From ambivalent spaces to spaces of reciprocal encountering: developing classroom culture in an intercultural story exchange

ABSTRACT This paper describes the process of developing classroom culture that promotes reciprocal intercultural encountering in the primary school. Storycrafting, an oral storytelling method used for democratic encounters between participants, was used to connect classes in Scotland, Finland and Belgium over three cycles of design-based research. Abductive analysis produced a theoretical framework of seven key dimensions that influence the learning space for reciprocal encounters: power, knowledge, relatedness, purpose, structure, continuity and meaningfulness. These dimensions were used to identify four kinds of learning spaces that occurred during the implementation of the story exchange: the ambivalent space, the space of information, the transitional space and the space of reciprocal encountering. The intervention created a platform for children’s encounters that reinforced mutuality and reciprocal interactions rather than Othering the exchange partners.


Introduction
Classroom culture influences how children engage in intercultural encountering and whether they encounter with reciprocity and openness.This article describes insights from empirical work that will help primary school teachers and researchers to understand how classroom culture can be developed together with students so that it supports encountering.We analyse empirical data from an intercultural exchange of stories, and develop theory about what constitutes a learning space of reciprocal encountering.
Classroom culture consists of a shared understanding of social norms and knowledge repertoires between the members of a classroom community, including the teacher(s) and the students (Kumpulainen and Renshaw 2007; Marton and Tsui 2004).It evolves over time as the community creates new shared experiences that replace previous ways of doing, thinking and knowing.Therefore, we understand classroom culture as a continuum of learning spaces, which are situations where the conditions allow a particular kind of learning to occur (Dewey 1938(Dewey /2015;;Marton and Tsui 2004).
We define intercultural learning as educative growth towards understanding, encountering and conflict resolution between individuals from two or more distinct communities (adapted from Dewey 1938Dewey /2015;;Kaikkonen 2004).One area of intercultural learning is reciprocal intercultural encountering, which means mutual, dialogic interaction between people from distinct communities.The communities can be national communities, but they could equally be communities based on locality, organisation, language, religion, interest, or any other commonality shared by group members.In this paper, when we say 'encountering', 'intercultural encountering' or 'reciprocal encountering', we are referring to reciprocal intercultural encountering.
The aim of this study is to identify how classroom culture can be developed so that it supports reciprocal intercultural encountering.Empirical data will be analysed to answer the following research questions: (1) What are the qualities of a space of reciprocal encountering?(2) How did teachers and students co-create a learning space of reciprocal encountering during an intercultural story exchange?(3) What other kinds of learning spaces were co-created by teachers and students during the development of an intercultural story exchange?
Reciprocal intercultural encountering is necessary for peaceful cooperation between local as well as global communities.Intercultural learning is incorporated into recommendations of intergovernmental organisations such as UNESCO (2013UNESCO ( , 2014)), the OECD (2018), and the Council of Europe (2001Europe ( , 2016Europe ( , 2018)), as well as into the national curricula of countries such as Finland and Scotland, and international curricula such as the International Baccalaureate.However, despite the prolific occurrence of the concept in educational policies and curricula, more work is needed to develop curricula and teaching practices that promote viewing cultural Others as neighbours and engage students in joint, transformative processes (McCandless et al. 2020).We need to better understand the challenges that schools and teachers face in implementing these types of practices, so that educational structures and school cultures can be developed to support intercultural learning.
There is evidence that international exchange projects which highlight cultural differences can unintentionally reinforce stereotypes and lead to a static view of culture (Gorski 2008;Lau 2015;MacKenzie, Enslin, and Hedge 2016;Royds 2015).To achieve the goal of reciprocal encountering, a story exchange project was developed through design-based research, where the focus shifted away from culture as a static entity to viewing culture dynamically as a process (Piipponen and Karlsson 2019).Three iterations took place so that the design and theory could be developed over time (Anderson and Shattuck 2012).The qualitative data were produced through narrative and ethnographic methods, to allow us to investigate the contextual nature of the learning spaces in detail.During the project, the children exchanged stories that were told using the Storycrafting method, where children freely tell oral stories to an interested scribe (Karlsson 2013).Through telling and listening to each other's stories, the children learned how to connect to their exchange partners experientially.Although the groups never met face to face, they could imagine each other through the stories (Piipponen and Karlsson 2021).

