Cognitive Development 20 (2005) 526–544

Episodic development in preschool children’s

play-prompted and direct-elicited narratives
Hande Ilgaz, Ayhan Aksu-Koç ∗
Department of Psychology, Boğaziçi University, 34342 Bebek, Istanbul, Turkey

Abstract
This study investigated the premise that action, manifested here through pretend play, is a semiotic
arena that can enhance narrative development. It was hypothesized that children would produce structurally
more complex narratives in play-prompted elicitation than in direct elicitation conditions, and that this
competence would increase with age. Thirty Turkish 3-, 4-, and 5-year-olds, with 10 children per age group,
participated in the study. Children’s narratives were analyzed according to a story grammar framework
that differentiated between pre-episodic and episodic structures. A repeated-measures ANOVA revealed a
significant effect between age and type of elicitation condition. Narratives displaying an episodic structure
with temporal–causally integrated components were observed in the play-prompt elicitation condition by
4-year-olds and in the direct elicitation condition by 5-year-olds, supporting the hypothesis that action
helps enhance narrative development. It is concluded that the emerging competence of children to construct
narratives with episodic structures can best be observed in play contexts where action and objects scaffold
both the conceptual organization and its expression.
© 2005 Elsevier Inc. All rights reserved.

Keywords: Play-prompted elicitation; Direct elicitation; Episodic structure; Narrative development; Play

1. Introduction

Narrative and play are prominent research areas that provide valuable insight to aspects of early
cognitive development. Both language and play emerge in the second year of life as manifestations
of the symbolic function that marks the shift from the plane of action to that of representation
(Piaget, 1962). Parallel developments in language and pretend play have been identified start-
ing as early as the second year of life (e.g., Bates, Benigni, Bretherton, Camaioni, & Volterra,
1979; McCune-Nicholich, 1981). Nelson (1996) regards action, in general, and play in particu-
lar, as a semiotic arena that facilitates linguistic development, specifically, thinking in language.

∗ Corresponding author.
E-mail address: [email protected] (A. Aksu-Koç).

0885-2014/$ – see front matter © 2005 Elsevier Inc. All rights reserved.
doi:10.1016/j.cogdev.2005.08.004
 H. Ilgaz, A. Aksu-Koç / Cognitive Development 20 (2005) 526–544 527

Nicolopoulou (in press), who also underscores the close connection between action in play and
narrative, argues that they stem from different sources and develop on different tracks, but are con-
nected and integrated with attainment of enough mastery. Although these views have met with
certain acclaim, empirical research comparing play and narrative to investigate their proposed
interplay has been scant. The present study compares preschool children’s play-prompted and
direct-elicited narratives to examine whether pretend play acts as a semiotic arena for narrative
development between 3 and 5 years of age.

1.1. Narrative development

Narratives can be defined as units of discourse representing a sequence of temporal–causally

related events. While temporality is the default relation and is characteristic of descriptions or
scripts (Nelson, 1996), causal coherence between events is fundamental to well-formed narra-
tives (Rumelhart, 1977; Stein & Glenn, 1979). The literature on children’s developing narrative
competence shows that a commonly used analytical framework for identifying young children’s
narratives is the story grammar approach which proposes that stories are constituted of episodes
(Mandler & Johnson, 1977; McCabe & Peterson, 1991; Peterson and McCabe, 1983; Stein, 1988;
Stein & Glenn, 1979). Story grammar approach focuses on the semantic relations between propo-
sitions in a narrative and the functions they fulfil. A complete episode has to include at least
the following three components: the protagonist’s goals, efforts to achieve these goals, and the
outcomes of these efforts (Peterson & McCabe, 1983). These minimal constituents of an episode
are held together by temporal–causal relations.
Trabasso and co-workers (Trabasso, Secco, & van den Broek, 1984; Trabasso & van den
Broek, 1985; van den Broek, 1990) propose that in processing and making sense of narratives
one organizes the incoming information into a series of causal chains. Similarly, in developing
narrative competence, children form expectations about the unfolding of information in stories.
This set of expectations bound with temporal–causal relations, constitutes a type of story schema
that guides both production and comprehension (Johnson & Mandler, 1980). Studies of children’s
production skills show that the structural complexity of narratives (personal or fictional) increases
with age as they start to incorporate more episodic components (Hudson & Shapiro, 1991; Peterson
& McCabe, 1983), and as the relations between events shift from temporal to causal.
Children do not display full competence in producing narratives with episodic structures before
school years (e.g., Bamberg, 1987; Berman & Slobin, 1994; McCabe & Peterson, 1991; Nelson,
1989). Although 5-year-olds can produce some narratives with episodic properties, their perfor-
mance is unstable across genres and methods of elicitation, with the quality of their narratives
fluctuating between pre-episodic and episodic structures. Around 7 years, children construct sto-
ries that are primarily episodic in character, which is further bolstered by literacy acquisition and
school experience (e.g., Berman & Slobin, 1994; Nelson, 1996).

1.2. Play and narrative

To overcome difficulties of eliciting narratives from young preschool children (Ely, Wolf,
McCabe, & Melzi, 2000) narrative researchers have utilized different methods ranging from
recordings of spontaneous productions (e.g., Küntay, 1997) to direct elicitation by asking for a
story (e.g., Leondar, 1977) to using story pictures (e.g., Berman & Slobin, 1994) and story stems
(e.g., Merritt & Liles, 1987). And some studies have relied on pretense play as a facilitative context
for story production. These studies have either integrated action as a dramatic component into
528 H. Ilgaz, A. Aksu-Koç / Cognitive Development 20 (2005) 526–544

children’s narrative expression by having them perform the stories (Nicolopoulou, in press; Paley,
1981; Pellegrini & Galda, 1982; Williamson & Silvern, 1991) or have used toy props to facilitate
production of narratives (Benson, 1993; Eckler & Weininger, 1989).
Evidence indicates that play with functionally explicit toys such as miniature versions of real life
items scaffold children’s productions yielding longer and more integrated narratives (Pellegrini,
1987). Whether play with replica toys should be classified as ‘pretend’ has been debated on the
grounds that it does not require object transformations and may not involve a shift into the ‘as
if’ mode of functioning (Pellegrini, 1986). Although children in such play do not make object
transformations, they nevertheless have to construct relations between the toy figures and make
them function within the constraints of an imaginary realm. The enhancing effect of this form
of pretend on narrative production is believed to stem from the reduced information processing
demands such play contexts present (Pellegrini & Jones, 1994).
While one is interested in finding the best method for narrative elicitation, our own concern
is also theoretically motivated. Both narrative construction and pretend play require the child to
operate in the representational medium. Definitions of pretend play converge on recognizing “the
projecting of a supposed situation onto an actual one” (Lillard, 1993, p. 349), that is, the imaginary
or ‘as if’ quality of play (Garvey, 1977; McCune-Nicholich, 1981; Perner, 1991). Evidence
indicates that children’s early pretense play is strongly based on action. Action is a defining and
essential feature of pretense play prior to the full emergence of its representational quality (Perner,
1991; Nelson, 1996). Furthermore, pretense play, similar to stories, requires an organization of
events around a plot. Play functions as a semiotic arena by facilitating plot construction as well
as allowing for expression simultaneously in action and verbalization. Paley asserts that “play
. . . [is] story in action just as story telling is play put into narrative form” (1990, p. 4).
A similar view underscoring the parallelism between pretense and story episodes is set forth by
Harris and Kavanaugh (1993). They claim that comprehension of pretense episodes is a process
analogous to text or story comprehension. Briefly, in pretense episodes children need to keep track
of the pretend identity of make-believe props while in story comprehension they have to keep
track of the referring expressions, and in both they have to keep these referents in temporary focus
for the duration of the episode. Furthermore, pretend comprehension, like story comprehension,
calls for elaborative causal inferences that integrate successive parts of an episode. In both types
of comprehension, such constructive processing yields an interconnected and retrievable repre-
sentation of the episode. In a series of experiments, Harris and Kavanaugh found that 2-year-olds
can hold in mind the stipulated make-believe identity of an object, refer to an imaginary sub-
stance, refer to pretend actions and their implied outcome, and thus describe what is happening
in a fictitious world. These abilities indicate that 2-year-olds can engage in causal extrapolation,
coordinate the successive parts of a pretend episode, and use one symbolic system (language) to
talk about the referents of the other (a pretend episode) (1993, pp. 56–57).
Production studies also show that children generate play plots in pretense around 2 years
(McCune-Nicholich, 1981) and attribute sensory, perceptual, emotive, and cognitive experiences
to their agents in play (Wolf & Gardner, 1979) between 2½ and 3½ years of age. By age 4,
children can handle more than two characters (Rubin & Wolf, 1979), and make attributions about
their internal states (Wolf, Rygh, & Altshuler, 1984). In sum, young preschool children can
create an imaginary realm in doll-play and make the figures interact within the constraints of this
pretend world. Thus, for preschool children play presents itself as a context that would support
the construction of episodically structured narratives.
Pretend play was used by Eckler and Weininger (1989) to elicit play narratives from 4- to 8-year-
olds. These researchers provided children with two sets of themed replica toys, and asked them to
 H. Ilgaz, A. Aksu-Koç / Cognitive Development 20 (2005) 526–544 529

