Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Let’s pretend: an investigation of the language used during children’s collaborative pretend play D’Orazio, Diana Patrizia 1994

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata


831-ubc_1994-0260.pdf [ 2.96MB ]
JSON: 831-1.0094721.json
JSON-LD: 831-1.0094721-ld.json
RDF/XML (Pretty): 831-1.0094721-rdf.xml
RDF/JSON: 831-1.0094721-rdf.json
Turtle: 831-1.0094721-turtle.txt
N-Triples: 831-1.0094721-rdf-ntriples.txt
Original Record: 831-1.0094721-source.json
Full Text

Full Text

LET’S PRETEND: AN INVESTIGATION OF THE LANGUAGE USED DURING CHILDREN’S COLLABORATIVE PRETEND PLAY by  DIANA PATRIZIA D’ORAZIO B.A., The University of Calgary, 1991  A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTERS OF SCIENCE in THE FACULTY OF cRADUATE STUDIES (Department of Audiology and Speech Pathology)  We accept this thesis as conforming to the required standard  THE UNIVERSITY OF BRITISH COLUMBIA April 1994 Diana Patrizia D’Orazio, 1994  In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission.  (Signatu  Department of  peech  The University of British Columbia Vancouver, Canada Date  DE-6 (2/88)  / t apv  cil 2 /  .>ci ec  ABSTRACT  This research examined preschool children’s use of language during social pretend play. Specifically, the ways in which language serves to (a) create and maintain a fantasy world, and (b) maintain a conversational interaction were compared across two age groups. Twelve children were paired and the six dyads were split into two age groups  --  the mean age for each group was 3;1O and 5;1, respectively. Transcribed  pretend play conversations were analyzed at the utterance level for specific content and for the children’s speaker identity. At the turn exchange level, the dialogue was examined in terms of discourse continuity and nature of collaboration. Preschool children use language to coordinate their independent pretend ideas. They create and sustain a novel, fictional reality by responding to, and building on, their partner’s ideas and actions as the play unfolds. Similarities in the use of language among the children in the two age groups provided information about the character of social pretend play. The majority of language used during pretend play was devoted to its construction and maintenance, rather than to the enactment of the pretend play. All children demonstrated the ability to talk about pretending and the pretend world while simultaneously assuming a fictional identity, as well as while holding their everyday identity. Most in-role collaboration consisted of elaboration of the play by expanding the partner’s contributions. In contrast, out-of-role collaboration involved negotiation for the resolution of disagreements. During pretend play, all children demonstrated the ability to relate to one another by providing a contingent response about half of the  11  time, and by obliging the partner to respond approximately one third of the time. Additionally, the results suggested a developmental progression in the children’s ability to direct the play while simultaneously assuming a fictional identity. Older children produced a higher proportion of in-role communication for the construction and maintenance of the pretend world than younger children. As well, older children collaborated within the play frame more than younger children.  111  TABLE OF CONTENTS  ABSTRACT  ii  TABLE OF CONTENTS  iv  LIST OF TABLES  vii  ACKNOWLEDGEMENTS  viii  CHAPTER 1  2  INTRODUCTION  1  Overview  1  Outline  1  The Emergence of Pretend Play and the Integration of Pretense into Peer Interactions  2  Properties of Social Pretend Play with Peers  3  The Language of Social Pretend Play  12  The Language of Sociodramatic Play: Two Perspectives on Sociodramatic Play  22  The Current Research Project  24  METHODOLOGY AND DESCRIPTION OF CODING TAXONOMY  30  Overview  30  Subjects  30  Setting and Apparatus  31  Procedure  32  Data Preparation  34 iv  3  Introduction to the Coding Scheme and Considerations for Use  36  Category Description for Coding Purposes  43  Summary of Coding System  74  RESULTS  76  Overview  76  Use of Language to Construct and Maintain the Pretend Play  76  Use of Language to Serve Pretend Play Interaction  85  Collaborative Interaction  4  Summary of Major Results  95  DISCUSSION  99  Overview  99  The Language of Sociodramatic Play  99  Pretend Communication to Create and Maintain the Pretend Worlds  100  Pretend Communication as Conversation  113  Collaboration  118  Methodological Observations  125  Sociodramatic Play as a Context for Learning  126  Directions for Future Research  128  Conclusion  130  V  REFERENCES  131  APPENDIX A  Play Style Ratings Scale  136  B  Flow Chart of Coding Decisions: Pretend Communication about the Pretend World and Pretending  137  C  Flow Chart of Coding Decision: Pretend Play Communication as Conversation  139  D  List of Codes  141  E  List of Specific Coding Conventions  145  vi  LIST OF TABLES  TABLE 1  Mean Proportion (and Standard Deviation) of Metaplay Utterances by Content Type of Young and Old Dyads  2  Mean Proportion (and Standard Deviation) of Metaplay Utterances by Speaker Identity Position of Young and Old Dyads  81  3  Mean Proportion (and Standard Deviation) of Metaplay Utterances by Content Type and Speaker Identity Position of Young and Old Dyads  83  4  Mean Proportion (and Standard Deviation) of In-Role Metaplay Utterances grouped by Content Type of Young and Old Dyads  84  5  Mean Proportion (and Standard Deviation) of Responsive Turns of Young and Old Dyads  86  6  Mean Proportion of Reponsive Turns following Obliging or Nonobliging Turns  87  7  Mean Proportion of Turns at Various Degrees of Continuity of Young and Old Dyads  8  Mean Proportion (and Standard Deviation) of Collaborative Exchanges by Type of Young and Old Dyads  91  9  Mean (and Standard Deviation) of Collaborative Exchanges by Speaker Identity Position of Young and Old Dyads  93  10  Mean Proportion of Elaborative and Negotiative Exchanges for each Speaker Identity Position  94  vii  ACKNOWLEDGEMENTS  I wish to sincerely thank the following people: Judith Johnston, my main supervisor, for her guidance, patience, and insight over the past year, for her ability to make anything seem possible when everything seemed impossible, and for instilling in me more confidence than I ever thought I had. Lisa Avery and Carolyn Johnson, members of my supervising committee, for the time they dedicated to this project. Especially, for their positive words and helpful hints, for their thoughtful questions and comments, and for their careful editing. The children, parents and staff of Southview preschool, and Elizabeth McLeod and friends, for their cooperation. My classmate and friend, Lisa Wollitzer, for “volunteering” to code a portion of the transcripts for reliability purposes. Neither a small nor simple task! My thesis-mates for their understanding and inspiration. Their contribution to this project is greater than they know. My friends, new and old, for helping me keep a smile on my face. Finally, I wish to extend special thanks to the family I left behind, and the one I became a part of three years ago, for their generous support and encouragement. And most of all, to my mom and dad for teaching me to appreciate the rewards of hard work.  vi”  CHAPTER 1 INTRODUCTION Overview Children are commonly engrossed in pretend activities alone, with an adult or with another child. Although seemingly simple, pretending is a rather sophisticated task, especially social pretend play among preschool peers (Garvey, 1992, Howes, 1992). Social pretend play has received considerable research interest from a variety of theoretical branches in the past, particularly in terms of child development (Fein, 1981). Behavioral changes in the pretend play of children provide insight into the development of cognition, socialization and communication. In fact, the context of social pretend play is viewed as a rich and valuable context for learning  --  both for children and  investigators of child development (Levy, 1984, Faver, 1989). Upon observation, one thing immediately apparent is that talk, or language, is an integral part of this activity. Language is a tool used by the players to communicate during pretend play. Children use language to jointly create, establish, and sustain the fictional world in which to play. They formulate and express their own play ideas, attend to their partners’ play ideas, and coordinate the activities of the ever-changing play world. The purpose of this study is to investigate and compare how children between the ages of 3 and 5 verbally collaborate with peers in pretend play.  Outline This chapter reviews the literature related to the area of interest, and reports on research  1  pertinent to the design of the study. A brief overview of the character and development of sociodramatic play is provided first, followed by a discussion of the language used in this context which includes a review of the existing classification schemes and their application. Then a reflective interpretation of research that motivates this project is presented. The chapter ends with a list of the research questions and their accompanying hypotheses. Chapter two describes the methodology used to collect and analyzed the data. The results of the study are reported in chapter three, and a discussion of the results can be found in chapter four.  The Emergence of Pretend Play and the Integration of Pretense into Peer Interactions The play of infants and toddlers is primarily exploratory and manipulative during the sensorimotor stage of development. Solitary play involving sensorimotor practice begins to take the form of pretend play at about 12 months of age. At this time, play begins to replicate behaviours observed in real-life activities, e.g., going to sleep or drinking from a cup. (See Fein, 1981; Lyytinen, 1991; Nicholich, 1981; or Rubin, Fein & Vandenberg, 1984 for a review of the components and development of early symbolic play.) The cognitive development underlying this behavioral change is the transition from sensorimotor to representational thought (Piaget, 1962). During the third year, two toddlers may engage in pretend play side by side, and may even attempt to assume a role (e.g., “mommy”), but with little awareness or understanding of the playmate’s part in the game. With increased social knowledge and competence, older two-year olds begin to assume complementary roles (e.g., “mommy” and “baby”) and share in  2  make-believe themes (e.g., playing house, or gas station) (e.g., Forys & McCune Nicholich, 1984; Dunn & Dale, 1984). Cooperative social pretend play gradually emerges around age 3 and reaches its  peak during the late preschool years (e.g., Fein, 1981; Garvey, 1977/1992; Howes, 1992; Piaget, 1962; Rubin et al., 1984; Wall, Pickert, & Gibson, 1989). The predominant features of social pretend play are: (a) the transformation of one’s personal identity (i.e., role-taking) and (b) the verbal and nonverbal interaction between the participants (e.g., Fein, 1981; Garvey, 1984; Smilansky, 1968). Howes (1985) notes two important elements of social pretend play that differentiate it from social play: (a)  an understanding that the partner is acting out a role as part of a story and (b) communication about the meaning of one’s actions.  Properties of Social Pretend Play with Peers This section begins with definitions of terms commonly used in discussions of this phenomenon and outlines their application within this paper. Then an overview of the properties and development of social pretend play is presented.  Definitions and Terminology Pretend Play Since pretense is central to this investigation, it seems appropriate to begin with a definition of the theoretical construct and its behaviourial characteristics.  Garvey  (1977/1992), a leading researcher in the area of pretend play, defines the construct of  3  pretense as behaviour in a simulative, nonliteral or “as if” mode. Similarly, McCune Nicholich (1981, p.52) describes the dual quality of the behaviour as “the juxtaposition of a real action and the intended make-believe meaning.” The “transformational quality” of the children’s pretend acts is the defining attribute of such behaviour (Fein, 1975). The identification of an activity as pretend relies on the contrast between the simulation and the behaviour from which it is derived (e.g., “cooking” is simulated by moving a spoon around in a pot that sits on top of a stove). Upon observation such a contrast is easy even for a naive onlooker. Several behavioral criteria have commonly been employed by investigators (e.g., Mathews, 1977; McLoyd, Warren & Thomas, 1984; McCune-Nicholich, 1981) to identify children’s behaviour as pretend: (a) performance of an everyday activity in the absence of necessary materials, e.g., bathing a doll in a plastic bucket without water; (b) performance of an action usually done by someone else, e.g., washing the dishes; (c) attribution of animacy to inanimate objects, e.g., feeding a doll; (d) performance of an activity in the absence of the usual outcome, e.g., a child waves good-bye, but does not leave the room; (e) substitution of one object for another, e.g., a stick is used as a magic wand; (f) exaggerated gestures, animated demeanour, or affective tone used to signal the pretend behaviour (Fein, 1981; McCune Nicholich, 1981). A variety of terms, such as imaginative play, make-believe play, fantasy play, dramatic play, and fictional play, have been used in discussions of pretend play, both solitary and social. More specifically, though, pretend play involving more than one  4  child is commonly referred to as social or interactive pretend play, or sociodramatic play. In the present discussion, these terms are used synonymously.  Metacommunication/Metapretend Communication During sociodramatic play children state or allude to the intended meaning of their actions through words and signals. Statements such as “You be the cooker” or “Pretend this is my magic earring” provide information about how nonliteral, pretend behaviours should be interpreted. Bateson argued, “this phenomenon, play, could only occur if the particular organisms are capable of some degree of metacommunication, i.e., of exchanging signals which would carry the message ‘this is play” (1955, p. 41). From an anthropological perspective, Bateson viewed play as communication, and as such, labelled signals or talk about the play as metacommunication . t  Many  researchers interested primarily in the language of social pretend play have adopted Bateson’s view of metacommunication. However, considering that play is not ordinarily construed as communication in essence, and that metacommunication is now typically reserved for talk about talk, the term metacommunication will not be adopted in this discussion. As it would be unwieldly to use the phrase “language that serves to construct, manage and sustain the fictional play reality” in lieu of metacommunication,  the term metaplay communication was coined. Metaplay communication serves a variety of purposes, ranging from initiating the  1  The prefix “meta” is commonly understood as “an awareness of” or “thought or talk about”.  5  play, to making plans, transforming objects and identities, and signalling the end of play (e.g., Garvey & Berndt, 1977; Goncu & Kessel, 1984). The general category of “talk about the play” or metaplay communication can be split into two types: those utterances that explicitly refer to the pretense (i.e., transformations of the here and now), and those that simply initiate, manage or end the play activity. Most researchers have included both types in their discussion and analyses; therefore, unless otherwise stated, metaplay communication refers to the more general category.  Negotiation and Enactment In the literature, the term negotiation has often been used in contrast to the term enactment (e.g., Connolly & Doyle, 1989; Garvey & Krammer, 1989). These terms refer loosely to talk about the pretend play, and talk for the performance of the pretend play, respectively. In the current research project, however, these terms are defined in a more narrow sense, and used technically as category labels for coding purposes.  Negotiation consists of talk used only to settle disagreements, for example, “No I’ll cook, you’re not old enough.” The term enactment is reserved for utterances which function purely to perform the drama, as exemplified by the “waiter’s” question, “What can I order for your meal?”  The Development of Sociodramatic Play This section provides an overview of the developmental progression of sociodramatic play. Before describing developmental changes related to social pretend  6  play in terms of thematic content, social interaction and communication, the controversy surrounding the typical age range associated with such play is discussed. The integration of pretense into social interactions with peers has been reported to appear in the latter part of the third year (e.g., Garvey, 1982; Howes & Unger, 1992; Sachs, Golman & Chaille, 1984). Conflicting reports regarding the earliest age at which children participate in sociodramatic play reflects individual differences between children, as well as differences in methodology employed by the researchers. Factors such as criteria used for the identification of social pretense (Howes, 1985), the age and familiarity of the play partner (Doyle, Connolly, & Rivest, 1980; Dunn & Dale, 1984), and structure of the play setting and task all have an impact on the results of the study. Reliable identification of early social pretend play depends on a clear definition of social interaction that includes the requirement of verbal interaction. Despite discrepancies regarding the age at which rudimentary skills of interactive pretend play are first observed, there is overwhelming agreement that once mastered, the amount and complexity of this kind of play increases with age (e.g., Field et aL, 1984; Fein, 1981; Garvey, 1992, Garvey & Berndt, 1977; Howes, 1992; Sachs et aL, 1984; McLoyd et al., 1984; Rubin, Fein & Vandenberg, 1984). However, findings regarding the shape of the developmental curve and the age at which sociodramatic play declines are inconsistent. Rubin et al (1984) reported that some researchers have found an inverted-U curve between the ages of 3 and 6 for make-believe play involving actual objects in the environment (e.g., playing house), while others have found a linear increase for fantasy play (e.g., “superheros”) from 1 to 5 years of age. Still other  7  investigators have reported a continued increase in the frequency and complexity of fantasy play up to age 6 1/2 (e.g., Wall et aL, 1989).  Structural Properties of Sociodramatic Play The notion that play becomes increasingly dramatic and increasingly social is apparent along a number of veins. Changes in the thematic content of the play, object substitutions, role-taking, the interpersonal structure of the social interaction, the use of language during pretend play, and to a lesser extent the relationship between these components, are reported in the literature. A common methodological strategy adopted by researchers in the past has been to observe children during free play in a naturalistic environment, rather than in a structured experimental setting. This section reviews those studies which have explored the structural properties of free play among peers. Two forms of social pretend play have been differentiated on the basis of the roles and activities of the players. The terms associated with these two distinct types  are sociodramatic and thematic fantasy play (Rubin et al., 1984). In a strict sense, sociodramatic play refers to play in which children adopt realistic roles and engage in activities that are familiar and related to real life (e.g., domestic or stereotypical occupational play). Realistic roles include: (a) functional or behavioral roles defined by the player’s behaviour, e.g., “truck driver”, (b) relational roles (i.e., family roles, e.g., “daddy”) or (c) character roles (i.e., stereotypical occupations, e.g., “policeman”) . 2  2  The terminology referring to different role types is taken from Garvey and Berndt (1977). In other discussions of role playing (e.g., McLoyd, 1984), researchers adopt an earlier set of terms, anticipatory and fantastic roles, from Stone (1962). 8  Thematic fantasy play, on the other hand, involves roles based on the fictional characters of stories or movies (e.g., “Batman” & “Robin”, “Aladdin” & “Jasmin”) and the children engage in events further removed from their everyday experiences. The earliest forms of play involve roles that children have experienced or might encounter later in life (e.g., “husband” and “wife”, “doctor” and “nurse”) (Garvey & Berndt, 1977; Mathews, 1977). Older children (4 to 5 year olds) continue to adopt such realistic roles, but also assume roles seldom or rarely experienced in real-life (e.g., “pirates” or “princesses”). Saltz, Dixon, & Johnson (1977, as reported in Rubin et al., 1984) implied that thematic fantasy play is more mature because it is supported less by the concrete, immediate environment and places greater demands on symbolic representation and imagination than sociodramatic play. While role-taking, children engage in activities in accordance with representations of the social roles adopted. Investigators (Garvey & Berndt, 1973; Mathews, 1977) have noted that while the activities performed by younger players were closely associated with the roles being enacted, older children were more likely to depart from these conventional activities (e.g., the “mother” may clean the house and repair the broken stove). Furthermore, violation of role-appropriate behaviours by younger children often resulted in the abandonment of the role or the breakdown of play. For instance, if the player who has adopted the role of “baby” wants to push the doll carriage, that player may announce that she is now the “daddy.” Interactive role playing often involves reciprocal roles, whereby the social relationship of the assumed role is complementary (e.g., “doctor” and “patient”). It is  9  likely that the role relation of reciprocal or complementary roles facilitates interaction between the players. Iwanaga (1973) extended Parten’s (1932) findings that social participation increases with rising age among 3 to 5 year olds by considering the nature of children’s cooperation. Interpersonal play structures, or the ways in which children structure their play interactions with peers, was found to progress from complementary to integrative with age. The 4 year olds of Iwanaga’s study engaged in complementary play, whereby the players independently enacted different (often reciprocal) roles. These children interacted only to the extent necessary to achieve the goal of the game. For example, two children adopted the roles of “conductor” and “passenger” and basically enacted each role on their own, except for the times when the “conductor” stopped the train for the “passenger” to get on or off. Whereas, the 5 year old children paid much more attention to the behaviours of the partner, and adjusted their own behaviours in response to the partner. Such play constituted an integrative interpersonal play structure. A similar developmental pattern has been reported by Howes (1985, 1987; Howes et al. 1989). However, the coding scheme developed by Howes did not capture the increased frequency and complexity of complementary play between the ages of 3 and 5 because only the presence or absence of turn-taking was considered. Before moving on, a brief summary of the developmental changes described thus far is provided. Research shows that social pretense becomes progressively more interactive and that the representations become increasingly complex as children get older. The thematic content of the drama becomes more diversified and further removed from the immediate environment as children adopt fantastic roles, and perform  10  activities that are divergent from conventional role expectations (Connolly et al., 1980; Field et al., 1982; Garvey & Berndt, 1977; Mathews, 1977; McLoyd et al., 1984; Sachs et aL, 1984). Parallel to the shifts in cognitive maturity, improvements in social competencies are evident by the fact that the structure of children’s play becomes more integrative (Howes et al. 1989; Iwanaga, 1973; Mcloyd et al., 1984). Advanced forms of sociodramatic play appear to be enhanced by language (e.g., Field et al., 1982; McLoyd et al., 1984). Some authors have suggested that language is crucial in this shift towards greater cooperation among the players (e.g., Sachs et al., 1984). In fact, several studies of children’s pretend play dialogues have shown age-related changes in its structure during the preschool years. For instance, Garvey (1974) identified two primary ways in which preschoolers respond to their partners’ play turn: they either repeat or complement the prior turn. According to Garvey, repetitions acknowledge the partner’s contribution, as in Child A: “Bye Mommy,” Child B: “Bye Mommy,”. Complementary turns imply both accurate interpretation of the partner’s intentions, and appropriate extension of the play. For instance, Child A may announce, “I have to go to work,” to which Child B responds, “You’re already at work,” followed by a rejection by Child A, “No, I’m not.” Garvey concludes that in the age range of 3 and 5 years, older children are more likely to complement, and younger children are more likely to repeat, their partner’s utterances.  11  The Language of Social Pretend Play Prior to Garvey and Berndt’s (1977) study of the organization of pretend play, very little attention was given to the communication of pretend play (Rubin et at. 1984). More recently, studies of social pretend play have focused on how language establishes  and maintains the pretend state and the social interaction. In particular, differences in the ways language is used in this context by children between the ages of 3 and 5 have become a focus. The literature related to such research interests is presented below.  