Method
Contrary to the usual convention, the research methods are discussed prior to the theoretical framework because the framework was produced following the data analysis.The methodology adheres to the principles of design-based research, which involves a cyclical process of alternately developing practice and theory (Anderson and Shattuck 2012).

Research context and participants
Three cycles of a story exchange project were organised between partner classes in primary schools.The first author conducted the first cycle of data production in School A in Scotland in spring 2015, where she was teaching full-time as Teacher 1.Her class exchanged stories with two partner classes in Finland (from Schools B and C), which were selected due to the teachers' expertise in the Storycrafting method and willingness to participate.
During the second cycle in spring 2016 and the third cycle in the academic year 2016-2017, the first author continued the classroom-based research while working in an international school in Belgium (School D).There the fifth grade teachers (1, 4 and 5) decided to co-teach the three classes together, with the first author coordinating the project.One of the partner class teachers in Finland (Teacher 2) and her next class continued in the exchange project for the second and third cycles.The research participants are described in Table 1.All the participants were selected based on convenience.

Ethics
This research has followed the ethical recommendations established by the Finnish Advisory Board on Research Integrity (2009Integrity ( , 2012)), which were in force at the time of the data production.This includes acquiring informed consent from the relevant municipalities, head teachers, children, parents of participating children, and the participating teachers.The research process was integrated into the daily school routines and caused minimal additional exertion to the teachers and children.The design, data production, analysis of data and reporting of the research has been conducted by paying special attention to the child perspective to ensure the child participants are valued and treated with dignity.All names have been replaced with pseudonyms.

Data production
In the first cycle of research, Class A exchanged stories with Classes B and C. The children told oral stories using the Storycrafting method, where an attentive scribe writes down the teller's oral story exactly as it is told (Karlsson 2013).Once a story is finished, the scribe reads it aloud and the teller can make changes or corrections (Karlsson 2013).Sometimes the whole class told a story by taking turns to tell and the teacher scribed.At other times, the children worked in pairs or small groups, where the children scribed for each other.Occasionally the teachers scribed for an individual student or a small group.The project developed organically to fit into the everyday life of the classrooms.The stories were translated into Finnish or English by the first author and sent to the partner class via email.Sometimes more than one story was sent at a time.They were read out in the school language of the partner class, then in the original language.To provide the children with an opportunity to respond to each story sent by the partner class, the first author arranged open-ended class discussions, which were audio-recorded in School A. In School B, the class teacher recorded individual students' comments, and in School C, no commentary was produced.The partner classes continued the exchange in this manner for 7 weeks.The first author kept a researcher diary to record contextual information about the sessions.To further triangulate the data, consent was also granted to collect the email correspondence between the first author (Teacher 1) and Teacher 2.
After the first cycle, the design of the story exchange was reviewed by the first and second authors and by getting written feedback from Teachers 2 and 3.For the second and third cycles in Schools B and D, the class discussions and children's recorded comments were discontinued and replaced with a drawing activity, where the children were told to draw a picture about the story they had just heard (see Figure 1).These drawings were photographed or scanned and sent back to the partner class along with a new story.At the end of the third cycle of research, upon hearing that the exchange was coming to an end, the children in the teacher-researcher's class of School D sent a thank you letter and class photo to the partner class.This was reciprocated by the class in School B. Teacher 5 in School D was also interviewed following a semi-structured interview schedule after two years of participating in the project.The data produced are further described in Table 2.

