Copyright 1994 by the American Psychological Association, Inc.
0012-I649/94A3.00
Developmental Psychology
1994, Vol. 30, No. 2,283-292
Specificity in Mother-Toddler Language-Play Relations Across the
Second Year
Catherine S. Tamis-LeMonda and Marc H. Bornstein
Relations among specific aspects of language (comprehension and production, semantics and utterance length) and relations between language and symbolic play were evaluated when children were
13 and 20 months of age. The contributions of maternal stimulation to toddler performance and
whether associations among toddler abilities might be explained by maternal behaviors were also
examined. Although measures of toddler language covaried, language-play associations in toddlers
were specific to semantic aspects of language. Associations between mother and toddler behaviors
emerged and tended to be specific: Maternal language related to toddler language, and maternal
play related to toddler play. Moreover, relations among toddler abilities maintained after maternal
influences were partialed. The multidimensional structure of language and specificities in languageplay associations were discussed with reference to models of early representational development.
Developmental psycholinguists theorize that nonlinguistic
representations of the world contribute to, as well as constrain,
the structure of child language and the course of its development
(e.g., Bloom, 1973; Brown, 1973; de Villiers & de Villiers,
1992). This cognitive model of language acquisition evolved
from Piaget's (1962) claim that language emerges out of nonlinguistic sensorimotor intelligence and is one of several manifestations of a general underlying symbolic capacity. However, empirical investigation has not supported the strong cognitive
claim that sensorimotor intelligence is prerequisite to all aspects
of language learning (see Corrigan, 1978, 1979; Curtiss, 1989;
de Villiers & de Villiers, 1992, for reviews). For example, children's performance on Piagetian tasks of object permanence
does not relate to global measures of language (e.g., Bates, Benigni, Bretherton, Camaioni, & Volterra, 1979; Corrigan,
1978), whereas specific cognitive achievements (e.g., Piagetian
invisible displacement tasks) have been found to predict specific
language achievements (e.g., verbal expressions of disappearance or nonexistence; Tomasello & Farrar, 1984).
Accordingly, questions such as whether linguistic achieveCatherine S. Tamis-LeMonda was supported by a New 'York University Fellowship, a Helbein Scholarship, an Intramural Research Training Agreement Fellowship from the National Institute of Child Health
and Human Development (NICHD), and by Research Grants
HD20559, HD20807, and MH48915-02. Marc H. Bornstein was supported by Research Grants HD2O559 and HD20807 and by Research
Career Development Award HD00521 from the NICHD.
We thank L. Baumwell, L. Cyphers, K. Dine, R. Kalman-Kahana, J.
Hampson, J. McClure, A. Melstein-Damast, K. Nelson, A. O'Reilly, B.
Wright, and M. Zamor for comments and assistance. We also thank two
anonymous reviewers, whose careful attention to and insightful comments on an earlier version greatly improved the manuscript.
Correspondence concerning this article should be addressed to Catherine S. Tamis-LeMonda, New York University, Department of Applied
Psychology, 239 Greene Street, 5th Floor, New York, New York 10003,
or to Marc H. Bornstein, Child and Family Research, National Institute
of Child Health and Human Development, Building 31, Room B2B15,
9000 Rockville Pike, Bethesda, Maryland 20892.
283
ments are associated with developments in nonlinguistic domains have been replaced with questions such as whether there
are specific or identifiable linguistic-cognitive associations at
specific periods in development (see Gopnik & Meltzoff, 1984,
1987). This new theoretical orientation is especially relevant to
the study of toddlers' play and language. Both play and language
might be conceptualized as multidimensional. The nature of
their interrelations and relations to other mental functions in
the child will depend on the measure of language or play that is
used (see Tamis-LeMonda & Bornstein, 1993). That is, language-play associations are most likely specialized rather than
global.
In this longitudinal study, we revisit a cognitive model of early
language development by exploring specificities in languageplay associations when children were 13 and 20 months while
simultaneously considering mothers' referential language and
symbolic play with their toddlers. This study makes two unique
contributions to this area of research. First, the developmental
approach we take permits examination of concurrent and predictive language-play relations at two points in the second year.
Second, by considering mothers' concurrent and predictive influences in these same areas of development, we ask whether
mothers' interactions mediate identified language-play associations in toddlers. To date, the majority of studies that explore
nonlinguistic correlates of language have based assessments on
a single age period, and virtually none has considered mothers'
contributions to children's developments in these areas.
In these assessments, we adopt a multidimensional approach
to language at both ages, exploring how different aspects of language interrelate and how they differentially relate to symbolic
play. Past research has shown that different measures of language are associated with one another during the second year,
but reported relations are usually moderate in size. Therefore,
some aspects of language may share variance whereas others
may not, and some aspects of language may relate to play
whereas others may not (see Bates, Bretherton, & Snyder, 1988;
Bretherton & Bates, 1984; Tamis-LeMonda & Bornstein,
1990).
284
CATHERINE S. TAMIS-LEMONDA AND MARC H. BORNSTEIN
As a part of these investigations, we examined mothers' contributions to specific toddler abilities. American mothers encourage verbal development and play sophistication in their
children, so it is possible that links between child language and
play—whether concurrent or lagged, specific or general—are
influenced by maternal stimulation (Belsky, Goode, & Most,
1980; Bornstein, 1985; Carew, 1980; Fiese, 1990; HardyBrown, Plomin, & DeFries, 1981; Olson, Bates, & Bayles, 1984;
Tamis-LeMonda & Bornstein, 1989, 1990; Tamis-LeMonda,
Bornstein, Cyphers, Toda, & Ogino, 1992; Teti, Bond, & Gibbs,
1988).
We focused on second-year achievements in toddlers because
during this period children make extraordinary developmental
strides in these two important arenas of mental functioning. We
initially observed children at 13 months, because at this time
most children have produced their first words and are already
advanced in language comprehension. Children were seen again
at 20 months. By this time, (a) most children have demonstrated
a spurt in productive vocabulary; (b) children are substantially
regular in the general semantic meanings they use to express
possession, location, action, and so forth; and (c) many children
are producing theirfirstword combinations (Bates et al., 1988;
de Villiers & de Villiers, 1992). At both ages, these measures of
language are characterized by substantial variation, and these
individual differences appear to be valid indexes of later linguistic or cognitive functioning (Bates et al., 1988; Bates, O'Connell, & Shore, 1987; Braine, 1976; Slobin, 1982).
