Development of intercultural competence among a class of 5 graders using a picture book

This paper presents a study in a Norwegian English as a Foreign Language class of 5th graders aimed at investigating the learners' ability to decentre a skill within intercultural competence, through the use of the picture book The Soccer Fence by Phil Bildner. Based on Byram's theory on intercultural competence and Piaget's theory on decentring, a literary context was created around a one-group pretest-posttest design. The participants’ ability to decentre was measured before and after an intervention period. During this intervention the group worked with the narrative and explored the different perspectives of the two main characters through various tasks. The results were then checked for statistical significance. The article also includes results from group interviews, which were carried out to either strengthen or challenge the pretest and posttest findings. Results showed statistical significance, indicating that the ability to decentre did increase after the intervention. This was also supported by the interviews. This study argues that because of the diversification of linguistic and cultural backgrounds in Norwegian classrooms, intercultural competence should favourably be explored more intentionally. It argues that literature, and the picture book specifically, provides interesting and relevant opportunities for teachers to incorporate intercultural aspects into their classroom contexts.

communication and perspective-taking through literature has played a central role in both understanding other cultures, as well as helping children from other cultures to better integrate into Norwegian classrooms. Language learning is culture learning, and the combination of the two is how we make sense of the world that surrounds us. This article argues the role of literature in providing exciting opportunities for developing intercultural competence and can be adapted to suit virtually all age groups.

The aim of the project
This article presents a study of the development of intercultural competence in the English subject among a Norwegian class of 5th graders. Intercultural competence is defined as the "ability to interact effectively with people from cultures that we recognize as being different from our own" (Guilherme, 2000, p. 297). This definition reflects the view that language cannot be taught independently from culture and the communicative situation. Thus, it is not a question of why intercultural competence should be taught, but rather what exactly intercultural competence is, and how should it be taught. This current study reports on how the picture book The Soccer Fence by Phil Bildner was used to influence the pupils' ability to address the different perspectives of fictional characters. Decentering is the ability to see reality from different perspectives than ones' own. The participants' ability to decentre was measured before and after an intervention period. During the intervention, the group read the story and explored the different perspectives of two characters through various tasks. Based on Byram's theory on intercultural competence and Piaget's theory on decentring, a literary context was created around a one-group pre-test-post-test design. Melvin Feffer's Decentering Scoring system was used as a main tool of analysis, as well as his suggested activities that sought to increase decentring skills during the intervention, as exemplified through Burkman (2008) and Fincher (2012).
Decentring is an element within intercultural competence, located under what Byram (2008) calls savoir être, concerned with openness and attitudes. Fincher defines decentring as "the ability to take another person's perspective", an element necessary to be able to act appropriately in an intercultural context (2012, p.1). Previously, most research on decentring has occurred in clinical settings, in psychopathology, and studies on metacognitive processes (cf. Bernstein et al., 2015;Kessel et al., 2016;Lebois et al, 2015). One field that actively explores different roles and changing perspectives is drama studies. Perspective-taking is inherent in drama studies, and other subjects can benefit from this field, including the foreign language classroom. One example is the research project Flukten fra Syria [The Escape from Syria] -a process drama with role and perspective exchanges in mathematics (Allern & Drageset, 2017). This project investigated how pupils in the 5th grade in a Norwegian primary school responded to exploring different perspectives and roles, such as refugees, human smugglers, and Italian coastguards. The study showed that practising perspective-taking leads to profound learning, and it also showed enhanced pupil engagement. Process drama has also successfully been used in a small-scale study in EFL Teacher Education to foster empathy and multiperspectivity (Savíc & Normand, 2017).
Despite previous relevant research projects within other fields, there is a need for further research within the English subject on how decentring can deliberately be developed to improve intercultural competence. The purpose of the study's design was to create a literary context in a Norwegian English classroom, where participants were to interpret and take positions on situations different from the ones they were used to. Thus, the research question was as follows: Do a group of 5th-grade pupils in Norway improve their decentring skills after an intervention period working with a picture book from different perspectives?