play a game of pretend recounting everything they did while playing. The elicited play narratives
were classified as either episodic or non-episodic, defined according to story grammar criteria. As
noted above, to qualify as episodic, a narrative must include the causally related components of
goal, attempt at action, and consequence. The episodic narratives were further classified into three
complexity structures: a single episode, two episodes in sequence, or three episodes two of which
are simultaneous. Eckler and Weininger reported that children as young as 4 years produced at
least one episode in their play narratives and that the number of episodes increased with age, thus
revealing significant development in the structural complexity of play narratives. Benson (1993)
compared children’s pretend play narratives elicited with a set of non-themed toys with narratives
elicited by several drawings of figures. She observed a significant change in structural complexity
with age, whereby 5-year-olds displayed a ‘much stronger mental model of a story’ (1993, p.
219) than 4-year-olds, thus confirming Eckler and Weininger’s findings (1989). Interestingly, this
development was observed in the story-telling session rather than in the play session.
In view of these findings (Benson, 1993; Eckler & Weininger, 1989; Harris & Kavanaugh,
1993; Pellegrini, 1987), we hypothesize that symbolic play serves as a semiotic arena in which
the child can more efficiently solve the complex problem of mapping hierarchically composed
thoughts to the linear structure of language. In play, toy props and action are contextually available
to support the child’s plot construction activity. Furthermore, a complete causal chain can be
specified simultaneously by action and verbalization. However, in the case of storytelling, the
only available means for representation of agents, actions, and relations is language itself, posing
a greater mental demand than play narratives.
The present study compared play-prompted narratives with direct-elicited narratives produced
by 3- to 5-year-old children. It was hypothesized that play-prompts would help the child move
to the imaginary realm, and provide the context for plot construction and expression by the
exploitation of both action and language. Play-prompted elicitation condition was predicted to
yield narratives that have more complex episodic structures than the direct elicitation condition.
In addition, it was expected that an increase in overall structural complexity of narratives would
be manifest at an earlier age in the play-prompted than direct-elicited narratives.

2. Method

2.1. Participants

Thirty Turkish monolingual, children five boys and five girls at each of three age groups (3-year-
olds, M = 3.8, range: 3.3–3.11; 4-year-olds, M = 4.7, range: 4.3–4.11, 5-year-olds, M = 5.6, range:
5.1–5.11) participated in the study. The children were drawn from three middle-class preschools
in a major city in Turkey and they were selected based on their willingness to participate.

2.2. Materials

2.2.1. Prompt-elicited narratives

Two sets of toys were used as play prompts. One set consisted of a farmhouse with animals,
pumpkins, a stack of hay, a farmer and a tractor. The animals were two horses, two cows, two
sheep, a pig and a chicken, where one animal was bigger than the other, suggesting a parental
relation. The second set consisted of a house with furniture and human figures. The furniture set
consisted of a bed, a couch, an armchair, a swinging chair, a table, four chairs, a coffee table, a
cupboard and a computer. The human figures were suggestive of a family, and consisted of an
adult male, an adult female and two children, a girl and a boy.
530 H. Ilgaz, A. Aksu-Koç / Cognitive Development 20 (2005) 526–544

2.2.2. Direct-elicited narratives

A Crayola coloring book, Crayola magic pens, coloring pens and scrap paper were used for
the warm-up period. Each page of the Crayola book depicted pictures of familiar animals. The
magic pens appeared white and changed color in a few seconds. The set was selected to allow for
interaction with the child for an entertaining warm-up.

2.3. Procedure

Each child was tested twice, once to elicit a toy-prompted narrative and once for a direct-
elicited narrative. The elicitation conditions were counterbalanced and were presented 1 week
apart. Only three children were tested 2 weeks apart due to school absences. All sessions were
videotaped. The experimenter visited the classroom and told the children that she was interested
in their stories, and that she had sets of toys and interesting coloring pens to share with them.
The children who volunteered were taken individually to an adjacent small testing room with the
appropriate toys or materials already arranged. In the play-prompted condition, the house and the
farmhouse were spread out on the floor next to each other. The materials of each set were lined in
front of the houses. The camera was set facing this display. In the direct elicitation condition, a
table and two chairs were arranged and warm-up materials were placed on the table. The camera
was set facing the child.
In the prompt-elicited play, the experimenter explained to the children that she wanted to make
a great book containing the stories children tell and had to video-record the session to be able to
write their stories later on. She then instructed the children to explore and play with the toys for
10 min. Warm-up started after the children understood what was expected of them. This warm-up
was used because during a pilot study it was observed that children took time to explore the toy
sets and asked questions about them. These interactions also helped establish a relationship with
the experimenter. After the warm-up period, the experimenter said to the child: “You have looked
at all the toys, now it is time to tell me a story using the toys. Can you tell me a story using these
toys?” Some children dropped the toys and started telling a tale without using them as props.
These children were instructed to enact their story using the toys and show to the experimenter
who was doing what. After the elaboration of the instruction none of the children had further
difficulties. Throughout the narrative sessions, the experimenter gave non-specific prompts such
as “Huh-huh”, “Then?” or verbatim repetitions of a portion of the child’s utterances. At times,
when the children’s speech got incomprehensible (e.g., voicing characters, whispering or talking
too fast) they were asked to repeat what they had just said. On occasions, when the child’s actions
were not in the camera’s range the experimenter voiced the actions if the child did not. Children
were given 10 min to tell a story, and if they were not done within this time period, they were
told to wrap up their stories. If the children protested, they were allowed some extra time, but the
recording was finished after 20 min of story telling. Children were thanked for “such a nice story”
and were given a sticker for their participation.
In the direct elicitation session, the children were allowed to color either the animal pictures
in the book or make drawings as a warm-up activity for 10 min. Then, the experimenter told the
children “My watch shows me coloring time is over, now I want you to tell me a story. Can
you please tell me a story?” If the children said that they could not tell a story, the researcher
asked whether their parents read or told them stories, and asked them to tell their favorite story.
If the children still had difficulty, they were told that they could tell a story about the animal they
had colored in the warm-up session. Only one 3-year-old boy refused to tell a story by direct
elicitation. Children were thanked and given a sticker for their participation.
 H. Ilgaz, A. Aksu-Koç / Cognitive Development 20 (2005) 526–544 531

2.4. Transcription and coding

All videotapes were transcribed by the experimenter, omitting the warm-up periods. The
prompt-elicited narratives were transcribed for (a) the child’s utterances, (b) the child’s actions
and (c) the experimenter’s utterances in the exact order they occurred. Narratives in the direct-
elicitation condition were transcribed for the child’s and the experimenter’s utterances. The
transcripts were coded in three steps: first, narrative units were demarcated in the discourse
produced by each subject in each condition. Second, the, specific propositions comprising such
units were coded as episodic components. Finally, these strings of episodic components were
identified as representing one of the structural complexity categories.