Communication about the Pretend World and Pretending In social pretend play, as in all social interactions, the participants must develop a shared understanding of the contextual circumstances of their interaction. Unlike other social interactions, however, the context of pretend play is often novel and relatively free from the current situation. Thus, the players must jointly define the frame that surrounds the imaginary world by transforming time, space, persons, objects, and actions, and by establishing the rules effective during the interaction (see Goffman, 1974, for a detailed explanation of framing). Signals and statements that send the message “this is play/pretend” mark a change in the form of the interaction from realistic to pretend, and outline the frame in which the nonliteral behaviours are to be perceived. Bateson (1955) argued that metacommunicative signals stand apart from the within-frame verbal exchange. In essence, the play frame, like a picture frame, draws the observer’s (partner’s) attention to the content within the frame, thereby excluding everything outside the frame. For  12  some researchers, metacommunication establishes the context of the social interaction (e.g., Giffin, 1984; Goncu & Kessel, 1984). Following Bateson’s (1955) theory of play and metacommunication, researchers have contrasted utterances that initiate and manage the pretend play with those that function to actually perform the play (Garvey & Krammer, 1989).  For them,  metaplay communication functions to frame the reality of the imaginary world, while enactment functions to substantiate that reality. Nevertheless, investigators recognize that the message, “this is pretend,” is implicit in the words and behaviours expressed within the play frame (Goncu, 1993b). That is, when the child who is pretending to be the “patient” says to the “doctor,” “Oh I’m sick,” two messages are delivered: one about the nature of the activity, and another about the content of the pretend play. But, talk for enactment within the play frame is not technically metaplay communication. The understanding that enactment is “just pretend” comes from other metaplay communication.  Schemes for Classifying Social Pretend Play The discussion that follows describes a variety of classification schemes which have been generated in the past to describe the language of sociodramatic pretend play. Those which form the basis of the taxonomy used in the current study are described in greater detail in the next chapter.  Henceforth the term metaplay communication will be used in lieu of metacommunication. Recall that metaplay communication refers to communication that serves to construct, manage and sustain the pretend reality. 13  Considering the entire range of communication related to the play theme, Garvey  and Berndt (1977) described five types of pretend communication. Some messages indicate a state of pretending, such as (a) enactment (i.e., use of words and actions characteristic of the adopted role); and (b) signals that mark and define the behaviour as pretend (e.g., smiles, giggles and facial expressions). Other messages indicate a transition to or from a state of pretending: (c) negation of pretend, including termination of the pretend play; (d) procedural and preparatory behaviour, including invitations to play; and (e) explicit mention of pretend transformations, such as roles, plans and objects. Schwartzman (1978), who investigated the verbal interaction among peers as an index of children’s social group status, expanded Garvey and Berndt’s list to nine categories. Whereas, Forbes, Katz and Paul (1986) suggested a distinction among three phases of play: (a) set-up, in which the elements of the drama are introduced, negotiated  and assembled; (b) enactment, in which the drama is played out; and (c) transition, sequences which connect two pretend episodes. The categories used by Garvey and Berndt (1977), Schwartzman (1978), and Forbes et al. (1986) suggest that play directing  --  metaplay communication  --  is outside  the play frame; it functions as part of the everyday world about the pretend world. Whereas, play acting  --  talk for the dramatic performance or enactment  --  is inside the  frame; it functions as part of the pretend world. This simple dichotomy between setting-up or directing the pretend play in the real world and performing the pretend play in the imaginary world, however, does not appreciate the fact that children are capable of directing and directing the play simultaneously.  14  Using data collected in a natural setting, Giffin (1984) generated a variety of categories to describe how children direct and coordinate the play activities. Both verbal and nonverbal modes of interaction were considered. Each component of the behaviour was judged independently as either functioning inside or outside the play frame. As a result, certain types of metaplay communication were viewed as functioning partially in-frame and partially out-of-frame. For example, a child may prompt his partner by using a lowered intensity voice while enacting his character role nonverbally. Because the nonverbal behaviour serves to maintain the reality within the play frame, the flow of the pretend play is not disturbed. The main contribution of Giffin’s taxonomy was the recognition that children have the option of directing the play from within the play frame. The categories were ordered along a continuum 4 ranging from completely in-frame (i.e., both verbal and nonverbal components functioned inside the pretend world), to partially in-frame and partially out-of-frame, to completely outof-frame. Talk about the pretend reality while simultaneously assuming a fictional identity was also a focus of Auwarter’s (1986) investigation of the puppet play of 3 to 10 year old girls. Utterances were coded along two dimensions: scope of validity (i.e., utterances may pertain to the make-believe reality, to real-life, or to a transition between the two realities) and speaker identity (i.e., child talks as himself or as an adopted identity). The intersection of the categories along these two dimensions made  Giffin’s categories consisted of only utterances which functioned as implicit or explicit pretend transformations. ‘  15  up “nine levels of reality.” It seems, then, that children manage and sustain the pretend experience by directing the play from within the play frame, as well as from outside the play frame (e.g., Auwarter, 1986; Giffin, 1984).  Developmental Changes in the Use of Language Most research about metaplay communication is descriptive in nature and based on qualitative analyses (e.g., Garvey & Berndt, 1973; Giffin, 1984, Sachs et al., 1984). These reports are valuable, nonetheless, because they provide a basis for generating classification schemes for the communication of the pretend play and suggest general developmental shifts in behaviour. For example, Garvey and Berndt (1973) noted differences in the content of metaplay communication as a function of age. Specifically, communication about the partner’s role (e.g., “Are you going to be the bride?”) and plans (e.g., “Pretend you hated baby fish.”) were found to increase with age. Auwarter (1986) described age-related changes with regard to the speaker identity position assumed by the children while constructing the play. Younger children (aged 3; 1 to 3; 8) only talked about the puppet play while assuming their own personal identity. Utterances that served to build-up the fictional reality from an in-role position were used most by children around 4 and 5 years of age. Whereas, the oldest children (aged 7;8 to 10) were observed to generate the fictional reality by mere presupposition, which presumes a high degree of shared knowledge about the script. General information about the changes in use of language during sociodramatic play between the ages of 2 and 5 also comes from quantitative studies. For example,  16  most 2 year olds observed by Fein, Moorin, and Enslein (1989) were reportedly less likely to use language to serve a metaplay purpose than older children. At around 3 years old, children were observed to explicitly express their plans and desires for the make-believe, e.g., “I’m gonna call my boyfriend now” (Field et al., 1982; Miller & Garvey, 1984). Also, the progression towards announced pretend play whereby children use phrases such as “Let’s pretend” has been reported to occur between 3 and 5 years old (e.g., Garvey & Krammer, 1989). Goncu and Kessel (1988) attempted to address Garvey and Berndt’s (1973) qualitative findings that older children more than younger children mentioned joint plans or partner-role transformations. These descriptive reports were found not to be statistically reliable by Goncu and Kessel (1988). However, the proportion of self object claims used by younger children was significantly higher than that used by older children. Besides the limited information about age-related changes in the language of pretending, there is little information about the intersection of the various dimensions of the metaplay communication, such as specific function types and speaker positions. The current investigation examines precisely such relationships: Are plans, object transformations, play descriptions, etc., more likely to be produced before or after a fictional identity has been assumed? Coding utterances in terms of metaplay function, however, does not convey much information about the structural properties of the conversational exchange. Information about how children respond to their partner’s contributions or how the partner’s involvement is extended in subsequent turns is not provided. As noted by  17  Goncu and Kessel (1984) there is little information available about how children maintain conversation with peers (as opposed to studies which examined the amount of social interaction [e.g., Howes, 1985], or the use of communicative strategies, such as calls for attention or tag questions [e.g., Faver, 1989]). A few studies (Garvey, 1977; Garvey & Hogan, 1973; Keenan, 1974; Mueller, 1972), however, have shown that children as young as three do engage in connected conversation. A brief review of research that investigated the conversational properties of social pretend play will be provided next.  Pretend Play Communication as Conversation Mueller (1972) observed previously unacquainted peers, between the ages of 3;6 and 5;6, engaged in dyadic pretend play. The “social effect” of individual utterances on the listener’s behaviour was judged. A successful exchange was one where the listener clearly provided a contingent response, such as appropriate nonverbal gestures (e.g., shaking one’s head “yes” or “no”), carrying out an action requested by the speaker (e.g. “Get me that doll”), or requesting repetition (e.g. “What did you say?”). For Mueller, success did not necessarily involve a highly efficient exchange of information. Contrary to initial predictions, conversational success was not found to improve with increasing age. Overall, the majority (62%) of the utterances received a definite response, 23% attracted the listener’s attention and only 15% failed to elicit a response. Factors found to facilitate or impede the success of a social exchange reflect characteristics of the preceding utterance, such as clarity of the utterance, use of  18  attention getting devices, and distance between the players. In Mueller’s study, the antecedent behaviours were coded for predictability of success and frequency of occurrence. The two most powerful predictors of success were: (a) nonspeaker’s attention and (b) direct response by the speaker to the nonspeaker’s prior turn. When frequency of occurrence was also taken into account, form of the utterance (e.g., command or question) and nonspeaker’s attention had the greatest overall utility. Mueller concluded that children as young as 3 1/2 have acquired the processes found to be important in the maintenance of verbal exchanges. Garvey and Hogan (1973) compared the amount of talk or nonverbal behaviour that was strictly adapted to behaviour of the partner used by children between 3;6 and  5. As in Mueller’s study, children were found to use language to achieve and maintain social contact with each other. Although the overall mean percentage of social speech  (59%) was similar to that reported by Mueller, Garvey and Hogan found that the proportion of social speech increased with age (young: 21 %  -  64%, old: 48%  -  77%).  (Perhaps, Mueller’s criteria for “social effect” was more lenient than Garvey’s definition of “social speech.”) Further age-related differences were evident in the length of continuous sequences produced by dyads in each group. Although single exchanges were by far the most frequent for all dyads, a larger proportion of older dyads produced longer sequences. For example, 11 of the 12 older dyads produced sequences of six exchanges in length; whereas, only three of the six younger dyads produced sequences of this length. Unlike Mueller (1972) and Garvey and Hogan (1973), who examined only the  19  consequence of turns on the following turn, Goncu and Kessel (1984) considered a  turn’s connection to the preceding and following turns. Categories generated by Torrenace and Olson (1982, as reported in Goncu & Kessel, 1984) were adopted. The four categories in order of decreasing complexity were: turnabouts (responding to and extending the partner’s prior turn, e.g., “Mom, can I go out to our river?” “Yeah sure and bring your binoculars”), responses (acknowledging and adapting to the partner’s prior turn, e.g., “Do you want hotdog for dinner?” “No”), mands (directing the partner’s future contribution, e.g., “Get those dishes over there”), and unlinked utterances (e.g., “This phone isn’t real” “I’m building a house”). The investigation of 3 and 4 1/2 year olds’ pretend play dialogues revealed a developmental trend towards greater continuity and coherence; the older children made significantly more turnabouts and fewer unlinked utterances than the younger children. These findings complement Garvey’s (1974) observation that the older children were more likely to interpret and extend their partners’ turns, while younger children mostly repeated their partners’ utterances or actions. The characteristics of preschoolers’ conversations during pretend play are summarized here. During the fourth year, children begin to master rudimentary conversational skills as evidenced by the use of social strategies which facilitate the success of a social communicative exchange (Mueller, 1972) and the ability to maintain a conversational interaction beyond a simple exchange (Garvey and Hogan, 1973). A developmental progression towards the use of more complex turns to maintain the conversation has also been suggested (Garvey, 1974; Goncu & Kessel, 1984, 1988).  20  The current study will adopt a coding scheme similar to that of Goncu and Kessel (1984, 1988) in an attempt to replicate the findings of these investigators.  Pretend Play Conversation  --  An Experience in Collaboration  Next, the focus of the discussion will shift slightly to examine how the pretend play is collaboratively achieved. Like adult conversation or children’s nonplay conversation, pretend play conversation consists of expansion and/or incorporation of the partner’s ideas, and negotiations about conflicting ideas. Reported results suggest that older children jointly plan, establish, and enact more sophisticated pretend play episodes by way of extensive verbal negotiations and elaborations (Field et aL, 1982; Howes, 1992; Garvey & Bemdt, 1977; Giffin, 1984; Goncu & Kessel, 1984; Sachs et aL, 1984). Only a few taxonomies have included categories which capture the notion of negotiation at the utterance level. For example, Garvey and Berndt (1973) identified a variety of negation statements, and Goncu and Kessel (1984) included terminations, negation statements and acceptance statements. More recently, Goncu (1993b) has paid particular attention to the social nature of children’s collaborations. Establishing a consensus about the pretense can be accomplished in different ways: by expanding the play, by expressing agreement or disagreement with the partner’s ideas, by emphasizing one’s own ideas, or by making comments which are not relevant. The relative frequency of “expansion acts” was found to increase in the conversations of older children (4 1/2 year olds), while the relative frequency of “emphases” was found to decrease. This was expected because the  21  play of older children was observably more elaborate and harmonious, thus consisting of more expansions of the play and greater agreement between partners.  Summary In summary, the study of sociodramatic play highlights developments in verbal  and social skills. Around age 3, children begin to engage in shared pretend play and collaborate to create the pretend experience.  At this time, children talk about the  representation of the pretend elements and about the organization of the pretend interaction. Preschool children discriminate between speech used  frj the enactment of  pretend and speech used about the enactment of pretend. As well, they are able to express their own intentions and relate to one another’s contributions within the pretend activity with increasing degrees of contingency. Furthermore, children reach a mutually satisfying consensus about the social pretend activity through the use of social conversational functions, such as negotiation and elaboration of play topic.  The Langauge of Sociodramatic Play: Two Perspectives on Sociodramatic Play Research shows that language is instrumental in constructing and maintaining the pretend play experience among peers. Various dimensions of this language have been investigated in the past: its content (e.g., plans, roles, object), its verbal and nonverbal components, the speaker identity position from which it is projected (i.e., personal identity or assumed identity), its discourse function (e.g., acceptance and rejection statements, or elaboration and negotiation) and its relation to preceding and following  22  discourse. The purpose of examining the language of sociodramtic play along these dimensions has been to better understand (a) how children express their ideas about the make-believe world, (b) how children maintain a continuous conversation during the interaction, and (c) how children coordinate the independent play ideas of the participants. Language used by children during sociodramatic play serves both communicative and social purposes. When studying the language of play, one can assume at least two different perspectives. First, if sociodramatic play is viewed an activity whereby children create an imaginary world, then language serves to mark the nonliteral quality of children’s behaviours and to define the context in which the pretend event is embedded. Inquiries as to whether the utterance is “about the play,” “for the play,” or both is of interest. Secondly, if sociodramatic play is viewed as an example of early peer conversation and social interaction, then language functions to maintain the social interaction. Emphasis is placed on whether or not the turns are thematically contingent to preceding and following discourse. For each perspective, various dimensions of language function may be studied separately. But an utterance or communicative act functions in many respects simultaneously. For example, an utterance such as, “I’m going on a trip,” serves to propose a plan. This utterance may be followed by a plan produced by the partner, “Yeah, let’s pretend we’re taking a plane to Disneyland!” Besides functioning as a plan, the utterance of the second turn is also a contingent response which serves to elaborate the play. Thus, the ultimate function of language in sociodramatic play is best be determined by integrating all the dimensions of the  23  language.  The Current Research Project In line with the literature reviewed, this research seeks to characterized the language used in sociodramatic play, and to establish differences in the use of such language by children between the ages of 3 and 5. The current research views the task of generating the pretend play experience as a continuous and ongoing process rather than as a phase or function distinct from the pretend acting/role-taking. Considering the entire range of communicative functions that language can serve during play, talk about  the play or communication for the construction and maintenance of the pretend world is of most interest.  At one extreme, the sole purpose of language is for the enactment of  an assumed role (“talk for pretend”). At the other extreme, the language is used for purposes completely unrelated to the pretend activity (“talk for real”). Language used to communicate about the pretend play lies in between these extremes. Construction and management of the drama can occur from an in-role position or from an out-of-role position. The language used to construct the play while simultaneously assuming a fictional identity (i.e., talk about the pretend play within the play frame) can be contrasted with the language used to construct the play as one’s self (i.e., talk about the pretend play outside the play frame). With this contrast in mind, the specific functions of metaplay communication (e.g., planning, role transformation, organization, etc.) will be examined within each of the two streams.  The general question surrounding this  aspect of the language is, “How does language serve the pretense?”  24  Following Goncu and Kessel’s (1984, 1988) studies of conversational continuity of children’s pretend play dialogues, this study also asks, “How are conversational turns tied together or linked?” Conversational continuity, however, provides little information about how children negotiate or elaborate the play. Therefore, the means by which children collaborate (i.e., through elaboration and negotiation) will be assessed as part of the social interactional functions served by language. In short, this study assumes two perspectives on the language of sociodramatic play: (a) The pretend-communication perspective focuses on the ways in which language is used to establish and preserve the pretense. (b) The social-conversational perspective focuses on the ways in which language functions to maintain conversation during pretend play.  questions and Hypotheses The research questions and accompanying hypotheses are presented next. 1. Are there age-related differences in the distribution of metaplay utterances by content  types, i.e., identity transfonnation, object transformation, planning, play description, and organization? Hi  -  The proportion of planning and organization utterances will increase with  age, while the proportion of play description utterances will decrease with age.  Planning the activities of the pretend play, and organizing the play set by distributing objects involve active participation by both players. Whereas, play  25  description, a simple commentary of on-going play events, demands less verbal interaction between the players. Older children, who are more experienced in jointly organizing and representing fictional reality, will use proportionally more planning and organizing utterances, and proportionally fewer play description utterances than younger children.  2. Are there age-related differences in the proportion of metaplay communication projected from different speaker positions, i.e., one’s own personal identity (our-of-role)  and an assumed fictional character identity (in-role)? H2  -  The proportion of in-role metaplay communication will increase with age.  Utterances which serve to explicitly construct and manage the pretend world while simultaneously enacting a fictional role convey multiple levels of meaning. At one level the content of the utterance is “just pretend” and at another level the function performed by the utterance is “for real.” If we assume that in-role metaplay communication is more complex than out-of-role metaplay communication, older children, who are cognitively more mature, should use a higher proportion of in-role communication than younger children.  3. Is there a relationship between speaker positions and specific content types? H3  -  The utterance types that function to initiate or terminate (e.g., role  transformation, and organization) the play will occur with higher frequency out of-role than in-role, and those that function to maintain or advance the play  26  (e.g., planning, play description) will be used more in-role than our-of-role.  4. Are there age-related differences in children’s responsivity to their partners’ contributions? H4  -  The proportion of responsive turns provided during pretend play  conversation will increase with age. Continuous social pretend play requires that the player provide relevant and contingent contributions to the conversation. Older children, who are more experienced conversational partners (i.e., attend to the behaviours of the partner and adjust their own behaviour in response to the partner), will provide more contingent responses than younger children.  5. Is responsivity related to the obligatory value of the preceding turn? H5  -  A contingent response will more likely be provided following an obliging  utterance than a nonobliging utterance. The form (e.g., command or question) of an utterance predictably facilitates the success of a social exchange. Turns or utterances which engage the partner through obligation will be followed by a response more often than those which are nonobliging.  6. Are there age-related differences in the types of turns which constitute continuous  pretend play conversation, i.e., initiation, simple response, complex response. H6  -  Turns which are linked to both the preceding and following turns will  increase with age. Older children, who are more aware of their partner’s role in  27  achieving the shared goal, will produce a higher proportion of turns which both respond to the prior turn and set up an expectancy for the partner’s involvement in the following turn than younger children.  7. Are there age-related differences in the degree to which children collaborate, and in the nature of that collaboration, i.e., elaboration and negotiation? H7 (a)  -  The proportion of collaborative exchanges will increase with age.  Collaboration consists of continuous turns which function to elaborate the play or resolve disagreements in the children’s representations about the play. Older children, who spend a greater proportion of time in social pretend play, will produce more collaborative exchanges than younger children.  H7 (b)  -  The proportion of collaborative exchanges which involve reconciliation  and compromise (i.e., negotiation) will increase with age.  8. Are there age-related differences in the proportion of in-role and out-of role collaborative exchanges? H8  -  Collaboration among the players while assuming a transformed identity will  increase with age. When children engage in joint role-playing, the partners must share a representation of both roles. Older children, who possess greater knowledge of social roles, will engage in more collaboration than younger children.  