Analysis method
The analysis process has been developed from the frameworks of design-based research and abductive analysis.Design-based research not only develops new practices, but also generates new theory over several iterative cycles of research: during a cycle, initial theoretical ideas are tested by designing an intervention in the classroom, and the theory is revised based on empirical findings (Anderson and Shattuck 2012).According to Timmermans and Tavory (2012), abductive analysis is a suitable approach when the researcher has a sufficiently broad knowledge of theories that can be used to produce new theory, where existing theories are unable to explain an observation.Abduction involves going back-and-forth between the theory and data to refine the theory and close the gap between the theory and empirical evidence (Timmermans and Tavory 2012).
Initially, a general coding framework was created, which was based on a breadth of theoretical literature on how teachers create conditions for student engagement (e.g.Daniels 2016Daniels , 2017;;Dewey 1916Dewey /1997Dewey , 1938Dewey /2015;;George and Richardson 2019;Kelly et al. 2014;Kumpulainen and Renshaw 2007;Marton and Tsui 2004;Mercer and Littleton 2007;Reeve 2013;Ryan andDeci 2000, 2017;Turner et al. 2014).The data were organised into primary as well as secondary data (see Table 2), which were cross-referenced with the primary data to identify more latent content (Graneheim, Lindgren, and Lundman 2017).Atlas.tisoftware was used to code the data extracts into general categories, including student engagement, student initiatives, teaching practices, educational structures, meaningfulness, relationality, pedagogical discourses.Then subcodes were created to specify how each general category was constituted in the data: for example, student engagement included sub-codes such as 'listening attentively' and 'agreeing with someone'.
The first author was responsible for the bulk of the data analysis.As the teacher-researcher, she had specific contextual knowledge about the situations that took place during the data collection, which allowed her to identify latent content in the data.To triangulate the analysis process and supply further interpretive perspectives, the second and third authors reviewed a representative cross-section of the first cycle of data using the coding framework provided by the first author.Based on the discussion, code categories were revised and added to so that they formed a coherent system.This system was used to review and recode all three cycles of data.Where necessary, some new codes were introduced at this stage as well.
The goal of the second phase of analysis was to identify patterns and relationships in the coded data that explained how the teachers created conditions for student engagement.These were dubbed pedagogies and first organised as a table by describing how each of the coded categories (e.g.educational structures, teacher beliefs, teacher practices, student beliefs, student engagement) manifested themselves in the data patterns.The table of pedagogies and the related data extracts were used to write up a description of the following pedagogies: ambivalent approach, pedagogy of information, and pedagogy of reciprocal encountering.At this stage, all three authors came together to review the table and descriptions.The discussion concluded that the code categories used in the analysis used were not fully capturing the dynamic nature of the classroom interactions that produced the learning space, where both teacher and students had agency.
In the final phase, the first author returned to the descriptions and the related coded extracts, and recategorised them into seven dynamic, process-based dimensions (see Figure 3) that influenced how the students and teachers together co-created a learning space.For example, the codes under teacher and student beliefs were restructured under the dimensions of knowledge and purpose.Furthermore, codes related to power were collated under one dimension, as they had been scattered across various code categories.This illustrates how one code could be included under two or more dimensions, as the dimensions interacted in the learning space.Further theoretical reading on the dimensions took place at this point (e.g.Aro, 2012;Foucault 1977Foucault /1995;;Freire 1970Freire /1996;;Hännikäinen and Rasku-Puttonen 2010;Kaikkonen 2004;Lanas 2017;Layne, Trémion, and Dervin 2015;Tornberg 2013).Their theoretical strength was tested by returning to the data to reanalyse the coded patterns, which were renamed learning spaces, and by revising the findings table and the written descriptions.The transitional space was identified at this stage.All three authors met again to evaluate how well the proposed theoretical framework fit with the empirical data, and further adapted it into the context of intercultural encountering in the classroom.In this way, the abductive analysis process produced both a suitable theoretical framework and a related set of findings to answer the research questions.

Theory
A theoretical model was created in an abductive analysis process (see previous section).It describes the formation of a classroom culture over time and learning spaces in the moment, showing how power, knowledge, relatedness, purpose, structure, continuity and meaningfulness influence the teachers' and students' co-creation process.

Continuity of classroom culture: bounded by extrinsic conditions
Classroom culture changes over time as new experiences in each learning space lead to new shared understandings and norms within the class community (adapted from Dewey 1938Dewey /2015)).A learning space is a single situation, whereas classroom culture exists as a continuum of situations (see Figure 2).Extrinsic conditions, such as school culture, national legislation or global trends, create boundaries for a classroom culture's development (Dewey 1938(Dewey /2015)).These extrinsic conditions play a part in explaining why teachers can find it challenging to create spaces of reciprocal encountering with their students.
Many societal, cultural and institutional forces affect the day-to-day life of the classroom.Laws establish a teacher's and students' basic rights and responsibilities within an education system, national and regional curricula influence the content of learning, and initial teacher education and continuing professional development opportunities set the bar for teachers' pedagogical abilities.The physical school environment and the available material resources can support or hinder particular types of learning, as can scheduling and time constraints (Connors 2016;Daniels 2016).School leaders, teacher colleagues and parents may also have certain expectations that guide how teachers and students enact teaching and learning in the classroom (Connors 2016;Daniels and Pirayoff 2015).Although this article focuses on life inside the classroom, the impact of extrinsic conditions on intercultural learning should not be underestimated.They contribute to the dimensions of power, knowledge, relatedness, purpose, structure, continuity and meaningfulness, which in turn create the learning space.