In parallel with this transition in language competencies,
second-year play progresses in sophistication as it moves from
exploratory manipulation toward expressions of pretense in
which children enact experiences and events through symbolic
gesture. Before thefirstyear, sensorimotor acts predominate in
children's play activity; children examine objects to produce
effects and to explore their concrete functions (e.g., pressing
buttons on a toy telephone). Around 13 months, incipient symbolic acts appear as children use gesture in representation, for
instance, pretending to talk on a toy telephone or to drink from
an empty teacup. Over the next several months, certain concrete, forms of play diminish in frequency and are replaced by
more cognitively based pretense actions. By 20 months, most
children are capable of a diversity of symbolic activities that
include incorporating others in pretense scenarios (e.g., putting
dolls to sleep) as well as sequencing acts of pretense (e.g., stirring
in empty teacups and then pretending to drink; see Belsky et
al., 1980; Belsky & Most, 1981; Bornstein & O'Reilly, 1993;
Fein, 1981; McCune, 1993; Tamis-LeMonda & Bornstein,
1991, 1993).
With respect to these developments in language and play, our
study had two main goals. The first concerned exploration of
the differentiated nature of language-play associations. How do
different aspects of language relate to symbolic play across the
second year? We hypothesized that specific correspondences
would exist between semantic aspects of language and symbolic
play. Piaget (1962) and Vygotsky (1967) both opined that representational thinking does not so much concern the development
of speech per se as much as the child's use and understanding of
meaning.
We also hypothesized that at 13 months children's language
comprehension would predict symbolic play better than would
their language production. At 13 months language production
is limited. It has been speculated that variability among children in productive ability at this time indexes a propensity toward talking (e.g., some children wait to talk until they feel they
can "get it right"; Bates, Thai, Fenson, Whitesell, & Oakes,
1989) or phonological-articulatory abilities (e.g., Huttenlocher,
1974), rather than underlying representational ability. In support of this notion, prior research in our laboratory uncovered
specific relations between children'sflexible(context-free) language comprehension and their symbolic play at 13 months
(Tamis-LeMonda & Bornstein, 1990; see also Vibbert &
Bornstein, 1989). The present study is based on subsequent
evaluation of the same children and their mothers when the toddlers reached 20 months of age, and this report includes an evaluation of toddlers' 13-month language and play in relation to
their 20-month language and play.
By 20 months, we expected measures of productive language
to relate systematically to measures of symbolic play, but again
to do so in a specific way. On the basis of the hypothesis that
language and symbolic play partly reflect children's ability to
conceptualize and represent experience, we hypothesized that
semantic diversity, the variety of meaning categories children
express (e.g., possession, agency, and location), would be a
stronger correlate of symbolic play than vocabulary size or utterance length.
Furthermore, we expected to identify moderate stabilities in
toddler behaviors over time. We anticipated that 13-month language comprehension would predict semantics in language at
20 months and that 13-month language production would predict productive vocabulary size and utterance length at 20
months. These predictions are based on the hypothesis that stabilities in language will exist to the extent that common underlying processes are tapped at both ages. If language comprehension is considered an index of linguistic representation at 13
months, it ought to predict the expression of meaning later. In
contrast, Bates et al. (1988) found that total production at early
stages of language (e.g., 13 months), rather than comprehension, was a stronger predictor of mean length of utterance
(MLU) at 20 months. We also expected to observe predictive
relations from language to play and from play to language between 13 and 20 months, again with particular attention to
those aspects of language that are rooted in meaning (i.e., language comprehension at 13 months and semantics in language
at 20 months were expected to relate to symbolic play at both
ages).
The second main goal in this study was to examine maternal
contributions to toddlers' language and play. Most researchers
who have explored links between language and play in early
childhood have ignored the potential role of environmental factors, most notably mothers' own interactive language and play
(but see Bornstein; Vibbert, Tal, & O'Donnell, 1992; Vibbert &
Bornstein, 1989). Variation among children in language and in
play, as well as relations between the two abilities, might be explained by exogenous factors, perhaps maternal verbal stimulation or symbolic representations. To assess mothers' contributions to toddlers' language and play, as well as their contributions to associations among language and play measures within
toddlers, we differentiated two facets of maternal activity to parallel our foci on toddler language and play. In the language do-
LANGUAGE AND PLAY
main, we assessed mothers' referential language because this
form of stimulation is associated with various measures of child
language, including number of nouns and non-nouns in comprehension, MLU, noun phrases, and verbs per utterance (e.g.,
Furrow, Nelson, & Benedict, 1979; Nelson, 1973; Newport,
Gleitman, &Gleitman, 1977;Tamis-LeMondaetal., 1992; Tomasello & Todd, 1983). In the play domain, we also assessed
mothers' demonstrating and soliciting specific toddler play actions. Measures of maternal play are consistently valid predictors of children's play sophistication across populations and cultures (e.g., Belsky et al., 1980; DeLoache & Plaetzer, 1985;
Lockman&McHale, 1985; Power &Parke, l982;Rogoff, 1990;
Slade, 1987; Tamis-LeMonda & Bornstein, 1993; TamisLeMonda et al., 1992). By differentiating maternal language
stimulation from maternal play stimulation, we sought to examine specificities in mother-toddler associations within and
across the second year and to explore whether maternal behaviors mediate language-language and language-play associations
in toddlers.
285
Toddler Measures at 13 and 20 Months
Thirteen-month language. We interviewed mothers about their children's productive and receptive vocabularies using the Bates et al.
(1988) language inventory. We elected to examine early language with
this inventory because it distinguishesflexiblefrom restricted aspects of
child language. During the language interview, the experimenter read
approximately 100 common words or phrases to the mother and asked
whether her child demonstrated productive or receptive proficiency
with each. If a mother indicated that her child produced or understood
an item or both, the experimenter further probed to determine whether
any specific gestural, vocal, spatial, or temporal cues were required for
the child to display competence with the word or phrase. Language data
at 13 months were scored according to procedures developed by Bates
et al. (1988) and revised and detailed in Tamis-LeMonda and Bornstein
(1990). In summary, children were credited with production if they
demonstrated consistent and appropriate use of a sound that approached the phonetic version of the adult target word (e.g., baba for
bottle), and they were credited with comprehension if their behavioral
response was appropriate to a particular word or phrase (e.g., toddling
over to mother if she said "Come here"). Production and comprehension were classified as restricted when contextual or other cues were
necessary; items that were independent of contextual cues were classified as flexible (see Snyder, Bates, & Bretherton, 1981). Analyses here
Method
focus on flexible measures of language production and comprehension.