28  9. Is there a relationship between the nature of collaboration and speaker identity? H9  -  Negotiation which involves resolution of personal differences will occur  more out-of-role than in-role, while elaboration which involves building-up and expanding the play will occur more in-role than out-of-role.  29  CHAPTER 2 METHODOLOGY AND DESCRIPTION OF CODING TAXONOMY  Overview The purpose of this study was to investigate the nature of children’s collaborative pretend play narratives. The language of pretend play was the focus of this research.  Two perspectives on language function in sociodramatic play were considered: (a) the pretend-communication perspective and (b) the social-conversational perspective of language. The goal of examining socially interactive fantasy play from these positions was to illuminate developmental changes in the way children construct and maintain their imaginary play.  Subjects Twelve preschool children were paired and the six dyads were split into two groups. The young group was composed of two mixed-sex dyads and one same-sex female dyad. These children ranged in age from 3;7 to 4;2 (M  =  3;lO). The old  group was composed of one mixed-sex dyad and two same-sex dyads (one male and one female) whose ages ranged from 4;3 to 5;5 (M  =  5;l).  Candidates for the younger age group were initially observed at their preschool during free play. Based on these observations and the recommendations of the preschool teacher, only children who were observed to engage in socially interactive pretend play were selected to participate. Children were then assigned to a dyad  30  according to age, familiarity and typical play style. Within each dyad, the children differed in age by less than six months. All partners had known each other for at least six months, attended the same class and were judged to be comparable using the “Play Style Rating Scale” (See Appendix A). This scale was developed as means for the teacher to describe a child’s typical play interaction style, for example, whether or not the child engaged in group play and was originator of play ideas. The older children (5 year olds) were not observed in play prior to participating in the study. However, these children were judged to be suitable candidates according to parental and teacher report using the “Play Style Rating Scale.” Partners attended the  same kindergarten class and were familiar play mates for at least six months. Although the procedures and criteria followed in selecting the subjects may limit the generalizability of the results somewhat, subjects were carefully selected to ensure that an adequate sample of the language used during sociodramatic play was collected. This study was not interested in whether children at these age levels were capable of engaging in social pretend play, or in the amount of time spent in social versus nonsocial play. Instead, the language of sociodramatic play was the main focus.  Setting and Apparatus All dyads, except two, were observed in a playroom that was part of the preschool they attended. Three corners of the room were arranged as kitchen, dress-up and block corners. The kitchen corner was equipped with a child-size table, two to four chairs, a stove and sink. In addition, there were kitchen utensils, menus, aprons and a  31  small table cloth. The dress-up corner contained a closet, a large mirror, a variety of men’s and women’s clothes, fancy hats, jewellery, purses, and shoes. Wooden and plastic blocks of different sizes, tool kits, two hard hats, two fireman hats, and two safety vests were placed together in the block corner. Other objects were also available to the children: a doll carriage, a doll cradle, pillows, a large blanket, three dolls, telephones, a suitcase, two Superman capes, fancy capes, three crowns, wizard-like hats, and wands. The materials ranged from highly structured, which were suggestive of a particular theme, to minimally structured. These props were selected to maximize the likelihood of both pretend play and social interaction. The play of the other two dyads was observed in one player’s home. Similar materials were used in this setting and the room was arranged to resemble the preschool set-up as much as possible. Each session was recorded on videotape and audiotape. A video camera (Panasonic AG-190) was set up in the same room to record the children’s play. Verbal communications were recorded using a 2-channel Marantz tape recorder and two Sanyo TDK AR-X 90 wireless microphones affixed to each child’s clothing.  Procedure Each dyad was escorted to the arranged playroom by the investigator. The children were given a few minutes to familiarize themselves with the arrangement of the playroom and the toys. Then the investigator gave the following instructions:  32  Let’s pretend that this room is a magical place where make-believe things can happen and you two have special powers to pretend anything you want. You can play with all the toys and dress-up clothes in this part of the room (designated area delineated by carpet and rope). I’m going to make a movie of you while you’re playing. Then, when you’re finished we can watch some of the movie.  The investigator was in the playroom at all times during the session to operate the equipment. If the children did not engage in social pretend play naturally, then the investigator encouraged the children to use their imagination and play together by modelling examples of such play. During the first session, the investigator often intervened by suggesting possible themes, roles or events; proposing a transformation; or drawing the children’s attention to his/her partner or to an interesting prop. Remarks such as “Hey you guys could pretend to be at a restaurant. Who’s gonna be the cook?” or “Oh no the restaurant is burning!” were used.  Interventions of this type occurred as  many as five or six times during the first 30 minute session. In the second session, however, there was less investigator involvement. In that session, only occasional prompts such as, “What’s gonna happen next in the story/movie?” were given. The use of such prompts was intended to structure the play slightly, as a means of eliciting the particular behaviour being investigated. When interruptions within a session were necessary (e.g., restroom breaks), children were brought back to the playroom for completion of the session. A dyad was allowed to play together for approximately 30 minutes at a time. Each dyad participated in two sessions. Sessions were separated by one week, except for two dyads. For these dyads, the second session was delayed by one week. The difference 33  in time lapse between sessions was not felt to be of significance.  Data Preparation Transcription Written transcripts of the twelve sessions were prepared from audiotapes according to SALT (Miller & Chapman, 1991) guidelines for analysis. Segments in which the children played together were examined to address the questions of this study. These included the children talking about what and where to play, as well as the actual make-believe play episodes. The portions of the sessions which were judged as not relevant to the study were excluded; for example, child-investigator interactions, nonpretend interactions between the children, or nonsocial play. The children’s verbalizations were transcribed verbatim, using regular orthography. Information about each child’s behaviour was recorded from videotapes. Supplemental information included: (a) paralinguistic cues, such as vocal quality (e.g, use of a “baby-voice”), intonational contours (e.g., rising intonation to indicate a question) or facial expressions; (b) nonlinguistic signals (e.g., symbolic gesturing like offering); (c) descriptions of the children’s actions (e.g., waving good-bye, eye gaze); and (d) other pertinent details about the context of the play (e.g., sound effects or child’s use of imaginary objects). Verbalizations were segmented into utterances according to intonational contours  34  and meaning. The utterances were then numbered consecutively throughout the transcript and divided into turns. Following the conventions for segmenting pretend play turns (Garvey & Berndt, 1973; McLoyd et aL, 1980), a turn was defined as all verbal and nonverbal communicative behaviours produced by one participant prior to a transition (i.e., partner’s turn or a pause). Turns and transition points will be discussed further in the section regarding language coding.  Coding All data were coded according to the descriptive categories generated for the analyses. The coding taxonomy used is introduced in the following section. Coding judgements were based on information provided on prepared transcripts. Verbal communication and, to a lesser extent, nonverbal communication (particularly those nonverbal behaviours which were (a) intentionally communicative or (b) related or responsive to the communique of the partner) was the focus of analysis. Judgements required both reading the transcripts and watching the videotapes to insure that the play text was interpreted in light of the context in which dialogue belonged.  Reliability To establish interceder reliability for pretend communication coding, five minutes of three transcripts were independently coded by two individuals. Overall percentage agreement was 87%. Intercoder reliability for the social conversational  aspects of language for a five minute sample was 83% agreement.  35  Introduction to the Coding Scheme and Considerations for Use The current taxonomy was motivated by two analytical perspectives on the communication between peers during spontaneous social pretend play. The first perspective focuses on the communicative functions of the language used to establish and preserve a shared understanding of the pretend play experience’. The second perspective focuses on the conversational functions of the language used in social interaction. Each perspective is considered in turn.  Pretend Play Communication to Create and Maintain Worlds In social pretend play, as in all social interactions, the participants must develop a shared understanding of the contextual circumstances of their interaction. Unlike other social interactions, however, the context of pretend play is often novel and relatively free from the current situation. Thus, the players must jointly define the frame that surrounds the imaginary world by transforming time, space, persons, objects, and actions, and by establishing the rules effective during the interaction (see Goffman, 1979, for a detailed explanation of framing). Collective construction of the play context involves communication about the fictional world. Such communication related to the pretend play functions (to a greater or lesser extent) to outline the boundaries between fantasy and reality, transform the “here and now” reality, prepare the set, orchestrate the event, organize the players,  1  Language and communication are used synonymously to refer to both verbal messages and nonverbal signals and behaviours. 36  clarify the disagreement and conserve the pretend experience. It is apparent that language serves numerous, not to mention multiple, functions in the construction and maintenance of a sociodramatic play episode. Garvey (1974, 1977, 1982, 1984, 1990; Garvey & Berndt, 1977) noted in her research that a “good deal of talk is directed to creating, clarifying, maintaining, or negotiating the social pretend experience” (1977, p.34.). It is precisely this aspect of the social pretend play that is addressed in the first component of the coding scheme. Design of the current taxonomy is motivated by the work of other researchers. Garvey and Berndt (1973) were interested in identifying all techniques used to “communicate pretending” by children between the ages of 3 and 5. The children used five types of communication to indicate a state of pretend or a transition to or from that state. These types were: negation of pretend, enactment (i.e. an overt representation of the vocal quality, content of speech, physical gestures, attitude, actions associated with the role played by the child), signals or markers that signify “this is pretend,” procedural or preparatory behaviours for the organization and management of the play,  and explicit mention of pretend transformation. This last category consisted of seven subcategories related to the content of the utterance (e.g., plans, roles, objects). Despite minor differences in category labels and definitions, the classification system devised by Goncu and Kessel (1984) was highly similar to that of Garvey and Bemdt. (The Goncu & Kessel scheme was based on a combination of schemes  --  Garvey & Berndt, 1973, and Schwartzman, 1978.) For instance, play utterances were again classified as plans, object statements, transformations, negations and termination  37  statements. However, Goncu and Kessel, unlike Garvey and Berndt, included invitations and acceptance statements, and excluded enactment in their coding. The collapse of these two taxonomies results in nine distinct categories: invitations, preparatory behaviours, enactments, plans, role transformations, object statements, acceptance statements, negations, and signals. Motivated by the same general interests as the research mentioned above, Giffin’s research (1984) concentrated specifically on communication processes used to talk about the elements of the pretense. Emphasis on the mode (i.e., verbal or nonverbal) differentiated Giffin’s coding scheme from earlier schemes. Seven descriptive categories were generated from the data: enactment, ulterior conversation (directing the play from a fictional character position), underscoring (verbal emphasis of nonverbal pretend behaviour), storytelling/narration, prompting the partner, explicit pretend structuring statements without acknowledging the pretense (e.g., “You hafta choose something to eat”), and overt proposals to pretend (e.g., “Let’s pretend that .“).  These categories ranged along a continuum from completely within-the-play-  frame to clearly outside-the-play-frame. Children either employed the words and actions of a pretend world character to covertly direct the play, or they exposed the pretense by making explicit announcements to pretend. The point at which each category fell along the continuum was determined by (a) the mode of communication and (b) the world to which each component of the behaviour belonged.  For example,  in ulterior conversation both verbal and nonverbal components of the behaviour are within the frame of the pretend world, while in story-telling, the nonverbal component  38  is in-frame but the verbal component is outside the frame of the pretend world. Finally, Auwarter (1986), was interested in exploring an even wider range of communication used during sociodramatic play, including nonpretend talk. He classified pretend communication into nine “levels of reality” using two dimensions: (a) the speaker’s role identity (i.e., child as self, as a neutral observer, or as a fictional character) and (b) the validity claimed for a specific utterance (i.e., everyday reality, make-believe reality or a transition between the two states). Four of these levels (constituted by the intersection of categories on the two dimensions) are of particular interest to the current project: complete stage fiction (similar to Giffin’s enactment), construction of the fiction as a fictional character, construction of the fiction as oneself, or as a neutral identity (structuring and overt proposals). The main feature to be noted about Auwarter’s taxonomy is the differentiation of distinct speaker positions. It is apparent from the four taxonomies described above that communication about pretend play can be characterized along several dimensions. The current taxonomy attempts to clarify and expand previous taxonomies by treating each dimension of the communication separately. All communication identified as social pretend (according to preliminary judgements described in a later section) was first coded as belonging to one of three primary function categories: enactment, construction of pretense/metaplay communication, or other. Then two dimensions of the construction of pretend communication 2 were analyzed: specific content, identity of the  2  Construction of pretense involves communication that defines elements (e.g., roles, plans, objects) and communicates about the structure and organization of the elements and the players. Researchers sometimes distinguish these subcategories of communication: The 39  speaker. Below is an outline of the categories and subcategories for the first component of the current taxonomy. A flow-chart of the coding decisions can be found in Appendix B. Definitions for the various categories are provided in a later section. I. Pretend Communication about the Pretend World and Pretending Unit of analysis Communicative Act -  A Preliminary Judgements 1. Pretend status a. Pretend b. Nonpretend c. Uncodable 2. Communicative status a. Communicative b. Noncommunicative B. Primary Function 1. Enactment ofpretend 2. Construction and maintenance ofpretense/metaplay communication 3. Other C. Construction and Maintenance of Pretense Communication 1. Specific content a. Role/Identity transformation b. Object transformation c. Plan d. Play description e. Organization f. Uncodable 2. Identity of speaker a. In-role b. Out-of-role i. Pretend announced c. Ambiguous  former strand is commonly referred to as metaplay communication or metacommunication, and the latter as preparatory or managerial communication. In the current discussion, construction and maintenance of the pretense/metaplay communication encompasses both “meta” and managerial communication.  40  Pretend Play Communication as Conversation Sociodramatic play can be viewed as one type of social situation. Analysis of pretend play interaction as “communication to create and maintain worlds” sheds light on how language serves pretense, but does not characterize the continuity of the conversation. The second component of the coding scheme serves to remedy this deficiency. The goal of this level of analysis was to characterize the ties between conversational turns that occur during children’s pretend play. Researchers (e.g., Garvey and Hogan, 1973; Goncu and Kessel, 1984) have attempted to described children’s developing ability to engage in sequenced social interaction and/or continuous conversation. Garvey and Hogan (1974) defined “social speech” as “utterances which were adapted to the verbal and nonverbal behaviour of the partner” (p. 563). In an attempt to determine the extent to which speech was social, the result, or consequence, of each turn was placed into one of five categories: no apparent consequence, unrelated speech, attending behaviour (e.g., change in body posture or eye gaze), appropriate nonspeech, and appropriate speech. The connection between players’ turns in a sequence was further explored by Goncu and Kessel (1984). Based on the work of Torrance and Olson (1982), Goncu and Kessel categorized turns as turnabouts (i.e., turns that contained both a response and a new expectancy), responses, mands (i.e., a new expectancy without reference to prior turn), or unlinked utterances. As illustrated by the category schemes of these earlier studies, consideration of  41  the degree of contingency between turns provides information about how children establish conversational continuity. The second component of the current taxonomy explores precisely those aspects of the pretend play conversation. A combination of the methods and categories used in prior studies and conventional methods of discourse analysis were employed. All turns which were composed of pretend utterances were first coded for social status. Then each social turn was coded for continuity in two ways: firstly, for contingency with the prior turn and secondly, for continuity relations with both preceding and following turns Below is an outline of the categories and subcategories for each dimension of conversational aspect. A flow-chart of the coding decisions is represented in Appendix C. Further definitions will be provided in a later section. II. Pretend Play Communication as Conversation Unit of analysis Social Turn -  A. Preliminary Judgements 1. Pretend status 2. Segmentation of turns 3. Interactive status a. Social b. Nonsocial c. Uncodable B. Turn Function 1. First-part of pair a. Obliging b. Nonobliging 2. Second-part of pair a. Response i. Complex Response ii. Simple response b. Nonresponse i. Initiation 42  Collaboration The focus shifts slightly now to the social nature of children’s verbal interaction while jointly constructing the pretend play. This level of analysis examined the degree to which children collaborate and the type of collaboration. Building on the classification of turn continuity, collaboration is defined as an interactive exchange in which the discourse topic is continuous across two turns. As reported by Goncu (1993a), establishing a consensus about the pretense can be accomplished in different ways: by expanding the play, by expressing agreement or disagreement with the partner’s ideas, by emphasizing one’s own ideas, or by making comments which are not relevant. In the current analysis, only two of Goncu’s categories were adopted  --  expansion and agreement categories. These categories were  relabelled/redefined as elaboration and negotiation, respectively. The outline below and the flow-chart in Appendix C summarize the collaboration level of coding. Further definitions will be provided in a later section. II. Pretend Communication as Conversation (continued) Unit of analysis Collaborative exchange -  C. Type of Collaboration 1. Elaboration ofplay 2. Negotiation about play 3. Other  Category Description for Coding Purposes An overview of the conceptual framework developed for categorizing the children’s utterances during their sociodramatic play is illustrated in Appendices B and C. The dimensions and categories used are described in the following section and 43  summarized in Appendix D.  Communication about the Pretend World and Pretending Preliminary Judgements Language is the tool used to communicative during sociodramatic play, as well as during nonpretend interactions between children, and to express ideas in nonsocial contexts. In light of the interests of this study, only language used to communicate a message to the play partner in relation to the pretend play episode was analyzed. Relevant data sections of the transcripts were identified on the basis of two preliminary judgements: Was the verbalization or nonverbal behaviour related to the pretend play? and Was the verbalization or nonverbal behaviour communicative?  Degree of pretense Firstly, each utterance was categorized on the basis of the player’s orientation to either (a) the actual situation of everyday reality, or (b) the constructed and enacted situation of an imaginary play reality. A description of each category is provided below. Pretend. In the current analysis, all behaviour (usually verbal behaviour) which functioned within or referred to the pretend play reality was coded as pretend. This category included utterances used to perform the drama, as well as those which served to create/maintain the fictional context. A detailed description of the various types of  pretend communication in sociodramatic play are provided in the next section. At this  44  point it is sufficient to note that all communicative acts which were realized within the imaginary world or which functioned to generate the context of the make-believe play were labelled pretend. For example, in the following sequence a server (S) and a customer (A) interact: S: Kay, here’s your dinner {places the prepared dish on the table in front of A). And there’s your spoon and you’re gonna get your drink pretty soon, okay {walks to stove to retrieve cup}? A: Okay. Non-pretend. Nonpretend utterances were completely literal and referred to the realities of the preschool classroom (or the children’s real environment). They neither functioned within, nor referred to, the pretend play reality. Children spoke in their usual tone of voice and addressed their partners by their real names. Words like “for real” might be used to mark the distinction between reality and make-believe. Examples of nonpretend talk observed in this study included: comments accompanying the exploration and manipulation of objects which had not been given pretend characteristics; descriptions of unanticipated occurrences during the play session; and discussions surrounding manipulative play with materials such as blocks. The following sequence of utterances exemplifies nonpretend speech: A: A: A: A: A:  What’s this {holds up a hat)? What is this? A hat {tries it on}? Yup {looks in the mirror). It’s a hat, right? S: Ya, it is a hat. 45  Recall that extended nonpretend play interactions had been excluded from the transcripts prior to analysis. Therefore the transcripts contained a minimal number of nonpretend utterances.  Uncodable. Utterances which were impossible to reliably code as either pretend or nonpretend were marked uncodable and excluded from further analyses. The pretend value of an utterance was uncodable if the child’s intentions were unclear, or if a substantial part of the utterance was unintelligible. Utterances directed at or in response to the experimenter were not of primary interest, and in most cases, were categorized as  uncodable as a convenient way of excluding them from further analysis. In a few instances, however, the utterance functioned to assign meaning to the child’s behaviour with reference to the imaginary play. Such cases were marked as pretend play and were analyzed by the same conventions as other pretend play utterances.  Communicative intent The next step involved judging the communicative intent of the speaker for all  pretend behaviour. Noncomrnunjcatjve. Utterances that were noncommunicative were excluded from further analysis . Utterances were coded as noncommunicative when they were marked 3  There is one exception to this rule. During a play session, a child assumed two different roles concurrently Batman and the bad guy and enacted an interaction between both characters. This dialogue, although noncommunicative as defined in this discussion, was coded as enactment. Considering that the partner watched at first and then joined the play by teaming up with Batman to poison the bad guy, it was determined that violation of the coding rule was necessary. --  --  46  by lowered vocal volume, or by physical posture and movement directed away from  partner. Questions that lacked a pause for the partner to respond, or in which the speaker responded as though talking to him/herself were also interpreted as noncommunicative. Such speech appeared to be intended not for the partner, but for the child him/herself. For example, S was across the room putting on a shawl. A, after failing to keep a hat from falling off her head, dropped the hat and said: A: Where’s my crown {looks around}? Oops this is my crown {picks up crown}.  