2.4.1. Demarcation of thematic clusters as narrative units

Each transcript was read by the first author several times in order to define the narrative
units produced by the child during each condition. A narrative unit was defined as a thematic
cluster provided through continuity of the protagonist’s point of view, as well as continuity of
temporal–causal relations. In a given session, a child could have produced one or more narrative
units.

2.4.2. Coding of propositions into episodic components

Peterson and McCabe’s (1983) adaptation of Glenn and Stein’s story grammar was used. Each
narrative unit was separated into propositions. A proposition was defined as a clause or an action
unit that contains a predicate. In the direct-elicited narratives, only the verbal productions were
treated as propositions. In the prompt-elicited narratives in addition to the verbal productions,
actions executed by the protagonist but not verbally expressed were also counted as propositions.
Only those actions that were attributable to the characters were considered. The propositions
within each narrative unit were coded into the following episodic components: introduction,
abstract, setting, action, event, motivating state, reaction, attempt, consequence, evaluation, and
closing (see Appendix 1 for the coding scheme of propositions into episodic components).

2.4.3. Coding for structural complexity

The third step of coding involved assigning the strings of episodic components into one of
the 10 hierarchically ordered complexity structures (see Table 1). Complexity structures 1–7 are
pre-episodic and 8–10 are episodic in nature. For a complexity structure to qualify as episodic,
the requirement was the presence of a protagonist with a goal, who attempts at action that results
in an outcome. These three narrative components had to be explicitly stated and causally related
although the causal connections that bind them together did not have to be verbalized by use
of causal connectives. Any complexity structure which did not meet the criteria for a complete
episode as defined above was considered as a pre-episodic complexity structure. The placement
of a narrative unit into a particular complexity structure subcategory was carried out according to
the definitions in Table 1.

2.5. Coding reliability

Both the coding of propositions into narrative components and their assignment into complexity
structures were carried out by the first author. Eighteen randomly selected narratives, nine prompt-
and nine direct-elicited from three children at each age group were coded and classified by a second
532 H. Ilgaz, A. Aksu-Koç / Cognitive Development 20 (2005) 526–544

Table 1
Narrative complexity structure hierarchy
Pre-episodic categories
1 Descriptive sequence Describes character(s), surroundings, and habitual actions with no
temporal–causal relationship
2 Script sequence Gives temporal and/or enabling relations between actions that constitute
routine events
3 Action sequence List of actions that are temporally rather than causally related
4 Reactive sequence Set of changes that automatically cause other changes with no planning
involved
5 Interactive reactive sequence Describes extensive interactions between two or more characters without
any evidence of planning, intentionality or goal-oriented action
6 Incomplete episode Gives all components of a complete, complex, or interactive episode except
the requisite consequence
7 Abbreviated episode Describes aims of the protagonist, but the attempts to obtain the goal gener-
ally must be inferred
Episodic categories
8 Complete episode Encompasses at least three of the components of event, motivating state,
attempt and consequence. The consequence category is compulsory
9 Complex episode Elaboration of a complete episode in one of four ways by including: an
embedded reactive sequence; an embedded complete episode; a multiple
plan application (i.e., repeated attempts); or a multiple plan application with
an embedded complete episode
10 Interactive episode Describes one set of events from two perspectives where both characters
have goals and influence each other. Interactive episode can be completed
from each protagonist’s point of view

rater. These constituted 30% of the overall corpus. Inter-rater reliability using kappa was 0.90,
p < 0.001 for structural components and 0.92, p < 0.001 for the complexity scores.

3. Results

The complexity scores ranged from 1 for the simplest descriptive sequence to 10 for the most
complex interactive episode. Typically, in each condition children produced more than one type
of episodic structure; but they were given one score corresponding to the highest level complexity
structure found in their narrations. All 30 subjects produced narratives in the prompt elicitation
condition and 29 subjects produced narratives in the direct elicitation condition. One subject, a
3-year-old boy did not produce any narratives in the direct elicitation condition. The distribution
of means and standard deviations of play-prompt elicited narrative (PEN) and direct-elicited
narrative (DEN) scores by age and gender is presented in Table 2.
It was hypothesized that there would be an age-related increase in the complexity of children’s
narratives regardless of elicitation condition and that children would obtain higher complexity
scores in the prompt elicitation than in the direct elicitation condition. To test these hypotheses,
a mixed design ANOVA on structural complexity scores, with age and gender as the between-
subjects factors, and type of condition as the within-subjects factor was used. There were no
significant effects of condition (F[1,23] = 1.98, p = n.s.), age (F[2,23] = 1.73, p = n.s.) and gen-
der (F[1,23] = 0.14, p = n.s.). There were no significant interaction effects for condition and age
(F[2,23] = 1.68, p = n.s.), condition and gender (F[1,23] = 0.38, p = n.s.) and for condition, age and
gender (F[2,23] = 0.04, p = n.s.). Since gender was not significant, and there were no theoretical
considerations for it, it was eliminated from further analyses.
 H. Ilgaz, A. Aksu-Koç / Cognitive Development 20 (2005) 526–544 533

Table 2
Distribution of means and standard deviations of structural complexity scores for prompt-elicited narratives (PEN) and
direct-elicited narratives (DEN) by age and gender
Age PEN DEN
M (S.D.) M (S.D.)

3
Female 6.40 (2.61) 6.60 (2.19)
Male 5.80 (2.78) 6.25 (3.59)
4
Female 8.80 (1.10) 7.00 (3.32)
Male 8.60 (1.14) 6.40 (2.79)
5
Female 8.00 (2.92) 8.40 (0.89)
Male 8.20 (0.84) 7.60 (2.70)

While these results did not support our hypothesis, the pattern of means showed that there may
be an interaction between age and condition (see Table 2). For prompt elicitation condition, there is
a shift in level of structural complexity at age 4, while in direct-elicited narratives this is observed
only at age 5. To probe the data further we decided to reduce age variability from a whole year
to a half year. This yielded a sample of 20 children: six 3-year-olds (3.5–3.11, M = 3.10), eight
4-year-olds (4.5–4.11, M = 4.8) and six 5-year-olds (5.5–5.11, M = 5.9). A mixed-design ANOVA
on structural complexity scores, with age as the between-subjects factor and type of condition as
the within-subjects factor did not yield significant effects for type of condition (F[1,17] = 1.27,
p = n.s.) or age (F[2,17] = 1.17, p = n.s.) but showed a significant condition by age interaction
(F[2,17] = 3.80, p = 0.05), confirming the trend observed in the overall sample. The means and
standard deviations are presented in Table 3.
To further explicate the nature of the observed shift we examined how children’s narratives
are distributed in terms of their episodic quality in the two elicitation conditions by age. For this
purpose, children’s narratives that had been identified as one of the 10 complexity structures were
combined into pre-episodic or episodic categories. Table 4 presents the distribution of children
in terms of the type of complexity structure and its corresponding episodic category produced in
the prompt elicitation and direct elicitation conditions by age. It is observed that in the prompt
elicitation condition, the 3-year-old group has a greater variability while the 4- and 5-year-olds
have produced mainly episodic narrative structures. In the direct elicitation condition, however,
greater variability is observed in the scores of the 3- and 4-year-olds; 5-year-olds are clustered
around higher complexity structures.