Two coders independently scored audiorecordings and transcripts of
Sample
10% of the language interviews, randomly selected, so that coding reliability could be assessed. Agreement for numbers of words and phrases
Forty-one toddlers (20 boys and 21 girls) and their mothers particiin production and in comprehension was calculated by dividing the
pated in two home observations scheduled 7 months apart. Subjects
number of agreements by the number of agreements plus diswere recruited from private obstetric and pediatric groups in New %rk
City. Toddlers averaged 13 months 1 week (range = 13 months 1 day to agreements. Reliability averaged 97% for production (range = 86%100%) and 92% for comprehension (range = 83%-98%); agreement for
13 months 2 weeks) at the time of thefirstobservation and 20 months 2
the flexible-restricted language distinction averaged 92% for production
weeks (range = 20 months 1 week to 20 months 3 weeks) at the time of
(range = 83%-100%) and 90% for comprehension (range = 83%-98%).
the second observation. Their mothers averaged 34 years of age (range
= 28 to 42). Subjects were from middle-socioeconomic-status houseTwenty-month language. Three aspects of children's language were
holds (M = 60 on the Hollingshead [ 1975] Four-Factor Index; Gottfried,
assessed. Lexical development reflected the magnitude of children's ab1985), and the majority of mothers (92%) had completed 4 years of
solute productive vocabulary size; utterance length reflected children's
college. Four additional toddlers were seen at the two ages, but their data
use of word combinations, plurals, inflections, auxiliaries, articles,
were eliminated from final analyses because they were univariate and
prepositions, and so forth; and semantics reflected the diversity of
bivariate outliers on language measures (Tukey, 1977). Two had extreme
meanings children express in speech. Utterance length and semantics
high scores (i.e., greater than three standard deviations beyond the samwere determined from transcripts of all child and mother utterances
ple mean) on receptive language at 13 months, one had extreme high
made during the play sessions. All transcripts were coded twice by indescores on productive language at both ages, and one had extreme high
pendent coders and were cross-checked a third time by a different conscores on productive language at 20 months.
stant coder. Any utterances that were unintelligible to coders were not
counted, and utterances whose content coders could not agree on were
also omitted from analyses. Agreements among coders were calculated
Procedure
separately for mother and toddler language and for words and phrases.
Percentage agreement averaged between 90% and 99% for all categories.
Home visits were scheduled during times when mothers felt their todProductive vocabulary size was estimated by using the Bates et al.
dlers would be alert and rested. At both ages, dyads were videotaped for
(1988) vocabulary checklist and was collected at the conclusion of the
15 min of free play, and mothers were interviewed about their children's
20-month visit. This checklist includes nearly 700 words taken from
language. Mothers were asked to remain with their toddlers and to disgeneral categories, such as animals, clothes, actions, and foods.
regard the experimenter's presence as much as possible. They were instructed to do whatever they would ordinarily do when their toddlers
Utterance length was based on spontaneous utterances coded for morengage in play. No further instructions were given. At both ages, a set of
pheme length following traditional guidelines (e.g., Bates et al., 1988;
toys was placed on the floor in front of toddler and mother. These toys
Bloom, Beckwith, & Capatides, 1989; Bloom & Capatides, 1987;
included a teapot and cover, a clown doll, spoons, cups and saucers, a
Brown, 1973). Only children's spontaneous utterances were coded, betoy telephone, a book, a ball, blocks, a nesting set, and a toy vehicle.
cause imitative and responsive utterances might alter estimates of children's ability, for example, in the case of children whose mothers ask
Data on children's language derived from both interviews and sponmany yes-no questions. Mean length of longest utterances (MLLU),
taneous speech in the play sessions. Children's play competence, materwhich is based on the average morpheme length of the child'sfivelongest
nal language stimulation, and maternal play stimulation were coded
spontaneous utterances, served as the index of utterance length.
from the play sessions. Toddlers' language and play and mothers' language and play were evaluated by different coders at each of the two ages
Semantic diversity was also based on children's spontaneous utterso that coders of each data set were unaware of the nature of all other
ances exhibited during the play session. Each utterance was classified
data sets.
into 1 of 14 semantic categories (e.g., agent, action, possession, recipi-
286
CATHERINE S. TAMIS-LEMONDA AND MARC H. BORNSTEIN
ent, or locative).' These semantic categories derive from existing studies
of semantics in second-year language and included categories signifying
functional relations (e.g., recurrence) as well as grammatical relations
(e.g., agent; see Bates et al, 1988; Bloom, 1970; Brown, 1973; Nelson,
1973, 1981; Schlesinger, 1971). In assessing semantics, coders watched
the videotaped session and used contextual cues to interpret the function of children's utterances. In those cases in which an utterance expressed two semantic categories, the child was credited with both. Each
child received a semantic diversity score that reflected the number of
different semantic categories used; the maximum score a child could
achieve was 14. Categories of speech that were not included in this code
consisted of routines, objections, words or phrases expressed with no
obvious or immediate contextual referents or meaning, as well as those
expressed solely to service symbolic play, for example, saying "Hi!" into
the toy telephone. (These latter utterances were not considered so as to
avoid confounding of the language-play correlations.) Interrater reliabilities of semantic classifications were calculated by having three independent raters code 15% of the sample, randomly selected. Kappa
reliability across semantic categories averaged .88.
Thirteen- and 20-month play. Toddler play in the 15-min videotape
was coded over 60 contiguous 15-s intervals. During each interval, the
coder noted which of eight levels of play the child achieved. The eight
levels were defined as follows: (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 (see the Appendix for operational definitions and examples). During a given interval, a child could be credited with more than one different level of
play.
All categorizations of toddlers' play actions, including symbolic play,
were determined by children's actions and nonverbal behaviors rather
than by their language. Thus, a child could look into a cup and say "hot"
but would not be credited with symbolic play; alternatively, if the child
tipped the cup to his or her lips pretending to drink, he or she would be
credited with symbolic play. Basing judgment of play on action rather
than language obviated inflating language-play correlations.