Communicative. An utterance or nonverbal behaviour was coded as communicative when it was clearly directed to the partner. Communicative intent could be marked by eye gaze or gestures, by lexical address, or by the fact that the semantic content of the message was contingent on prior conversation.  Before moving on, it should be noted that only those coded as pretend and  communicative were analyzed further.  Communication about the Pretend World and Pretending Primary function Two types of primary function were of interest: (a) communication for the enactment of the pretend play, and (b) communication to create and maintain worlds (i.e., metaplay communication). Categorical descriptions of the codes applied for this  47  dimension are provided below. Enactment. Enactment consists of bringing the imaginary scene to life with the actions and words of make-believe characters. The player’s vocal quality, content of speech, physical gestures, attitudes, and actions are all characteristic of the adopted identity (Garvey, 1974; Garvey & Bemdt, 1977). As all necessary transformations and plans have been previously established, the players are simply acting “as-if” they are in a state created and defined by the players themselves. The purpose of this type of communication is to realize the relationships and activities already specified by the play definition, not to negotiate or elaborate the play script (i.e., theme and plot). Utterances interpretable entirely within the established context were coded as enactment. For example, the following exchange between mother (S) and child (A) was enacted at the kitchen corner. S: Do you want celery and pizza and cheese and fries and oranges? A: Pizza. A: A piece of pizza. S: Just pizza? A: Yeah.  Manipulation of miniature objects (e.g., figurines) accompanied by vicarious interaction through the objects was also interpreted as enactment. In the following sequence K and S interact via two dinosaurs, enacting their drama. K: Uh can you be nice to me {speaks with a different 48  voice while holding a dinosaur next the dinosaur held by S}. S: No! K: How come? S: ‘Cause I’m a bad guy.  Construction of the pretense! metaplay communication. Recall that spontaneous social pretend play requires communication about the nature and structure of the pretense. All elements of the episode need to be introduced, negotiated, and/or organized by the players. Children talk about the symbolic representation of the situation, objects and identities and jointly work out a plan of action. The following sequence of utterances illustrates the negotiation of a setting. S: Pretend we’re camping already. A: No! S: I wanna pretend. A: Kay, pretend we’re camping.  Constructing the pretend context involves more than transforming entities or directing events. Preparatory conversation regarding costumes or props, for instance, can also function to bring about the realization of make-believe. In the example below, two children negotiate the distribution of objects. S: I hafta have these two {refers to the wands she holds}. A: No we both have one. S: No I hafta have two. A: No pretend we don’t. S: Pretend that I hafta have two.  49  For the purposes of the current investigation, all utterances which functioned to define the contextual parameters of the fictional world, to provide details about the imaginary world, or to organize and manage the play were coded as metaplay. Examples of this type of talk found in the transcripts were: clarifications of players’ rights, suggestions of possible activities, descriptions of antecedent or subsequent events, expressions of specific details about the fictional reality, transitions to and from the pretend state (i.e., invitations to play and negation! termination of pretend). It should be noted that all types of speech acts and utterance forms may occur in this category. No distinctions were made between requests, assertions, declarative, promises, etc.  Other. This category served to classify communication that was neither enactment nor metaplay, but which was none-the-less related to the pretend play. For instance, a player’s request for clarification of his/her partner’s utterance or nonverbal behaviour is not a true structuring strategy but it is necessary for the play to proceed. All discourse level repair strategies (e.g., requests for clarification/repetition and accompanying responses; clarification of own actions or statements, repetition of overlapped utterances, requests for affirmation! confirmation, etc.) were categorized as  other. Additionally, utterances that served to get a player’s attention were also coded as other. This category encompassed all communication which functioned to continue the play, without contributing new information to the play context.  Uncodable. All utterances for which primary function could not be reliably coded were assigned the label uncodable.  50  Communication judged as metaplay was analyzed in greater detail, as these acts were the focus of the study. The specific function and content of the metaplay utterance was coded first, followed by an analysis of the speaker’s identity.  Construction of the Pretense/Metaplay Communication Content Pretend play is organized around various transformations of reality. The ultimate fictional reality is built-up and substantiated with details surrounding the basic pretend elements. To capture this aspect of the play communication every metaplay was coded for its propositional content. Five content categories were coded: role transformation, object transformation, plan, play-description, and organizational comments. An utterance could contain one or more of these content types. Definitions for each category and accompanying examples are presented below: Role/identity transformation. An utterance which functioned to change a child’s true identity to that of a fictional character was coded as role transformation. The transformation could involve changing one’s own role, the partner’s role or both. The exchanges between A and S exemplifies negotiation of A’s role: A: Pretend I was a teacher and pretend xxx teacher of the parade. S: You’re my kid and the teacher. A: I’m a teacher of the parade.  Object transformation. Use of one object to represent another, or changes in the  51  state or quality of an object (e.g. “The table is broken.”) were classified as object  transformation. The invention of an imaginary object also constituted an object transformation. Talk about objects did not always concern transformations. For instance, the assignment of an object to a player was coded as an organizational comment rather than a transformation. (See category description below).  Plans. Action plans for the events and activities of the imaginary play originated from, and were conducted by, the players themselves. Talk about “what to do” as a means of directing (as opposed to explaining or describing) the verbal and nonverbal behaviour of the player(s) was coded as a plan. Such communication included the announcement of present and future activities for the player him/herself, as well as commands, suggestions and prompts for the player’s partner. Utterances containing a joint plan for both players or about what should happen in general were less typical. Illustrated below are examples of both plans and object transformations. K and S were pretending to heal a sick dinosaur. In the following sequence, S proposed a new plan and followed through. Then K opposed S’s implicit invention of the contents of the empty tea pot by declaring her own object transformation: S: And we need tea to make her feel better {gets tea in pot}. We hafta give him tea in his mouth. Give’m tea {holds the dinosaur upright and tips the tea pot as though pouring}.  52  K: No! S: {pouring sound effects}. K: No that’s water.  Play Description. The next category was reserved for comments about events on the pretend play world. As the speaker must have a certain degree of distance from the immediate on-going pretend in commenting on it, this type of talk cannot be interpreted as the enactment of a specific role. Despite its close resemblance to enactment, it is more accurately labelled play description or narration. Such utterances served to elaborate the play by describing or explaining some on-going activity. For example, explanations of obvious behaviour often served to underscore the performer’s actions but were not necessarily directed towards or intended for the other player. Sound effects which represent specific actions symbolically were another example. A characteristic feature of this type of communication was the stringing together of many phrases in one turn, creating monologue-like talk. K expressed the following chain of utterances independently: K: Yup he’s out there again. Now he’s gone, you missed that time, he’s gone now. Know what? He’s gone, he wandered off. He didn’t see anything good to eat in here. He doesn’t like children, all he wants is to eat food.  53  This category differed from plans because play description did not function to  direct the behaviour of the characters, but rather served to support and give meaning to the drama being enacted. The younger child often directed such comments at the experimenter.  Organization. Comments about the play which did not involve any  transformation, planning, or stomy description, but which nonetheless functioned to structure and define the pretend play world, were coded as other. This category  primarily included language used to organize and synchronize the individual player’s contributions by: (a) determining objects/materials to be used; (b) assigning possession of objects to players; (c) discussing a player’s appearance in a costume; (d) clarifying player’s rights (e.g., turns, use of objects, etc.); and (e) beginning and ending play episodes with invitations and termination statements. Examples of this category can be found in the following sequence: S: You hafta wear a dress or something, I think {both children look in dress closet}. A: No I think I found my cape {picks up a shawl}. Yes I did find it Mum. S: {looks at A). A: My shirt. A: What goes with it Mum? S: A necklace.  Not Coded. Utterances with content that did not fit into any of the above categories were not coded.  54  Identity of the player One might imagine that all constructive contributions are expressed before the children have transformed their identities at the onset of a pretend episode. But actually, generation of the play context is an on-going task because players often need to redefine certain elements or wish to elaborate the story or setting. The players are capable of constructing the scenario from different speaker positions  -  as self or as an  imaginary character. Out-of-Role Communication about Pretense. While still projecting their everyday personal identity, children can plan and manage the novel reality prior to and after imaginary enactment has begun. In the latter instance, children abandon their adopted identity momentarily and return to their everyday identity. If the children behaved and spoke as themselves, then the communication operated outside the play frame and the children were judged to be out-of-role. Out-of-role communication about the pretense occurred in normal voice, and had one or more of the following characteristic features: (a) the words and actions were inappropriate for the imaginary situation; (b) partners were addressed by everyday name; (c) proposals were followed by requests for confirmation (e.g., tags “okay?”); (d) words such as “remember” or “supposed to” were used; (e) shifts to the third person pronouns following use of first or second person pronouns. Utterances related to social script knowledge or to the practical requirements of the on-going fantasy were also viewed as out-of-role. Prior to the exchange shown below, A and S decided to dress-up as brides. A  55  wanted a purse as part of her costume, but S informed her that a bride could not carry a purse. A: Actually I need xx {reached for a purse}. S: You can’t have purses in weddings. S: Don’t have purses in {utterance interrupted by A}. A: Yes you can!  Out-of-role messages which verbally acknowledged the pretense with phrases such as “let’s say that” or “pretend that” were differentiated from other out-of-role communication and received an additional code. The following interaction illustrates enactment followed by out-of-role structuring. T was the customer and K was the server/cook. T: Urn, I’ll have the Kentucky chicken. K: Where’s the Kentucky chicken {looks in cupboard}? T: Just get a chicken out. T: If there’s none, just pretend.  In-role communication about pretense. Once the pretend enactment has begun, the characters of the story can continue to structure and develop subsequent imaginary events. Such was judged to be in-role when there was no evidence of a transition from the pretend state immediately prior to the utterance, and the child continued to enact the imaginary part nonverbally. In essence, when the children behaved as though they were  56  enacting with an “ulterior purpose to create and orchestrate new events” (Giffin, 1984,  p.80), the communicative acts were coded as in-role communication about the pretense. This category differed from enactment because the communication seemed to be interpreted at the border of the actual pretend frame. The child subtly directed the play by prompting his/her partner, changing the conditions of the established situation, specifying details, or adding new information, etc. Such communicative acts were meant to be viewed by the partner as stage management, with the speaker acting and directing simultaneously. For example, A and S were engaged in pretend play at the kitchen and dress-up corners. S acted as Mother and A as Child. A advanced the play without stepping out of her established fictional role. A: Mom? S: Yeah. A: Can I go out to our river and see what’s there? S: Ok and bring your binoculars.  Linguistic markers which distinguished in-role talk included the use of nonhabitual voice quality, use of partner’s fictional name, reference to transformed objects  and situation, and/or content that was associated with the transformed identity. Skillful negotiation of setting, while in-role, is exemplified by T and K in the sequence below: T: Yeah, I’m in my hotel. T: I think I’ll just unpack my stuff. T: xx lay down. K: That’s my house {emphasis on y} so you  57  can stay with me now. T: Uh okay, I’ll share my home. K: Alright we’ll share your home.  As communication of this type is partially within and partially outside of the play frame, in some cases the speaker’s identity could not be disambiguated on the basis of the clues available for a particular communicative act. In such cases, judgements reflected the coding of preceding communicative acts. For instance, if the child was incostume and her identity had been transformed prior to a given utterance, then she could be assumed to still be in-role. The convention established was: For the same speaker, assume the preceding speaker identity unless otherwise marked. Utterances coded as in-role by this convention were given a differentiated subcode because such a rule has the potential effect of over-estimating the complexity of the child’s intentions and skill. In the following sequence, S had adopted the role of “mother” and A that of a “princess/daughter.” S, while clearly in-role, prepared a meal, invited A to eat and then set two places at the table. As S returned to the stove to get a drink for A, she informed A of which setting was for whom. Initially S’s speech and nonverbal behaviour clearly indicated in-role talk but towards the end of the sequence the signals became more ambiguous, and thus the last two utterances of the sequence were coded as assumed to be in-role. S: Come on honey. A: {follows}.  58  S: You have a red plate and napkin and yours dinner {sets two plates full of food on the table, then goes to stove briefly}. S: And yours spoon {gives A a spoon}. S: And you’re gonna get yours drink pretty soon, okay {returns to the stove and looks in the cupboard}? A: Kay {begins to eat}. S: The hotdog’s mine hotdog, okay {points to hotdog on the table}? A: Kay.  In a small number of additional instances, specific speakers routinely failed to make explicit announcements of identity transformation. Nevertheless, the content of the child’s speech, the child’s affective tone, and the child’s nonverbal behaviour, all suggested that the child was acting as though she existed in another world. These cases  were also coded as assumed to be in-role. In this example, K is assumed to speak from an in-role position because she is able to relate to the feeling of a character of an imaginary world. First she discovers that the dinosaur has broken body parts and then she performs the role of dinosaur-healer without mention or signal of a change in identity. K:  Uh oh, look at what’s happening {looks up at partner and experimenter). He broke his leg. You’re breaked {looking down at dinosaur}! He almost broke this one too. Well, he broke his tail too, so we hafta take him somewhere to see what’s happened {takes dinosaur to the table}. Yup something happen really badly.  59  Ambiguous. In cases where the speaker’s identity could not even be assumed on the basis of formal markers and/or behavioral signals of the preceding activity, the utterance was coded as ambiguous. These utterances are important because it is likely that this ambiguity exists for the other participant, as well as for the researcher. The consequence, if any, of this ambiguity may effect how the children achieve a shared understanding of the make-believe situation. In the sequence below, dyad A-S was enacting a dinner time scenario, where A was the daughter and S was the mother. 1. A: Here Mum {hands S a plate). 2. A: Sorry you get the red plate today. 3. A: Here. 4. S: (That that) but that red plate’s cracked.  5. A: It’s okay. 6. A: Pretend, okay? 7. S: I’ll pour some tea for us {holds teapot). 8. A: No, no, no Mom. 9. A: I’ll do it, cause I made dinner. 10. S: I’ll get the cups out, okay {puts tea pot down and picks up cups)? 12. A: Okay.  Judging from the seriousness of S’s vocal quality and facial expression, she had stepped out of her role in utterance 4 to comment on the condition of the plate she has  60  been offered. The fact that the plate really is cracked supports the likelihood that she is not pretending at this point. Furthermore, A’s announced pretend transformation of the plate’s condition leaves little doubt that the children are interacting in the reality of the preschool classroom (utterance 4 through 6). Although it is clear that A has reassumed her fictional identity in utterances 8 and 9, S’s speaker position is not as evident. This example demonstrates how the children can implicitly move in and out of role. It is precisely such subtlety that creates confusion for the interpreter. A few further specific coding conventions were applied regarding combinations of codes. These can be found in Appendix E. So far the codes which were applied to communication about the pretend world and pretending taxonomy  -  -  the first component of the  have been defined. In short, all pretend and communicative utterances  were first coded for primary function, then two dimensions of the communication used to construct the pretense were coded: content and speaker identity. Codes which belong to the second component  --  the conversational aspects of communication  --  are  described next.  Pretend Play Communication as Conversation As stated above, the focus of the social-conversational perspective was the sequential organization of children’s pretend play conversation. “Turn-taking seems [to be] a basic form of organization for conversation” (Sacks et al. 1974, p.700). As such, a turn was taken to be the basic unit of analysis for this level of coding. Close examination of the contingency and relevancy between conversational turns was a means  61  of measuring the continuity the dialogue and the nature of the collaboration.  Preliminary Judgements The conversation was first segmented into turns. Then based on the judgements made about the pretend status of utterances, only turns that were composed of pretend utterances were judged for social status.  Segmentation of conversation into turns In keeping with other researchers (e.g., Garvey, 1974; Garvey & Hogan, 1973;  and McLeod et al., 1984), a turn was defined as a stretch of one person’s (verbal or nonverbal) behaviour. The point at which one turn ends and another begins was termed a “transition-relevant place” by Sacks et al. (1974). Such transition relevant points were most obviously marked by the contribution of a new speaker, but also could be marked by silence. In the latter case, the same speaker might begin a new turn following a pause (greater than three seconds). Once the interaction was segmented into turns, each turn was coded for the degree of interaction between children. Turns were classed as social, nonsocial, or uncodable as defined below.  Degree of interaction  Social. A turn was coded as social if at the transition-relevant place either partner intentionally directed his/her behaviour towards the other child. Social intent  62  could be marked by eye gaze at the partner, mutual attention to an object, invitational gestures or verbal initiations (e.g. address). The sequence below contains social turns. S: Katie, I wanna be the customer. K: What would you like {Stands beside the table and looks at S}? S: I’m have chocolate and...  Ordinarily, the speaker’s intentions established or maintained the social status of an interaction, but in some exchanges the nonspeaker perceived and responded to possibly nonsocial speech as though it was social. In the following exchange both turns were coded as social based on the K’s reaction. T’s intentions were ambiguous. However, K responded as though T’s behaviours were directed at her. T: I’m the king. T: Anybody that knows about this project answer to me {proudly walks around the play space without attending to K who is looking at the baby clothes}! There is a short pause between speakers. K: Your highness, I guess nobody knows about you.  Even if the nonspeaker fails to respond verbally to preceding social speech, indications of social interest with eye gaze or nonverbal behaviour (e.g., a compliant action or a head nod) were sufficient for a turn to be coded as social. 63  Although each turn was analyzed and coded individually, the analytical decisions demanded consideration of the turn within an exchange or sequence. Typically, once  social speech had been established, subsequent speech was also considered social, despite lack of explicit social markers, unless otherwise marked (e.g., lower volume voice); i.e., a turn was coded as social unless it was clearly marked as nonsocial. A turn containing utterances which were coded as noncommunicative in terms of pretend was coded as social if the majority of the utterances were intended for the partner.  Nonsocial. A turn in which neither child directed his/her verbalization or behaviour at the other child was coded as nonsocial. For example, this code would apply when one or both children independently explored an object or engaged in an independent imaginative activity. In the following sequence, both children were engaged in play of their own: K was by the stove looking at the food while S was near the sink pretending to be riding a motorcycle (a chair). Their turns were coded as nonsocial. S: I fell off the motorcycle {lays on the ground}. S: I did {looks at experimenter}. K: No I’m not going to {looking downward, acts as though she is talking to someone else}. K: But who will look after Shadow? S: I fell off the motorcycle {to experimenter}.  Uncodable. All turns which did not fit into one of the above categories were  64  uncodable. An uncodable turn would be one in which the speaker’s social intent was unclear and the turn was not responded to by the partner. All turns directed at, or in response to, the experimenter were also coded into this category.  All of the social turns were coded along two other dimensions: interactive structure and nature of the collaboration.  Function of a Turn The next dimension of the social interactional aspects of coding attempted to capture the continuity and coherence of the dialogue. To address this topic, the relation between consecutive utterances of an interactive exchange was examined. Schegloff and Sacks (1973, as cited in Sacks et al. 1974) used the term “adjacency pair” to describe adjacent utterances produced by different speakers. Pairs are ordered, as a first part and a second part, and are typed, so that a particular first part may require a particular (or range of) second part. Adjacent utterances are connected first by mere juxtaposition  and maybe further connected by the interdependence of their content. The notion of “adjacency pair” was broadened to accommodate the interests of this investigation. Thus, consecutive turns rather than utterances were analyzed as the adjacency pair unit. It is important to note, however, that for the purposes of this analysis an adjacency pair was defined first by juxtaposition alone, i.e. there was no requirement for consecutive turns to show continuity. This is different from Sacks and Schegloff’s definition of the term.  65  In the sequence below consecutive turns do not form a continuous sequence of conversation. S: I’m trying to make your house ‘cause your house is broken. A: There’s a fire! S: {looks at A, but does not provide a verbal response} A: Oh, I hafta fix this tighter {whispered, winded toy drill}. A: Something else is broken. A: The table!  In this study, each turn was coded for its relationships to both the preceding and following turns, i.e. as both a “first pair-part” and a “second pair-part”. The only exception to this was a turn which acted strictly as a response to the demands of the prior turn (e.g., okay, yes, no), in which case it was coded in a separate category. A more detailed description of the codes applied to each turn is provided below.  