Table 3
Distribution of means and standard deviations of structural complexity scores for prompt-elicited (PEN) and direct-elicited
narratives (DEN) by age
Age

3 (N = 6) 4 (N = 8) 5 (N = 8)
M (S.D.) M (S.D.) M (S.D.)

PEN 6.29 (2.75) 8.87 (0.99) 7.67 (2.31)

DEN 6.17 (3.37) 6.63 (2.88) 7.15 (2.76)
534 H. Ilgaz, A. Aksu-Koç / Cognitive Development 20 (2005) 526–544

Table 4
Number of children producing episodic vs. non-episodic complexity structures in prompt-elicited (PEN) and direct-elicited
narratives (DEN) by age
Type of complexity structure Age
PEN DEN

3 4 5 3 4 5

Pre-episodic
Descriptive sequence 0 0 0 1 0 0
Action sequence 3 0 1 1 2 1
Reactive sequence 0 0 0 0 2 0
Interactive reactive sequence 1 0 0 0 0 0
Incomplete episode 2 0 0 0 0 0
Abbreviated episode 0 2 1 4 1 0
Episodic
Complete episode 1 1 3 1 1 6
Complex episode 3 5 3 2 2 1
Interactive episode 0 2 2 0 2 2

Episodic formations comprising a protagonist’s motivating states, attempts to reach a goal, and
the consequences of these attempts were observed as early as 3 years of age (see Fig. 1). These
were the dominant type of structure by 4 years in the prompt elicitation condition and by 5 years
regardless of condition. These findings indicate that children’s narrative productions changed in
organization from non-causal structures to causal structures between 3 and 4 years of age in their
prompt-elicited narratives, but between 4 and 5 years of age in their direct-elicited narratives. The
nature of the age-related changes observed in the mean structural complexity scores in the different
elicitation conditions thus reflects a shift from pre-episodic to episodic narrative structures (see
Tables 2 and 3).

3.1. Stories that illustrate complexity structures typical for each age group

To display the nature of the shift in narrative complexity in the two elicitation conditions, we
present typical examples from each age group.

Fig. 1. Percent of children who produced episodic narratives in direct elicitation condition (DEN) and prompt elicitation
condition (PEN) by age.
 H. Ilgaz, A. Aksu-Koç / Cognitive Development 20 (2005) 526–544 535

3.1.1. Three-year-olds: predominantly pre-episodic pattern

Three-year-olds produced predominantly pre-episodic structures in both direct-elicited narra-
tives and prompt-elicited narratives. The percentage of episodic structures was almost equally
low in the two conditions: 40% in prompt-elicited and 33% in direct-elicited. Six out of the nine
narratives produced in the direct elicitation condition bore explicit influence of some well-known
fairy tale, namely the Red Riding Hood, Sleeping Beauty and Snow White. In fact, two of the
three stories with an episodic structure reflected the influence of these fairy tales. Examples (1)
and (2) are pre-episodic structures produced by the same 3 year old in her direct-elicited and
prompt-elicited narratives, respectively.1
(1) Elif (3;3); DEN, Abbreviated Episode (Utterances 1–20, Total number of utterances: 20)
(1) Bir varmış, (2) bir yokmuş. (3) Evvel zaman içinde, (4) kalbur saman içinde (5) Bir
büyük bir pamuk prenses varmış. (6) O pamuk prenses de annesini bulamamış. (7) Gitmiş
(8) bahçeye bakmış (9) Oraya, çiçeklere su döküyormuş. (10) Annesi de . . . Annesi de gitmiş
(11) Bakmış (12) Kızı gelmiş eve. (13) O da geldikten sonra annesi onu bulmuş. (14) Kızı
onu bulmuş. (15) Sonra da eve götürmüş annesi. (16) Cumartesi günü de tekrar diğerlerini
sulatmış. (17) Sonra suları onları açıyor.Açmış. (18) Sonra bişeyler yapmış yapmış yapmış.
(19) Sonra bişeyler yapmış yapmış yapmış. (20) Bu masal da bitmiş.
(1–4) Once upon a time. (5) There was a Big Snow White (6) That Snow White could
not find her mother. (7) She went (8) looked in the garden. (9) There, she was watering
the flowers. (10) {her mother}The mother, too, went. (11) Looked (12) her daughter has
come home. (13) After she came too, her mother found her. (14) Her daughter found her.
(15) Then her mother took her home. (16) On Saturday she made her water the others
again. (17) Then they are opening the waters, them. They opened. (18) Then she did some
thing, she did, she did. (19) Then she did some thing, she did, she did. (20) This tale ends
here.
[Introduction (1–4)-Setting (5)-Event (6)-Attempt (7–8)-Action (9)-Attempt (10–11) Con-
sequence (12–14)-Action (15–20)-Closing (21)]
Elif builds her narrative around a common fairy tale heroine, Snow White. This narrative has
a clear complicating action where Snow White seeks her mother and as a consequence mother
and daughter are united, yet the planning of attempts and how the goal is obtained is depicted
ambiguously. It is difficult to ascertain how the consequence is reached; therefore, this is qualified
as an abbreviated episode.
(2) Elif (3;3); PEN, Action Sequence (Utterances 16–23; Total number of utterances: 143)
(16) (E.Çocuk)Sokaklarda dolaşıyor. (17) Oraya yatmışlar sokağın ortasına. (18) Ortasına
yatmışlar. (19) (Çocuk)Düştüm sandalyeden. (20) Sandalyeden düşmesi çok tehlikeli biliy-
ormusun? (21) (Çocuk)Gelmiş. (22) Davet etmişler onlar da. (23) Ortalarda dolaşıyor.
(16) (The boy) He is wandering in the streets. (17) They lied down there, in the middle of
the street. (18) They lied in the middle of it. (19). (The girl):“I fell off the chair”. (20) (to

1 In the examples that follow, the actions children performed in the prompt-elicited condition without verbalization are

marked with an (*) asterisk. False starts in narrative are enclosed in ( { } ) curly brackets. The protagonist of the narrative
is underlined. Information in the prompt-elicited narratives that was not verbalized yet indicated with gesture or displayed
by the play situation is presented in parenthesis.
536 H. Ilgaz, A. Aksu-Koç / Cognitive Development 20 (2005) 526–544

the experimenter) It is very dangerous for her to fall off the chair. (21) He came. (22) They
invited him then. (23) He is wandering around.
[Action (16–19)-Evaluation (20)-Action (21–23)]
The excerpt above is an action sequence which is the most complex structure found in her
prompt-elicited narrative.

