Construction of Self in Children’s Daydreaming Narratives: Story of Two Generations1

This paper reports on the research results regarding children’s positioning in their daydreaming narratives. A special feature of the research was its cross-generational character, i.e. we have collected and analyzed written accounts of primary school children’s daydreams in grade 4, which were produced in the year 1984 and the year 2009. One of the major insights from our research is that children’s daydreaming narratives broadly belong to the category of “preferred self” constructions. We classified these preferred self-constructions in three categories, namely “fictional selves”, “future selves” and “actual selves”, which emerged from repetitive themes in children’s narratives. Each of these broad categories was further “unpacked” into more specific constructions of preferred selves. Based on the qualitative data we developed hypotheses about generational shifts in preferred self constructions and used nonparametric statistical tests to evaluate these hypotheses.

Daydreaming is one of the terms resisting consensual and precise definition even though it is rather meaningful to the majority of people.Some of the problems in defining daydreaming are the result of its complexity and multidimensionality.In different definitions of daydreaming three dimensions represent the main points of divergence, focusing on the question whether daydreams are: 1) realistic or fanciful; 2) task-related or task-unrelated; and 3) deliberate or spontaneous (Klinger, 2008).For example, daydreaming may be defined as a form of imaginative thought, but also as a feature of naturally occurring stream of consciousness (Singer, 1999).Also, daydreaming may overlap with fantasy, as some daydreams are forms of mental activity departing from reality.Early research on daydreaming was sparked with Freud's explorations of the relationship between daydreaming and creativity and for the Corresponding author: pavlovich.jelena@gmail.com 1 Note.This article is a result of the project "Education for knowledge-based society" No 149001 (2006)(2007)(2008)(2009)(2010), financially supported by the Ministry for Science and Technological Development, the Republic of Serbia.
first half of the 20 th century a major source of evidence into daydreaming was provided by patients in psychoanalysis (Singer, 2003).Creation of the Imaginal Processes Inventory (IPI; Singer & Antrobus, 1970) marked the beginning of a line of studies which used questionnaires as a principal data gathering technique.Repeated factor analyses of IPI produced three second order factors that referred to different styles of daydreaming: 1) Positive-constructive; 2) Guilt and fear of failure; 3) Poor attentional control (Huba, Singer, Aneshensel, & Antrobus, 1982).These factors mapped variations in daydreaming styles from a positive attitude toward daydreaming as a worthwhile phenomenon facilitating problem solving and idea generation, and providing warm feeling and pleasant thought, to depressing and frightening qualities or tendency toward mind wandering and loss of interest.A number of studies have also asked participants to describe daydreams that are remembered retrospectively and to provide written accounts of their daydreams (Gruis, 2005).These descriptions were then analyzed and quantified, providing insights into the frequency of the particular types of settings, main characters and their actions.
Studies of daydreaming have mainly investigated daydreams of adult population and a majority of studies focused even more narrowly on college students' daydreaming (Gruis, 2005).There are significantly fewer studies of daydreaming among children, probably due to the methodological limitations researchers are confronted with when using questionnaires with children population.Apart from these methodological difficulties in studying children's daydreaming, development of the Imaginal Processes Inventory for Children (IPI-C) was confronted with unsatisfactory reliability in non-American samples (Vooijs, Beentjes, & Van der Voort, 1992).However, factor analysis of IPI-C enabled researchers to gain insights into three kinds of fantasies that emerge in children's daydreaming: 1) Positive-intense daydreaming; 2) Aggressive-heroic daydreaming and 3) Dysphoric-aggressive daydreaming.The positive-intense daydreaming style was characterized by vivid and pleasant daydreams, similar to the positive-constructive style found in adult population.The distinction between aggressive-heroic and dysphoric-aggressive daydreaming styles was important because of the different role of aggression (Young, 1997).In aggressive-heroic daydreams, aggression was not the main focus, but the morality of the tale.
Though qualitative methods have been widely used in the study of creativity and imaginative activities (Deacon, 2000;Trotman, 2006;VanSledright & Brophy, 1992), there have been few qualitative studies of daydreaming.This relative lack of qualitative studies of daydreaming may be seen as somewhat surprising given that qualitative methodology is often referred to as suitable for research of phenomena in which complexity and process are more relevant than accurately quantifying relationships (Silverman, 2000).It may be argued that methodological burdens of IPI-C and relatively few studies of children daydreaming make the use of qualitative methods even more suitable when it comes to the research of children's daydreaming.