The total number of different play levels (i.e., Levels 1 through 8)
achieved was calculated within each interval; these totals were then
summed across the 60 intervals. Similarly, the number of different symbolic play levels (i.e., Levels 5 through 8) was calculated within each
interval and then summed across the 60 intervals. A comprehensive
measure of symbolic play was calculated by dividing the symbolic sum
by the total sum. In these calculations, only toddlers' spontaneous symbolic play was considered (i.e., play acts in which the child exhibited a
symbolic activity that was neither demonstrated nor solicited by the
mother in the same coding interval or the preceding interval; see also
Tamis-LeMonda & Bornstein, 1989,1990,1991). This was done so that
toddlers' play scores were not confounded with mothers' immediate
symbolic play. Moreover, by considering toddlers' spontaneous play
only, assessments paralleled those for toddlers' language that were based
on spontaneous utterances (see earlier). These percentages were arc sin
transformed for statistical analyses. Intercoder reliabilities for toddler
play were obtained by having independent coders score toddler play for
24% of the sample, randomly selected. For each subject scored, observers were said to agree when they both coded the same level(s) of play for
a given interval. Kappa agreement for the eight levels of play across
intervals and subjects averaged .77.
Maternal Measures at 13 and 20 Months
Referential language. When infants were at both 13 and 20 months,
mothers' referential language was coded from the 15-min free-play session. During each of 60 successive 15-s intervals, the coder noted
whether the mother verbally referred to an object, activity, or event in
the environment by describing, labeling, or asking her toddler about the
unique qualities of the referent (e.g., "That's a spoon" or "What color
is the spoon?"). Comments that were based on actions with the provided
toys (e.g., "You're drinking," to a child pretending to drink from the toy
cup) were also counted as referential language. Thus, over the entire 15min play session mothers could receive a maximum score of 60.
Although mothers did much talking during the sessions, referential
language was distinct from sheer talkativeness. There were many utterance types that were excluded from this category. For example, comments or questions about nonpresent events, people, or objects (e.g.,
"Daddy will be home soon" or "Do you remember going to the zoo
yesterday?"), affect (e.g., "You're happy" or "Mommy likes to see you
play"), non-object-oriented actions (e.g., "You're waving"), praise (e.g.,
"Good!"), commands (e.g., "Sit down"), reprimands (e.g., "Don't do
that"), affirmations and exclamations (e.g., "Yes!"), and play solicitations or demonstrations (e.g., "Feed the doll" or "Mommy is feeding the
doll"; see later) were not counted as referential language. Intercoder
reliabilities for referential language were obtained by having independent observers code 17% of the sample, randomly selected. For each
subject scored, observers were said to agree when they both coded referential language in the same interval. Kappa agreement across intervals averaged .73.
Play. During each 15-s interval, coders noted whether mothers exhibited a play action by modeling specific play activities for their toddlers or by prompting toddlers to engage in specific play activities themselves. Each maternal play act was also coded for its level of sophistication with the same play scale as that developed for toddlers. As for
toddlers, one comprehensive measure of mothers' symbolic play was
calculated by summing across the frequencies of intervals for Levels 5
through 8 and dividing by the total frequency obtained for Levels 1
through 8; these percentages were arc sin transformed before analyses.
Intercoder reliabilities for mother play were obtained by having independent coders score mother play for 17% of the sample. Kappa
agreement across the eight levels of play for matched intervals averaged
.86.
Results and Discussion
We first report descriptive statistics for toddler and mother
language and play at the two ages. We next present mothertoddler language-play associations at 13 and at 20 months, followed by predictive mother-to-toddler and toddler-to-mother
language-play associations between the two ages. We then report concurrent relations among toddler measures at 13 and at
20 months and predictive associations between toddler 13- and
20-month performance. Finally, we reevaluate within-toddler
associations after partialing maternal stimulation to assess
whether language-play links within toddler are explained (in
part) by maternal stimulation.
Because measures of language production appeared skewed
at the two ages, square-root and logarithmic transforms of production variables were inspected where appropriate. Results
did not differ from untransformed data; reported results are
based on untransformed data.
Descriptive Statistics and Mother-Toddler Correlations
at 13 and 20 Months
Table 1 presents descriptive statistics (means, standard deviations, and ranges) for toddler and mother measures, and Table
1
A comprehensive list of the semantic categories coded along with
their operational definitions can be obtained from Catherine S. TamisLeMonda or Marc H. Bornstein.
287
LANGUAGE AND PLAY
Table 1
Descriptive Statistics of Toddler and Mother
Variables at 13 and at 20 Months
Variable
13 months
Toddler
Receptive vocabulary
Productive vocabulary
% symbolic play*
Mother
Referential language
% symbolic play*
20 months
Toddler
Productive vocabulary size
Mean length longest utterance
Semantic diversity
% symbolic play*
Mother
Referential language
% symbolic play"
M
SD
Range
38.7
6.2
10.0
18.8
6.0
9.0
10-75
0-27
0-32
29.6
35.0
10.1
15.0
8-48
9-78
162.1
2.6
7.7
24.0
116.8
1.0
2.9
18.0
31.2
46.0
9.7
20.0
8-434
1-5.4
1-12
2-70
9-53
14-96
* Nontransformed percentages; percentages were arcsin transformed in
statistical analyses.
2 presents concurrent and predictive correlations between
mother and toddler measures. Toddler and mother measures
showed substantial variation at both ages. Across age, toddlers
changed dramatically in their language as well as symbolic play,
ft(40) = 8.81 and 6.00, ps < .001, respectively. For example,
toddlers' symbolic play increased nearly 250%. In contrast,
mothers showed little change in their referential language, f(40)
= 1.03, but a modest 33% increase in their symbolic play, #40)
= 3.31, p < .01. For both language and play, toddlers showed
relatively greater change over age than did their mothers, as indicated by strong partner by age interactions, Fs(l, 40) = 8.81
and 7.74, ps < .001. Although mothers' referential and symbolic
play interactions may facilitate toddlers' development in play
and language, progress in these domains must also be influenced
by a variety of other factors.
Table 2 shows that in half of the correlations, variability in
children's abilities was associated with variability in mothers'
stimulation, and specificity in mother-toddler associations was
apparent at both ages. Table 2 (top left) shows 13-month concurrent correlations. Toddlers' receptive and productive vocabularies were significantly associated with their mothers' referential language, and, similarly, toddlers' symbolic play was significantly associated with their mothers' symbolic play.
Mother-toddler language-play cross-correlations were not significant. Thus, in both domains, it was not absolute amount of
maternal stimulation per se that predicted language or play in
toddlers, but instead the nature of that stimulation that was telling (e.g., mothers who talked more about objects did not have
children who played more, but mothers who specifically
prompted play did have children who played more). Not shown
is the finding that mothers' referential language was independent of mothers' symbolic play (r = -.07).