Learning spaces: promoting intercultural encountering
A learning space is a situation which enables a particular type of learning to occur (Marton and Tsui 2004), involving the social, cultural and physical environment.Teachers and students are considered active agents in its co-creation.This understanding is derived from an amalgamation of  theoretical perspectives, including Dewey's (1938Dewey's ( /2015) ) concept of a situation, Vygotsky's (1986) zone of proximal development and Marton and Tsui's (2004) space of learning, as well as drawing from self-determination theory (Ryan andDeci 2017, 2000).Following abductive data analysis (Timmermans and Tavory 2012), we have formed a theoretical framework that describes a learning space through seven key dimensions: power, relatedness, knowledge, purpose, structure, continuity and meaningfulness.They can be used as analytical tools to explain why a learning space unfolds the way it does.Figure 3 illustrates how the dimensions not only influence the learning space, but also interact with each other.For example, if power is distributed democratically in a classroom, it will likely have an influence on how the class members relate to one another and how the learning space is structured, and vice versa.
The next section will describe how each theoretical dimension contributes to creating a learning space that promotes intercultural encountering in the primary school classroom.

Power
Power in a learning space can range from authoritative to democratic, and this influences whether reciprocal encountering can take place.Authoritative power, derived from the educational institution, enables teachers to control who speaks and what can be said (Aro, 2012;Foucault 1977Foucault / 1995;;Mercer and Littleton 2007).In contrast, power is distributed more democratically when children are seen as agentic (Karlsson 2020) and can shape encounters by drawing from their own cultural repertoires (Gutiérrez and Rogoff 2003;Piipponen and Karlsson 2021).A democratic learning space enables children to voice their thoughts and take action in transformative ways (Freire 1970(Freire / 1996)).In practice, teachers distribute power by supporting students' autonomy (Ryan and Deci 2017;Turner et al. 2014) and social participation (Karlsson 2020), and students use agentic engagement to proactively influence the teacher and their learning environment (Reeve 2013).

Knowledge
The way knowledge is conceived in broader educational structures and discourses affects what kinds of knowledge are privileged in schools (e.g.Gorski 2009;Portera 2008;Tornberg 2013).Knowledge is defined on a continuum as dynamic, experiential and situated at one end and static, cognitive and universal at the other (Dewey 1938(Dewey /2015)).Understanding culture as a static, unchanging entity will likely lead to designing learning experiences that focus on concrete aspects, such as dress, food and festivals, which tend to reinforce cultural stereotypes and essentialism (Alvaré 2017;Piller 2017).However, if culture is understood as dynamic, layered and actively created by its members, there is more room for a nuanced understanding to develop (Dervin 2011;Layne, Trémion, and Dervin 2015;Piipponen and Karlsson 2019;Piller 2017).

Relatedness
Relatedness is a basic need, which means 'feeling connected and involved with others' (Ryan and Deci 2017, 87).The lack of relatedness produces fear, suspicion and a relational hierarchy.Intercultural encounters are characterised by feeling connected to one's own community as well as finding a way to relate to the other (Jaatinen 2015).A warm, inclusive classroom culture is thus necessary for students to feel heard when engaging in intercultural encountering (Karlsson 2020;Lanas 2017).Teachers promote relatedness in practice by modelling and encouraging mutual respect and by creating conditions for students to collaborate productively (Turner et al. 2014).

Purpose
The purpose of intercultural education historically began as developing sojourners' (e.g.diplomats, business and military personnel) communicative competence in preparation for going abroad (Piller 2017), but alongside this instrumental view has developed the more holistic purpose of understanding and acting in a diverse society (Abdallah-Pretceille 2006;Layne, Trémion, and Dervin 2015).Viewing intercultural learning as simply an instrumental tool that 'helps us to get along' ignores the complex sociohistorical contexts that have an impact on intercultural encounters, so we argue that learning should not be limited to decontextualised skills and knowledge, but aim for holistic growth of the learners (Dewey 1938(Dewey /2015;;Lanas 2017;McCandless et al. 2020;Tornberg 2013).