Background
The development of learners' intercultural competence and understanding is an important goal in education. Intercultural education challenges learners to take responsibility for their neighbour, both close and far, shifting the focus both between the global and local responsibilities. UNESCO highlights the importance of intercultural education not only being an add-on to the existing curriculum but also influencing the learning material, the methods of learning, languages of instruction etc. The inclusion of multiple perspectives and voices is emphasised, especially those from non-dominant groups in society (UNESCO, 2006, p. 19).
The Norwegian Directorate for Education and Training (NDET) has explicitly formulated aims for the English subject regarding intercultural competence, about exploring ways of living and traditions of different societies in English speaking countries (NDET 1, 2020). It is also highlighted that teaching should help pupils reflect on their identity and cultural belonging (NDET 1, 2020). The new curriculum of 2020 divides the English subject into the focus areas "communication", "language learning" and "meeting English speaking texts," specifying the need for competence within intercultural understanding and citizenship (NDET 2, 2020). Intercultural competence is a natural and necessary part of language competence, simply because English is used worldwide. Language teachers have an important responsibility to ensure that foreign language learners acquire the different skills needed to be competent in the target language (Byram, 2008, p. 83). These skills do not only involve the basic skills of listening, speaking, writing, and reading, but also literary and cultural knowledge and critical literacy. There is ample support in the Knowledge promotion of 2006, which still applies for 10th grade, Vg2, and Vg3, for teaching intercultural competence. The core curriculum emphasizes the importance of knowledge about other cultures and people, which in turn gives us the chance "to foster mutual understanding and respect" (NDET 3). In the English subject curriculum, the aims regarding culture, society, and literature after year 10, include developing specific knowledge, but also discussing, describing, and reflecting on the current situations in English speaking countries (NDET 4). Thus, we can see that being active, seeing things from different perspectives, as well as being critical, is quite central.
The current article discusses how a picture book about being a black boy in South Africa in the 1990s can be used to develop the skill of decentring, a central element in intercultural competence. There is ample support for the use of literature and narratives in EFL teaching, as literature can engage and evoke feelings in a very special way (cf. Day & Bamford, 1998;Ibsen, 2000;Ibsen & Wiland, 2000;Birketveit et al., 2013;Hoff, 2013). It contains relatable characters and describes difficult situations through the eyes of these characters. Thus, stories "help participants to decentre from their own values", and what they daily take for granted (Barrett et al., 2014, p. 40). Using a narrative presents an alternative, and, in many ways, a favourable approach to the teaching of historical events. By identifying with different characters or experiencing different perspectives in a narrative, pupils have to interpret and mediate between different cultures, which are skills needed in appropriate and effective foreign language communication. As argued by Hoff (2017), literature can be seen as an intercultural experience in itself. Meaning is created in the interaction between text and reader, as the reader navigates the complex identities of the text. Fenner's (2000;2001) contribution is relevant here as she describes how the act of reading fiction in a foreign language can be seen as a personal encounter with the foreign culture, which requires adjustment of attitudes between individuals and between cultures.
A picture book was chosen as reading material for this study. As discussed by Birketveit et al. (2013), there are several reasons why picture books are ideal texts for EFL teaching as picture books have at least one picture on every double opening. Thus texts come with ample visual support that will enable learners of English to cope with more advanced texts than they usually would. By reading picture books, they are exposed to authentic stories and authentic English offering a wider vocabulary than course books or graded readers do.
Picture books tell the stories through the iconotext, the interdependence of pictures and words (Birketveit et al., 2013, p.19). Artistic illustrators often make picture books into objects of art, e.g. through the symbolic use of colours and foregrounding of characters or objects in addition to clever use of endpapers and title pages. Another advantage of using picture books in EFL is that they often contain little text. Thus, they offer the reader the pleasure of reading whole stories instead of extracts of stories much used in course books. There is something deeply satisfying about having read a whole book, and it is likely to give children a sense of accomplishment. The picture book that was chosen for this intervention study, The Soccer Fence, is an example of a book that contains symbolic use of colours and artistic illustrations.
A thorough presentation of the book will follow in the methods section.