Table 2 (bottom right) shows 20-month concurrent correlations. Variation among toddlers in productive vocabulary size
and semantic diversity was associated with their mothers' referential language, and toddlers' symbolic play again related to
their mothers' symbolic play. Mothers' symbolic play also related to semantic diversity in toddler speech. Relations among
mothers' referential language and toddlers' MLLU and play, as
well as toddlers' vocabulary and MLLU and mothers' play, were
not significant. In mothers, referential language related significantly to their symbolic play (r = .28, p < .05).
Table 2 (top right) shows predictive mother-to-toddler relations. Mothers who referred to objects more at 13 months had
children with larger productive vocabularies, higher MLLU,
and greater semantic diversity at 20 months.2 In contrast, mothers' earlier symbolic play did not predict their toddlers' later
2
We analyzed mothers' referential language with respect to noun versus non-nouns in children's comprehension and production at the two
ages. (In this article, we report language totals.) There was no systematic
prediction to noun classes only. For example, mothers' referential language at 13 months relates to common nouns in toddler comprehension
.40; non-nouns, .43; and total comprehension, .35, and predicts 20month noun production .53 and non-noun production .44.
Table 2
Concurrent and Predictive Relations Between Mothers' Stimulation and Toddlers' Performance at 13 and 20 Months
Toddler 20 months
Toddler 13 months
Mother
Receptive
vocabulary
Productive
vocabulary
Symbolic
play
Productive
vocabulary
.35*
.09
.32*
-.15
-.02
.30*
.50***
-.30*
Toddler-to-mother relations
20 months
Referential language
Symbolic play
.13
.32*
Note. MLLU = mean length of longest utterance.
*p<.05. **/;<.01. ***/><.001.
.06
.40**
Semantic
diversity
Symbolic
play
Mother-to-toddler relations
13-month concurrent correlations
13 months
Referential language
Symbolic play
MLLU
,41**
.00
.44**
.01
.05
.17
20-month concurrent correlations
.08
.33*
.29*
.21
.08
.08
.30*
.40**
.04
.64***
288
CATHERINE S. TAMIS-LEMONDA AND MARC H. BORNSTEIN
symbolic play performance. A curious and unexpected finding
is that mothers' symbolic play at 13 months showed an inverse
relation to toddlers' later productive vocabulary size. One possible explanation is that at 13 months, high levels of symbolic
play, in which objects are vehicles of representation (i.e., they
are principally used to reenact past experiences and to stand for
things they often are not), may confuse children. In contrast, the
relation between mothers' early referential language, in which
things are more concretely labeled for what they are, and toddlers' later vocabulary, is .50.
Table 2 (bottom left) shows toddler-to-mother predictive correlations. Neither receptive nor productive vocabulary nor symbolic play in toddlers at 13 months predicted mothers' later referential language. In contrast, all three early toddler measures
predicted mothers' later symbolic play: Specifically, toddlers
with greater receptive and productive vocabularies, and who
had more symbolic play at 13 months, had mothers who engaged them in more symbolic play at 20 months.
Two points might be made about these toddler-to-mother associations. First, at a broad level, they suggest reciprocal influences of dyad members on one another. Although these data are
correlational and do not directly test causal influences, mothers
and children appear to influence each others' behaviors both in
the immediate context as well as over time. Second, the results
highlight the specific nature of toddlers' influence. Toddlers predict mothers' symbolic play, rather than her referential language. The toddler-to-mother prediction may be attributed to a
context-appropriate adjustment of mothers at 20 months to
their relatively precocious 13-month-olds. That is, by 20
months, a more sophisticated pattern of interaction in a play
session might be for a mother to engage her toddler in symbolic
themes that extend the limits of reality. Such engagements
would be reflected in high numbers of symbolic demonstrations
and solicitations by mothers but would not be reflected in high
levels of labeling and questioning about present toys (referential
language). Thus, at 20 months a more advanced interactive style
would be one that uses toys to service symbolic play (e.g., "Let's
have a tea-party") rather than one in which toys are the topic of
discussion (e.g., "That's a cup"). The outcome would be that
children at 13 months, who are already advanced in language
and play, may lead mothers to later prompt and maintain pretense-based exchanges.
Within-Toddler Correlations at 13 and 20 Months
Figure 1 shows a correllelogram of within-toddler associations at each age and across the two ages. Both zero-order correlations among the toddler variables and partial correlations
among variables after covarying mothers' stimulation are
shown (partial re appear in parentheses). For partial correlations, mothers' 13- and 20-month referential language and
symbolic play were partialed from toddler language and toddler
symbolic play at each age, respectively. Specifically, languagelanguage correlations partialed mothers' language at the associated age(s); play-play correlations partialed mothers' play at the
associated age(s); and language-play correlations partialed both
mothers' language and play at the associated age(s). Hence, the
number of maternal covariates ranged from one (when relating
child performance within age and within domain) to four (when
relating child performance across age and across domain).
Zero-order within-toddler correlations. At 13 months, receptive vocabulary related significantly to productive vocabulary and to symbolic play; language production did not relate to
symbolic play (Tamis-LeMonda & Bornstein, 1990). Interrelations among the language variables at 20 months indicated
that productive vocabulary, MLLU, and semantic diversity all
covaried positively. As hypothesized, only semantic diversity related to symbolic play at 20 months (see Figure 1).
Predictively, toddlers' productive vocabulary size, but not
play, was stable between 13 and 20 months. Toddlers' 13-month
receptive vocabulary predicted semantic diversity at 20 months;
productive vocabulary predicted MLLU, semantic diversity,
and symbolic play; and children's symbolic play predicted only
their semantic diversity. Given that 13-month language comprehension and production and 20-month productive vocabulary
size are based on maternal report, concurrent and predictive
relations between these aspects of language at 13 and 20 months
and observed language and play at these two ages suggest that
report measures may provide accurate estimates of children's
early language proficiency (see also Bates et al., 1988). Neither
early receptive vocabulary nor symbolic play predicted productive vocabulary nor MLLU at 20 months, nor did early receptive vocabulary predict later play.
Partial within-toddler correlations. It would be premature
to characterize stability or predictive validity of toddler measures as reflecting endogenous processes without considering
the potential role of mother. First, concurrent associations between toddler and mother measures were generally significant
(Table 2, top left and bottom right). Second, mothers' early referential language predicted all three dimensions of children's
later language and all toddler measures predicted mothers' later
play (Table 2, bottom left and top right). Moreover, mothers
demonstrated significant stability in their referential language
and symbolic play over this 7-month period (rs = .32 and .49,
ps < .05 and .001, respectively). Thus, it is possible that lagged
associations in toddler competencies could be ascribable to maternal stimulation. Accordingly, we reanalyzed concurrent as
well as predictive within-toddler associations after removing the
correlative contribution of maternal stimulation.