SELF-REPRESENTATIONS, DAYDREAMING AND NARRATIVES
Self-representations have been a specific line of interest in research on daydreaming.According to Singer (2003), people are almost continuously engaging in daydreams that involve judgement about their actual selves and how these relate to ideal or socially accepted selves.This line of research has therefore focused on the influence of self-representations on daydreaming patterns.In other words, self-representations are perceived as one of the determinants of the content of daydreaming.Yet, another research option is to look at daydreaming as an important "vehicle" for the construction of selfhood and consider what sort of self-images are produced in daydreams.This line of thinking is suggested by the narrative approach, which we adopted in our study of children's daydreams.
According to the narrative approach, telling stories is about making sense of ourselves and the world around us (Bamberg, 2008;Davies & Harre, 1990;Pavlović, Džinović, & Milošević, 2006).In this approach, stories are not perceived as mere expressions of our experience, but as constitutive forces in creating experience.One important aspect of this experience is our sense of ourselves.Narrative researchers assume that the sense of ourselves is produced in stories we tell.They use the notion of positioning to refer to the discursive process whereby selves are produced and located in stories (Bamberg, 1997;Davies & Harre, 1990;De Fina, Schiffrin, & Bamberg, 2006).The notion of positioning implies that when telling different stories, we are taking different identity positions in conversations.For example, one person can take several different identity positions during the same conversation or within the same story.This discursively produced self is thus multiple, discontinuous and fragmented, rather than unitary, inherently given and stable.Instead of the traditional metaphor of self as "located under the skin", narrative researchers adopt a metaphor of self-as-a-story.Self-as-a-story is built from cultural symbolic resources and its positioning depends on the opportunities and restrictions set by these resources.This means that availability of different cultural resources plays an important role in creating our sense of ourselves.Consequently, in the encounter with narrative, we are shaping and refining not only our understanding of personhood, but also of the cultural resources that enabled certain self-constructions.Narrative analysis aims to elaborate on the way identities are constructed and performed in stories told in a particular cultural context.In other words, narrative analysis aims at exploring both the construction and the function of narratives.Construction of narratives refers to the linguistic resources in terms of images and metaphors underlying a story.Function of narratives refers to the question of how persons use language to perform different kinds of actions in an interaction or a local discursive context (Bamberg, 2004;Potter & Wetherell, 1995).For example, a person may use language for apologizing, accusing, justifying, negotiating intimacy, attributing, giving a positive image of self to other people, legitimizing their versions of events or maintaining a credible stance in interaction (Burr, 2003:60).In this functional sense, narratives are both enabling and constraining, working both in inhibiting and productive ways, designating both exclusions and choice.However, this distinction between construction and function of narratives should not be painted to sharply, since they are usually combined and mark different shades of research emphasis (Potter & Wetherell, 1995).
In this study, we adopted a narrative approach to study children's daydreams and based our research on the assumption that selves may be researched as constructed and performed in narratives.We assumed that daydream narratives allow us to analyze how children's selves are "positioned" along the story lines.This notion is opposed to the idea of children's daydreaming narratives as manifestations of some internal, hidden and unconscious dynamics in their personalities.Instead, we assumed that specific self-images created in these narratives are important object of analysis in their own right.Images and metaphors around which these self-images are being built inform us about their worldviews, ways in which they wish to present themselves, as well as about important functions of these self-constructions.
Our choice of the narrative framework in study of daydreaming may be justified by several arguments.First, there is evidence that daydreams naturally comprise a continuous sequence of images that unfold like elaborate stories (Klinger, 1990).In other words, it may be argued that daydreams are naturally recollected in narrative form.Furthermore, narrative is often referred to as an important way of conceptualizing experience and organizing memory (Eubanks, 2004).If we assume that research on daydreams, especially among children, deals with under-conceptualized experience and with memory that is not regularly accessed, then narrative may be seen as an appropriate research tool for this topic.Finally, qualitative analysis of this kind allows informants to illuminate their complex and unique frames of reference, rather than to conform to the categories of researchers.This is of special significance since there are few studies on children's daydreams and therefore not so many known ways of categorizing children's experience.