First pair-part  Obliging First Pair-Part. According to Sacks, Schegloff, and Jefferson (1974, p.7l’7) a “first pair-part” may “set constraints on what should be done in a next turn (e.g., a ‘question’ making ‘answer’ specially relevant for next turn).” The definition and example provided by Sacks et al. illustrate how a typical “first pair-part” can establish an expectancy for the nonspeaker to respond in the subsequent turn. An obliging first pair-part can function to initiate an interactive exchange, but first pair-parts within an interactive sequence do not necessarily initiate a new topic 66  entity. For instance, contingent queries or tag questions act as obliging first pair-parts which require a response but are not normally employed as initiators of conversation. Examples of this class of units are greeting, invitation, or offer which set expectations such as greeting, acceptance/decline, and acceptance/refusal, respectively. A turn was coded as an obliging first pair-part if it establishes a strong expectation for a response from the next speaker, either a verbal reply or a nonverbal social act. For example, imperative (e.g., “Get me that pot.” or “You hafta ...“) and some declarative forms (e.g., “Pretend you are a witch.”) also impose expectations on the non speaker. D attempted twice to involve his partner with two different types of obliging turns, neither of which were successful and only one of which received a response. D: How’bout you build me in a castle? N: Na. D: Build the castle around me. N: {no response}  Nonobliging. First pair-parts differ in the degree to which they establish expectancies for a specific sort of “second pair-part”. Some, like the examples just given, require quite specific types of responses, other first pair-parts may not require a response at all, e.g., descriptive comments. A turn was coded as a nonobliging first pair-part if it did not strongly require a specific sort of response. Note that a turn which is nonobliging can nevertheless be part of a continuous adjacency pair if the propositional content of the following turn is contingent on that turn. 67  After K announced her plan for where to play, S followed her lead and decided to also play at the block centre. K: I’m just gonna play with drills. S: I’m gonna do drills too with you. S: I’m gonna be the man with this {picks up a screw). Second pair-part A second pair-part of an adjacency pair contains the partner’s presumed “response” to the conditions established (if any) in the first part of the pair. These “responses” vary from disregard to strict adherence to the discourse expectations. The appropriateness of a second pair part was determined by whether the content of the turn  was continuous with the prior turn in terms of discourse topic. Keenan & Schieffelin (1976) defined discourse topic as “what is being talked about” or “the question of immediate concern”. Topic can be maintained in a variety of ways. In this study, simple acknowledgments (verbal and/nonverbal), compliances, acceptances, refusals, rejections, as well as exact or modified imitations of the prior proposition, all satisfied the requirements of a continuous response. More complex responses involved the contribution of new information which was topically contingent to the prior proposition. All second-pair parts were judged according to two criteria: (a) continuity in terms of discourse topic and (b) degree of obligation to respond. The following types were coded:  Response-obliged. A turn in which the content was continuous and the second pair-part was obliged, as in the exchange between A and S:  68  1. S: Do you want celery and pizza and cheese and fries and orange? 2. A: Pizza. A piece of pizza.  Response-nonobliged. A turn in which the content was continuous but the  second-pair part was not obliged. For example, S responded to A’s turn of the above example with: 3. S: Just pizza?  Nonresponse-obliged. A turn in which the content was not continuous and the  turn did not meet the established obligation. A missed turn, where the children disregarded the partner altogether, also fit into this category. For example, K disregards S’s question. S: Wanna come to the restaurant? K: Do you know how to open this {a suitcase} Stephanie?  Nonresponse-nonobliged. A turn, following a nonobliging turn, in which the content was not continuous. For example, D had invited N to be Superman’s helper and was looking for the costume. N, who was not interested in the D’s plan, expressed his own plan. D: Cause I need a helper. N: Let’s pretend that cape was magical and it helped you. This cape {holds up the cape for D to see}.  69  Turn Type Although it is convenient to conceptualize the first and second part of an adjacency pair as separate entities, empirically these two components often coincide, with a given turn functioning as a first pair-part for the following turn and as a second pair-part of the preceding turn. Three initiation-response combinations for a turn were of interest:  Complex response. A complex response turn was one that incorporated some aspect of the previous turn, and also established a new expectancy for the following turn. Consider the sequence below: 1. T: I want this for dessert if you get it in, okay? 2. K: Airight you can have that. 3. K: Now you hafta choose something for your meal. 4. T: Urn I’ll have diet coke for my drink and can I have the menu? Utterance 1 is a first pair-part and utterance 2 functions as its second pair-part. However, K’s turn continues; the first pair-part of the next adjacency pair is established in utterance 3.  The example above illustrates the most complex type of turn exchange demonstrated by the dyads of this study. The exchange between A and S contains another complex turn. It is slightly less sophisticated because, although S directed A’s nonverbal behaviour, the propositional content of utterance 4 was not original. 1. A: Oh this is hard. 2. A: We’ll hafta take the sink out 70  and make two more different edges, okay sir? 3. S: Okay. 4. S: Take off the sink.  5. A: Yeah, I did. A’s turn ended with a request for affirmation from S. In terms of conversational function, utterance 2 carries the first component of the adjacency pair, and places constraints on what should be done by the next speaker. The second pair-  part is provided in utterance 3 and the first part of a new pair in utterance 4.  Simple Response. A simple response turn was one in which the child merely responded to the expectations of the prior turn without contributing anything further. A short verbal acknowledgment/acceptance or rejection, back channel-like responses and nonverbal compliant acts were the most common types of simple responses. K cooperates with her partner by following his lead and verbally agreeing with his plan: T: It the last flight. T: The plane’s ready to board, hurry up! K: Alright {runs towards the closet ‘the plane”}! -  Initiation. A turn which was not linked to the prior turn (i.e., not a response) but which provided an opportunity for the partner to respond was classified as an  initiation. For instance, D and N enacted a short Superheroes scenario, the last two utterances of which are shown below (utterances 1 & 2). Immediately following 71  this pretend interaction, N introduced a new play theme (utterances 3 & 4).  N: You broke the wall. D: I broke that what was coming down on you. N: Now what kind of urn stuff do you want to eat? Hamburgers and cheese? D: Urn uh chicken please.  Unlinked turns. Not all turns were closely linked to their surrounding turns. That is, the mere juxtaposition of turns did not necessarily result in continuous collaborative conversation, for instance, if a turn was followed by a missed or nonsocial turn, or if the two adjacent turns were discontinuous in terms of discourse topic.  Type of Interactive Exchange The next level of coding applied only to collaborative exchanges. A collaborative exchange was defined as one that was continuous in terms of discourse/pretend play topic. That is, the second turn of an exchange contained a responsive utterance. All collaborative exchanges were classified by function into two types: elaboration and negotiation. Definitions and examples are presented below:  Elaboration. A collaborative exchange that involved maintaining the play interaction by adding information to the ongoing activity was coded as elaboration. Examples of elaborative responses are: contingent comments which integrate new information, exact or modified repetitions of the prior turn, or non-intrusive feedback (e.g., head nods). 72  The sequence below contains utterances which elaborated the play. S: I’m gonna be a bad dinosaur, okay? S: I’m gonna be the bad one. S: I’m gonna be the bad dinosaur. K: Okay. K: Kay I’m walng around that guy. S: I’m hungry {different voice) S: I’m hungry. K: Eat! K: Go ahead eat. K: That’s a good one. S: {pretend to eats}.  Negotiation. Collaborative exchanges that surrounded disagreements about the play between the players were coded as a negotiation. Such exchanges included rejections and revisions of the partner’s ideas, or attempts to resolve conflicting ideas about play, as illustrated below: K: Trevor I want you to be a baby. T: No I don’t wanna be anything. K: Oh remember it’s lots of fun. K: Remember you get to go “gugu gaga”. T: I don’t want to! K: Trevor!  Other. Collaborative exchanges that did not function to elaborate or negotiate the pretend play were coded as other.  73  Summary of Coding System In summary, an outline of the current taxonomy is presented.  I. Communication about the Pretend World and Pretending A. Preliminary Judgements 1. Pretend status a. Pretend b. Nonpretend c. Uncodable 2. Communicative status a. Communicative b. Noncommunicative B. Primary Function 1. Enactment ofpretend 2. C’onstruction and maintenance ofpretense/metaplay 3. Other 4. Uncodable C. Construction and Maintenance of Pretense Communication 1. Content a. Role/identity transformation b. Object transformation c. Plan d. Play description e. Organization f. Not coded 2. Identity of the speaker a. Out-of-role b. In-role c. Ambiguous  II. Pretend Play Communication as Conversation A. Preliminary Judgements 1. Pretend status 2. Segmentation of turns 3. Interactive status 74  a. Social b. Nonsocial c. Uncodable B. Turn Function 1. First-part pair a. Obliging b. Nonobliging 2. Second-part pair a. Response i. Complex Response ii. Simple Response b. Nonresponse i. Initiation C. Type of Interactive Exchange 1. Elaboration of play 2. Negotiation about play  75  CHAPTER 3 RESULTS Overview This investigation of collaborative pretend play attempted to address the following question: Do children between the ages of 3 and 5 years show age related changes in their ability to construct and sustain pretend play? To address this question, two perspectives on the language of pretend play were assumed: (1) The pretendcommunication perspective focused on the ways that language and nonverbal behaviour  served to construct and maintain pretend play. (2) The social-conversational perspective focused on the conversational functions of language during sociodramatic play. Assuming these two perspectives, three aspects of the communication used during interactive pretend play were examined: pretense, conversational continuity, and coordination of children’s independent play ideas.  For each aspect of pretend play  studied, one or more specific questions were asked. To answer each question, subjects were divided into two groups (young: 3;7 months, M  =  -  4;2 months, M  =  3;lO, and old: 4;3  -  5;5  5; 1), and the Student’s t-test comparison or the Pearson’s chi square test  of independence were conducted.  Use of Language to Construct and Maintain the Pretend Play The first set of analyses focused on how children used language (and nonverbal behaviour) to create and develop the pretend play episodes. The task of setting-up the play demands communication about the play organization and the pretense, itself. The  76  questions surrounding communication used the construction of the pretend play are: 1. Are there age-related differences in the distribution of metaplay utterances by content type’ ? 2. Are there age-related differences in the proportion of metaplay communication projected from different speaker positions? 3. Is there a relationship between speaker positions and specific content types?  Children use language to create an imaginary world and to organize the players and activities. As most pretend play episodes are novel and spontaneous, it is expected that much of the communication related to pretend play will be devoted to its construction. Indeed, for both groups of children, a substantial proportion of the pretend communication was devoted to this purpose (young M  =  69% and old M  70%), rather than simple enactment of the drama.  Specific Content The propositional content and purpose of an utterance determined its specific function in the construction of pretend play. Five content types were studied: identity  transformation, object transformation, planning, play description, and organization. It was expected that these communicative functions would not occur equally. Since  1  An utterance or nonverbal communicative act was the unit of analysis for this set of questions. Since the data consisted primarily of verbal communication, the term utterance will be used to refer to both verbal and nonverbal behaviour. Recall that metaplay communication refers to communication for the construction and maintenance  ofpretense. 77  generation of the pretend play context is an ongoing task, those utterance types that function to maintain the pretend play experience (e.g. organization, and planning) should be used more that those that serve to initialize the pretend event (e.g., role  transformation). It also seems possible that the distribution of communicative content types would vary with age. Planning “what to do next” and organizing the props and players seem to involve greater consensus between the partners than play description, which is a simple commentary on the unfolding events. If so, older children, who are more capable of verbally interacting with a partner, might use a greater proportion of  planning and organizing utterances than younger children. To determine whether the specific content of metaplay utterances varied with age, the proportion of utterances for each specific content type was calculated as a function of total metaplay utterances. Means (and standard deviations) for these proportions for both age groups are presented in Table 1. The distribution of specific functions was highly similar for both groups and indeed, statistical tests (t-tests) revealed no reliable group differences. Approximately 25% of the communication was devoted to organization, 50% to planning and description of the activities and events,  and only 14% to object and identity transformation.  78  Table 1 Mean Proportion (and Standard Deviation) of Metaplay Utterances by Content Type of Young and Old Dyads Content Type  Young  Old  Organization  25 (9)  30 (14)  27.5 (11.5)  Planning  21(5)  23 (4)  22.0 (4.5)  Play Description  29 (10)  22 (10)  25.5 (10)  Identity Transformation  9 (6)  6 (3)  7.5 (4.5)  Object Transformation  6 (4)  5 (4)  5.5 (4.0)  Not Coded  9 (4)  13 (4)  11.0 (4.0)  Average  Speaker Identity Position Children may communicate about the pretend play while projecting their everyday identity or while enacting a fictional role. Talk about the pretend experience from an in-role position is more complex than from an out-of-role position. At one level, the content of the utterance is “just pretend”; at another level, the intentional function of the utterance is “for real.” Because in-role communication must convey multiple levels of meaning (symbolic and meta), children may be less likely to use this type of communication than out-of-role communication, especially younger children. The next analysis determined whether there were age-related changes in the proportion of metaplay utterances projected by the child while in-role and out-of-role.  79  Each child’s utterances were categorized by speaker position, and the proportion of metaplay utterances for each category was computed. The categories were: in-role, assumed in-role and out-of-role. Table 2 presents the group means (and standard deviations) for the proportion of metaplay utterances for each speaker identity category. For both age groups, the proportion of out-of-role utterances was greater than that for utterances clearly marked as in-role. When the categories of in-role and assumed inrole were combined, however, it was no longer the case for the older children, that out-  of-role metaplay utterances were used with higher frequency than in-role metaplay. T test comparisons of the group means revealed a significant increase in the use of clearly  in-role utterances (t  =  7.34, df  =  10, p  =  .00) and a complementary tendency for oui  of-role communication to decrease as children get older (t  =  1.79, df  =  10, p =.10).  Considering the small group size and the high degree of variability among the subjects in the young group, statistical treatment of these differences should not be over interpreted. The results are, nevertheless, valuable in that they provide information about the character of social pretend play.  80  Table 2 Mean Proportion (and Standard Deviation) of Metaplay Utterances by Speaker Identity Position of Young and Old Dyads Speaker Position  *  Young  Old  Out-of-Role  54 (18)  42 (5)  Assumed In-Role  34 (18)  34 (4)  In-Role  8 (3)  Ambiguous  4 (4)  19  (3)*  5 (3)  p < .025, experimentwise p < 1.0.  Relationship between Content Type and Speaker Identity Position Having analyzed metaplay communication by speaker identity and content type independently, the next logical question to ask was: Is there a relationship between these two variables? For this analysis, the speaker identity categories in-role and assumed inrole were collapsed, and only the most frequently used content types (i.e., organization, planning and play description) were included. The other content types were not included because it was felt that proportions of use were too small to subdivide further. To determine if there was a relationship between the three most frequently employed content types and the two distinct speaker positions, each utterance was categorized by content and speaker identity and the proportions of metaplay utterances for each category were calculated. The six speaker position-content combinations were:  in-role organization, out-of-role organization, in-role planning, out-of-role planning, in-  81  role play description, and out-of-role play description. It is important to note that the two dimensions, content type and speaker identity, are logically independent. According to the definitions provided earlier, each content type can be employed both in-role and out-of-role. For instance, children can express “what to do next” (i.e., plan) both before and after adopting a different identity;  describe antecedent play activities from within or outside the make-believe world (e.g., out-of-role: “Let’s say that you can’t see the dinosaur anymore, because’s already gone” as opposed to in-role: “No, the dinosaur’s gone now. You can’t see him anymore. Too bad”); and negotiate the apportionment of objects (i.e., organize) as themselves or as fictional characters (e.g., out-of-role: “You could wear a dress or something” or in-role: “Honey, you hafta wear a dress for dinner”). In fact, the overall and group means reported in Table 3, support this independence since all combination were employed by the children. Close examination of the means of each combination revealed two relationships at both ages levels and overall: organizational talk occurred most when the player was  out-of-role; whereas, play description occurred most when the player was in-role (X 2 (1, N  =  12)  =  13.2, p < .01). There also seemed to be a tendency for greater  proportions of planning and organization to occur in-role as age increased.  82  Table 3 Mean Proportion (and Standard Deviation) of Metaplay Utterances by Speaker Identity and Function Type Group  Content Type  Speaker Position In-Role  Out-of-Role  Young  Organization Planning Play Description  3 (5) 9 (6) 21 (10)  22 (8) 10 (8) 6 (3)  Old  Organization Planning Play Description  11 (4) 15 (3) 15 (5)  17 (9) 8 (3) 7 (5)  Overall  Organization Planning Play Description  7 (6) 12 (6) 18 (8)  20 (9) 9 (6) 6 (4)  To explore the latter relationship further, the proportion of in-role communication for each content type was calculated as a function of the total number of  in-role utterances. T-test comparisons of the group means revealed significant agerelated differences for the categories tested. These values are summarized in Table 4. While in-role, use of organizational utterances (t  =  2.58, df  planning utterances (t  2)  increased with age; whereas, use of  =  2.77, df  =  10, p  =  .02  =  10, p  descriptive comments about the pretend play decreased with age (t  =  =  .02) and  4.22, df= 10, p  =.00).  2  The group variances were significantly different; however, the result of this test is determined to be robust and unbiased by the asymmetry of distribution (Boneau, 1960). 83  Table 4 Mean Proportion (and Standard Deviation) of In-Role Metaplay Utterances Grouped by Content Type of Young and Old Dyads  *  Content Type  Young  Organization  8 (10)  Old (9)* 22  Planning  20 (7)  29 (4)  Play Description  48 (9)  28 (8)  p < .025, experimentwise p <.075.  Summary of Results: Construction and Maintenance of Pretense Based on the results of the analyses described above, all the children observed communicated a great deal about the construction and management of the pretend play. There were age-related differences in the use of language to achieve this task. The older children demonstrated greater ability, as reflected in their language, to manage and develop the pretend play while maintaining a transformed identity (i.e., assumed the role of actor and director simultaneously) than the younger children.  Not only did the  younger children use less in-role communication to construct the pretense, this type of talk was qualitatively different from that of the older children. While in-role, the younger children tended to simply describe the pretend activities, whereas the older children continued to plan and organize the play.  Use of Language to Serve Pretend Play Interaction Analysis of the data from a conversational perspective addressed questions which 84  focused on the degree to which children’s pretend play conversations were continuous. Continuity of the pretend conversation is ultimately determined by the responsivity of players to their partners’ contributions. Recall that continuous conversation was defined as a sequence of adjacent turns for which the discourse topic was the same. A complete turn was the unit of analysis. The questions for this level of analysis were: 1. Are there age-related differences in children’s responsivity to their partners’ contributions? 2. Is responsivity related to the obligatory value of the preceding turn? 3. Are there age-related differences in the distribution of conversational function across turns?  Response Rate The contribution of each speaker may be continuous (responsive) or discontinuous (nonresponsive) with the discourse topic established in the preceding turns. The possibility that older children, who are more socially adept and more skilled at maintaining topic, are likely to produce more responses and longer continuous sequences than younger children (Garvey & Hogan, 1974) is explored next. Are there developmental changes in the continuity and maintenance of children’s pretend play conversations? To determine whether responsivity varies with age, each child’s turns were categorized for degree of responsivity, and the proportions of turns for each category was calculated as a function of the total number of turns. The categories used were:  85  response-obliged, response-nonobliged, nonresponse-obliged (both failure to respond at all, and thematically discontinuous contributions), and nonresponse-nonobliged. Table 5 presents the group means (and standard deviations) for the proportion of responsive turns by type. A t-test comparison of the group means revealed no significant agerelated difference for overall responsivity (t  =  1.22  ,  df  =  10, p  =  .52). Children in  both age groups were responsive in approximately 50 % of their turns (young M old M  =  =  51,  54), and approximately half of these responses followed an obliging turn.  Table 5 Mean Proportion (and Standard Deviation) of Responsive Turns of Young and Old Dyads Response type Response Overall Response Obliged Response Nonobliged  Young  Old  51(6)  54 (5)  23 (10) 28 (8)  23 (8) 31(2)  Since successful maintenance of a joint pretend interaction is also reflected in the length of continuous sequences, the number of responsive turns in each sequence was calculated. A comparison of the length of sequences revealed that all dyads produced sequences at least four exchanges long. The longest sequence produce by the older dyads was twice as long as that of the younger dyads, 14 and 7, respectively; but the mean length of sequences was only slightly greater for the older dyads (young M=3, old M=4).  86  Obligation as a Source of Responsivity The next analysis further explored the possibility that responsivity was related to the nature of the partner’s prior turn. The previous analysis established that half of the child’s responsive turns were obliged, but did not determine the relative rates of response to obliging and nonobliging turns. This analysis addressed the question: Do obliging turns increase the likelihood that a response will follow? It was expected that children would be more likely to respond if they were obliged to do so by the content or form of the preceding utterance. To determine whether responsivity was related to the nature of the partner’s prior turn, each child’s turns were categorized as obliging or not and the immediately following partner turns were categorized as responsive or not. The results are summarized in Table 6. Rate of responsivity was significantly higher following an obliging turn (X 2 (1, N  =  12)  =  11.13, df  =  1, p < .001), and there were no reliable  age differences.  Table 6 Mean Proportion of Responsive Turns Following Obliging or Nonobliging Turns  Speaker Obligation +  69  Partner Response Rate  87  45  Given the higher response rates to obliging turns, it seemed important to ask whether older children produce more of these opportunities to respond. Statistical comparisons of proportion of obliging turns revealed no age-related differences. For children in both groups, about 30% of the turns obliged a response.  