Structure
Structure refers to supports and opportunities in a learning space that enable learning to take place (George and Richardson 2019).Language is a key tool for structuring learning (Kumpulainen and Renshaw 2007; Marton and Tsui 2004;Mercer and Littleton 2007).Language is not a neutral communication device; instead, it focuses attention on certain perspectives and transmits certain ideological discourses (Freire 1970(Freire /1996)).A structure that supports reciprocal encountering enables students to personally experience encountering someone from another culture, with learning taking place if the students engage dialogically with each other (Kaikkonen 2004).More commonly, intercultural learning is structured for students to learn facts about another culture, perhaps because imparting knowledge through class discussions is so habitual to teachers (Howe and Abedin 2013;Kaikkonen 2004).However, learning information is different to engaging in reciprocal encountering, because it imparts a static, objectifying understanding of cultural Others (McCandless et al. 2020).Dewey (1938Dewey ( /2015, 35) , 35) describes the connectedness of situations in different points in time as 'continuity of experience', because our experiences flow into one another as we interpret new situations through the tinted glasses of memory, prior knowledge and future expectations.Thus, students' and teachers' previous experiences of similar situations define how they frame a new learning situation (Dewey 1938(Dewey /2015;;Marton and Tsui 2004).Children's personal experiences and familiarity with different languages and cultures promote openness to encountering in the future (Ben Maad 2016).Not all children have opportunities to experience cultural diversity in their own environment; therefore, intercultural learning should be integrated into school curricula and everyday practices, so that students have repeated opportunities over time to develop their abilities (Ragnarsdóttir and Blöndal 2015).Encountering others with reciprocity cannot be learned during a one-off lesson, but develops gradually as children's experiences of encountering accumulate.

Meaningfulness
Meaningfulness explains the value of or interest towards a task, and supports intrinsic motivation (Turner et al. 2014).Learning is experienced as meaningful when it authentically connects to children's lives.It is particularly important to make intercultural encounters feel meaningful for childrenencountering is a deeply personal endeavour, because it involves experiencing the other side as well as opening up the self (Buber 1947(Buber /2002)).Participatory practices are a way for teachers to learn what children think, how they learn and what is meaningful to them (Hännikäinen and Rasku-Puttonen 2010).This involves developing the classroom culture so that it values and supports children's contributions and initiatives (Karlsson 2020).

Findings
Four types of learning spaces were identified over the course of the three cycles of design-based research: an ambivalent space, a space of information, a transitional space and a space of reciprocal encountering.The spaces followed a rough progression so that cases of the ambivalent space were most common during the first cycle of research and the space of reciprocal encountering was most common during the last cycle.With growing experience, the teachers were able to guide the learning spaces more effectively towards reciprocal encountering.However, if the key dimensions of power, relatedness, knowledge, purpose, structure, continuity or meaningfulness were not well aligned for encountering, sometimes classes reverted to one of the other three spaces.

Ambivalent space
An ambivalent space was formed particularly when the exchange was new to both the teacher and children, so that power, knowledge, purpose and structure fluctuated.The ambivalent space featured most prominently in the beginning of the first research cycle during class discussions in School A. The following extract took place after the teacher-researcher read aloud a story from a partner class.The extract shows how the teacher-researcher (Teacher 1) used an open-ended question to create a democratic space for the children to share their thoughts, but when the children began to analyse the form of the story, she steered the conversation to the content (security cameras).Her purpose was to discuss how the children experienced the story, but ironically, the outcome was a typical authoritative class discussion where the teacher decides what should be discussed and the children answer in short phrases (Mercer and Littleton 2007).The space was structured (unreflectingly) like a language arts lesson, where evaluating the form of texts is typical, so Jimmy commented on the length of the story, and even after the teacher tried to redirect, Neil criticised the scant use of description.The children's responses show that the continuity of classroom culture influences strongly how they interpret the learning space (Dewey 1938(Dewey /2015)).
Critiquing the story did not support reciprocally encountering the exchange partners, because the children's responses were actually intended for their teacher.They demonstrated that they have the right knowledge that the teacher typically assesses for in the context of language arts (Aro, 2012).It was unproductive to document the children's responses to the partner's stories through a class discussion, because talking about experiences is a meaningful method for an (adult) teacher, but it does not account for the breadth of ways in which children express their experiences.