3.1.2. Four-year-olds: mixed complexity pattern

Four-year-old children produced more episodic structures in their prompt-elicited (80%) than
in their direct-elicited narratives (50%). All those who produced episodic structures in their direct-
elicited narratives also produced these structures in their prompt-elicited narratives, while three
children who did not produce episodic structures in the direct elicitation condition did so in the
prompt elicitation condition. Hence, the availability of objects and an action arena appears to
affect the episodic quality of the 4-year-olds’ narratives.
(3) Kayra (4;5); PEN, Complex Episode (Utterances 20–50, Total number of utterances:
157)
(20) *Kuzu eve yürüdü. (21) *Evin üst katına çıktı. (22) *Kitaplıktan kitap aldı. (23) Düşmüş,
yırtılmış. (24) *Kuzu çiftliğe yürüdü. (25) *Ahırına girdi. (26) *Düştü. (27) *Kalktı. (28)
*Kız yürüdü. (29) *Evin 2. katına çıktı. (30) K:“Ben pembe kitabımı alayım.” (31) *Kitaplığa
baktı. (32) Ama pembe kitabını yırtmış. (33) K:“Pembe kitabım nerede?” (34) *Kız odada
yatan babasının yanına gitti. (35) K:“Baba pembe kitabım yok” (36) *Baba kitaplığa baktı.
*Baba kitabı bulamadı (37) *Tekrar yattı.. (38) B: “Aaa belki birisi almıştır.” (39) *Kız
çiftliğe yürüdü. (40) K: “Hemen gideyim bakayım, sorayım.” (41) *Kız kuzunun ahırının
kapısını açtı. (42) *Ahıra girdi. (43) *Ahırın kapısını kapattı. (44) K: “Ne yaptın sen benim
kitabıma?” (45) Kuzu: “Yırttım.” (46) K: “Tamam ama bir daha yapma!” (47) *Ahırın
kapısını açtı. (48) *Kız çıktı. (49) *Ahırın kapısını tekrar kapattı. (50) Kuzu: “Bir daha
yapmayacağım” demiş.
(20) *The sheep walked to the house. (21) *He went upstairs. (22) *He took a book from
the bookcase. (23) He fell. It tore. (24) *The sheep walked back to the farm. (25) *He went
into his barn. (26) *He fell. (27) *He got up. (28) *The girl walked. (29) *She went upstairs.
(30) The girl: “I’ll take my pink book” (31) *The girl looked in the bookcase. (32) But he
tore her pink book. (33) The girl: “Where is my pink book?” (34) *The girl went near her
father who was lying down in the next room. (35) The girl: “Daddy my pink book is gone”.
(36) *The father looked at the bookcase. (37) *He lied down again.—He could not find the
book—. (38) Father: “Aaa maybe someone took it.” (39) *The girl walked to the farm. (40)
The girl: “I’ll go and look right away, I’ll ask.” (41) *The girl opened the door of the sheep’s
barn. (42) *She went into the barn. (43) *She closed the door. (44) The girl: “What did you
do to my book?” (45) The sheep: “I tore it.” (46) The girl: “Okay, but don’t do it again!”
(47) *She opened the door of the barn. (48) *She went out. (49) *She closed the door. (50)
The sheep said: “I won’t do it again”.
[Event (20–27)-Motivating State (30)-Attempt (28–29, 31)-Consequence (32)-Motivating
State (33)-Attempt (34–37)-Consequence (38)-Attempt (39–44)-Consequence (45–50)]
In the prompt-elicited narrative of this 4-year-old girl, the contribution of enacted propo-
sitions to episodic structure is clearly observed, as the child does not verbally render all the
 H. Ilgaz, A. Aksu-Koç / Cognitive Development 20 (2005) 526–544 537

actions of the characters. It could be argued that habitual actions such as walking or opening
and closing doors were not mentioned, not to interrupt the flow of the story. Some of the non-
verbalized actions, however, such as the sheep’s travel to the house, his tearing of the book
and the father’s quest for it, are the key components of this episode and serve to elaborate
and clarify the narrative. As depicted in this example, the prompt elicitation condition provides
children with an arena to express their narrative competence through the action medium, when
they cannot sufficiently express it verbally. Example (4) presents the same child’s direct-elicited
narrative.
(4) Kayra (4;5); DEN, Action Sequence (Utterances: 1–14; Total number of utterances: 14)
(1) At ormanda koşuyormuş. (2) Koşarken (3) bir sürü at arkadaşını görmüş. (4) Okula
gitmişler beraber. (5) Öğretmenler yazı yazmışlar öğretmenler. (6) Öğrenmişler (7) eve
gelmişler. (8) Sonra at arkadaşına gitmiş. (9) “Hastasın di mi?” demiş. (10) “Hayır”
demiş. (11) O zaman gitmiş (12) odasına çıkıp (13) televizyon izlemiş. (14) Bu kadarmış.
(1) The horse was running in the forest. (2) While he was running (3) he saw a lot of his horse
friends. (4) They went to school together. (5) The teachers wrote writings, the teachers. (7)
They learned (8) came home. (9) Then the horse went to his friend. (10) “You are sick,
aren’t you?” he said. (11) “No” he said. (12) Then he went (13) went up to his room (14)
watched television. (15) That was all.
[Setting (1–2)-Action (3–5)-Reaction (6)-Action (7–9)-Reaction (10)-Action (11–13)-
Closing (14)]
In contrast to her prompt-elicited narrative, Kayra’s direct-elicited narrative does not display
an episodic structure. The propositions are ordered temporally without displaying any causal
relations.

3.1.3. Five-year-olds: predominantly episodic pattern

Narratives produced by 5-year-old children in either the prompt elicitation or direct elicitation
conditions did not show a marked difference: 80% produced episodic structures in the former,
while 90% did so in the latter condition. Episodic structures were observed to be the predominant
structure for 5-year-olds regardless of elicitation condition.
(5) Berkay (5;7); PEN, Complex Episode (Utterances 61–82; Total number of utterances:
151)
(61) Sonra (çiftçi atın yanına) gelmiş. (62) (Çiftçi):“Neden bunlar çalınmış” demiş buna.
(63) (Büyük at): “Çünkü çocukla babası, kız annesi onları götürmüşler. (64) Bir hayvanat
. . . bir yerine koymuşlar.” (65) *İneği dışarı atın yanına koydu. (66) Çifçi de bunları (hay-
vanları) almış geri. (67) *Domuzu dışarı koydu (68) Sonra bunu (kuzuyu) da almış. (69)
Buraya (çiftliğin önüne) koymuş. (70) Bunu (küçük atı) da almış. (71) Buraya (çiftliğin
önüne) koymuş. (72) Bunu (küçük kuzuyu)da almış. (73) *Çiftliğin önüne koydu. (74) Bunu
(Büyük ineği)da almış. (75) Kapıları kapatmış. (76) Böyle çocuk da bakmış. (77) (Erkek
Çocuk): “Burada şey yok demiş. Yani kuzu. Kuzu.” (78) Aşağı da bakmış. (79) Erkek Çocuk:
“Hiç kimse yok” demiş “burada.” (80) Babası da, köpeği de (aşağı) atlamış. (81) Çiftçi de
korumuş onları. (82) *Köpek ve çocuk uzaklaştılar.
(61) Then (the farmer) came near (the horse). (62) (The farmer) said: “Why are these
stolen?”. (63) (The big horse): “Because the child and his father, the girl and her mother
538 H. Ilgaz, A. Aksu-Koç / Cognitive Development 20 (2005) 526–544