Conversational Continuity Thus far, conversational turns have been analyzed for their links to prior  to  following turns. The final analysis explores the emergence of turns that contribute to continuity by linking both preceding  an following  discourse. Three degrees of  continuity were coded: (a) initiation, a turn that was thematically connected to the following turn only, (b) simple response, a turn that was connected to the prior turn only, and (c) complex response, a turn that was connected to both the preceding and following turns. It was expected that older children would produce a greater proportion of complex turns, in which the speaker attends and responds to the preceding turn, and also provides the partner with a structured opportunity to contribute to the conversation. To determine whether the degree of continuity varied with age, each child’s turns were categorized, and the proportion of turns for each category was computed as a function of total turns. The proportions are presented in Table 7. The proportions of turns in each category appeared to be similar for both groups and, indeed, statistical tests revealed no reliable age-related differences.  88  Table 7 Mean Proportion of Turns at Various Degrees of Continuity of Young and Old Dyads Turn Type  Young  Old  Initiation  20  18  Simple Response  38  39  Complex Response  12  16  Summary of Results: Conversational Continuity Analysis of the data from a conversational perspective provided a general description of the dialogue that supports social pretend play, but revealed few agerelated differences. For children aged 3 to 5, half of the turns in pretend play maintain the theme of prior turns, and 30% oblige their partner to respond. Response rates are higher when the prior turn has an obliging character. Although it seemed likely that the older children would be able to both respond and structure an opportunity for response in a single turn, no age-related differences in such complex turns was found, and they comprised only 15% of turns overall. There was some indication that the older dyads were more likely to create longer stretches of connected dialogue, but this difference was not reflected in other responsivity and continuity measures.  Collaborative Interaction The first set of analyses explored the way individual children used language to set-up the pretend play. The second set of analyses explored the way they used  89  language to maintain the social interaction. The final set of analyses combines these themes to explore how children collaborated to establish and maintain the joint pretend experience. The unit of analysis for this exploration was an interactive exchange, that is, the transition points between speakers/turns. An interactive exchange in which the second turn was contingent and responsive to the first was defined as collaborative. The questions with respect to collaboration were: 1. Are there age-related differences in the degree to which children collaborate  and in the nature of that collaboration? 2. Are there age-related differences in the proportion of in-role and out-of-role collaborative exchanges? 3. Is there a relationship between the nature of collaboration and speaker identity?  Degree and Nature of Collaboration by Age To determine whether the proportion of collaborative exchanges varied with age, each dyad’s social exchanges were categorized as collaborative or noncollaborative, and the proportion of collaborative exchanges was calculated. In line with the earlier responsivity analyses, the older group used a slightly greater proportion of collaborative exchanges than the younger group (young: M  =  52%, old: M  statistical tests revealed no reliable group difference (t  =  =  1.45, df  57%); however, =  10, p  =  .18).  Recall that collaborative construction of the pretense involves both negotiation about and elaboration of the social play. Through negotiations children reach a  90  consensus about their independent ideas; through elaboration they develop the established play further. Since other studies reported that negotiation about the roles  and themes begins to emerge during the fourth year, this type of collaborative interaction was expected to occur less frequently in the younger dyads. To determine whether the nature of the collaboration varied with age, each dyad’s collaborative exchanges were coded by type, and the proportions of collaborative exchanges for each category were computed as a function of the total number of collaborative exchanges. Descriptive statistics are presented in Table 8. A t-test comparison of group means revealed no significant age-related differences; both groups demonstrated a greater proportion of elaboration than negotiation.  Table 8 Mean Proportion (and Standard Deviation) of Collaborative Exchanges by Type of Young and Old Dyads Type of Collaboration  Young  Old  Elaboration  51(3)  47 (10)  Negotiation  32 (8)  33 (14)  Other  17 (8)  21(7)  Collaborative Exchange by Speaker Identity Position Generation of the pretend play is an ongoing task, and as such the players must continuously collaborate outside the frame (as themselves) and within the pretend frame (as new characters). Therefore, the next analysis examined the relationship between collaborative exchange types and speaker position. Prior analyses established that older 91  children produced more in-role metaplay utterances, but also that these in-role utterances were likely to be play description  --  a function that often involved  monologues. Would older children still produce more in-role language when only collaborative exchanges were analyzed? To determine whether the proportion of collaboration for each speaker position  varied with age, each dyad’s collaborative exchanges were categorized by speaker identity, and proportions were calculated as a function of total collaborative exchanges. Collaborative exchanges were coded in-role if the second speaker was in-role; the first speaker may or may not have been in-role. Assumed in-role turns/utterances were again collapsed with in-role turns/utterances. Turns in which speaker identity for the first utterance was coded as ambiguous were excluded from this analysis. Table 9 presents the group means (and standard deviations) for the proportion of collaborative exchanges by speaker identity for both age groups. A t-test comparison of group means for each category revealed significant age-related differences. The proportion of out-of-role collaborative exchanges was found to decrease significantly with age (t 10, p  =  =  2.83, df  =  .02), and there was a complementary tendency for in-role collaboration to  increase with age (t  =  1.42, df  =  10, p  =  .18). Additionally, there was a tendency  for all children to engage in greater proportions of in-role collaboration than out-of-role.  92  Table 9 Mean Proportion (and Standard Deviation) of Collaborative Exchanges by Speaker Identity Position of Young and Old Dyads  Speaker Position  *  Young  Old  Out-of-Role  33 (2)  25 (7)  In-Role  41 (11)  48 (6)  Ambiguous  26 (5)  27 (6)  p < .025, experiment-wise p < .075.  Relationship between Nature of Collaboration and Speaker Identity Position Having analyzed collaboration by type and speaker identity independently, once again the next logical question to ask was: Is there a relationship between these variables? Since the distribution for the proportions of each collaboration type was highly similar for both groups, the data were collapsed across age. To determine whether the type of collaboration was related to the child’s speaker position, all elaborative or negotiative exchanges were grouped by speaker identity and the proportions for each category were calculated as a function of total exchanges. The  combined categories were in-role elaboration, out-of-role elaboration, in-role negotiation, and out-of-role negotiation. These categories are logically independent by definition. Since conificts and disagreements arise both within and outside the pretend play, children may negotiate from both speaker positions, in-role and out-of-role. At the onset of play, negotiations 93  focus on reaching a consensus about “who will do what” and “who will use this or that,” but unexpected disagreements which arise while role playing are often settled without abandoning the fictional identity. Similarly, before transforming their identities, children establish the basic play outline and add sufficient details for the moment, and then continually elaborate the unfolding play while assuming a pretend role. The reconciliation of differences in play ideas or agendas, while simultaneously maintaining a character role, seems cognitively more complex than out-of-role negotiation. Considering this possibility in light of the results that children (a) collaborate more in-role than out-of-role, and (b) generally use more elaborative exchanges than negotiative exchanges, it is expected that in-role collaboration will be devoted to elaboration rather than negotiation. Statistical tests revealed two significant  relationships: elaboration occurred more in-role than out-of-role; and negotiation occurred more out-of-role than in-role (X 2 (1, N  =  12)  =  24.11, p < .001). These  values are summarized in Table 10.  Table 10 Mean Proportion (and Standard Deviation) of Elaboration and Negotiation for each Speaker Identity Position Nature of Collaboration  Speaker Position In-Role  Out-of-Role  Elaboration  66  21.5  Negotiation  33.5  52.5  94  Summary of Results: Collaborative Interaction Results of the last set of analyses revealed few quantitative differences in the degree or type of collaboration between the young and old children of this study. For all children, two tendencies were true: (a) a greater proportion of collaborative exchanges were devoted to elaborating the play than to negotiating about that play; (b) collaboration tended to occur more in-role than out-of-role. Further analysis of the relationship between the nature of the collaboration and the speaker’s identity confirmed expectations that out-of-role collaboration is devoted to the negotiation of differences in play ideas, and in-role collaboration serves to elaborate the shared meaning of the pretense.  Summary of Major Results To summarize the results of this study, we return to the hypotheses stated in the Introduction. It should be noted that the generalizability of these results is limited by the small sample size.  Hypothesis 1  -  The proportion of planning and organization utterances will increase and  the proportion of play description utterances will decrease with age. Result  -  This hypothesis was not confirmed. Interestingly, though, organizing and  planning were the most frequently used function types for both groups. The distribution of specific content types was as follows: 50% of utterances were used for organization and planning, 25% for play description, and 14% for object and identity transformation.  95  Overall, approximately 70% of the language used in pretend play was devoted to its construction and maintenance.  Hypothesis 2  -  The proportion of in-role nietaplay communication will increase with  age. Result  -  This hypothesis was confirmed by the fact that the younger children used  significantly less in-role communication for the construction of pretend play than the older children.  Hypothesis 3  -  Utterance types that function to initiate or terminate (e.g., role  transformation and organization) the play will be used with higher frequency out-of-role than in-role; and those that function to maintain or advance the play (e.g., planning and  play description) will be used more in-role than out-of-role. Result  -  This hypothesis was confirmed by the fact that out-of-role communication  served to organize and plan the play, while in-role communication most frequently functioned to describe the play. Additionally, age-related differences for the content of  in-role metaplay communication were revealed. Older children planned and organized the pretend play while simultaneously assuming a new role significantly more than younger children; younger children used descriptive comments about the pretend activities while in-role significantly more than the older children.  96  Hypothesis 4 Result  -  -  Responsivity rate will increase with age.  This hypothesis was not confirmed. At both age levels, approximately half of  the turns in pretend play maintained the theme of prior turns.  Hypothesis 5  -  Obligation to respond will increase the likelihood that a contingent  response will be provided. Result  -  The fact that response rates were higher following obliging turns than  nonobliging turns confirms this hypothesis.  Hypothesis 6  -  Turns that are linked to both the preceding and following turns will  increase with age. Result  -  This hypothesis was not confirmed. There were no differences in the  proportional distribution of turn types for the two age groups. The continuity of the conversation was maintained most frequently by use of a simple response (40%).  Initiations (20%) and complex responses (15%) were used less frequently.  Hypothesis 7  -  The proportion of collaborative exchanges will increase with age, and  the proportion of collaborative exchanges which involve reconciliation and compromise (i.e., negotiation) will increase with age. Result  -  This hypothesis was not confirmed. Although the amount of collaboration was  not significantly different for the two age groups, negotiations were less frequent than  elaborations.  97  Hypothesis 8  -  Collaboration among the players while simultaneously assuming a  transformed identity will increase with age. Result  -  There was a tendency for in-role collaboration to increase with age, and the  amount of out-of-role collaboration was found to decrease significantly with age. For both age groups, the proportion of in-role collaboration was greater than for out-of-role collaboration.  Hypothesis 9  -  Negotiation involving the resolution of personal differences will occur  more out-of-role than in-role, while elaboration involving the expansion of play ideas  will occur more in-role than out-of-role. Result  -  This hypothesis was confirmed by the fact that elaboration occurred  significantly more in-role than out-of-role, and negotiation occurred significantly more  out-of-role than in-role.  98  CHAPTER 4 DISCUSSION Overview This study set out to investigate how children coordinate their independent pretend ideas to create and sustain a novel, fictional reality. Emphasis was placed on the functional properties of language used in this situation. Specifically, this project was designed to (a) illustrate the ways that language serves to construct and maintain the pretend play experience while sustaining a conversational interaction, and (b) identify developmental changes in the use of pretend language between the ages of 3 and 5. Various dimensions of the language used during sociodramatic play were explored. Several interesting and enlightening facts about the character of pretend play in general, and about age-related differences, were revealed. The results related to communication about the pretend world will be discussed first, followed by those regarding the pretend play interaction as a collaborative conversation.  The Language of Sociodramatic Play It is apparent to anyone who has seen or heard children pretending that such play contains a high level of verbal activity. What may not be transparent, though, is the amount of the talk that is devoted to the construction of the pretend world in relation to the actual performance of the drama. Approximately 70% of the language used during this activity was found to serve the former purpose. This is compatible with existing empirical evidence that social pretense demands extensive verbal preparation (e.g.,  99  Auwarter, 1986; Garvey & Berndt, 1977; McLoyd et al. 1984; Sachs  Ct  al., 1984).  Time spent by children who are completely enthralled with their own inventiveness is obviously pleasing, as signified by the laughter and positive feelings that accompany such social play. It is assumed that the purpose of social pretend play is for children to experience the make-believe reality with a partner. It seems, then, as though the pleasure of the activity stems not only from acting out the make-believe events, but also from reaching a mutually satisfying consensus about those events. The inherent pleasure of creating the experience together must justify the effort required to accomplish the joint pretend endeavour.  Pretend Communication to Create and Maintain the Pretend Worlds Although children’s pretend play unfolds with seemingly great ease, the achievement of this multifaceted task should not be depreciated. The clear preponderance of talk devoted to building up the pretense suggests that communication is central to the creative process. This section discusses the ways that language functions to construct and maintain the pretend play. As described earlier, the language of sociodramatic play may be characterized along two dimensions. The content of metaplay communication specifies the purpose  served by the utterance: for example, to transform the child’s identity, to invent an object, or to prompt the partner about what to say next. The speaker position from which the utterance is produced refers to whether the proposal was explicitly stated by the children as themselves, or implied in the words and actions of the character roles  100  assumed by the children.  Communicative Purpose As expected, the various content types did not occur equally. Peers devoted the majority of their conversation about the pretend play to transformation of the situation  and activities, the players’ identities, and objects (75%). This finding is consistent with previous research findings (e.g., Doyle, Connolly, & Rivest, 1980; Goncu & Kessel, 1988; Matthews, 1977). In the present study, plans and play descriptions together constituted most transformations, followed by role and object transformations. In comparison, talk used to manage the play activity (i.e., organization utterances, e.g., attribution of costumes and objects, or marking boundaries of episodes through invitations and termination statements) was less frequent (25%). The following series of sequences contains examples of the different content types of metaplay utterances: Sequence 1. A: I’m gonna be the king {puts on a crown}. [role transformation] S: Ok. A: And you’re gonna be the beautiful girl. [role transformation] S: I’m a beautiful one but I have this on me {refers to a fancy hat}. [organization]  101  A: Does this look nice on me {refers to the fancy hat she is wearing)? [organization] S: Yeah.  Sequence 2. K: Mom could I cook dinner today? [plan] S: Yup you can honey and I hafta take baby somewhere. [plan] K: Pretend this is gonna be the delicious soup I made {refers to contents of the empty miniature pot on the stove). [object transformation] S: Okay {laughs}. Sequence 3. 5: No, pretend (that you) you hafta be my boss {points at K). [role transfonnation] S: Here’s your phone. [organization] K: Okay {picks up the phone). S: Hello {talks into phone}. K: Hello. S: No, you hafta ask me something. S: You hafta say, “Fix my stove!” [plan] K: Fix my stove {smiles).  In light of the evidence that play becomes less self-oriented with age (Garvey & Bemdt, 1977; Goncu & Kessel, 1988), it was expected that older children might use  organization and planning functions, which involve greater active participation by both 102  players, more than play description. Besides the fact that the most frequently used content type was different for the two age groups, the distribution of metaplay utterances did not vary significantly with age. The profiles for each group in descending order are as follows: younger children  --  play description (29%),  organization (25%), planning (21 %), identity transformation (9%), and object transformation (6%); older children  --  organization (30%), planning (23%), play  description (22%), identity transformation (6%), object transformation (5 %). This finding is actually consistent with previous findings (e.g., Field et at, 1982; Garvey & Berndt, 1977; Sachs et al., 1984; Goncu & Kessel, 1984; McLoyd, 1984). Perhaps, then, it is more accurate to think about the distribution of content types as a function of the pretense, rather than of the social interaction. Successful pretending with a partner relies on the assumption that each player knows who the partner is (role), what the partner is doing or intends to do (plan), where the partner thinks he is  (setting), and what the partner is using (object) (Garvey & Berndt, 1977). The key to this success is the communication of such meanings. Whether the information is expressed as a plan or play description is secondary. Furthermore, the relative distribution of utterances may reflect the nature of the activity itself. Because the actions performed by the character need to be coordinated regardless of whether they  are conventional or diverse, the players’ actions require greater talk about planning and organization than role transformations. As this need does not change with age, such play requirements may explain the common distribution of utterance types across the age groups.  103  Creating and Being In the preceding section, utterances that served to construct and maintain the  play were considered in terms of their specific purposes. In this section, these utterances are discussed in relation to those that functioned to enact the drama. The language of sociodramatic play may be viewed as serving two primary functions  --  to create and maintain the pretend experience, and to enact the reality of the  make-believe world. In keeping with Giffin’s (1984) continuum of out-of-frame to within-frame communication, it seems that communication about pretending and the pretend world takes place most clearly and overtly outside the play frame (out-of-role metaplay communication), and that communication for the realization of the pretend events and activities is embedded within the play frame (enactment). However, the distinction between creating and experiencing the pretend play is not that simple, as noted by past researchers (Auwarter, 1986; Giffin, 1984, Kane & Furth, 1993). This simple dichotomy between out-of-role metaplay communication and  enactment creates tension in coding the language used during social pretend play. The tension experienced in coding reflects tension in the play phenomenon itself  --  the  paradox inherent in creating and experiencing the pretend world . Communication 1 about the pretense from a position within the play frame is necessary to maintain the fictional reality (defined by out-of-frame/out-of-role communication). In other words,  1  Giffin identified a system of rules that influence the use of particular metaplay communication types. The use of in-role metaplay communication can be explained by the “Illusion Conservation Rule.” This rule, as formulated by Giffin, states “When constructing make-believe play, players should negotiate transformations with the least possible acknowledgement of the play frame” (1984, p.88). 104  some communication functions as both communication for the construction and  maintenance of the pretense, and for enactment of the pretense. Such communication fits first into the larger category of construction and maintenance ofpretense, and then is classed as in-role along the dimension of speaker identity (i.e., in-role metaplay  communication). The following sequence is an excellent example of how the children maintain their fictional role identities  --  mother and child  --  while introducing new plans  and transformations. Prior to the verbal exchange presented below, S interacted with the experimenter while A stood near the stove and pretended to cook. Following S’s announced role transformation, A skillfully used in-role speech for directive purposes by implicitly proposing a plan for herself.  S: Actually, I’m the mom and I have this {a crown} on me. A: Mom {stands behind S}? S: Yeah. A: Can I go out to our river and see what there is? 5: Okay {pause} and bring yours binoculars. A: I am bringing them {looks at S and then looks through the binoculars}. 5: Okay you come right back. A: {walks across the play area as though going to the river). S: {notices and suggests a possible transformation of setting) S: (You could pretend) you could pretend this is it, okay? A: Kay. 5: {goes back to the mirror).  105  A: {looks around the room through the binoculars). A: Mum, there was only a shark that bite me {walks over to S}. S: A shark {looks up)? A: Umhm. S: The shark.. .{A interrupts) A: He was crazy. S: Did he bite you at all? A: A: A: A:  No he won’t. See, he won’t bite you. Know why? He’s my friend shark.  S: Oh.  The general finding with regard to distribution of metaplay utterances by speaker identity showed that all children used a higher proportion of out-of-role metaplay utterances than in-role metaplay utterances. It was predicted that in-role communication, a more sophisticated mode of generating the pretense, would be used more by older children than younger children. As illustrated in the example above, in-role communication allows children to develop the pretend episode by acting “as if” they are merely responding to the unfolding events. As children get better at thinking and talking about pretend play, they demonstrate the ability to covertly, but nonetheless intentionally, develop the drama while assuming the identity of a character within the play frame (e.g., Auwarter, 1986). The fact that the proportion of out-of-role metaplay utterances decreased clearly with rising age in favour of in-role metaplay utterances suggests that out-of-role communication for the construction of the pretense developmentally precedes in-role communication. An alternate view of the children’s identity while constructing the play is put 106  forth by Kane and Furth (1993). These authors place speaker identity at the heart of children’s construction of the pretend reality. For them, the identity assumed by the child (either one’s own, or an assumed fictional identity) while talking about the play is a primary indication of the child’s degree of engrossment in the play activity. In their view, however, “identity should not be treated as an intentional posturing that requires a conscious, or metacognitive, decision. Rather, it is engulfed in the doings of the child” (p. 201). But, this analysis does not appreciate the complexity of in-role metaplay  communication. In the following sequence K used her role as “waitress” to gain command of the play situation. Such in-role communication calls Kane and Furth’s impressions into question. By limiting the availability of the menu items, there is little doubt that K intentionally and consciously acted with an ulterior motive to dictate her partner’s choice of food. This is especially evident in the rationale provided by K to justify her denial of T’s request (underlined). Despite T’s persistence, and seriousness in addressing K by her real name, K did not change her mind . 