Space of information
Like the ambivalent space, the space of information typically occurred during the class discussions in the first cycle of research in School A. The space was created when the children grew curious of their exchange partners after hearing a story in Finnish.The questions they asked ranged from language, clothes, make-up, sweets and sports to school, currency, wildlife and uniquely Finnish words, drawing from the children's interests and cultural repertoires.In other words, they connected the questions to their personal experiences and knowledge, which made the discussion meaningful.
However, by answering the children's questions, the teacher-researcher positioned herself as a knowledgeable expert and thus implicitly communicated to the children that the purpose of the lesson was to increase their knowledge about Finland.The discussion became a question-andanswer session where the students were positioned as apprentices learning from the expert teacher.This structure did not support encountering between the exchange partners, but rather objectified, even essentialised them.The question-and-answer sessions upheld a static conception of culture, as the teacher-researcher described in the researcher diary: Overall, the questions were very concrete, based on superficial observations.What they [the students] got out of the conversation was a list of facts, rather than a deep understanding.(Researcher diary, 18/10/2015) A better course of action would have been to collect the children's questions and send them to the partner class to be answered.Without the ability to engage directly with the exchange partners, the children were left to make comparisons between their own lives and the 'facts about Finland' as described by their teacher.Whilst a comparative approach to cultural knowledge could in some cases help students to view the world from different perspectives, in this case it reinforced a juxtaposition between 'us' and 'them'.After hearing about some of the features of the school system in Finland, such as a later start-age for school than Scotland, not having school uniforms and calling teachers informally by the first name instead of surname, several of the Scottish children deduced that school must be less rigorous for their Finnish counterparts.This inference was further confirmed by how short several of their exchange partners' stories had been.This shows how easy it is to make judgements of another culture when cultural 'facts' are removed from their context, and the need to support relatedness between the exchange partners, which cannot be achieved in a space of information.

Transitional space
The transitional space occurred usually at the beginning of each research cycle.The children wanted to test the limits of the new space: how far is it possible to go in telling about taboo topics before someone objects?Is the power to tell really with the storyteller?Rebellious or taboo content appeared in the very first story of a class in School D in the second cycle of research.A child began the story with the taboo of death: 'Once there was a guy named Bob and he died'.Several children then during their turns proceeded to challenge the classroom norms of appropriateness by mentioning drunkenness, excrement, vomiting, infidelity and killing.Similar taboo topics emerged in a few stories in each participating class, but the example above was the most poignant for incorporating so many different taboos in one tale.The teachers did not comment or flinch at these, but simply wrote down each story as it was told.In traditional school culture, taboo topics are typically reprimanded, but the teachers and children were able to negotiate an exception to this rule, as Teacher 2 explains: One child said that poop stories are not allowed in school because there is an agreement not to tell stories with poop or guns, but now this story [from School D] DID mention poop.I said that when Storycrafting, you can say whatever comes to mind, that's the rule of Storycrafting.So this time we must deviate from that school rule.After that [the child] happily drew a picture of the bird that POOPED on Bob.(Teacher email, 7/4/2016) The space is named 'transitional', because it occurs very briefly, if at all, when the class community transitions from the status quo classroom culture to a democratic space, where children's knowledge that is not typically valued in school becomes visible.However, the looser structure called students to push boundaries.As the new space provided unusual freedom, many students in School A wanted to use their voice as much as possible: Many of the children want to have a turn, but get bored listening to others and start chatting with the people close to them.On the other hand, often they are so involved in the story that they shout out suggestions even when it is not their turn.(Researcher diary, 29/3/2015) There were 27 children in this class (13 had research consent), so they were used to having to compete for space to speak.At this stage, the teacher-researcher took the role of a scribe and listener, and reaffirmed that she would write down any story the children chose to tell.She did not revert to an authority figure striving to control, but allowed the children to experience the sense of freedom and her interest in their voices.Once the children trusted that the democratic space was, indeed, for them, the sense of urgency subsided and the class as a community learned a new way of being, listening, telling and laughing together.