As indicated in Figure 1, both within and across the two ages,
nearly all the zero-order correlations among toddler variables
remained significant after mothers' contributions were taken
into consideration. An exception to this was the relation between toddlers' 13-month productive vocabulary size and toddlers' 20-month play. The zero-order correlation of .42 was attenuated to a nonsignificant .24 after partialing mothers' earlier
and later contributions.
Conclusion
In the present study, we examined associations between features of language and play, as well as their stability and predictive validity, during the second year. We distinguished several
aspects of early language, such as comprehension and production, semantics and utterance length, to explore specificity in
concurrent and predictive language-play relations. Our findings
lend support to a modified cognitive model of language repre-
289
LANGUAGE AND PLAY
Receptive vocabulary
.33*
(•34*)
.40"
{.33')
.59"* (.54"
Productive vocabulary
.42"
(.24)
.10
(.16)
Symbolic play
13 months
•Symbolic play
20 months
Figure 1. Correllogram of zero-order and partial correlations (presented in parentheses) among toddler
measures at and between 13 and 20 months. Partial correlation coefficients represent relations among
toddler competences over and above maternal contributions. For these, language-language correlations
partialed mothers' language at the associated age(s); play-play correlations partialed mothers' play at the
associated age(s); and language-play correlations partialed both mothers' language and play at the associated age(s).
sentation in which specific linguistic abilities relate to specific
cognitive attainments at specific periods in development. The
nature of the language-play associations we identified suggests
that relations between linguistic and other cognitive achievements will exist when common principles and structures are
required (e.g., Corrigan, 1978,1979;Curtiss, 1989). Finally, the
present study extended prior research in the area of toddler play
and language by documenting specificities in mother-toddler
relations as well as continued specificity in associations among
toddler competencies after mothers' contributions to play and
language were considered.
In summary, ourfindingsshow that toddler language ability
does not generally predict toddler play; instead, only certain aspects of language relate to play. At 13 months, language comprehension, rather than production, related to symbolic play; at
20 months, only semantic diversity in language related to symbolic play. Moreover, 13-month symbolic play predicted semantic diversity in language at 20 months. These findings indicate
that symbolic play is associated specifically with aspects of lan-
guage that concern children's use of meaning as opposed to children's sheer productive quantity or utterance length (Piaget,
1962; Vygotsky, 1967). It could be that conceptual and structural aspects of language are functionally independent when relations to nonlinguistic domains of intelligence are examined.
In accord with this perspective, Curtiss (1989) presented data
from the case studies of two groups of individuals with selective
language impairments. She first described case studies of linguistic isolates, in which semantic and conceptual features of
language are found to be intact but syntactic competencies are
impaired. These individuals are contrasted with persons who
show pervasive cognitive deficits and severe impairments in semantic features of language yet are capable of fully elaborated
morphological and syntactic language structures.
A question arises as to why semantic diversity in particular
should relate to symbolic play. Semantic diversity largely captures children's use of functional and grammatical relations.
As noted by Bloom (1973), semantic categories depend on and
derive from meaning relations between words. Children do not
290
CATHERINE S. TAMIS-LEMONDA AND MARC H. BORNSTEIN
have a cognitive notion of agent or object of action, but rather,
these categories represent relational experiences such as things
doing things to other things and things changing with respect to
other things. These word usages differ from the performative use
of words or phrases (e.g., bye-bye, clap hands; Bates et al.,
1988), which make reference to classes of objects and events
based on formulae and are highly dependent on the gestural
routines that accompany them, as they do from substantive
word forms (e.g., labels; Bloom, 1973), which make reference to
classes of objects and events on the basis of perceptual features.
Symbolic play, like semantic categories, shares this feature of
relational meaning. Pretense activities are enactments of relations among agents, actions, and objects; like semantic language, symbolic play reflects children's construction of predications about such objects and events.
The fact that semantic diversity, but not productive vocabulary size, relates to symbolic play further supports the idea that
different linguistic-cognitive abilities are tapped by the different
measures of language. Further analysis of the language-play relations shows that semantic diversity continues to be predicted
by 13-month symbolic play and to be concurrently related to
20-month symbolic play after controlling for productive vocabulary size (rs = .46 and .45, ps < .01, respectively). This suggests
that semantic diversity will continue to relate to symbolic play
in the presence of either high or low productive vocabulary. In
fact, a post hoc cluster analysis of toddlers on semantic diversity
and symbolic play measures identified four groups of children.
One group was low on all language and play measures; a second
group was average on all measures; and the third and fourth
groups scored high on semantic diversity and high on symbolic
play. However, what distinguished the latter two groups was
their productive vocabulary sizes. Group 3 had high productive
vocabularies at both 13 and 20 months, whereas Group 4 had
relatively low productive vocabularies at the two ages. The results of this analysis suggest that children might be high on the
representational abilities shared by symbolic play and semantics, independent of the size of their lexical vocabularies. In
short, talking about the names of a lot of things may be different
from talking about how things relate to one another.
In contrast to the significant symbolic play-to-language crossage correlation, early language did not predict later symbolic
play. One possible interpretation of this asymmetry is that certain nonlinguistic attainments of representation (e.g., symbolic
play) are necessary prerequisites (though not sufficient) to the
attainment of certain linguistic achievements (e.g., semantic diversity; see also Bates et al., 1979; Veneziano, 1981). Alternatively, it may be easier for children to express themselves
through gesture than through language at early stages of symbolic emergence once requisite cognitive skills are in place
(Goodwyn & Acredolo, 1993).
Considering maternal contributions to children's performance in play and in language (Bornstein et al., 1992; Vibbert &
Bornstein, 1989), results point to the existence of specific
mother-toddler relations, and they lend support to a model of
environmental specificity (Bornstein, 1989; Wachs, 1987). In
this model, child cognition and environmental influence are
each characterized as multidimensional. Experience does not
influence child abilities globally, but rather certain experiences
operate selectively in relation to certain child abilities. We found
that mothers' language tends to covary with toddlers' language,
mothers' play tends to covary with toddlers' play within age,
and mothers' language predicts toddlers' language across age.
However, most relations between toddler language and play also
maintained over and above the significant contributions of select maternal activities. This suggests that the specific languagecognition relations in children that were identified here are not
solely mediated by maternal language and play. This approach
provides key support for the hypothesis that variation among
children in emerging aspects of language and symbolic play
might reflect variation in an underlying capacity to understand
and express experiences in several modalities, through language
and through representational action.
