations, lending further support to the functional distinction
of these two maternal styles.
Stability in Toddler and Mother Play Between 13 and 20 Months
The next analyses examined 7-month stability of the play in toddlers and
mothers independent
of partner play (i.e., with variance ascribable to
partner play at 13 months and at 20 months removed). Specifically,
mothers’ earlier and later nonsymbolic and symbolic play were partialled
from stability in toddlers’ nonsymbolic
and symbolic play, respectively,
and toddlers’ earlier and later play were partialled from stability in mothers’
158
TAMIS-LEMONDA
AND
BORNSTEIN
play. Because cross-age correlations in the ranking of toddler and mother
play might be maintained in part by partner activity, partialling toddler
play from mother play stability, and the reverse, allowed assessments of
pure stability in mother and toddler respectively. Table 6 shows that
under these constraints toddler play was not stable. In contrast, mothers
were stable on all dimensions of play, independent of toddlers, and their
stability coefficients were medium to large in effect size.
Thus, mothers’ stability in play is not solely maintained
by their
toddlers’ activities but might also reflect factors such as maternal affect
or activity level. Alternatively,
individual
mothers might have distinct
attitudes and beliefs regarding the role of stimulation in children’s play
development
that continue over the toddler’s second year and translate
into behavioral stability in interaction.
Unique Predictive Relations Between Toddler and Mother Play
Thus far, two patterns of mutual partner correspondences in play have
been discussed. First, at both 13 and 20 months, toddler play and mother
play are keyed to one another. Second, changes in individual mother and
toddler play between 13 and 20 months are strongly associated with
matched changes in partner play. The final set of analyses examined the
extent to which one partner’s 13-month play predicted the relative ranking
of the other’s play at 20 months. To assess unique predictive influences
of mothers’ play on toddlers’ play, we conducted a series of hierarchical
regressions which, for the criterion of 20-month toddler play, first entered
1Imonth
toddler play, second the 20-month concurrent play of mother,
and last 13-month mother play. The same format was used to examine
Toddler
and
Mother
Ploy
TABLE
Stability
Ploy
Mother
Nonsymbolic
Symbolic
.46”
S5”’
.53”’
.61 l
Demonstrations
Nonsymbolic
Symbolic
l
p
<
.05.
l *
p
ploy
<
.Ol.
covoried
l
** p
from
<
.OOl.
**
31’
33’
Solicitations
Nonsymbolic
Symbolic
matched
ploy.
20 Months
.28
.20
Symbolic
0 Mothers’
mothers’
13 ond
StobilityO
Toddler
Nonsymbolic
from
6
Between
.33’
.40”
toddlers’
ploy;
toddlers’
matched
ploy
covoried
PLAY IN TODDLERHOOD
159
toddlers’ predictive influences on mothers. The final step of these regression
equations provided the change in R2associated with the unique influence
of one partner’s 13-month play on the other partner’s 20-month play. In
these analyses, the few significant predictive correlations that were obtained
at the zero-order level attenuated to nonsignificance. For example, mothers
who solicited more symbolic play from their toddlers at 13 months had
toddlers who showed more symbolic play at 20 months, r(43) = .28, p
< .05; however, after entering toddlers’ earlier symbolic play and mothers’
later symbolic solicitations, early symbolic solicitations explained less than
1% of unique variance in toddlers’ later play.
CONCLUSIONS
This study represents one step in the more complete characterization
of
individual variation, mutual correspondence, and developmental
change
and stability in toddler and mother play in the second year. In order to
describe parallels in toddler and mother play comprehensively,
we classified
maternal play activities according to their level of sophistication
based
on extant coding systems which, to date, have been applied exclusively
to toddler play. We then examined toddler-mother
correspondences in
specific play levels, both concurrently and with attention to the particular
changes that occurred in these levels for both partners across the second
year. Assessing similarities and differences between partners in levels of
play engagement is central to understanding
the role of scaffolding and
guidance in dyadic interaction.
Before proceeding to discuss these findings, it should be noted that
these data are based on dyadic interactions during a naturalistic free-play
session. Thus, the reported findings might apply uniquely to the play
activities of toddlers and mothers in this context. Moreover, the participating families were relatively restricted in terms of socioeconomic status
and education level, thereby limiting the generalizability
of findings to
other settings and populations.
Nonetheless, even within this relatively homogeneous population,
toddlers and mothers exhibited enormous variability
in how much they
engaged in play and in their relative play sophistication
at the start and
near the end of the second year. At both ages, moderate to strong
concurrent associations emerged in the play sophistication
of partners.
Moreover, changes in play across the second year for individual toddlers
closely matched changes in mothers’ play, and vice versa. Thus, partners’
play activities were linked during short-lived interactions, and they demonstrated corresponding patterns of development across the second year.
Changes in mother play and toddler play remained even after covalying
partner play, indicating that developments
in toddler and mother play
were also somewhat independent
of partner activities. Together, these
160
TAMIS-LEMONDA
AND
BORNSTEIN
findings suggestthat a comprehensive model of play development in the
second year needs to take into account activities both specific and not
specific to play, activities exhibited by both mother and toddler.