Method
The aim of our study was to explore children's self-constructions in written accounts of their daydreams.In line with the twofold focus of narrative psychology on both construction and function of narratives, we addressed the following research questions: 1) What sort of self-images are constructed?2) What is the function of these self-constructions?3) In what ways do children's constructions of their selves vary in narratives produced in the past and nowadays?
Participants and data production.We analyzed written anonymous accounts of daydreams of primary school children in grade 4 (age 10).These written accounts were produced in the year 1984 2 and the year 2009 in the same primary school.With the help and support of the 2 Data from 1984 were produced in a study by one of the authors (J.Šefer) which has not been published so far.
school psychologist, we approached the participants during their school classes.Children were invited to produce written accounts of their daydreams.We gave them a somewhat general instruction to describe what they usually daydream about ("Please write down what you usually daydream about.This is not an assessment of your writing skills and there will be no marks.Write down what you usually daydream about freely as if you were telling it to us.").This general instruction was further clarified in line with additional questions children may have asked (e.g."How much text they should write", "How much time they have available for the task").The instruction to write as if they were telling us and additional clarifications that they have the whole class period available and that they should write as much as they wish allowed children to produce stories of their daydreaming rather than just listing elements and topics.Children completed this task in conditions somewhat similar to a regular class (they were sitting at their desks in the classroom and writing down in relative silence).In some classes the school teacher was present, while in some classes not.However, we wanted to stress that the narratives produced would not reach the school teacher and would not be assigned school grades.In this way, we intended to stimulate children's expression and avoid their concerns about grammar, teacher's evaluations etc.The above mentioned instruction was accompanied by further explanations of our understanding of the term "daydreaming".Daydreaming was referred to as an imaginative activity in which they engage in their waking state as opposed to their night dreams.Narratives from year 1984 were produced by 66 boys and 66 girls.Contemporary narratives were produced by 53 girls and 33 boys.In total, there were 218 children participants.In both cases, narratives were produced in similar research conditions -all classes of one generation of children were invited to participate; children were addressed as a group in their classroom; children were given the same instruction.In 2009 there were somewhat fewer participants in general and fewer boys in particular.Although there are differences in the sample structure between the two generations, we wanted to maintain the original data production conditions -addressing the whole classroom as a group and researching one generation in one particular school.
Analytic procedure.Written accounts of children's daydreaming were analyzed in line with the principles of narrative positioning analysis (Bamberg, 1997) and grounded theory (Strauss & Corbin, 1990).We developed a coding scheme to capture patterns of actions and states in given self-constructions.Our coding scheme was not a matter of assigning predefined categories, but it emerged during the analytic process.First, we performed open coding (Strauss & Corbin, 1990) by breaking down data into key themes: self-referential images, accompanying actions, as well as our interpretations of their functions.By doing so, we created a record of open coding, which included 36 pages of text and enabled us to keep the coding process transparent.This record was then used as a basis for axial coding, which was more focused, directed and conceptual than open coding.While in the open coding phase we identified and named certain concepts, during the axial coding phase we reduced these concepts to a smaller number of categories and reassembled the data by relating concepts to categories.During the selective coding phase we scanned open codes to check for fit with the identified categories and we further theorized these categories by identifying patterns of relationships between them in terms of their similarities and differences, their functions, generational variability, etc.In other words, we subsequently developed broader categories of self-referential images and integrated our line-by-line codes into a set of hypotheses about the main themes and their relationship.Although we adopted a qualitative approach, we also used counting and nonparametric statistical tests to evaluate some of the hypotheses which emerged during the coding phase.