2 K: Okay what would you like? T: I said this {insistently, points at trail mix). K: Nope {pushes hand away). T: I want it Kendall {almost begging). K: No! T: Well, well I K: You can’t have it! K: (It’s not) it’s not in.  2  The symbols < > mark overlapped speech. 107  T: <(1)1 want it Kendall>! K:  <We we only had that> we only had that (that for for urn foil for lots of people have a party here. T: I want this {points to trail mix).  K: No you may not have it! K: You have to choose this, this, this, this, this, this {points at a variety of items). T: I want this {p} for dessert. T: If you get it in, okay? K: Alright you can have that {trail mix}. K: Now you hafta choose something for your meal.  The Director versus The Director/Actor Attention is now turned to the specific purposes served by metaplay communication for each speaker identity position. All content types occurred in both speaker positions, but the distribution of utterances by content type was not symmetrical for the two speaker positions. Overall, organization utterances were produced most often while the children were out-of-role (out-of-role  =  20%, in-role  =  7%);  conversely, play description occurred most often from an in-role position (out-of-role 6%, in-role  =  =  18 %). The proportion of plans was similar for both speaker positions.  The complementary relationships between specific functions and speaker position --  organization utterances occurred primarily out-of-role and play description occurred  mainly in-role  --  are of particular interest from a developmental perspective. In terms  of general development, specific language functions which are later developing are often  108  used less frequently at first, than functions which emerge earlier. Does this imply that language functions that occur with low frequency (e.g., in-role organization) are later  developing or developmentally complex? There are at least two reasons why a particular language function may occur with low frequency: (a) the language function is not highly used by adults in general, and (b) the language function is complex and not well learned yet by children. If the former explanation accounts for frequency of occurrence, then mere low frequency of occurrence does not mean that the language function develops later because it always occurs with low frequency. However, if a developmental pattern is observed, i.e., there is a shift in the frequency of use over time, then increased frequency of use speaks to the growing competencies of children. For example, contingent queries as conversational management features are used with low frequency by adults. Nevertheless, the use of this discourse management device is observed to increase with age (Garvey, 1977). Assuming this, and also assuming that in-role metaplay communication is later  developing than out-of-role metaplay communication, one might speculate that in-role organization, which was the least frequently employed in-role communication type, is a later developing function than in-role play description. This interpretation is confirmed by the fact that the younger children, who are developmentally less mature used significantly less in-role organization than the older children who demonstrated a shift towards higher use of this more advanced function. Reconsideration of the nature and purpose of these two content types (i.e., how the utterances fit into the different realities) offers another possible explanation for these  109  relationships. “Structuring jobs,” or organization of the play, such as the distribution of play materials, or the clarification of children’s personal rights, seem to be closely associated with the real world, and thus are produced while the children assume their own identity (out-of-role); whereas play descriptions closely resemble enactment, and as such the commentary is produced as viewed through the eyes of the character in the imaginary world. Such associations call into question the logical independence of these variables, and the potential impact of these decisions on the results. But as noted earlier, the categories and dimensions are logically independent by definition and by the fact that examples of all combinatorial categories (e.g., in-role organization and out-of-  role play description) were found in the data. The content of the children’s in-role utterances, by definition, is usually characteristic of the adopted identity and appropriate for the pretend situation, but the specific communicative function of utterances is not restricted to a particular type. In fact, the frequency with which specific content types (e.g., play description) were used while assuming a fictional identity varied with age. All children used play description most often as a means of directing the play from an in-role position, but younger children employed this function type significantly more often than the older children. On average, 48% of the young children’s in-role metaplay communication consisted of  play description, whereas only 28% of the older children’s in-role talk about the play was descriptive. Qualitative differences also emerged. The in-role play description of younger children could best be described as narration of the pretend episode through the eyes of  110  someone who lives in the pretend world. The children explained the unfolding event of the imaginary world, but the children themselves were not necessarily involved in the activity, or at least, their actions were not the focus of the description. The following sequence of utterances exemplifies how a child in the younger group used play description. All of K’s utterances were judged as play  description. It should also be noted that prior to this sequence there was no announced identity transformation on K’s part, as such her identity was coded as  assumed to be in-role. K: {slams the cupboard door and looks up} K: What shots {uncertain about the last word}? K: Something is so noisy {looks around}. S: {stands up to see what K talks about) K: It’s a dinosaur beating on the window {looks at window}. K: But I don’t think he’s there anymore. K: Maybe I should look out to see if he’s still there {goes to window). K: Might not be there anymore {lifts blinds}. S: {watches} K: Nope he’s not there {looks back at S}. K: His face was there a minute ago but we weren’t looking. 5: {sits by stove again) K: {picks up two wands and begins to twirl them}  111  The older children, in contrast, seemed to be talking to themselves while communicating the pretend transformations to anyone who happened to overhear them. The play description of older children could best be described as explanations, or emphases, of their own behaviours within the pretend world. An example of such play description (underlined) is presented below. The episode begins with the announcement of T’s plan to travel. Nearby, K struggled to put on a cape. In the meantime, T talked to himself as he prepared for his trip. When K overheard T’s plans, she proposed to join him. T: I’m going on a trip. T: There I brought my hat with me {picks up a safari hat and takes it over to the suitcase). K: {tries to put cape on again} T: There I brought my hat with me {picks up safari hat}. T: Put that in the suitcase {puts something in suitcase). T: Get binoculars. K: (Can I) can I come with you ‘cause I’m going on a trip too {watches T}. T: No you may not {changes voice quality). K: I can so! T: {continues to pack} K: {requests help from experimenter to put on the cape) K: {looks out the window). T: Yeah I’m in my hotel {sits on a mattress near the window). K: {watches T}. T: I think I’ll just unpack my stuff. T: I’ll put this down {takes hat off).  112  K: That’s in my {emphasized} home so you can stay with me now {stands on chair near T and looks down}. T: Oh okay I’ll share my home.  Pretend Communication as Conversation Continuity Turning now to a discussion of the results pertaining to pretend play communication as conversation, recall that the questions and analyses related to this aspect of language focused on the ties between conversational turns, not merely on the  social interaction between players. This study revealed that young children between the ages of 3 and 5 demonstrated the ability to engage in continuous verbal exchanges with a peer. The children in both groups provided a thematically continuous contribution about 50% of the time. This is consistent with the recurrent finding of past research that children have acquired the interactional skills necessary to carry on conversation by age 3 (Garvey, 1974, 1977; Garvey & Hogan, 1977; Keenan, 1974; Mueller, 1972). In examining the extent to which children’s speech was contingent on their partner’s contribution, it was expected that the older children would show improvements in conversational continuity. This prediction was not confirmed. Children in the two age groups provided thematically continuous turns at the same rate. There are two potential explanations for this discrepancy in expected and actual findings. The first appraises the implication of definitional differences and the second considers the effects of insufficient sample size. 113  The operational definitions for response, and consequently simple and complex response, used in this study were compared with those of similar past studies. Upon initial review, it appeared as though the criteria used to identify such turns were similar in all studies. Often the examples of the codes and categories provided in other studies were prototypical cases in which coding decisions were apparent. However, empirically, coding decisions are not always clear-cut. In ambiguous or less typical cases, it is possible that different conventions were applied by the investigators. One might expect that such cases would be in the minority and not influence the results. However, if the sample size is small, as it was in the present study, relatively few discrepancies in coding could significantly affect the results. As well, within group individual/dyadic differences could also skew the results when the sample size is insufficient. The following sequence illustrates the potential effects of both coding decisions and insufficient sample size. Prior to the sequences presented below the children were engaged in a restaurant scene, with K as “waiter/cook” and T as “customer”. While waiting for K to prepare the order, T attempted to expand the play by proposing to order a pizza for K. K rejected T’s plan. Many of the exchanges in this sequence were coded as discontinuous because according to the operational definition established, the turns were not thematically coherent.  1. T: She wants to talk to you {passes phone to K}. 2. K: {continues to dial using another phone}  114  3. T: I’m not giving you the right number. 4. K: Hello {into phone, pauses}. 5. T: 222 5555 {pushes buttons on phone). 6. T: She need it, she need it {talks on phone as though speaking to same third person as K}. 7. T: Don’t listen to her. 8. T: I’m speaking in both phones and I’m just saying hello and I want the hottest anchovies but {distracted briefly}> 9. K: I don’t need it thank you {on phone}! 10. T: Yeah that was just <pizza>. 11. K: <And> don’t listen to him! 12. T: No don’t listen I’m gonna shut her off {tries to hang up K’s phone). 13. T: Ya bye {on own phone).  According to the operational definitions applied in the current study, turns 2, 4,  5, 9, 10, and 11 were coded as nonresponsive. In terms of the propositional content, the turns were not continuous, but it is apparent that the intent of each speaker’s contribution was to convince the partner to agree with his/her own idea. If the topic was judged to be “what to do next”, then the above sequence could be categorized as continuous. From this example it is evident how coding practices may influence final results. As well, in this short example, there were six discontinuous turns. Although this is a relatively small amount when considering the equally small number of total turns, the proportion of responsive turns can be reduced by such cases. Furthermore,  reduced responsivity rate for one dyad can affect the overall group mean.  115  Both explanations offer plausible ways to account for the discrepant fmdings regarding age differences. A comparison of the group means of the present study, with the findings of Garvey & Hogan (1972), illustrates this point. Garvey and Hogan studied the language of children in the same age range using a similar research design  and coding scheme as the present study. The overall mean percentage of turns coded as forming a continuous exchange was similar in both studies (Garvey and Hogan present  -  -  59%,  53%). In the Garvey study age group differences were found, but means  actually ranged from 48% to 77% among the older children and from 21% to 64% among the younger. Group means for the present study fell into the low and high end of these ranges for the old (54%) and young dyads (5 1%), respectively. Since there were only three old dyads, it seems possible that the lower mean of the old dyads could be accounted for by the discontinuous “pizza argument” sequence in dyad K-T. However, when the pizza argument sequence is recoded as continuous, the group mean is not affected by recoding. Therefore, the definitional issue seems to be the best explanation to account for the divergent results of the present study.  Obligation as a Source of Responsivity Having considered the extent to which children’s pretend play dialogues are continuous, one might next wonder, how do children achieve this continuity? One means can be seen in the data regarding obligation. The children were observed to use obligation as a strategy for securing the partner’s future involvement approximately 30% of the time, and responses to these strategies were 15% higher than otherwise. The fact  116  that children are more likely to respond when obliged is consistent with Mueller’s (1972) ranking of utterance form (e.g., command or question) among the predictors of a successful exchange. Like greetings or invitations often used by adults as a conversation opener, for example, obliging turns create a state of mutual engagement (Schegloff, 1968, as cited in Sacks et al., 1974). Although the use of such a strategy was highly successful, it was not a perfectly predictable means to securing a response. The number of missed turns or nonresponsive contributions following an obligation, observed in the conversations of children, was higher than would be expected for adults. Under similar circumstances, adults virtually always respond or seek to repair by providing an excuse for the missed turn. Adults also seem to understand the force of “obliging” utterances and their value in creating continuity.  For example, parents often ask children questions to check if  the children are listening.  In doing so, adults can be seen to manipulate obligation to  ensure the continuity of future discourse by pulling their conversational partners into the verbal interaction. The child’s emerging awareness of the need to structure the play and conversation so that the partner is more likely to join, and remain a part of, the interaction is an important and necessary conversational skill. This strategy was used 20% of the time as an initiation, and in combination with a contingent response to the prior turn (complex response) only 15 % of the time. In contrast, children provided simple responses to the prior turns approximately 40% of the time. This data suggest that children between the ages of 3 and 5 are not accomplished in setting up  117  expectancies to ensure the continuity of subsequent discourse. Perhaps a better picture of the developmental progression towards greater conversational continuity can be seen by exploring the use of obligation by school age children. As well, age-related differences may emerge by differentiating various types of obliging turns (e.g., request for information vs. requests for confirmation or tag questions).  Collaboration Pretend play conversations, similar to adult conversations, involve turn-taking, the establishment of a play theme/topic, and the continual integration and coordination of play ideas by elaborating the established understanding and/or negotiating ideational discrepancies about that understanding. With these characteristics of conversation in mind, the final set of questions explored how children collaboratively constructed the pretend play, through either elaboration or negotiation during the joint play experience. The children in this study provided a response that was contingent on the discourse topic of the preceding turn approximately half of time. It is not surprising, then, that collaboration occurred in similar proportions. The older children produced a slightly higher proportion of collaborative exchanges (57%) than the younger group (52%); however, this difference was not statistically reliable. Of greater interest, however, is the fact that collaboration occurred more while the children assumed a fictional identity than while they still held their own personal identity. The nature of the fictional roles adopted by the children provides some explanation for this result. Since the children often assumed complementary roles (e.g.,  118  “mom” and “child”, “cook/waiter” and “customer”), the activities of the partners were linked through appropriate reciprocal behaviours. Before they entered their roles, however, there was more opportunity for initiative and independent action. One might speculate that in-role collaboration is the result of role play that exhibits an integrative interpersonal structure. Iwanaga (1973) found that as children get older, their play becomes more integrative, reflecting the enactment of complementary social role relationships. Fein and Stork (as cited in Rubin et al. 1984) report a striking increase of such reciprocal relationships in the play of children between the ages of 3 and 5. Although the present study did not inspect the interpersonal structure of the dyadic play per  Se,  the findings  related to collaborative exchanges may shed some light on these earlier studies. It was found that the older dyads produced significantly less out-of-role collaborative exchanges (25%), than the younger dyads (33%), in favour of a shift towards increased  in-role collaboration (old M48%, young M=41 %). If, as argued above, this increase reflects the nature of the roles, the same explanation could apply to the reported increase in social reciprocity. Children are more reciprocal because the roles imply this style of interaction. The questions, then, are: Why these roles? And why more of them later? So far the findings related to the amount of collaboration in general, with respect to speaker identity, have been considered. But collaboration is of two sorts, negotiation and elaboration. Each will be considered in turn. Although the children observed spent over half of their time collaborating, the  119  play was not smooth and easy-flowing. The partners spent a fair amount of the time (about 30%) negotiating (in the sense of resolving disagreements). This type of collaboration occurred significantly more while children were out-of-role than in-role. The focus of arguments was often on personal concerns such as the possession of objects, or status (i.e., who is supposed to, or allowed to, do what). Two possibilities for the tendency for negotiation to occur mostly out-of-role are suggested. First, perhaps players often step out-of-role to negotiate because they need to be free from the constraints of the adopted role. As themselves, children do not have to contend with appropriateness or legitimacy of the argument from the position of the assumed character role. For example, “baby” is completely dependant on “mother” and has no control over situation. But the child pretending to be “baby” must have the opportunity to express his wishes for the play to be pleasurable. The child assuming the “baby” role can only be active in directing the play if the fictional role is dropped. Second, one might speculate that stepping out of the assumed role to settle disagreements reduces the complexity of thought involved.  When out-of-role the children can have  less to manage. The burden of assuming a fictional character-role is removed and the sole purpose of the talk is to convince the partner of one’s own personal wishes. This is not to suggest, however, that extended negotiations occur exclusively outside the play frame. Actually, the sequence below between A and S clearly illustrates both in-role (lines 1-9) and out-of-role (lines 13-22) negotiation. During the episode prior to the sequence presented below, “mother” (S) had prepared and served dinner for her “daughter” (A). The play was then frozen  120  momentarily, because S stepped out of her role to request a drink of water. Upon return to the play scene, S cleaned up the dishes while A took over the cooking task  --  a duty which was inconsistent with her role (i.e., children do not  cook). S rejected A’s violation of the role expectancy.  1. S: All the dishes <hafta  stay>. 2. A: <Mom I’ll> cook my other more. 3. S: No no I’ll cook it. 4. S: No <I’ll cook it>. 5. A:  <NoI’llcook>.  6. S: (You’ll) you’ll burn yours hands. 7. A: No I’m old enough to cook, remember. 8. S: You’re not thirteen. 9. A: No pretend <that I am>. 1O.S:  <I hafta> put all the food in there {tries to open cupboard}. 11. A: Okay {opens the cupboard door).  12. S: <Get all the food>. 13. A:  <But pretend> you let me cook, okay?  14. S: (Can you go) can you go get that food {points to food on table)? 15. A: {gets food and puts it in cupboard). 16. A: But still pretend you let me cook. 17.A: <Okay>? 18. S: <Urn you> put only the food in there {p} and I’ll put this in here {puts pot and plate in sink}. 19. A: But pretend still you let me cook {looks in the cupboard). 121  21. S: No no you could cook dinner for us. 22. A: I am {continues to look for things in cupboard under the stove.}. 23. S: {walks around the stove and sink}.  There are several intriguing elements about the negotiative exchanges and the progression from in-role to out-of-role worth noting. S and A persistently negate each other’s proposition and, initially, explicitly marked this rejection with the word “no” (line 3- 5). As the argument continued, S tried to discourage A from cooking by appealing to logical reasoning (line 6). Then, A contradicted S’s rationale with a plea to remember the pretend plan, as though the age of the “child” had been previously determined (line 7). A eventually realized that as long as she assumed the role of “child”, she would loose the argument because “child” is subordinate to “mother”. She reasserted the equality of her peer status in real-life by stepping out of her role, and used pretense in an argumentative sense to redefine the rules of play (line 9). Unlike the exchanges presented in the example above, the majority of collaborative exchanges functioned to elaborate the pretend play. Elaborative exchanges occurred with higher frequency while the children were in-role, than out-ofrole. Elaboration of the play in accordance to role expectations (i.e., performance of activities which closely resembled those conventionally associated with the real-life role)  are most accepted by the participants. One older child, however, violated the literal role expectations of the role he was enacting as a means of elaborating the play.  122  As illustrated below, elaboration of the play may evolve from or result in the  negotiation. The “pizza argument” described earlier is a good example of how children elaboration and negotiation are intertwined. Recall that T’s attempted elaboration resulted in a disagreement: T was pretending to be a “customer” and K was the “cook.” As T waited for K to prepare the food, he offered to order her a pizza, but K rejected his plan. Then T persisted with his elaboration by skillfully involving an imaginary third party.  T: {watches K briefly then reaches for phone). K: Two apples {sets two apples on a plate}. T: I’m gonna order you a pizza {puts phone to ear). K: No you’re not {looks at T}. T: Do you like anchovies in it? K: No thank you and I <don’t need pizza>. T:  <I’ll have a> large pizza and the hottest anchovies you can find {places order by phone then hangs up}. K: I don’t need anything to eat {turns to T, holds two prepared plates by stove}!  T: Well I already called them to drop it off. T: Apple soup {pretends to eat quickly)? T: Yum. K: {laughs and reaches for phone). T: Hey can I have the phone please { serious tone of voice, grabs phone from K’s hands).  123  K: No! K: I’mgonnacallit <back> {gives T the phone she was holding and picks up another phone on the table)! Children struggle for phone. T:  <Remember it’s no> remember it’s no fighting. K: I’m gonna call it back.  T: The hottest anchovies you can find {holds receiver to ear again}. T: She wants to talk to you {passes phone to K). K: {continues to dial using another phone} T: I’m not giving you the right number. K: Hello {into phone, pauses). T: 222 5555 {pushes buttons on phone). T: She need it, she need it {talks on phone as though speaking to same third person as K). T: Don’t listen to her. T: I’m speaking in both phones and I’m just saying hello and I want the hottest anchovies but {distracted briefly}> K: I don’t need it thank you {on phone)! T: Yeah that was just <pizza>. K: T: No don’t listen I’m gonna shut her off {tries to hang up K’s phone). T: Ya bye {on own phone}.  124  <And> don’t listen to him!  Methodological Observations Taxonomy Generating a taxonomy for the language of social pretend play was a formidable task accompanied by many frustrations and challenges. Reference to past taxonomies was initially beneficial in learning about the variety of categories associated with different dimensions of the language being studied. However, the application of those earlier coding schemes was difficult for two reasons. First, the categorical definitions provided were often general and vague. Secondly, the goal of each classification system was slightly different and, therefore, not particularly suited to the interests of the current project. Previous classification systems, nevertheless, formed the foundation of the current taxonomy. Many of the categories were drawn from existing schemes while others were generated to code finer distinctions in the ways language was used. The goal of the present research was to document how many times a certain behaviour occurred and to compare the frequency of occurrence across two age groups. The challenge lay in designing a coding scheme which would capture interesting age-related differences. Attempts were made to ensure that the codes were specific enough to capture qualitative differences between the age groups yet general enough to reveal the connection between the differences, as well as allow for the interpretation of developmental changes (i.e., to identify the emergent behaviours of the younger children which are earlier forms of those mastered by the older children). Another challenge stemmed from the application of what was felt to be a workable coding scheme, as  125  interpretation of the pretend play behaviour was subjective and qualitative in nature.  