Consider, first, influences on toddlers’ and mothers’ play that depend
on partner play activities. The concurrent associationsobtained between
toddler and mother play might arise from reciprocal awarenessand sensitivity between partners during ongoing dyadic communication. During
interactive play, for example, mothers might respondto the visual, verbal,
and tactual interest toddlers exhibit and to their successesor failures in
operating on objects. Mothers might appropriately interpret those cues in
ways which guide their own play-directed actions. Similarly, through
incidental observation of and direct interactions with their mothers, toddlers might imitate their mothers’ actions and use their mothers’ suggestions to acquire information about objects, to act upon objects, and to
achieve play goals.
Consider, too, mechanisms that might explain developments in play
that are independent of partner play. For toddlers, one plausible explanation is that children’s own exploration and interactions with the inanimate and social environments guide learning independent of explicit
didactic experiences(seealso Rogoff, 1990).For mothers, changesin play
that are independent of toddlers’ own play might be partly determined
by a priori notions about children’s changing play capabilities, ideas that
are developmentally appropriate and translate into mothers’ own play
with children. Alternatively, mothers might be sensitive not to toddlers’
play per se but, rather, to other characteristics of the child, such as
developing languageability or social cues children direct towards them
regarding the difficulty of the task and the interpretability of mothers’
suggestions.Strong links appear to exist between second-yearlanguage
and play that suggestthat competencies in the two domains reflect an
underlying transition in representationalability (Bretherton & Bates, 1984;
Piaget, 1962;Tamis-LeMonda & Bomstein, 1989,1990;Ungerer& Sigman,
1984).
In contrast to concurrent links between mother and toddler play, and
the finding that cross-agechangesin partner play are connected,neither
mothers’ nor toddlers’ early play uniquely predicted later forms of play
in the other. Thus, considering the play status of individual toddlers and
mothers at the start of the secondyear doesnot contribute to understanding
or predicting partner play status at the end of the second year. Instead,
for mothers, a moderately strong predictor of later play was mothers’
earlier play activities. Mothers were stable on all aspects of play even
after considering toddler play sophistication. This suggeststhat mothers
maintain their relative status in play sophistication even as they change
in their overall level of play. In contrast, toddlers’ earlier play did not
predict their later play sophistication. Thus, toddlers’ 20-month play was
PLAY IN TODDLERHOOD
161
neither uniquely informed by their own play nor by the play of their
mothers at 13 months. The lack of predictability
of toddler play might
reflect the ages examined; differential acceleration or deceleration rates
and/or consolidation periods in development
might lead to instability of
individual differences in toddler play, as well as a lack of maternal influence
on toddler play (e.g., Bretherton & Bates, 1984; McCall, 1979; Piaget,
1962). Alternatively,
other measures of early toddler abilities (such as
language proficiency) or other measures of maternal stimulation
(such as
responsiveness) might predict toddler play sophistication better than mothers’ or toddlers’ earlier play. We are currently investigating these alternative
influences on toddlers’ and mothers’ second-year play.
REFERENCES
Belsky. J., & Most, R.K. (1981). From exploration to play: A cross-sectional study of infant
free play behavior. Der&p~~n/a/
f~,~cho/og?~, 17. 630-639.
Bretherton. I.. & Bates, E. (1984). The development of representation from IO to 28 months:
Differential stability of language and symbolic play. In R.N. Emde & R.J. Harmon
(Eds.). Continuiries
and disconrinuilies
in developnm~.
New York: Plenum.
DeLoache. J.S.. & Plaetzer. B. (1985. April). Tea for two: Joint mother-child
symbolic play.
In J.S. DeLoache & B. Rogoff (Chairs), Collaboralive
cog&ion:
Parenls
as guides
in
cognirilse dev@opmenr.
Symposium conducted at the meeting of the Society for Research
in Child Development, Toronto. Canada.
Dunn. J.. & Wooding, C. (1977). Play in the home and its implications for learning. In B.
Tizard & D. Harvey (Eds.). The biology q/play
Philadelphia: Lippincott.
Fein. G. (1981). Pretend play in childhood: An integrated review. Child Developmenl.
52.
1095-l 118.
Garvey. C. (1977). Play. Cambridge, MA: Harvard University Press.
Gottfried. A.W. (1985). Measures of socioeconomic status in child development research:
Data and recommendations. Merrill-Palmer
Quar~erlv.
31. 85-92.
Hrncir. E.J.. Speller, G.M.. & West, M. (1985). What are we testing? Developmental
Psychology. 21. 226-232.
Killen, M.. & Uigiris. I.C. (1981). Imitation of actions with objects: The role of social
meaning. The Journal
c$’ Genetic
Psvchology.
138. 2 19-229.
Lockman, J.J.. & McHale, J.P. (1985, April). Iqfanr and marernal
exploration
q/ objecrs.
Paper presented at the meeting of the Society for Research in Child Development.
Toronto, Canada.
Lowe, M. (1975). Trends in the development of representational play in infants from one
to three years: An observational study. Journal
q/Child
Psvchology
and Psychiarry,
16. 33-47.