RESULTS AND DISCUSSION
Preferred selves as the central category in children's daydreaming narratives One of the major insights from our research, which we noted even during the open coding phase, is that almost all children's daydreaming narratives included self-images which were related to positive emotions, pleasant and desirable situations.We identified a total of 3 narratives (out of 214) in which negative and unpleasant self constructions were elaborated (e.g."Most often I daydream about my mathematics test.I get all sweaty daydreaming about it", "I daydream about ghosts that chase me and I try to run away", "I daydream about my parents being dead...I know it's wrong because I couldn't survive without them").Therefore, children's constructions in our research seem to fit into the concept of "preferred selves", which refers to the desirable stories about oneself or different ways in which persons would prefer to describe themselves (Fridman & Kombs, 2009).This major finding may also be associated with Freud's early interpretations of the wish-fulfilling functions of daydreams (Mueller, 1990).Rather than to rely on psychoanalytic concepts and the idea of unconscious dynamics, we have, however, interpreted this finding from the narrative perspective, pointing out children's daydreaming as an imaginative space for the preferred selves' elaboration.

Subcategories of preferred selves: fictional, future and actual selves
Based on the initial codes identified during the open coding phase, we developed the main subcategories of the preferred self-constructions: fictional, future and actual selves.This conceptualization of data emerged gradually during the phase of axial coding, as we classified concepts coded line-by-line.Underlying these subcategories is the hypothesis about some sort of "distance" between children's preferred selves and their actual "child identity".For example, fictional selves refer to self-constructions that step outside the realm of everyday experiences into the realm of fiction.Future selves refer to children's selfconstructions as adults, while actual selves refer to preferred self-constructions in which their "child identity" is preserved.Each of these subcategories was further "unpacked" into more specific constructions (Table 1).
It is worth mentioning that these categories sometimes overlapped, e.g.children sometimes referred to themselves in terms of some fictional character, but also some future character in the same narrative.Cases where more than one character was constructed in a single narrative made the total number of selfconstructions (233) greater than the total number of participants (218).Fictional selves Previous research on adult daydreaming pointed to the relatively low importance of fiction in daydreaming.Klinger (1990) estimated that no more than 20 percent of adult daydreams contain fictional actions.In our analysis, however, fictional characters represented the dominant construction of children's preferred selves.This finding raises a question whether fiction becomes less important in daydreams with age or adults tend to be more pragmatic and more oriented towards everyday problem solving.We classified and interpreted a variety of fictional themes that emerged in children's narratives into three subcategories (Table 2): (1) fighters against Evil; (2) change agents; (3) exploring selves.In light of previous research on children's daydreams, selves as fighters against Evil may be interpreted in line with the general aggressive-heroic pattern found in IPI-C, while selves as change agents and explorers may be interpreted as elaborations of the positive-intense daydreaming pattern.The subcategory of fictional selves as fighters against Evil refers to an image of a courageous self and representative of Good (e.g."I save the people and bust the bad guys", "In the middle of the battle a black spider shows up and everybody runs away except me").During the selective coding phase we developed a hypothesis that this subcategory is more dominant among boys.Using the chi-square test we found statistically significant differences in occurrence of this subcategory among boys and girls (χ 2 = 18.096, df = 1, p < 0.01).This is in line with the finding that daydreams of adult men resemble traditional stereotypes of masculinity and are determined by aggressive drives 3 "S" refers to narratives from the past, while "N" refers to contemporary narratives; ordinal number refers to the school class, while Arabic number refers to the code made for the specific participant.
and active lifestyles (Gruis, 2005).In narratives from the past, some elements (e.g.importance of national defence, ambition to follow the steps of national heroes who died for the country, etc.) point to the strong influence of dominant ideologies of the time on boys' construction of selves as fighters against Evil.The subcategory of selves as change agents refers to making a positive change in the world, mainly in form of fairies and wizards ("I would turn bad people into good ones", "I would bring happiness to people").A better world is constructed as a controllable world in which children have the ability to control natural forces, change their order or control the wellbeing of others.In narratives from the past, this image of a better world may also reflect values of a socialist society: equality and the world of brotherhood and unity ("I would make Tito alive", "I would bake bread for all the people in the world, so there would be no hunger").We found a statistically significant difference between boys and girls in terms of likelihood to construct an image of self as a change agent (χ 2 = 9.84, df = 1, p < 0.01).Previous research on daydreaming of women (Gruis, 2005) pointed out to their passive, nurturing and protective role, in line with the general image of feminine and maternal attributes.Our research supports this orientation towards nurturing and protecting, but it also points out the active role girls play in their daydreams.
The subcategory of selves as explorers refers to daydreaming about being an astronaut who explores the Universe, meets other civilizations, travels through time, discovers new planets etc. ("I would discover every single part of the Solar system", "As a researcher of the Universe, I would make friends with aliens").Almost complete disappearance of astronauts as preferred selves in contemporary narratives brought us to a hypothesis about a generational shift.The chi-square test confirmed that the likelihood of the subcategory of the exploring self is not the same in narratives from 1984 and 2009 (χ 2 = 9.218, df = 1, p < 0.01).This finding may be interpreted in light of the "race" for the Universe, which was once an issue of national progress and strategic positioning between the "Western" and "Eastern" blocks.Lack of explorers in contemporary narratives in general may also indicate decreased interest for discovery and experimentation.