Note that the language-play relations observed in this study
at and between 13 and 20 months might apply specifically to
the procedures and measures we selected. The variables chosen
for study here constitute only a sampling of the range of possible
language, play, and stimulation indexes available from toddler
and mother. Consider the present assessment of symbolic play.
Symbolic play included reality-based pretense (e.g., pretending
to feed doll with a spoon) and substitution pretense (e.g., pretending to feed doll with a stick) as well as both single and sequenced acts of pretense. A different definition of symbolic play
might show different relations to language. For example, Shore,
O'Connell, and Bates (1984) observed no significant relations
between MLU and symbolic play, when symbolic play collapsed
across reality and substitution play. However, when they further
analyzed symbolic play into levels of abstraction, significant associations between sequence length in substitution play and
MLU in language emerged. This result led Shore et al. to conclude that the transition to multiword speech involves information-processing changes that transcend language proper, for example, increased memory or combinatorial abilities. Our data
lend further support to such a cognitive model of early language
acquisition.
References
Bates, E., Benigni, L., Bretherton, I., Camaioni, L., & Volterra, V.
(1979). The emergence ofsymbols: Cognition and communication in
infancy. San Diego, CA: Academic Press.
Bates, E., Bretherton, I., & Snyder, L. (1988). Fromfirstwords to grammar: Individual differences and dissociable mechanisms. Cambridge,
England: Cambridge University Press.
Bates, E., O'Connell, B., & Shore, C. (1987). Language and communication in infancy. In J. D. Osofsky, (Ed.), Handbook ofinfant development (2nd ed., pp. 149-204). New York: Wiley.
Bates, E., Thai, D., Fenson, L., Whitesell, K., & Oakes, L. (1989). Integrating language and gesture in infancy. Developmental Psychology,
25, 1004-1019.
Belsky, J., Goode, M. K.., & Most, R. K. (1980). Maternal stimulation
and infant exploratory competence: Cross-sectional, correlational,
and experimental analyses. Child Development, 51, 1163-1178.
Belsky, J., & Most, R. K. (1981). From exploration to play: A crosssectional study of infant free play behavior. Developmental Psychology, 77,630-639.
Bloom, L. (1970). Language development: Form andfunction in emerging grammars. Cambridge, MA: MIT Press.
Bloom, L. (1973). One word at a lime: The use ofsingle word utterances
before syntax. The Hague, The Netherlands: Mouton.
Bloom, L., Beckwith, R., & Capatides, J. B. (1989). Developments in
LANGUAGE AND PLAY
291
the expression of affect. Infant Behavior and Development, 11, 169- Hardy-Brown, K., Plomin, R., & DeFries, J. C. (1981). Genetic and
environmental influences on the rate of communicative development
186.
in thefirstyear of life. Developmental Psychology, 17, 704-717.
Bloom, L., & Capatides, J. B. (1987). Expression of affect and the emerHollingshead,
A. B. (1975). The four-factor index of social status. Ungence of language. Child Development, 58, 15-22.
published manuscript, Yale University.
Bornstein, M. H. (1985). How infant and mother jointly contribute to
Huttenlocher, J. (1974). The origins of language comprehension. In R.
developing cognitive competence in the child. Proceedings ofthe NaSolso (Ed.), Theories in cognitive psychology (pp. 331-368). New
tional Academy of Sciences, 82, 7470-7473.
York: Halsted.
Bornstein, M. H. (1989). Between caretakers and their young: Two
Lockman, J. J., & McHale, J. P. (1985, April). Infant and maternal
modes of interaction and their consequences for cognitive growth. In
exploration ofobjects. Paper presented at the meetings of the Society
M. H. Bornstein & J. S. Bruner (Eds.), Interaction in human developfor Research in Child Development, Toronto, Ontario, Canada.
ment (pp. 197-214). Hillsdale, NJ: Erlbaum.
Bornstein, M. H., & O'Reilly, A. (Eds.). (1993). New directionsfor child McCune, L. (1993). The development of play as the development of
development: The role ofplay in the development of thought (Vol. 59). consciousness. In M. H. Bornstein & A. O'Reilly, (Eds.), New directions for child development: The role of play in the development of
San Francisco: Jossey-Bass.
thought (Vol. 59, pp. 67-79). San Francisco: Jossey-Bass.
Bornstein, M. H., Vibbert, M., Tal, J., & O'Donnell, K. (1992). Toddler
Nelson, K. (1973). Structure and strategy in learning to talk. Monolanguage and play in the second year: Stability, covariation, and ingraphs of the Society for Research in Child Development, 38(1-2, Sefluences of parenting. First Language, 12, 323-338.
rial No. 149).
Braine, M. D. S. (1976). Children's first word combinations. Monographs ofthe Society for Research in Child Development, 4/(1, Serial Nelson, K. (1981). Individual differences in language development: Implication for development and language. Developmental Psychology,
No. 164).
17, 170-187.
Bretherton, I., & Bates, E. (1984). The development of representation
Newport, E. L., Gleitman, H., & Gleitman, L. R. (1977). Mother, I'd
from 10 to 28 months: Differential stability of language and symbolic
rather do it myself: Some effects and noneffects of maternal speech
play. In R. N. Emde & R. J. Harmon (Eds.), Continuities and disconstyle. In C. Snow & C. A. Ferguson (Eds.), Talking to children: Lantinuities in development (pp. 229-261). New York: Plenum Press.
guage input and acquisition (pp. 109-149). Cambridge, England:
Brown, R. (1973). A first language: The early stages. Cambridge, MA:
Cambridge University Press.
Harvard University Press.
Olson, S. L., Bates, J. E., & Bayles, K. (1984). Mother-infant interaction
Carew, J. V. (1980). Experience and the development of intelligence in
and the development of individual differences in children's cognitive
young children at home and in day care. Monographs of the Society
competence. Developmental Psychology, 20, 166-179.
for Research in Child Development, 45(6-7, Serial No. 187).
Piaget, J. (1962). Play, dreams and imitation in childhood. New York:
Corrigan, R. (1978). Language development as related to stage 6 object
Norton.
permanence development. Journal ofChild Language, 5, 173-190.
Power, T. G., & Parke, R. D. (1982). Play as a context for early learning.
Corrigan, R. (1979). Cognitive correlates of language: Differential criIn M. Laosa & I. E. Sigel (Eds.), Families as learning environments
teria yield differential results. Child Development, 50, 617-631.
for children (pp. 147-178). New York: Plenum Press.