Mandler. J. (1983). Representation. In J.H. Flavell & E.M. Markman (Eds.), P.H. Mussen
(Series Ed.), Handbook
q/child
psvchology:
Vol. 3. Cognilise
developrnenr.
New York:
Wiley.
McCall, R. (1979). Qualitative transitions in behavioral development in the first two years
of life. In M.H. Bornstein & W. Kessen (Eds.), Psychological
deve/opmennt,fio/n
infancy:
From
image lo inlention.
Hillsdale. NJ: Erlbaum.
McCall. R.. Parke. R.. & Kavanaugh, R. (1977). Imitation of live and televised models by
children one to three years of age. Monographs
L$ the Sociery
./or Research
in Child
Development.
42(5, Serial No. 173).
162
TAMIS-LEMONDA AND BORNSTEIN
McCune-Nicolich.
L. (1981). Toward symbolic functioning: Structure of early pretend games
and potential parallels with language. Child De~doprno~t.
52. 785-797.
McCune-Nicolich.
L.. & Fenson. L. (1984). Methodologicdl issues in studying pretend play.
In T.D. Yawkey & A.D. Pelligrini (Eds.). Child’s play: D~~vloptt~mta/
and applied.
Hillsdale. NJ: Erlbaum.
O’Connell. B.. & Bretherton. I. (1984). Toddler’s play alone and with mother: The role of
maternal guidance. In 1. Bretherton (Ed.). .S~vt~ho/ic p/a~r. Orlando, FL: Academic.
Piaget. J. (1962). P/q: drmtvs
and ittti/ariot~
irz childhood.
New York: Norton.
Power. T.G.. & Parke. R.D. (1982). Play as a context for early learning. In M. Laosa &
I.E. Sigel (Eds.). Fatnilies
as /earni,lg environtuetus
@r- children.
New York: Plenum.
Rogoff. B. ( 1990). .-lpprentice.ship
in thinking:
Cognitive
doeloptuent
in social contest.
New
York: Oxford University Press.
Rogoff. B.. Malkin. C.. & Gilbride. K. (1984). Interaction with babies as guidance in
development. In B. Rogoff & J.V. Wertsch (Eds.). Children’s
learning
iu the “Zone
q/ Prositnal
Developtnent.
” San Francisco: Jossey Bass.
Rogoff. B.. & Wertsch. J.V. (Eds.). (I 984). Childrett ‘s /earning
in the “Zone
q/ Pro.\~ittta/
Development.
” San Francisco: Jossey Bass.
Ruff. H.A.. & Lawson, K.R. (1990). Development of sustained. focused attention in young
children during free play. Devdoptuental
Psychology,
2. 85-93.
Shore. C.. O’Connell, B.. & Bates. E. (1984). First sentences in language and symbolic play.
De~eloptnental
Ps,vchology,
20. 872-880.
Slade. A. (1987). A longitudinal study of maternal involvement and symbolic play during
the toddler period. Child Development.
55. 367-375.
Tamis-LeMonda.
C.S.. & Bornstein. M.H. (I 989). Habituation and maternal encouragement
of attention in infancy as predictors of toddler language, play. and representational
competence. Child Drveloptnent.
60. 738-75 I.
Tamis-LeMonda.
C.S., & Bornstein. M.H. (1990). Language. play. and attention at one year.
Infant
Behavior
and
Developtnent.
13. 85-98.
Teti. D.M.. Bond. L.A.. & Gibbs. E.D. (1988). Mothers. fathers, and siblings: A comparison
of play styles and their influence upon infant cognitive level. Internarional
Journa/
ST Behavioral
Devrloptnenr.
I I. 4 15-432.
Tukey. J.W. (1977). Esp/orutot:~~
data an&is.
Menlo Park. CA: Addison-Wesley.
Ungerer. J.. & Sigman. M.D. (1984). The relation of play and sensorimotor behavior to
language in the second year. Child Developtmnt.
55. 1448-1455.
Ungerer. J.A.. Zelazo. P.R.. Kearsley. R.B.. & O’Leary. K. (1981). Developmental changes
in the representation of objects in symbolic play from I8 to 34 months of age. Child
Development.
52, l86- 195.
Vibbert. M.. & Bornstein, M.H. (1989). Specific associations between domains of mother-child
interaction and toddler referential language and pretense-play. In&r
Behasior
and
Dec~eloptnent.
12. I63- 184.
Vygotsky. L.S. (1962). Thought
and language.
Cambridge. MA: MIT Press.
Vygotsky. L.S. ( 1978). Mind in society: The developtmnr
q/ higher psychological
processes.
Cambridge. MA: Harvard University Press.
Watson. M., & Fisher. K. (I 977). A developmental sequence of agent use in late infancy.
Child
De~elopn~ent.
48.
828-836.
Zukow. P.G. (1986). The relationship between interaction with the caregiver and the emergence of play activities during the one-word period. British Journal
q/ Developtnental
Ps~~chology,
4. 223-234.
12 July 1989; Revised 26 October 1990
n