Subjects In this study, as in others, the focus was on how children between the ages of 3 and 5 use language during social pretend play. The study was not interested in whether or not children of this age range are capable of this activity. Thus, it was important to select “master players” as subjects to get the best picture of language used in this situation. In doing so, one runs the risk of observing children whose skills fall into the upper extreme of the normal range of that age group. This risk was greater for the younger children because the skills being studied are emergent during the fourth year. If the younger children observed represented the upper end of the age distribution, but the older children were more representative of the general distribution, then it is likely that the age difference was reduced, and possibly masked, in some cases. This seems particularly likely since subject recruitment was not an easy task. Although findings in the literature suggest that 3 year olds engage in sociodramatic play, the generalizability of this finding is questioned by the difficulty that was experienced in locating subjects to fit the lower age group of this study.  Sociodramatic Play as a Context for Learning Evidence presented in this study shows that preschool children use their emerging language skills to coordinate their personal make-believe representations with a partner, and bring their expanding knowledge of the social world to the conversational  126  interaction. Past research (e.g., Auwarter, 1986; Garvey & Krammer, 1989) has indicated that pretend play may encourage the use of specific linguistic forms such as modals and quasi-modals to set up the play (e.g., “Let’s play house and you could be the baby.”), temporal and locative expressions to talk about the timing of events (e.g., “You better hurry up before the plane is gonna leave.”), clausal conjunctions to express complex sentences as in play description, etc. As well, social issues, like control and compromise, are expressed through negotiations within interactive pretend play (e.g., Howes, 1992). To ensure mutual satisfaction, both children must balance their desire to control the play activity and the need to compromise and incorporate the wishes of their partner. Thus, social pretend play provides a context for children to practice and master their communicative and social skills (see also Faver, 1989; Doyle & Connoly, 1989 for more examples). What is it about the context of social pretend play, though, that facilitates learning and practice of social and conversational skills? Children are free to manipulate the conditions of the make-believe reality within the general constraints of their script knowledge for an event; they are at liberty to revise plans, redefine the setting or their roles, and/or re-establish the conditions of the interaction to accommodate their representations about the pretend world. This freedom to control the situation, along with the social status often associated with assumed roles (e.g., “mother” and “child”), often allows children to dominate the activities of their partners. This is not necessarily true of nonplay peer interactions. Also, children are often less inhibited to persist with their own agendas even if displeasing to their partner because the negative, real-life consequences of such behaviours are not suffered.  127  In terms of compromise, on the other hand, the illusion of make-believe may allow children to more readily give up their own agendas in favour of the partner’s, because, in a sense, there is less at stake. As the events and activities are “just pretend”, social status and control may be more easily relinquished in this context as compared to real-life interactions. A child who is typically more dominate may succumb to the wishes of a more passive child in the context of make-believe more easily than in nonpretend play situations. Although pretend play is novel and spontaneous, the play events and sequence of activities are supported by the knowledge of social scripts. When the script is shared and highly familiar (e.g., dinner time), the children’s energy and concentration may be focused on elaborating the shared knowledge and negotiating variations of the theme, rather than on strictly adhering to the expectations of the script. Corsaro (1983) found that children engage in longer sequences of play interaction if the topic is a shared one. In these cases, the children are not burdened with the task of establishing thematic content and; therefore, they can focus on the social business of the play, that is, interacting with the partner.  Directions for Future Research Although it is important to examine the integration of different language functions, one advantage of treating the aspects of the language separately is to get a sense of the development and learning that occurs in different domains. Behaviourial changes in the play of children provide insight into the development of cognition,  128  socialization and communication. Looking more closely at the language used during sociodramatic play, we can observe the development of skills in different domains. From a cognitive/ideational standpoint, one might ask “What is the play about?” or “How complex is it?” If interested in the development of socialization, one might ask “How often does the child initiate a new theme during the play session?” or “How does the child interact with the partner?” It is difficult, however, to understand these distinct  areas of learning when all aspects of development are intertwined. A means of exploring the relative contribution of language skills, social skills  and cognitive skills is through the observation of children with a specific language impairment (SLI) engaged in sociodramatic play. These children experience a language delay in the absence of other developmental delays. For these children, social and cognitive skills are assumed to be similar to age-matched peers, but language abilities are comparatively reduced. On the other hand, children with SLI are socially and cognitively more mature than children with equivalent language abilities (i.e., younger typical children). The developmental gap in language and nonverbal cognitive abilities offers researchers an opportunity to tease apart the relative role of communicative competence in sociodramatic play. An area for future research would be to explore the pretend play of specifically language impaired children with peers. If the play of children with specific language impairment was compared to that of both age-matched and language-matched peers, would the children with SLI resemble his/her age-matched peers or language-matched peers?  “How does language impairment affect the amount  and complexity of collaborative pretend play?”  129  Conclusion Based on the observations of the current investigation, without a doubt, social pretend play is extremely language based. Children demonstrate the ability to engage in multiple levels of role-taking and they move subtly in and out of different roles. The roles assumed by children are reflected in the language used during the play activity. At one level, children may be viewed as directors of the drama while holding their personal, everyday identities. For example, before a child’s actions and words become representative of an identity other than his own, the child usually announces a role transformation (e.g.,”I’m gonna be Superman”). At another level, children can be seen as actors playing the parts of a characters in the pretend world. Here, the words and actions of the children serve purely to enact the role of the fictional identity. Children also may be seen to assume both positions, that is, they directs the pretend play activity as one of the actors in the drama. For example, the “waiter” may say to the “customer”, “Hurry get out! There’s a fire in the kitchen!” With this interpretation in mind, then, there are two streams for the construction  and maintenance of the pretend play  --  one that lies outside the make-believe world and  the other that lies on the inside. The dual role of director/actor, is probably more accurately one of co-director/actor because it is the stream within the pretend world in which children collaborate most. Pretend play which is mutually satisfying for the players is usually the product of both children’s contributions. Collaboration is the key to successful sociodramatic play as in most real-life social/conversational interactions.  130  REFERENCES  Auwarter, M. (1986). Development of communicative skills: The construction of fictional reality in children’s play. In 3. Cook-Gumprez, W. Carsaro, & J. Streek (Eds.), Children’s world and children’s language. (p. 205-230). Berlin: de Gruyter. Bateson, G. (1955). A theory of play and fantasy. Psychiatric Research Reports,  2, 39-5 1.  Black, B. (1989). Interactive Pretense: Social and symbolic skills in preschool play groups. Merrill-Palmer Quarterly, 3(4), 379-397. Bretherton, I. (1984). Representing the social world in symbolic play: Reality and fantasy. In I. Bretherton (Ed.), Symbolic play: The development of social understanding. (p. 144). New York: Academic Press. Connolly, J., and A. Doyle. (1984). Relation of social fantasy to social competence in preschoolers. Developmental Psychology, 2Q, 797-806. Corsaro, W. (1981). The development of social cognition in preschool children: Implications for language learning. Topics in Language Disorders, 2(1), 77-96. Corsaro, W., and G. Tomilinson. (1980). Spontaneous play and social learning in the nursery school. In H. Scwartzman (Ed.), Play and Culture. West Point, N.Y.: Leisure Press. Doyle, A., J. Connolly, E. Rivest. (1980). The effects of playmate familiarity on social interactions of young children. Child Development, j, 2 17-225. Doyle, A. , and J. Connolly. (1989). Negotiation and enactment in social pretend play: Relations to social acceptance and social cognition. Early childhood Research Quarterly, 4(3), 289 -302. Dunn, 3. and N. Dale. (1984). I a daddy: Two year olds collaboration in joint pretend with a sibling and with a mother. In I. Bretherton (Ed.), Symbolic play (pp. 131-158). New York: Academic Press.Faver, 3. (1992). Communicating shared meaning in social pretend play. Early Childhood Research Ouarterly, 2, 501-5 16. Fein, G. (1981). Pretend play in childhood: An integrative review. Child Development, 1095-1118. Fein, G. (1975). A transformational analysis of pretending. Developmental Psychology, 291-296.  131  ,  .11,  Fein, G., E. Moorin, & J. Enslein. (1982). Pretense and peer behaviour: An intersectional analysis. Human Development, , 392-406. Fein, G., & S. Schwartz. (1986). The social coordination of pretense in preschool children. In G. Fein and M. Rivkin (Eds.), The young child at play: Reviews of the literature. Volume 4. Washington, DC.: National Association for the Education of Young Children. Field, T., L. DeStefano, & I. H. Koewler III. (1982). Fantasy play of toddlers and preschoolers. Developmental Psychology, j, 503-508. Forbes, D. & G. Yablick. (1984). The organization of dramatic content in children’s fantasy play. In F. Kessel and A. Goncu (Eds.), New Direction for Child Development Research: Analyzing Children’s Play Dialogues. (p. 23-44) San Francisco: Jossy Bass. Forbes D., M. Katz, & B. Paul. (1986). “Frame talk”: A dramatistic analysis of children’s fantasy play. In E. Mueller & C. Cooper (Ms.), Process and outcome in peer relations (pp. 249 -265). New York: Academic Press. Forys, S.K.S., & L. McCune-Nicolich. (1984). Shared pretend: Sociodramatic play at three years of age. In I. Bretherton, (Ed.), Symbolic play. (pp 159-194). New York: Academic Press. Garvey, C. (1977). The contingent query: A dependant act in conversation. In M. Lewis and L. Rosenbium (Ms.), Interaction, conversation, and the development of language. New York: Wiley, 1977. Garvey, C. (1984). Children’s talk. Oxford: Fontana. Garvey, C. (1990). Riy. Cambridge, Massachusetts: Harvard University Press. Garvey, C., & R. Berndt. (1977). Organization of pretend play. Catalogue of Selected Documents in Psychology, 2, 1589, 1-23. Garvey, C. & R. Hogan. (1973). Social speech and social interaction: Egocentrism revisited. Child Development, 4.4, 562-568. Garvey, G. & T. Krammer. (1989). The language of social pretend play. Developmental Review, , 364-382. Giffin, H. (1984). The coordination of meaning in the creation of shared make believe reality. In I. Bretherton (Ed.), Symbolic play (pp. 73-100). New York: Academic Press. 132  Goffman, E. (1974). Frame analysis: An essay on the organization of experience. Cambridge MA: Harvard Unversity Press. Goncu, A. (1993a). Development of Intersubjectivity in social pretend play. Human Development, 3 (4), 184-198. Goncu, A. (1993b). Development of intersubjectivity in the dyadic play of preschoolers. Early Childhood Research Ouarterly, , 99-116. Goncu, A. & F. Kessel. (1984). Children’s play: A contextual-functional perspective. In F. Kessel and A. Goncu (Eds.), New Directions for Child Development: Analyzing children’s play dialogues. (pp. 5-22). San Francisco: Jossey-Bass. Goncu, A. & F. Kessel. (1988). Preschoolers’ collaborative construction in planning and maintaining imaginative play. International Journal of Behavioral Development, jj(3), 327-344. Flowes, C. (1985). Sharing fantasy: Social pretend play in toddlers. Child Development, 1253-1258.  ,  Howes, C. (1992). The collaborative construction of pretend: Social pretend play functions. Albany: State University of New York Press. Howes, C., 0. Unger, & L.B. Seidner. (1989). Social pretend play in toddlers: Parallels with social play and with solitary pretend. Child Development, , 77-84. Iwanaga, M. (1973). Development of interpersonal play structure in three, four, and five year-old children. Journal of Research and Development in Education, ., 71-82. Kane, S., & H. Furth. (1993). Children constructing social reality: A frame analysis of social pretend play. Human Development, (4), 199-214. Keenan, B. (1974). Conversational competence in children. Journal of Child Language, 1(2), 163-183. le Normand, M. T. (1986). A developmental exploration of language used to accompany symbolic play in young normal children (2-4). Child Care. Health and Development, j., 121-134. Levy, A. (1984). The language of play: The role of play in language development. A review of the literature. Early Child Development and Care, 17, 49-62.  Lyytinen, P. (1991). Developmental trends in children’s pretend play. Child Care. Health 133  Development,  fl, 9-25.  Mathews, W.S. (1977). Modes of transformation in the initiation of fanstasy play. Developmental Psychology, j, 212-216. McCune-Nicholich, L. (1981). Towards symbolic functioning: Structure of early pretend games and potential parallels with language. Child Development, , 785-797. McCune-Nicholich, L., & S. Carroll. (1981). Development of symbolic play: Implications for the language specialist. Topics in Language Disorders, 2(1) 1-15. McLoyd, V., D. Warren, and E. Thomas. (1984). Anticipatory and fantastic role enactment in preschool triads. Developmental Psychology, 2Q, 807-8 14. Miller,J., & R. Chapman. (1991). Salt: A Computer Program for the Systematic Analysis of Language Transcripts. Madison, WI: Language Analysis Laboratory, Waisman Center, University of Wisconsin-Madison. Mueller, E. (1972). The maintenance of verbal exchanges between young children. Child Development, 43, 930-938. Nelson, K. & S Seidman. (1984). Playing with scripts. In I. Bretherton (Ed.), Symbolic (pp. 45-72). New York: Academic Press. Pellegrini, A. (1985). The narrative organization of children’s fantasy play: The effects of age and play context. Educational Psychology, 5(1), 17-25. Piaget, J. (1962). Play. Dreams and Imitation in Childhood. New York: Norton. Rubin, K., Fein, G., & Vandenberg, B. (1983). Play. In P.H. Mussen (Ed.), Handbook of Child Psychology (4th ed.), Vol. 4, (pp. 693-774). New York: Wiley. Sachs, J., J. Goldman, & C. Chaille. (1984). Planning in pretend play: Using language to coordinate narrative development. In T.D. Yawkey and A.d. Pellergrini (Eds.), Child’s play: Developmental and applied. Hilsdale, NJ: Lawerence Erlbaum Associates. Sacks, H., E. Schegloff, G. Jefferson. (1974). Language, 5Q(4), 696-735. Singer, G., & J.L. Singer. (1990). The high season of imaginative play. In G. Singer & J. Singer (Eds.), The House of Make-Believe: Children’s play and the developing imagination. (pp. 64-88) Havard University press: Massachusetts. Smilansky, S. (1968). Sociodramatic play as a type of play phenomenon. In S. Smilansky 134  (Ed.), The effects of sociodramatic play on disadvantaged preschool children. New York: Wiley. Swchartzman, H. (1978). Transformations: The anthropology of children’s play. New York: Plenum Press. Wall, S., Pickert, S., & Gibson, W. 1989. Fantasy play in 5-and 6- year-old children. Journal of Psychology, (3) 245-256.  m  .2i  Yawkey, T.D., & T.J. Miller (1985). The language of social play in young children. In L Galda & A.D. Pellegrini (Eds.), Play. language. and story: The development of children’s literate behaviour. Norwalk, NJ: Abex.  135  APPENDIX A Play Style Rating Scale Please identify the phrase which best describes how (child’s name) plays with other children. Knowing that all children vary from day to day, please indicate the phrase that is true most of the time. For each item, place an X along the continuum (either near a number or between two numbers) to mark the child’s typical play behaviour. 1. When the child is playing alone and is invited to play by another child, he/she will: a.  b.  c.  d.  e.  a. immediately respond positively to the invitation. c. initially decline the invitation, but with some persuasion will join the other child. e. ignores or rejects the invitation completely. 2. When the child is playing alone and notices others engaged in play that appears fun, he/she will: a.  b.  c.  d.  e.  a. join the play and will redirect the play according to his/her own plan. c. join the play in progress and will conform to the “rules” established. e. not join the group/other child or simply engages in parallel play. 3. When the child is playing in a small group or with another child, he/she will: a.  b.  c.  d.  e.  a. originate most of the play ideas and direct the other child(ren) constantly. c. take turns being the leader. e. follow the other child’s lead and will rarely make any new suggestions. 4. When a child is playing with another child and he/she wants an object that his/her partner has, he/she will: a.  b.  c.  a. tell the other child that he/she must have that toy. c. ask the other child if he/she may play with that toy. e. walt until the other child has finished playing with it.  136  d.  e.  APPENDIX B Flow Chart of Coding Decisions Pretend Communication about the Pretend World and Pretending  L  All Utterances  4, Pretend Status  r_NonPretend ‘V  Pretend  1  Not analyzed  L._ Uncodable  1 Communicative Status I  SI,  —  Noncommunicative  ) Not analyzed  Communicative  I Primary Function  Enactment Metapretend  Other  (Construction and Maintenance of the Pretend)  Uncodable  ‘I,  Not analyzed  137  ‘I, Specific Content Identity/Role Transformation Object Transformation Planning Play Description Organization Not Coded ,  Identity of Speaker In-Role Ambiguous Out-of-Role  >  Pretend Announced Yes N  138  APPENDIX C Flow Chart of Coding Decision Pretend Play Communication as Conversation  All Turns  1 Pretend Status i— Nonpretend  Pretend 4.  .) Not analyzed I  I— Uncodable  Degree of Interaction  Nonsocial Social  Uncodable  -  F Not analyzed  I  --  i4r  Continuity of Turn — First-part of Pair  Obliging Nonobliging  — Second-part of Pair  Response (obliged and nonobliged)  L  Simple Response Complex Response —  139  Nonresponse (obliged and nonobliged) Initiation  4,  N  Type of Collaboration  H Elaboration Negotiation Other  140  APPENDIX D List of Codes I. Communication about the Pretend World and Pretending Unit of analysis Communicative Acts -  A. Preliminary Judgements 1. Pretend Status a. Nonpretend [NP] an utterance (or nonverbal behaviour) that referred to the reality of the preschool (or the children’s real environment). -  b. Pretend [PP]- an utterance that functioned within or referred to the pretend play reality. c. Uncodable [PU] an utterance that was impossible to reliably code as real or make-believe. These were excluded from further analyses. -  2. Communicative intent a. Noncommunicative [PPNC] not intended for the partner.  -  a pretend play utterance that was  B. Primary Function of Communication Act 1. Enactment of events [FE] the words and actions of make-believe characters used to bring the imaginary scene to life. -  2. Construction and Maintenance of the pretend [FM] an utterance that functioned to define the contextual parameters of the fictional world, organize and manage the play, and provide details about the imaginary world. -  3. Other [FO] an utterance that functioned neither to enact nor construct the pretend play, but which was none-the-less related to the pretend play. -  4. Uncodable [FU] be reliably coded.  -  an utterance for which primary function could not  141  C. Construction and Maintenance of Pretense Communication 1. Specific Content a. Identity/Role Transformation [CR] an utterance that functioned to change a child’s true identity to that of a fictional character. -  b. Object Transformation [COb]- an utterance that functioned to change the state or quality of an object, or to invent an imaginary object. c. Plan [CP] an utterance that functioned to direct (as opposed to explain or describe) the verbal and nonverbal behaviour of the players. -  d. Play Description [CStJ an utterance that functioned to elaborate the play by describing or explaining the on-going activity. -  e. Organization [CO] an utterance that functioned to structure and manage the pretend play, but did not involve any transformation or story description. -  f. Not Coded [CN] above categories.  -  an utterance that did not fit into any of the  2. Identity of Speaker a. Out-of-Role i. not announced [FMoutj the child behaved and spoke as him/herself while talking about the pretense. -  ii. announced [FMoutL] that” or “pretend that”.  -  use of phrases such as “let’s say  b. In-Role i. marked [FMin]- the child behaved and spoke as a character within the imaginary world. ii. assumed [FMinA] the child was assumed to be in-role based on the preceding speaker identity. -  c. Ambiguous -the speaker’s identity could not be determined.  142  II. Pretend Play Communication as Conversation Unit of analysis Turn -  A. Preliminary Judgements 1. Pretend Status 2. Segmentation of Turns 3. Degree of Interaction a. Social [TS]- a turn in which either partner directed his/her behaviour towards the other child. b. Nonsocial [TNI a turn in which neither child directed his/her verbalization or behaviour at the other child. -  c. Uncodable [TU] categories.  -  a turn that did not fit into one of the above  B. Continuity between Turns 1. First-part of pair a. Obliging [A lOb] next speaker.  -  a turn that required a response from the  b. Nonobliging [A1NOb]- a turn that did not require a response from the next speaker. 2. Second-part of pair a. Response-obliged [A2R+] response was obliged.  -  a turn that was continuous and a  b. Response-nonobliged [A2R-]- a turn following an obliging turn which did not meet the terms of the obligation. All responses: i. Complex Response [CRES] a turn that was connected to both the preceding and following turns -  ii. Simple Response [SRES]- a turn that was connected to the prior turn only. 143  c. Nonresponse-obliged [A2NR+J a response was not obliged.  -  d. Nonresponse-nonobliged [A2NR-J nonresponsive and not obliged.  a turn that was continuous but  -  a turn that was  All nonresponses: i. Initiation [INIT]- a turn that was thematically connected to the following turn only.  C. Type of Interaction Unit of Analysis  -  Turn Exchange  1. Elaboration [SELB]- a collaborative exchange that involved maintaining the play interaction by adding information to the on-going activity. 2. Negotiation [SNEG]- a collaborative exchange that surrounded disagreements about the play between the players. 3. Other [SOth]- a collaborative exchange that did not function to elaborate or negotiate the pretend play.  144  APPENDIX E List of Specific Coding Conventions  Exact or slightly modified repetitions of utterances were coded into whatever content category applied, but were also marked as repetitions. The purpose of this coding convention was to ensure that children were credited only for original structuring proposals.  2.  Occasionally, children spoke to an imaginary third person during the enactment of an event. These utterances received an enactment code. In most cases, the speech was neither directed at nor intended for the child’s partner and; therefore, it was also reclassified as noncommunicative. The most common example of this type of talk was an imaginary telephone conversation. For example, A called someone on the telephone to help her repair the house while S continued to repair the cupboard on her own. A: I hafta phone somebody on my phone {looks at S and then turns away}. A: Ah ha {speaks into the telephone). Hi {brief pause}. Oh could you build this house? I can’t do it so you may do it.  3.  Sound effects were produced from different speaker positions depending on the context and nature of the sound effect. If the child used onomatopoeia as representative of actions associated with role playing (e.g., eating or drinking  145  sounds), then the utterance was coded as enactment. If, on the other hand, the onomatopoeic instances did not represent actions produced by the character of the pretend play, then they were not treated as enactment because the child was perceived as having stepped out of the fictional character role to comment on the situation. For example, explosion or shooting sound effects were coded  as out-of-role play description. D and N make a subtle transition to and from the pretend state during the following sequence: N: Oh and this is a very special tea pot, yup. Guess why? See {shows D and makes explosion sound effects). D: {makes explosion sound effects). N: Bye-bye. D: That was a punch tea pot, I’d say.  146  


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items