took them away. (64) They put them in a animal place. (65) *He put the cow near the horse
in front of the barn. (66) The farmer took these animals back. (67) *He put the pig in front
of the barn. (68) Then he took this (the sheep). (69) He put her here (in front of the barn.)
(70) He took this (the little horse) too. (71) Put him here (in front of the barn). (72) He took
this (the little sheep). (73) *He put the little sheep in front of the barn. (74) He took this
(the big cow), too. (75) He closed the doors. (76) Like this the boy looked. (77) (The boy):
“There is nothing here” he said. “I mean the sheep, the sheep . . .” (78) He looked down.
(79) (The boy) said: “There is no one here”. (80) The father and the dog jumped down. (81)
The farmer protected them (the animals). (82) *The dog and the boy go away.
[Event (61–64)-Attempt (65, 67–74)-Consequence (66,75)-Event (76–80)-Attempt (81)-
Consequence (82)]
In this narrative, the first complete episode (the farmer taking back his stolen
animals—utterances #61–75) functions as an event motivating the second episode (boy and father
searching for the animals and farmer protecting the animals against them—utterances #76–82)
and represents a causal sequence. Despite the extensive use of ambiguous referencing, which
is clarified by the child’s actions, the causal relations inferable within the action arena renders
the story comprehensible. An episodic structure is also observed in this child’s direct-elicited
narrative.
(6) Berkay (5;7); DEN, Complete Episode (Utterances: 23–38; Total number of utterances:
56)
(23) Ama içinde bir tane de fil varmış. (24) O fil de gerçekmiş. (25) Ama küçükmüş. (26)
Ondan sonra o da fille yatmış. (27) Sonra fil . . . akşam olduğu zaman (28) fil buzdolabının
kapısını açmış. (29) Oradan hep yiyecekleri açmış (30) yemiş. (31) Ondan sonra da yedikten
sonra hepsini yemiş (32) ve kocaman olmaya başlamış . . . (33) Ondan sonra yedikten sonra
(34) annesi {çocuk} çocuğu {çocuk} yemiş sanmış. (35) Ama çocuk yememiş. (36) Çocuğa
kızmış (37) ama “ben yapmadım” demiş. (38) “O fil yaptı” demiş.
(23) But there was an elephant among them (the presents Santa Claus gave). (24) That
elephant was real. (25) But it was small. (26) Then he lied down with the elephant. (27)
<Then the elephant>When it was night time (28) the elephant opened the door of the
refrigerator. (29) From there he opened the food. (30) He ate. (31) After that, after he ate,
after he ate all of them (32) and he started to grow large. (33) After that, after he ate (34)
his mother thought, {the child}her child, {the child} ate them. (35) But the child hadn’t
eaten them. (36) She was angry at the child. (37) But “I didn’t do it” he said. (38) “The
elephant did it” he said.
[Setting (23–25)-Action (26)-Setting (27)-Event (28–31)-Reaction (32)-Setting (33)-
Motivating State (34)-Event (35)-Attempt (36)-Consequence (37–38)]
This is a complete episode in which an event (the elephant eating all the food—utterances
#28–31) causes an attempt (the mother scolding the child—utterance #36) and a consequence
(the child giving an explanation—utterances #37–38). Furthermore, Berkay’s use of direct speech
with a child’s intonation to disambiguate the speaker shows how he resorts to mimetic ways of
expression in discourse.
In summary, 3-year-olds typically produced pre-episodic narratives under both conditions.
At age 4, episodic structures became prevalent in children’s prompt-elicited narratives whereas
 H. Ilgaz, A. Aksu-Koç / Cognitive Development 20 (2005) 526–544 539

in their direct-elicited narratives half of the children still produced pre-episodic formations. The
performance of the 5-year-olds did not differ between elicitation conditions and episodic structures
dominated their narratives.

4. Discussion

This study investigated age-related changes in the structure of narratives produced by preschool
age children in play-prompted and direct elicitation conditions. Pretend play was hypothesized
as a context that would facilitate narrative production in this age group. It was expected that
children would attain higher complexity structures at an earlier age in prompt-elicited narratives
as compared to direct-elicited narratives.
The results confirmed our hypothesis that action can serve as a symbolic arena to promote nar-
rative development. We found that children’s performance differed across elicitation conditions
with age. A shift towards increased narrative complexity was observed at 4 years in the play-
prompted context and at 5 years in the direct elicitation condition. Further probing of the data
showed that this change in complexity was from pre-episodic to episodic structures. Three-year-
old children were able to produce episodic formations occasionally, yet these were not dominant
in their narratives. What looked like stories with an episodic structure were imperfectly mod-
eled after fairy tales that children were familiar with. In contrast, episodic structure formations
were dominant in the narratives of 5-year-olds regardless of elicitation condition. Four-year-olds
occupied a transitional phase in that these children were capable of episodic formations but could
express their competency noticeably better in the prompt elicitation condition than in the direct
elicitation condition. The findings bear our expectation that children can produce higher com-
plexity structures at an earlier age in play contexts that provide action and objects to scaffold their
narrative activity than in contexts without such support.
These findings can be explained by reflecting on the definition of episodic structure used in
the present analysis. The central notion underlying the story grammar framework is the require-
ment of the presence of causal connections between a character’s goals, the actions used to
achieve these goals and the consequences of these actions (Mandler & Johnson, 1977; Stein
& Glenn, 1979). In the present study, episodic and pre-episodic structures were differentiated
based on the presence or absence of these causal connections. Namely, pre-episodic formations
were characterized by either temporal relations or very simple causal relations that hold together
only two structural components of the story such as the motivating state/event and attempt, or
attempt and consequence rather than all three. Episodic structures, on the other hand, required
two explicit causal relations between the three constitutive components, namely the motivating
circumstances, the attempt and the consequence. The difference between the younger and older
children’s performance in terms of episodic structures can, therefore, be sought in their ability
to bind the core components of their stories in a causal chain. As just discussed, children could
formulate this causal framework at an earlier age in play-prompt elicited than in direct-elicited
narratives.
Studies focusing on children’s developing play competence have shown that children under-
stand such relations in play contexts well before the age of 4. In comprehending short-play
episodes, 2- to 3-year-olds can track the pretense identities of toy props, relate actions to out-
comes and use inferential constructive processes (Harris & Kavanaugh, 1993). By age 4 children
can construct play plots with more than two characters and refer to their internal states in order to
motivate their actions (McCune-Nicholich, 1981; Wolf et al., 1984). Evidence thus shows that by
the age of 4, children have the competence to comprehend and produce episodic structures in the
540 H. Ilgaz, A. Aksu-Koç / Cognitive Development 20 (2005) 526–544

context of play. In fact, in their study on structural parallels between play and narrative, Eckler
and Weininger (1989) have shown that this competence was extended to narrative productions in
play context. Using an identical procedure to our prompt elicitation condition with thematically
related toys, and an identical coding procedure where enacted propositions were coded along with
verbalizations as episodic components, they observed that half of the 4-year-olds and about 90%
of 5-year-olds could produce episodic narratives, similar to the results obtained in the present
study.
It seems that contexts of play facilitate children’s narrative productions by also decreasing
the on-line processing demands. In the present study, the toy-props were two themed sets, which
supplied children with settings, objects and characters at the perceptual level. The toy props
enabled children to conceive of a variety of relations between the characters and objects by
triggering general world representations associated with these settings. Toy props also acted as
cues and anchors in the progression of children’s narratives, reminding them of the preceding
happenings and reflecting the current situation in the narrative. Hence, their use had the effect of
reducing the memory load required in narrative telling.
In telling a story, the child has to conjure up the components of his story such as the characters,
the goals, the actions and their consequences, to organize them in accordance with a narrative
schema and to express this verbally. That is, story telling requires a mapping between the con-
ceptual and linguistic levels of representation. In the present study, both conditions required such
activity, yet the play-elicitation condition, which provided children with a medium of action,
supported both the cognitive construction and the expression of their narratives. The results can
be explained by postulating an age-related development of a guiding story schema by which
story components are organized. This development, which reflected a qualitative change from
pre-episodic to episodic structures, was observed earlier in play-prompted context, revealing the
presence of a guiding story schema at 4 years of age.
The only study that we are aware of that used the same design as ours is Benson (1993)
who obtained results opposite to those of the present study. She compared story telling and
pretend-play narratives in 4- and 5-year-olds, using a framework of analysis that differentiated
between descriptions, sequenced narratives and plotted narratives. Her results showed an age-
related increase in the production of stories with a plot; 5% of 4-year-olds were able to produce
plotted narratives compared to 33% of the 5-year-olds. Interestingly, of the plotted narratives
66% occurred in the story telling session and not in the play session. These discrepancies between
Benson’s and our findings might be due to several factors. Of primary importance is the coding
scheme used to analyze children’s narratives. Although the definitions of a coherent narrative unit
are alike in the two schemes, they differ in terms of levels of structural complexity. Furthermore,
in the present study enacted propositions were coded along with the verbalizations of the child
as a component of the narrative, whereas in Benson’s study, action was not considered as a part
of the narrative that added to coherence. Hence, more complex narrative units could be identified
in the present study. A third difference is that the replica toys used as prompts in Benson’s study
were not related thematically. In the light of these conceptual and methodological differences,
the discrepancy between the two studies is not surprising. The prompt elicitation condition in
our study was specifically designed to tap the highest level of competence children may have by
providing them with maximum scaffolding with themed toy sets and by taking action as a medium
of expression.
When compared with previous studies using story grammar analysis, the level of complexity
of children’s narratives observed in the present study is, again, considerably higher. Peterson and
McCabe (1983), whose method corresponds to our direct elicitation condition, found complete
 H. Ilgaz, A. Aksu-Koç / Cognitive Development 20 (2005) 526–544 541