Space of reciprocal encountering
The space of reciprocal encountering is created when children can playfully encounter each other through the partner's stories and drawings.The teacher assists in structuring the space so that it is democratic and inclusive, but resists being the recipient of the children's creations.She accepts the stories and drawings without attempting to correct or evaluate them.The authorship and power over the stories and drawings rests with the children, so the breadth of children's knowledge becomes visible and valued.Over time, the exchange partner classes can create a shared narrative culture (Piipponen and Karlsson 2021).
The children felt warmth and interest towards their exchange partners, as illustrated in the following teacher email: One of the children said that we should definitely travel to Belgium to see the storytellers.And that you should travel to Finland to visit our class.But in our school, grade 3 doesn't yet go on long trips.The tradition is that in grade 6 we could go, if the parents have collected money.(Teacher email, 1/4/2016) The email also illustrates how extrinsic conditions, such as school culture and cooperation with parents, create boundaries for teachers to enable children's initiatives.
In the second and third cycles of research at School D, the teacher-researcher (Teacher 1) supported relatedness and a democratic power structure in the class community more explicitly than in the first cycle.After she taught the Storycrafting method to the children, she asked them for suggestions on how to equally include everyone in the activity.The students' suggestions were discussed, and through a whole-class negotiation, some ground rules were agreed.
We agreed on some key rules: (…) The teacher chooses the person who starts.Each storyteller then chooses someone to follow.The children decided it should be boy-girl-boy-girl, so that boys and girls have an equal chance of being chosen.(Researcher diary, 31/1/2016) As the children tended to be friends with those of the same gender, alternating between boys' and girls' turns meant they couldn't favour only their friends, but everyone in the class had to be included.This way, the different friendship groups of the class came together and reinforced their shared group identity as a class.
The teacher's role was significant in introducing the expectations of the story exchange, but with time the children became more autonomous and introduced new initiatives in the second and third cycles of research.After becoming familiar with the Storycrafting method, the children often worked in small groups and scribed each other's stories.Many children in School D had enjoyed telling the class stories together, so they invented an alternating scribe role to allow each group member to tell a part of the story.Furthermore, often groups requested if they could read their finished story out to the entire class before sending it to their partners in Finland.The suggestion to write a thank you letter also came from the children.The children really enjoyed the sense of togetherness that permeated the exchange project, and so their initiatives further supported reciprocity and mutual respect within and between the class communities.The same is true of School B in Finland.Teacher 2 explained that Storycrafting became particularly important for children on the verges of the class community: In our class, a couple of artistic students withdraw from the so-called inside group of students.Additionally, a few are otherwise just left on the outside.It is these students who love telling and drawing during Storycrafting.Everyone was curious to hear and see your creations.(Teacher email, 7/6/2017) By focusing on the enjoyment of storytelling rather than an analytic approach to texts, the children were able to immerse themselves in the story worlds and experience the thrill of each other's tales in their imaginations, thus encountering the other with mutuality and openness.Table 3 summarises the findings.