Future selves
Previous research pointed out the role of daydreams in anticipating future involvement in life events (Gruis, 2005).In our analysis, children also constructed their preferred selves as their future selves (Table 3): (1) celebrity selves; (2) professional selves.If we refer to previous research on children daydreaming, we may link the celebrity and professional self narrative with the positiveintense daydreaming pattern in IPI-C and achievement oriented daydreams in IPI for adults.

FUTURE SELF Unspecified
The subcategory of celebrity selves refers to standing out of the crowd, being adored, admired and respected ("When I walk down the street, everybody recognizes me", "Everybody would applaud to me"), enjoying the company of other celebrities ("I would meet Angelina Jolie"), as well as being a wealthy person ("I have a lot of money and newspapers write about me").While selves as explorers almost disappeared in contemporary narratives, we noticed that celebrity selves became dominant.The chi-square test confirmed a difference in variability of this subcategory in narratives produced in 1984 and 2009 (χ 2 = 12.085, df = 1, p < 0.01).This finding may be seen as part of a broader shift towards the "celebrity culture", which found strong support in the development of various forms of mass media.Though the celebrity self-narratives are told in the "when I grow up" genre and represent some sort of children's future selves, they are very close to fictional characters in terms of their capacity to loosen the boundaries of children's actual selves.
The subcategory of professional selves refers to self-constructions which provide elaboration of future professional choices ("I daydream about becoming an architect"), experience of professional fulfilment ("I would be a farmer... and I would be so happy"), connection of children's current interests with their future work ("I am already into volleyball and I find it very interesting, so I would like to become a volleyball player one day"), etc.A theme that is worth mentioning as a prevailing one is boys' construction of professional selves as pilots.Becoming a pilot was a professional role that was particularly appealing in the boys' narratives from the past.This profession allowed boys to find their heroes not only in fictional characters, but also in real characters that may have been their neighbours.