Michael Byram's theory of intercultural communicative competence (ICC)
According to Byram, being intercultural is an activity, where a person acts as a mediator between cultures and the people in them, as he or she takes an external perspective on a situation (2008, p. 68). The process of acquiring intercultural competence involves five elements, which Byram calls savoirs. The French word "savoir" refers to knowledge or knowing, and the first element of intercultural competence is called just that, savoirs. Cultural differences are often the cause of disputes and communicative challenges, which is why thorough knowledge of how social groups and social identities and practices function is crucial. Savoir comprendre describes a person's skill in interpreting and relating. By this is meant the ability to look at an event or document from a different culture and relate it to one's own, trying to interpret it in the light of the knowledge one has. Savoir apprendre is the skill of discovery and interaction. It regards a person's ability to acquire new information about a culture and its traditions, and to use these in communication and interaction (Byram, 2008, p. 69).
Savoir's engager is often called critical cultural awareness or political education, as it can be seen as having the ability to look at a situation from a metaperspective. Generalizations and stereotypes that have developed over time tend to influence people, and critical cultural awareness highlights the important need to be aware of the assumptions and judgements we make that can be products of our society (Houghton et al., 2013, p. 1). The last and final savoir in Byram's model is called savoir être, which covers attitudes. The ability to decentre is an aspect within savoir être, along with empathy, openness and curiosity (Byram, 2008, p. 69). In terms of primary education, Byram states that this is the savoir most appropriate to consider in this context. Byram's ICC model has become very influential in foreign language teaching, but his model has also been criticised for concerning itself mainly with critical cultural competence, and not enough with the relational and dialogic importance of the intercultural development (Matsuo, 2015, pp. 3-4). Furthermore, Matsuo (2015) criticises Byram for creating an individual-oriented list-type model that does not explain the relations between its components (p. 15). Rather than accepting all of Byram's savoirs, this study focussed on savoir être as it is the most appropriate for the age group studied, although as demonstrated in the Results section, the pupils also showed abilities within the other savoir categories.
A central objective within savoir être is broadening one's perspective. Byram exemplifies this when proposing that pupils could engage in role-play or be explained a (hypothetical) situation and then try to look at it from different perspectives. Usually, one sees a situation or conflict in light of one's values and opinions. An important challenge to achieve growth within this area is to try to look at phenomena from different perspectives, both from familiar and unfamiliar vantage points, this being the ability to decentre or decentring.
According to Cromer (1971), Piaget first introduced the term when working with theories on cognitive development. He explained that the child has a developmental process that takes it from initial egocentrism, where everything revolves around the child's own body and actions, to a decentred state (Cromer, 1971, p. 353). Piaget defined decentring as "the ability to conceptualize multiple perspectives simultaneously, or being capable of observing a situation from more than one point of view" (in Fincher, 2012, p. 6). Also, Piaget suggested that the decentring ability increased with chronological age, and was not sufficiently developed until the formal operational stage (11-12 years of age). This age-claim is the main reason why learners from a 5th-grade class were chosen to be participants in the present study.
It should be noted that parts of Piaget's model have also been subject to criticism.
According to Cherry (2016), most researchers seem to agree that children develop many of the abilities at an earlier age than Piaget claimed and that they are far less egocentric than he held. Cherry (2016) also claims that even children as young as 4-or 5-years old have, to some degree, the ability to take the perspective of others. This researcher has seen decentring skills among children far younger than what is claimed by Piaget, and the present study, being carried out among 9 and 10-year-olds, also suggests than decentring is, in fact, a skill possible to develop below the formal operational stage (11-12 years of age).

The Intervention
Through the use of the picture book, The Soccer Fence, and based on Byram's ICC theory, the goal of the study was to measure the development of a class of 5th graders' abilities to decentre. The ability to decentre was measured before and after an intervention period.