Curtiss, S. (1989). The independence and task-specificity of language.
In M. H. Bornstein & J. S. Bruner (Eds.), Interaction in human devel- Rogoff, B. (1990). Apprenticeship in thinking: Cognitive development in
social context. New York: Oxford University Press.
opment (pp. 105-137). Hillsdale, NJ: Erlbaum.
DeLoache, J. S., & Plaetzer, B. (1985, April). Tea for two: Joint mother- Schlesinger, 1. M. (1971). Production of utterances and language acquisition. In D. I. Slobin (Ed.), The ontogenesis ofgrammar: Somefacts
child symbolic play. In J. S. DeLoache & B. Rogoff (Chairs), Collaborative cognition: Parents as guides in cognitive development. Sympo- and several theories (pp. 63-101). San Diego, CA: Academic Press.
Shore, C, O'Connell, B., & Bates, E. (1984). First sentences in language
sium conducted at the meeting of the Society for Research in Child
and symbolic play. Developmental Psychology, 20, 872-880.
Development, Toronto, Ontario, Canada.
Slade, A. (1987). A longitudinal study of maternal involvement and
de Villiers, P. A., & de Villiers, J. G. (1992). Language development. In
symbolic play during the toddler period. Child Development, 58,
M. H. Bornstein & M. E. Lamb (Eds.), Developmental psychology: An
367-375.
advanced textbook (pp. 337-418). Hillsdale, NJ: Erlbaum.
Fein, G. G. (1981). Pretend play in childhood: An integrated review.
Slobin, D. 1. (1982). Universal and particular in the acquisition of lanChild Development, 52, 1095-1118.
guage. In E. Wanner & L. R. Gleitman (Eds.), Language acquisition:
The state ofthe art (pp. 128-218). Cambridge, England: Cambridge
Fiese, B. H. (1990). Playful relationships: A contextual analysis of
University Press.
mother-toddler interaction and symbolic play. Child Development,
61, 1648-1656.
Snyder, L., Bates, E., & Bretherton, I. (1981). Content and context in
early lexical development. Journal ofChild Language, 5, 565-582.
Furrow, D., Nelson, K., & Benedict, H. (1979). Mothers' speech to children and syntactic development: Some simple relationships. Journal
Tamis-LeMonda, C. S., & Bornstein, M. H. (1989). Habituation and
ofChild Language, 6, 423-442.
maternal encouragement of attention in infancy as predictors of toddler language, play, and representational competence. Child DevelopGoodwyn, S. W., & Acredolo, L. P. (1993). Symbolic gesture versus
ment, 60,738-751.
word: Is there a modality advantage for onset of symbol use? Child
Development, 64, 688-701.
Tamis-LeMonda, C. S., & Bornstein, M. H. (1990). Language, play, and
attention at one year. Infant Behavior and Development, 13, 85-98.
Gopnik, A., & Meltzoff, A. N. (1984). Semantic and cognitive development in 15- to 21-month old children. Journal of Child Language, Tamis-LeMonda, C. S., & Bornstein, M. H. (1991). Individual varia495-513.
tion, correspondence, stability, and change in mother and toddler
Gopnik, A., & Meltzoff, A. N. (1987). The development of categorizaplay. Infant Behavior and Development, 14, 143-162.
tion in the second year and its relation to other cognitive and linguistic
Tamis-LeMonda, C. S., & Bornstein, M. H. (1993). Play and its redevelopments. Child Development, 58, 1523-1531.
lations to other mental functions in the child. In M. H. Bornstein &
Gottfried, A. W. (1985). Measures of socioeconomic status in child deA. O'Reilly, (Eds.), New directions for child development: The role of
velopment research: Data and recommendations. Merrill-Palmer
play in the development ofthought (Vol. 59, pp. 17-27). San Francisco:
Quarterly, 31, 85-92.
Jossey-Bass.
292
CATHERINE S. TAMIS-LEMONDA AND MARC H. BORNSTEIN
Tamis-LeMonda, C. S., Bornstein, M. H., Cyphers, L., Toda, S., &
Ogino, M. (1992). Language and piay at one year: A comparison of
toddlers and mothers in the United States and Japan. International
Journal ofBehavioral Development, 15, 19-42.
Teti, D. M., Bond, L. A.T & Gibbs, E. D. (1988). Mothers, fathers, and
siblings: A comparison of play styles and their influence upon infant
cognitive level. JnternationalJournal ofBehavioral Development, 11
415-432.
Tomasello, M., & Farrar, J. (1984). Cognitive bases of lexical development: Object permanence and relational words. Journal of Child
Language, 13, 495-505.
Tomasello, M., & Todd, J. (1983). Joint attention and lexical acquisition
style. First Language, 4, 197-212.
Tukey, J. W. (1977). Exploratory data analysis. Menlo Park, CA: Addison-Wesley.
Veneziano, E. (1981). Early language and nonverbal representation: A
reassessment. Journal of Child Language, 8, 541-563.
Vibbert, M., & Bornstein, M. H. (1989). Specific associations between
domains of mother-child interaction and toddler referential language
and pretense play. Infant Behavior and Development, 12, 163-184.
Vygotsky, L. S. (1967). Play and its role in the mental development of
the child. Soviet Psychology, 12, 62-76.
Wachs, T. D. (1987). Early experience and cognitive development: The
search for specificity. In I. C. Uzgiris & J. M. Hunt (Eds.), Infant
performance and experience (pp. 207-229). Chicago: University of
Illinois Press.
Appendix
Toddler Play Levels
Play level
Definition
Nonsymbolic
1. Unitary functional activity
2. Inappropriate combinatorial
3. Appropriate combinatorial
4. Transitional play
Symbolic
5. Self-directed pretense
6. Other-directed pretense
7. Sequential pretense
8. Substitution pretense
Production of an effect that is unique to single object (e.g.,
squeeze foam ball)
Inappropriate juxtaposition of two or more objects (e.g., put ball
in vehicle)
Appropriate juxtaposition of two or more objects (e.g., put lid on
teapot)
Approximation of pretense but without confirmatory evidence
(e.g., place spoon in cup, no stirring)
Clear pretense activity (e.g., eat from spoon or cup)
Clear pretense activity directed toward other (e.g., kiss or hug
doll)
Linking of two or more pretense actions (e.g., feed doll and wipe
mouth with sponge)
Pretense activity involving one or more object substitutions (e.g.,
pretend block is sponge and scrub floor)
Received August 13, 1992
Revision received July 26, 1993
Accepted August 3,1993