episodes to be present in all age groups ranging from 3½ to 9½, though with a prevalence rising
with age from 16 to 60%, respectively. Complex episodes were totally missing from the 4-year-
olds’ narratives. In contrast, in our direct elicitation condition, half of the 4-year-olds and almost all
of the 5-year-olds told narratives with complete episodes or more complex levels of organization.
Two factors may help explain these differences: genre and participants’ characteristics. Genre
has been found to play a determining role in the level of structural complexity of children’s
narratives. Peterson and McCabe’s data consist of personal narratives whereas our data, as well
as Eckler and Weininger’s data discussed above, consist of fictional narratives. Allen, Kertoy,
Sherblom, and Pettit (1994) who observe a difference between the two genres in data from
4- to 8-year-olds conclude that personal narratives exhibit more descriptive sequences, reactive
sequences and complete episodes whereas fictional narratives show action sequences and multi-
episodic structures. The different demands these two genres impose on narrators are also noted
by Ucelli, Hemphill, Pan, and Snow (1999). The participant’s characteristics we refer to is that
all participants in the present study came from middle-class families who reported reading to
their children extensively from a very early age onwards. Their preschool environments were also
very stimulating in narrative experiences. Peterson and McCabe’s participants, however, came
from working class homes, which according to the literature do not constitute highly stimulating
environments in this respect. The fact that such experience plays a critical role in the development
of narrative organization has been shown by studies comparing narratives of children from different
socio-educational backgrounds: middle-class children typically tell stories with more complex
episodic structures than their age mates from working-class homes where exposure of children to
storybooks is minimal (Aksu-Koç, 2005; Dickinson & Tabors, 2001; Goldman, Varma, Sharp, &
the Cognition and Technology Group, 1999).
Finally, it is useful to consider some aspects of the procedure that might have influenced the
quality of the narrative outcome in the present study. In both conditions, we started with a 10 min
warm-up session to allow the children to feel comfortable with the experimenter and explore the
toy props, which may have enhanced the quality of their narrative productions. In addition, the
instructions in our study explicitly asked for a story with the play-props whereas both Eckler and
Weininger (1989) and Benson (1993) required the child to have a pretend game with the toy-props
and tell the experimenter what they did throughout the play. This difference in instruction might
also have affected the results. The warm-up time and experimental instruction are among the
factors that may have contributed to the narrative productions of children in the prompt-elicited
condition.
In conclusion, the prompt elicitation method employed in this study aimed both at the practical
purpose of finding a means of eliciting narratives from preschool children effectively, and served
the theoretical purpose of demonstrating that action functions as a tool for children to express their
developing story competence. Most significantly, we found some evidence that action provides
an alternative means of expression, which can at times support and at times stand in place of
linguistic means. We, thus, propose that play functions as a semantic scaffold for the construction
of narratives with episodic structures by supplying the child with an action arena which (a)
enhances shift to a fantasy realm, (b) decreases memory demands and (c) provokes and enhances
the use of general world representations. As a last point, comparisons with other studies show that
despite the use of very similar frameworks of analysis, method of elicitation makes a difference
in revealing children’s potential levels of competence (for similar arguments, see Clark, 2004;
Vygotsky, 1978). Studies comparing play-prompted narratives with narratives elicited via different
methods to seek the cognitive and empirical differences these tasks instigate should provide fertile
ground to understand children’s developing story competence better.
542 H. Ilgaz, A. Aksu-Koç / Cognitive Development 20 (2005) 526–544

Acknowledgement

We would like to express our thanks and appreciation to Dr. Ageliki Nicolopoulou for her
constructive editorial comments on this paper.

Appendix 1. Coding scheme of propositions into episodic components

Peterson and McCabe’s (1983) coding categories were used in classifying verbal and enacted
propositions into episodic components with a few modifications. Introduction and closing com-
ponents were included in the present coding scheme since some children framed their story with
formal introductory and closing utterances. An abstract category was also added because, espe-
cially in the prompt-elicited narratives, some children gave a short outline of what would happen
next in the story before acting it out. The action and event categories were used to differentiate
between simple actions versus those that initiate a causal chain, respectively.

Introduction. Statements of formal story opening. (e.g., Once upon a time)

Abstract. Statements in which the child gives a summary of what is to happen next in the narrative
before proceeding to act it out or tell it.
Setting. Internal states, external states or habitual actions that serve to introduce the characters
and the physical or social environment.
Action. Behaviors of a character that follow each other temporally, but with no causal relation
between them.
Event. Natural occurrences, actions or environmental states resulting from actions that function
to initiate an episode.
Motivating state. Internal states, such as affects, cognitions or goals that motivate the protagonist.
Reaction. Either internal states that are precipitated by events, attempts or consequences and
do not motivate behavior, or purposeless actions that are precipitated by events, attempts or
consequences.
Attempt. Actions initiated by an event or a motivating state and are preparatory to goal attainment.
Consequence. Actions that directly achieve or fail to achieve a goal, or existing states once all
attempts have failed.
Evaluation. Statements in which the child steps out of the time frame and comments on the
narrated events.
Closing. The statements in which the child indicates that his/her story had to come to an end,
either with a formal story ending or a statement of closure.

Each utterance was classified as one episodic component. There were two exceptions to this
rule. (1) Dialogic utterances that expressed both a motivational state and an attempt at action
were coded as Motivating state and Attempt simultaneously. For example the utterance “I want
food” was coded as Motivating state and Attempt, since it expressed desire explicitly and attempt
implicitly. (2) Interactive episodes are characterized by the opposing struggles of two protagonists
and utterances have a different functional value from each protagonist’s point of view; therefore,
they are coded twice.