Actual selves
A majority of research on adult daydreaming points to their orientation towards presence.In our research on children daydreaming, however, fictional characters and grown up characters, both distant from children's actual selves, appear to be the dominant constructions.However, a variety of actual selfconstructions also emerged in our study (Table 4): (1) travelling selves; (2) playful selves; (3) actual selves in "repaired realities".We developed a hypothesis that the construction of preferred self as a child character is more dominant in girls' narratives than in boys' narratives.The chi-square test points to statistically significant differences in occurrence of actual selves among boys and girls (χ 2 = 16.862,df = 1, p < 0.01).This finding may be interpreted in terms of children's identity being closer to the girls' sense of preferred selves, than is the case with boys' identity.The subcategory of travelling selves refers to travelling around Europe, sightseeing, learning foreign languages and shopping ("To see the Eiffel tower would be the happiest day of my life").Similarly to exploring selves, these travelling selves also include the notion of exploring unknown territories.However, today's travelling selves are not curious and brave explorers of the Universe, but rather enjoy the pleasures of popular touristic sites, which may be seen as an aspect of the globalization process and the accompanying "touristic practices" indicating materialistic and hedonistic life orientation.
The subcategory of actual selves in "repaired realities" refers to a self construction in which children preserve their child identity, but redefine the social context around them ("Nobody is quarrelling, there is no envy", "Parents would have enough money to buy me my favourite toys") or their own role in it ("I dream of being the best pupil in the school").Finally, the subcategory of playful selves refers to self-constructions in which children enjoy the play with peers or animals (e.g."I would knit a dress for my doll", "I would play with my sister and we would plant flowers in our yard").It is interesting that although this self-construction seems to be most available to children in terms of its grounds in reality, it is one of the least dominant self-constructions.

Function of self-constructions in daydreaming narratives
During the selective coding phase each of these categories of selfconstructions was further "theorized" by looking at functions of the protagonists.The most prevailing function of children's self constructions in daydreaming narratives refers to acquiring respect and admiration.Children's positioning both as fighters against Evil and as celebrities seems to refer to imaginative elaboration of children's selves as important actors in the social world of adults.This type of functional orientation points to daydreaming narratives as a context for experimenting with subject positions which may otherwise be unavailable to children.In other words, these findings raise a question whether children manage to position themselves as respectable in everyday discursive spaces other than imaginative.
Another important function of children's daydreaming narratives is imaginative exercise of control over life.Girls' positioning as change agents points to the interactive business of controlling wellbeing of others and changing the order of natural forces in line with the socially desirable states (e.g.absence of illness and death).Again, we may raise a question whether, outside daydreaming narratives, children have other discursive spaces available for elaboration of the important topic of control versus lack of control in life.
Children's narratives also seem to point to the importance of daydreaming as a discursive space for elaboration of unknown (outer space, different planets and continents, etc.).Selves as explorers dominant in children's narratives from the past point to this important function of daydreaming in the given cultural milieu.Absence of this type of self-construction points to a hypothesis that this developmental theme may not be of such importance nowadays, or that technological advances in the last couple of decades have created discursive spaces for elaborating this theme.

CONCLUSION
The majority of self-constructions refer to the category of preferred selves, which indicates that children in both generations are active and constructive in their daydreams.These preferred self-constructions are more often fictional characters than realistic adult characters located in the future.Actual selves tend to be less dominant in children daydreaming narratives than in the research on adult population.We found a difference in variability of categories of fighters against Evil and change agents between boys and girls.Moreover, we found differences between boys and girls in the occurrence of self-constructions as actual selves.Finally, we found generational patterns in self-constructions as explorers and celebrity selves.
Disappearance of exploring self-constructions may indicate that curiosity and experimentation are no longer on the "list" of the preferred self-constructions, pointing out that science, discovery and other creative task oriented activities are not considered as popular as they were in the past.The rise of celebrity selfconstructions points out a cultural shift in values that occurred over time.While the influence of socialistic ideology was strongly present in the past, some of the generational shifts may also be understood as a reflection of dominant trends in social values indicating growth of a hedonistic, consuming and glory oriented style of life, especially prevailing in modern media.
In the end, we may conclude that our results indicate that fiction as a creative potential for self-constructions is alive nowadays as it has been in the past.Although daydreaming seems to be an individual and private business, it may also be seen as our "window" into social resources, which are available to children for the important business of constructing their identities.

Table 1 .
Overview of frequencies of children's self-constructions.

Table 2 .
Excerpt of the coding scheme: Fictional selves

Table 3 .
Excerpt of the coding scheme: Future selves.

Table 4 .
Excerpt from the coding scheme: Actual selves.