The Text
The Soccer Fence by Phil Bildner, illustrated by Jesse Joshua Watson, was published in 2014.
In this carefully crafted picture book, the verbal text and the artistic illustrations use the black boy Hector's struggle to play football with the white boys as a metaphor for the last stages of  (1996). All these events are part of the plot.
Hector is the narrator, and he loves to play football. Twice a month, together with his mother and his sister, he rides the PUTCO bus (main public transport for people of colour under apartheid) to a wealthier neighbourhood where his mother works. From behind the soccer fence, Hector watches some white boys playing football in the park. This is the 90s, and despite Hector's attempts to establish contact, the white boys do not take any notice of him. However, he gradually becomes visible to them as he one day kicks the ball back over the fence (Bildner, 2014, 6th opening). Finally, he is asked to play (Bildner, 2014, 15th opening). There are obvious parallels to the events dismantling apartheid. Thus, Hector's whisper: "One day, I am going to play on a field just like that. One day real soon" (Bildner, 2014, 3rd opening), is followed by the celebration of Nelson Mandela's release from prison (4th opening). When Hector kicks the ball back over the fence, the next opening shows queuing at the polling station for the first open election in South Africa (Bildner, 2014, 7th opening). Then we are told that President Mandela loved sports and how he used the hosting of the African Cup of Nations to rally people around the national team, Bafana Bafana.
During the finals, Hector and the blonde-haired boy, Chris, bond through waving the South African flag and leading the fans in celebrating the victory. When Hector visits the wealthy neighbourhood after the national team have won the African Cup of Nations, Chris asks Hector if he wants to play, and when he steps through the football fence, he metaphorically steps through the colour barrier. Sports and politics have come together.
High-quality picture books such as The Soccer Fence make skilful use of the paratext in telling the story. The title page shows the Johannesburg township in sketchy pencil drawings against an orange background. The houses have tinned roofs and an old tyre is lying outside the shabby-looking shacks. The washing hangs outside to dry. To the right, there is the wired fence, and the street is rocky. In the far distance, high-rise buildings can be seen.
The first double opening shows Hector playing football barefoot in the street using empty cartons as a goal. The next opening shows him sprinting along the wire fence that he calls "the out-of-bounds that separated our Johannesburg township from the rest of the world" (Bildner, 2014, 2nd opening). The main characters of the story are drawn in strong and clear colours with solid contours making them stand out as strong and beautiful. The township setting is always in sketchy grey pencil drawings rendering the background somewhat hazy, nondescript and unobstrusive, although the perceptive reader will notice the poverty. The part where Hector stands outside the football fence watching the white boys play (Bildner, 2014, 3rd opening) is especially interesting in the way it shows the spaces dedicated to white and black people. In the foreground, the white boys play football in colourful and proper football gear in the green park. Behind the fence, where Hector stands, people of colour descend from the PUTCO bus into the street, but both people and objects are nondescript and drawn in grey pencil details against a light blue background. Ironically and symbolically and in contrast to the white people's space, the space belonging to the people of colour is colourless.

Participants
The participants were a 5th-grade class, twenty-one in number, from an ethnically diverse school in the western part of Norway. The research was carried out in this school because the main researcher (Furnes Sindland) had existing contacts there who facilitated access. Because of the topic, the teacher and parents of the participant group were especially interested in participating. Countries represented were Ethiopia, the Philippines, Spain, Somalia, Iran, Sweden, Poland, and Germany. Each pupil was awarded a number that was used to identify their pre and posttest instead of using their names, to ensure anonymity. The main researcher (Furnes Sindland) taught the lessons, starting with a 15-minute collective reading of the book, explained more thoroughly in the following sub-chapter on research procedure. Table 1 The illustrated process of data collection

Research procedure
Step 1: Pretest Before both the pretest and posttest, a collective reading of the picture book took place. Here, each page was shown on the board, and both pupils and the teacher took turns in reading and translating the text. This was done to ensure that all the pupils both understood the plot and who the characters were. All pupils spoke Norwegian fluently, but their English competence varied, and the language of the book was too advanced for some of the participants. The participants were then given a sheet consisting of four questions. These questions were to be answered based on a picture from the book that was displayed on the smartboard. Although it was regarded as unlikely that the pupils would remember their pretest answers, a different picture was chosen for the posttest. The pupils answered in writing, and, because of the varying levels of English, they were allowed to answer the questions in either English or Norwegian. Most chose to answer in Norwegian, except three pupils.
Questions 1-3 are adapted from Feffer's studies on perspective-taking, which were based on the participants' reactions to pictures (Fincher, 2012, p. 20 Step 2: Group interview Rounds of group interviews were held in groups of five or six in the front of the classroom while the rest of the class worked individually on an apartheid-related task. Putting them in groups not only allowed us to see if they answered the same as on the written pretest and posttest, but also allowed us to see if they changed it based on other pupils' answers, and maybe witness their thought processes as they answered the different questions. The researcher wrote down the numbers of each participant as they answered so that their numbers matched with their pretest-posttest answers. Each interview consisted of the same four questions that were asked during the pretest and posttest, with the same pictures. The reason for using the same questions was to reveal any misunderstandings and gain insight into how the pupils arrived at their answers from the pretest. The participants took turns in answering the questions, and which participant was asked to answer first would vary. The main findings derived from the tests, but using group interviews provided a more nuanced picture of the participants' decentring skills. The interviews showed whether any participants changed their answer when put in a larger group, thereby suggesting whether their answers were reliable or if they changed under the influence of others. It is also possible that their conviction was not very strong or that they were persuaded. Gubrium and Holstein (2001) note that interviews with children can be successful, both to allow them to "give voice to their own interpretations and thoughts" instead of merely adopting adults' opinions, and to investigate themes "that are salient in their lives, but do not occur in daily conversations" (p. 181). During the research, I decided it might also be easier for the participants to express themselves orally with the opportunity to explain, and it would also pick up on information that the tests missed, e.g. if a pupil had misunderstood the question. All interviews were sound-recorded and transcribed.
Step 3: Intervention First (1), the participants were to write a comic strip where they retell the story from the view of Hector's parents. Secondly (2), they participated in a group discussion. They discussed the questions first in pairs and then contributed with their answers in a plenary conversation. This conversation was also based on the same four questions. These activities were based on both Feffer's first measure of decentring, the Role Taking Task (RTT) and the Thematic Apperception Task (TAT), which focus on retelling a story in different ways and from different angles (Leeper et al., 2007, p. 151 Thirdly (3), the pupils wrote a short text telling the story from the perspective of Chris, the second character. He is an important character in the story, but unlike Hector, his thoughts and feelings are not explained in the text.
Step 4: Posttest Identical to step 1, with a different image from the picture book. This choice will be further discussed under limitations in chapter 5.4.