References

Aksu-Koç, A. (2005). Role of the home context in relations between narrative abilities and literacy practices. In D. Ravid
& H. Bat-Zeev Shyldkrot (Eds.), Perspectives on language and language development. Dordrecht: Kluwer.
 H. Ilgaz, A. Aksu-Koç / Cognitive Development 20 (2005) 526–544 543

Allen, M. S., Kertoy, M. K., Sherblom, J. C., & Pettit, J. M. (1994). Children’s narrative productions: A comparison of
personal event and fictional stories. Applied Psycholinguistics, 15, 149–176.
Bamberg, M. (1987). The acquisition of narratives: Learning to use language. Berlin: Mouton de Gruyter.
Bates, E., Benigni, L., Bretherton, I., Camaioni, L., & Volterra, V. (1979). The emergence of symbols: Cognition and
communication in infancy. New York: Academic Press.
Benson, M. (1993). The structure of four- and five-year-olds’ narratives in pretend play and storytelling. First Language,
13, 203–223.
Berman, R. A., & Slobin, D. I. (1994). Relating events in narrative: A cross-linguistic developmental study. Hillsdale,
NJ: Erlbaum.
Clark, H. H. (2004). Variations on a ranarian theme. In S. Strömqvist & L. Verhoeven (Eds.), Relating events in narrative:
Typological and contextual perspectives. Mahwah, NJ: Erlbaum.
Dickinson, D. K., & Tabors, P. O. (2001). Beginning literacy with language. Baltimore: Paul H. Brookes.
Eckler, J. A., & Weininger, O. (1989). Structural parallels between pretend play and narratives. Developmental Psychology,
25, 736–743.
Ely, R., Wolf, A., McCabe, A., & Melzi, G. (2000). The story behind the story: Gathering narrative data from children. In
R. Menn & L. N. Bernstein (Eds.), Methods for studying language production (pp. 249–269). Mahwah, NJ: Erlbaum.
Garvey. (1977). Play. Cambridge, MA: Harvard University Press.
Goldman, S. R., Varma, K. O., Sharp, D., & the Cognition and Technology Group, Vanderbilt. (1999). Children’s under-
standing of complex stories: Issues of representation and assessment. In S. R. Goldman, A. C. Graesser, & P. Van den
Broek (Eds.), Narrative comprehension, causality, and coherence: Essays in honor of Tom Trabasso. Mahwah, NJ:
Erlbaum.
Harris, P. L., & Kavanaugh, R. D. (1993). Young children’s understanding of pretense. Monographs of the Society for
Research in Child Development, 58(1), Serial No. 231.
Hudson, J. A., & Shapiro, L. R. (1991). From knowing to telling: The development of children’s scripts, stories, and
personal narratives. In A. McCabe & C. Peterson (Eds.), Developing narrative structure (pp. 89–136). Hillsdale, NJ:
Erlbaum.
Johnson, N. S., & Mandler, J. M. (1980). A tale of two structures: Underlying and surface forms in stories. Poetics, 9,
51–86.
Küntay, A. (1997). Extended discourse skills of Turkish preschool children across shifting contexts. Unpublished Ph.D.
dissertation, University of California, Berkeley.
Leondar, B. (1977). Hatching plots: Genesis of story making. In D. Perkins & B. Leondar (Eds.), The arts in cognition.
Baltimore, MD: Johns Hopkins University Press.
Lillard, A. S. (1993). Pretend play skills and the child’s theory of mind. Child Development, 64, 348–371.
Mandler, J. M., & Johnson, N. S. (1977). Remembrance of things parsed: Story structure and recall. Cognitive Psychology,
9, 111–151.
McCabe, A., & Peterson, C. (1991). Getting the story: A longitudinal study of parental styles in eliciting narratives
and developing narrative skill. In A. McCabe & C. Peterson (Eds.), Developing narrative structure (pp. 217–253).
Hillsdale, NJ: Erlbaum.
McCune-Nicholich, L. (1981). Toward symbolic functioning: Structure of early pretend games and potential parallels
with language. Child Development, 52, 785–797.
Merritt, D. D., & Liles, B. Z. (1987). Story grammar ability in children with and without language disorders: Story
generation, story retelling and story comprehension. Journal of Speech and Hearing Research, 30, 539–552.
Nelson, K. (1989). Monologue as representation of real-life experience. In K. Nelson (Ed.), Narratives from the crib (pp.
27–72). Cambridge, MA: Harvard University Press.
Nelson, K. (1996). Language in cognitive development. Cambridge, UK: Cambridge University Press.
Nicolopoulou, A. (in press). The interplay of play and narrative in children’s development: Theoretical reflections and con-
crete examples. In A. Göncü & S. Gaskins (Eds.), Play and development: Evolutionary, sociocultural, and functional
perspectives. Mahwah, NJ: Erlbaum.
Paley, V. G. (1981). Wally’s stories: Conversations in the kindergarten. Cambridge, MA: Harvard University Press.
Paley, V. G. (1990). The boy who would be a helicopter: The uses of storytelling in the classroom. Cambridge, MA:
Harvard University Press.
Pellegrini, A. D. (1986). Play centers and the production of imaginative language. Discourse Processes, 9, 115–125.
Pellegrini, A. D. (1987). The effect of play context on the development of children’s verbalized play fantasy. Semiotica,
65, 285–293.
Pellegrini, A. D., & Galda, L. (1982). The effects of thematic–fantasy play training on the development of children’s story
comprehension. American Educational Research Journal, 19, 443–452.
544 H. Ilgaz, A. Aksu-Koç / Cognitive Development 20 (2005) 526–544

Pellegrini, A. D., & Jones, I. (1994). Play, toys and language. In J. H. Goldstein (Ed.), Toys, play and child development.
New York: Cambridge University Press.
Perner, J. (1991). Understanding the representational mind. Cambridge, MA: MIT Press.
Peterson, C., & McCabe, A. (1983). Developmental psycholinguistics: Three ways of looking at a child’s narrative. NY:
Plenum Press.
Piaget, J. (1962). Play, dreams and imitation in childhood. NY: Norton.
Rubin, S., & Wolf, D. (1979). The development of maybe: The evolution of social roles into narrative roles. In E. Winner
& H. Gardner (Eds.), Play, language, and fantasy in childhood. New directions in child development No.6 (pp. 15–28).
San Francisco, CA: Jossey-Bass.
Rumelhart, D. E. (1977). Understanding and summarizing brief stories. In D. Laberge & J. Samuels (Eds.), Basic processes
in reading: Perception and comprehension. Hillsdale, NJ: Erlbaum.
Stein, N. L. (1988). The development of children’s storytelling skill. In M. B. Franklin & S. S. Barton (Eds.), Child
Language: A reader (pp. 282–297). NY: Oxford University Press.
Stein, N., & Glenn, C. (1979). An analysis of story comprehension in elementary school children. In R. Freedle (Ed.),
New directions in discourse processing (pp. 53–120). Norwood, NJ: Ablex.
Trabasso, T., Secco, T., & van den Broek, P. (1984). Causal cohesion and story coherence. In H. Mandl, N. L. Stein, & T.
Trabasso (Eds.), Learning and comprehension of text (pp. 83–111). Hillsdale, NJ: Erlbaum.
Trabasso, T., & van den Broek, P. (1985). Causal thinking and the representation of story events. Journal of Memory and
Language, 24, 612–630.
Ucelli, P., Hemphill, L., Pan, B. A., & Snow, C. (1999). Telling two kinds of stories: Sources of narrative skill. In L. Balter
& C. S. Tamis-LeMonda (Eds.), Child psychology: A handbook of contemporary issues. Psychology Press: Taylor &
French Group.
van den Broek, P. (1990). The causal inference maker: Towards a process model of inference generation in text com-
prehension. In D. A. Balota, G. B. Flores d’Arcais, & K. Rayner (Eds.), Comprehension processes in reading (pp.
423–446). Hillsdale, NJ: Erlbaum.
Vygotsky, L. S. (1978). Mind and society. Cambridge, MA: Harvard University Press.
Williamson, P. A., & Silvern, S. B. (1991). Thematic fantasy play and story-comprehension. In J. F. Christie (Ed.), Play
and early literacy development (pp. 69–89). NY: State University of New York Press.
Wolf, D., & Gardner, H. (1979). Style and sequence in early symbolic play. In N. R. Smith & M. B. Franklin (Eds.),
Symbolic functioning in childhood. Hillsdale, NJ: Erlbaum.
Wolf, D., Rygh, J., & Altshuler, J. (1984). Agency and experience: Actions and states in play narratives. In I. Bretherton
(Ed.), Symbolic play: The development of social understanding (pp. 195–218). NY: Academic Press.