Discussion
This study has found that reciprocal intercultural encountering in the primary school classroom happens best when the classroom culture supports a democratic, experiential and inclusive approach, where learning is understood as holistic, exploratory, contextualised and cumulative.
In the beginning, the learning spaces of the story exchange were often ambivalent or transitional, or focused on learning information rather than encountering.As the children'sand especially the teachers'experiences of the exchange grew, the classroom culture shifted so that reciprocal encountering could be achieved.The theoretic framework developed in this study provides an analytic tool for teachers to use when designing spaces for reciprocal encountering.
Developing the story exchange produced important insights for the teachers.It turned out that shifting the focus from the product of learning (the stories) to the process of learning (engaging in a space for reciprocal encountering) required resisting some of the norms of traditional school culture, where the teacher controls discourses.Changing existing cultural practices may also require the teacher to engage in critical, reflexive action, as the teacher may be unintentionally upholding the very structures she is attempting to change (Bash 2014).In this study, the dual role of the teacher-researcher supported the transformation of practice, but also led to a professional transformation.Insights from theory and empirical work motivate the teacher-researcher to involve the students more actively in the co-creation of learning spaces (Niemi 2019).For example, not judging or evaluating the content of the student's stories was transformative action, because it shifted the purpose of the learning space from performative to relational.
The findings lead to important implications for teacher educators, policy-makers and administrators: if we are to fully embrace intercultural education, educational structures should encourage and enable teachers to develop a warm, participatory, democratic classroom culture (Hännikäinen and Rasku-Puttonen 2010;Karlsson 2020;Lanas 2017).Children should be seen as agents who participate in (re)creating classroom culture together with the teacher, and capable of transformative action (Freire 1970(Freire /1996;;McCandless et al. 2020).Instrumental, performative teaching practices can work for learning that is narrowly defined as knowledge-acquisition, but this style of teaching in the field of intercultural education is at best superficial and at worst counterproductive, even harmful (Gorski 2008;Lanas 2017;McCandless et al. 2020).Our study reconfirmed a frequent finding in the literature that good intentions do not always lead to good outcomes (Alvaré 2017;Gorski 2008;Lau 2015).Knowing 'facts' about a culture is not sufficient to help children develop empathy, reciprocal encountering skills or critical action (McCandless et al. 2020).
In the same vein, we would like to question authoritative assessment practices that are a tool to control students (Foucault 1977(Foucault /1995)).Should intercultural encountering be assessed formally?Our findings indicate that when the teachers resisted evaluating the children's stories and drawings and instead showed they valued the children's contributions, the focus of the exchange shifted away from analysing the learning products towards the process of encountering.However, we acknowledge that teachers may not have control over assessment practices, especially if a school system practises recurrent, high-stakes testing.High-stakes testing, a hierarchical management structure and instrumental goals in education systems are derived from the ideology of neoliberalism; we concur with prior research that this kind of focus on school performance narrows the purpose of education and the opportunities for intercultural learning (Gorski 2008;Lanas 2017;Portera 2008).
This study focused on developing a space for reciprocal intercultural encountering, but intercultural learning also includes conflict resolution (Gorski 2009;Kaikkonen 2004).Once the shared narrative culture and a sense of mutuality were established with the partner class, the teachers could have further created opportunities for children to engage with conflicts, such as dealing with stereotypes, and how to discuss differing beliefs.It would be worthwhile to conduct further research on how children can start developing critical self-awareness of their own beliefs and prejudices, and whether the Storycrafting exchange method could be further developed to address critical or decolonising interculturality (Gorski 2008(Gorski , 2009)).Some limitations have been identified.Given that this was a small-scale, qualitative, development research project, the theory and findings pertain to this specific research context only, and largescale, longitudinal research is warranted to make any generalisations.Furthermore, it would be valuable to do research on the intercultural story exchange in other parts of the world, in different school systems and cultures, especially between the Global North and South, as previous research points to particular challenges in projects that cross this divide (Dervin 2011;Gorski 2008;Lau 2015;MacKenzie, Enslin, and Hedge 2016;Royds 2015).
The participant (auto-)ethnographic approach posed some challenges for data collection.Working simultaneously as teacher and researcher meant that sometimes observation notes were fragmented, written up after a time delay, or just missing.However, the gaps can be filled by the multiple methods of data production as well as the teacher-researcher's intimate contextual knowledge of the research setting.The data may also be partially skewed in School A, because only 13 out of 27 students returned the consent form, whereas in the other schools, consent was received from the majority of students and their parents.Finally, the researchers' interpretations dominate the data analysis process.In upcoming research, interviews with the students will address this shortcoming.Future research will also give consideration to children's experiences of intercultural encountering between different micro-cultures, such as friendship, gender, interest, age or language groups.

Conclusion
Reciprocal intercultural encountering is more than simply being in contact with someone from a different cultural background; it requires deliberate effort from teachers to develop a constructive classroom culture over time, which fosters reciprocal encountering even between distant exchange partners.Although at first the children's learning from the story exchange seemed superficial, the ambivalent spaces and spaces of information that occurred in the classroom helped the researchers to become aware of how dimensions such as power, relatedness and continuity influence learning spaces.Using the resulting theoretic framework will hopefully help other educators to develop learning spaces that promote reciprocal intercultural encountering.

Figure 1 .
Figure 1.Classes from Schools B and D exchanged stories and drawings in the second and third cycles of research.The stories were translated by Teacher 1.

Figure 3 .
Figure 3. Dimensions of power, relatedness, knowledge, purpose, structure, continuity, and meaningfulness interact with each other to form a learning space.

Figure 2 .
Figure 2. Development of classroom culture over time as a continuum of learning spaces, influenced by extrinsic conditions on global, national and local levels.

Table 1 .
Description of schools and participants for each research cycle.

Table 2 .
Data produced over three cycles of research between 2015 and 2017.

Table 3 .
Learning spaces co-created by teachers and students during an intercultural story exchange.