Data analysis
Based on Feffer's scoring system for decentring, explained in Burkman (2008), an example table for the different scores and categories was created for this specific study (table 2, below) (Burkman, 2008, p. 13). Examples shown in this table are based on the pupils' actual answers from the pretest, and the table ranges from score 1 to 9. The tests were immediately scored, assigning them to a level of decentring, first by the researcher (Furnes Sindland) and after by two external scorers. Two graduate scorers were trained in the analysis plan, and they scored the answers independently, not knowing whether the answers they scored were from the pretest or posttest. Then, the three compared answers and agreed on a score between 1-9 for each participant. The t-test was used to check for significance and compare the mean scores of the group from the pretest and the posttest, based on the overall score of each participant, summarising the scores of each test to provide a total individual score. This value was checked for statistical significance using the paired samples t-test (N=21), computed through the R project for Statistical Computing. Scores are arranged on a scale from 1-9, 1 representing the lowest and 9 the highest level of decentring. Scores at a sequential level mean the participant provides answers based mainly on the actions in the story, and not on anticipation or interpretation of what the character might think or feel. Meanwhile, scores at a simultaneous level indicate that the participant has shown the capability to interpret the situation, as well as to think in more abstract and less egocentric ways. Specifically, a score of 1 indicates no decentring, and the pupil is not indicating a cause or reaction to the events in the picture, as he or she is only stating what can be seen ("He stands by the fence"). A score of 2 or 3 indicates that the pupil can see layers in the given situation, mostly in relation to the main character ("He stands by the fence and wants to play"). If the pupil involves the other characters in his/her explanation, seeing that the factors in the story can elicit reactions from other characters, this would award a score of 4 ("He stands by the fence and wants to play, but the boys do not ask him"). The scoring scheme draws a distinct line between 4 and 5, as this is where the pupils are beginning to notice layers to the story that goes beyond what is visible. They decentre into the story, taking the place of the characters to explain what goes on beyond the surface.
The higher levels of decentring (5-9) indicate an ability to internalise the properties of the characters. The character goes from mainly reacting to the actions to predicting the other character's behaviour, something that requires internalisation (Fincher, 2012, p. 12). Answers that anticipate, interpret, and thereby internalise another character's behaviour, are awarded scores between 5 and 7. At a score of 8, the participant is able to internalise more than the subject (Hector) and object participant (Chris), and at 9, the character can differentiate between the characters but maintains a relationship between them at the same time.
On the same day that the interviews were carried out, the interview data were transcribed, assigning the quotes to the correct participants. The answers were then scored by both the main researcher and the same two graduate scorers as on the pre and posttest. In this part as well, neither scorer was aware of whether they scored interviews from the pretest or posttest, to ensure validity.

Ethics, validity and reliability
Before data collection started, the project was reported to the Norwegian Social Science Data Services (NSD). Consent forms were collected from every participating pupil, soundrecordings were stored on a password-protected computer, and all data material was deleted after research ended. The scoring categories in question have previously been used in other both large and small studies (cf. Burkman, 2008;Fincher, 2012), and correspond to cognitive development (Leeper et al., 2007, p. 153). The validity still rests on the researcher's own ability to analyse the data as close to the scoring categories as possible, as well as the researcher's impact on the teaching situation. The participants already knew the main researcher and were familiar with the setting in which the research took place, meaning artificiality was not a general concern in the data collection. To check reliability in coding, two other master's students were approached to score examples of participant answers according to the categories on the adapted scoring system. They evaluated the scoring system and tested it, to see if the scores they gave were similar or if the system could be misunderstood to provide different results. Their scores were almost identical, suggesting that the scoring system would be reliable.

Results
For the pretest, the mean score for the 21 participants was 17.81 (SD = 4.45) out of a total maximum score of 36 per test (4 questions, 9 being the maximum score per question). The posttest yielded a mean score of 24.19 (SD = 6.19). The t-test showed that the difference in mean score between the two tests was statistically significant. The t-test provided a probability value below 0.05, and is thus considered by conventional criteria to be significant, with a confidence interval calculated from the mean at 95% CI. The mean difference in scores between the two tests is 6.38. From the provided statistics, one can therefore say that the decentring scores post intervention, compared with scores prior to the intervention, have increased in a statistically significant manner; t(df=20) = 5.84, p = < 0.01, 95% CI [-9.742, -3.018]. In addition, the effect size computed using Cohen's d equals 1.187 (95% confidence intervals: 0.259-2.114). This is a large effect size, providing sufficient foundation to argue that there is in fact a strong score increase between the pretest and posttest.

Distribution of scores from pre and posttest
There is a clear change from the pretest to posttest scores. Figure 1 shows the distribution of the pretest scores, where the four questions are abbreviated to Q1, Q2, Q3 and Q4. Figure 2 shows the distribution of scores from the pretest. There is a cluster around the border between sequential and simultaneous, indicating that the decentring skills of the participants before the intervention were average. This indicates that there was an existing level of decentring among the participants even before the project started, but also that it had room for improvement. When comparing this to the distribution of scores from the posttest, displayed

Figure 2 Distribution of scores in the participants' pretest answers
in Figure 3, one sees the increase. Very few participants received scores below 5, and the majority of scores are found within the top half of the scoring system. The cluster has moved from the middle of the scale to the higher scores of 5 to 9. One also sees that the highest score of 9 was awarded to a total of 9 participants on the posttest, while no participants reached this level on the pretest. These results would indicate that, in most cases, the overall mean score of the individual participants has increased.

Figure 3
Distribution of scores in the participants' posttest answers

Interview findings
The pretest generally indicated that the group had a lot of room for improvement in terms of their decentring skills. With a small number of exceptions, most participants merely explained what they observed in the picture, therefore receiving an average decentring score. Some participants did include more personal characteristics such as happy and excited, but the majority did not take into consideration the background of the story ( Question 2: What is Hector thinking in this picture? Pretest: "That he sees the blonde boy who plays on the field outside the garden of his mum's work." (P12) Interview: "I think he remembers the boy from outside his mother's yard, from the good field, and thinks that now…now they can maybe become friends and he can be allowed to play." (P12)

Question 4: Why do you think they act the way they do?
Pretest: "Because they are cheering." (P5) Interview: "Because they are at a football game and cheering." (P5) The second round of interviews showed almost as high a level of correspondence as the first, still above half of the group providing corresponding answers (12 or 13 on all questions), whereas the remainder provided partial correspondence, no correspondence, or gave a blank answer. However, fewer gave answers that did not correspond, as low as one participant on question 2. Interestingly, 8 participants gave answers during the interview that showed higher decentring skills than they did on the test. The interviews also showed that more participants took into consideration the context when answering. When explaining events in the picture, fewer pupils described exactly what is depicted, but, rather, elaborated upon the situation. The following statements are examples from the posttest, showing a clear increase in the ability to interpret and explain other factors than those visually available: Question 1: What is happening in this picture?
Pretest: "Hector is sad because he is not allowed to play football with the white boys who play a game on the field beside him." (P16) Interview: "Hector is shut out because before, black people did not have the same rights as the white, and he is shut out because of it." (P16) Question 2: What is Hector thinking in this picture?
Pretest: "I think he feels left out." (P1) Interview: "I think he is trying to join, but no one, uhm, talks to him, or no one responds, and he feels alone." (P1) Question 4: Why do you think they act the way they do?
Posttest: "Just because of apartheid, and because those with white skin had more rights than the black people." (P8) Interview: "Because it was, uhm, apartheid, and they had a lot more rights. But he kind of wanted to, but it was, like, other people around him had yelled and said no to that maybe, and then he did not dare to let that boy in to play with them, because he was scared of getting yelled at and stuff. But if not, he would actually have done it." (P8)

Discussion of findings
The main findings of this study show that there was a significant increase in decentring scores from the pretest, with a mean score of 17.81, to the posttest, which yielded a mean score of 24.19 for all four questions. The probability value was < 0.01, considered by conventional criteria to be highly significant. In addition, the interviews supported these findings, showing a high level of correspondence between pretest and posttest answers compared to the interview answers from the same participant. They even indicated that the participant group possessed even greater decentring skills, had the interviews been scored in the same way as the tests were. The ability to decentre was traceable even before the intervention, and was enhanced in the course of the three weeks.
There is evidence in both the tests and the interviews that show a clear improvement in decentring skills by the majority of the participants. When asked why they think the characters act the way they do in the picture, a large number of the participants managed to see the situation from both Chris's and Hector's points of view, taking the apartheid context into consideration when they explained their actions, especially during the interviews.
An example of this is the participant who explained: "Because maybe, when they grew up, their parents and others told them that the people who had black skin, they are not good people. So therefore they have kind of been taught not to care about them, and just…yeah, ignore them" (P9).
One statement during the plenary discussion exemplified a prime example of decentring. When the class was asked their opinions on the possibility that Chris actually thought black people were worth less, a heated discussion was sparked. However, participant 13 provided an especially interesting answer: "He is just doing it the way his parents…or his society thinks he should. How can I know that I would do it differently from him if it was like that here?" His question touches the core of decentring, as he has immersed himself so thoroughly into the situation of the character that he actually imagines himself in Chris' situation, and tries to transfer the situation into his own reality to understand his actions. If there was initial doubt that young children were not able to suspend their perspective in favour of someone else's, this example shows evidence of the opposite.

Decentring and Byram's savoirs
The ability to decentre is traditionally labelled an aspect of Byram's savoir être (Byram, 2008, p. 69 From this analysis, we see how decentring is located within aspects of each savoir.
The ability to change perspectives seems to require knowledge, the ability to understand and interpret, the ability to see one's own culture from a different perspective as well as the ability to question generalisations and to use the new-found information in interaction, such as the activities conducted during the intervention.

The Piagetian age limit
The significant results of an increase in decentring skills presented above show that decentring is a learnable skill, also in primary school. According to Piaget, decentring requires logic and reasoning, which is mainly developed in the preadolescent period (11-12 years). Some of the most important results of this study are those that challenge this claim.
Not only did the participants from the very beginning show a sequential ability to decentre even before the intervention, but this ability also increased significantly after working systematically with activities that challenged their decentring skills. The majority of the participant group received decentring scores at the simultaneous level post-intervention. This suggests that not only do they possess the ability to decentre below the Piagetian age limit (11-14 years of age), but many were also at one of the highest levels, based on Feffer's scoring scheme (Fincher, 2012, p. 2). Thus, the results stand to question the decentring age limit. A potential reason why there might be a disagreement about the age at which decentring is developed could be that cognitive development is highly individual and difficult to generalise (Burkman, 2008).

Decentring and fiction in the classroom
Based on the results of this study, it can be argued that fictional literature should play a more significant role in the foreign language classroom. Hoff (2017) and Fenner's (2000;2001) research provides support for literature as an intercultural experience in itself. In the present study, the reading of the picture book and the intervention tasks encouraging the participants to explore different perspectives in the narrative are likely to have affected the participants' increased decentring skills.
Working with The Soccer Fence challenged the participants to remove themselves, temporarily, from what they viewed to be obvious truths. The level at which this was achieved varied among the participants, but the ability of the group to change perspectives increased noticeably. Although picture books have previously been used mainly for the development of literacy, this study has provided a clear example of how versatile the picture book can be in the foreign language classroom, especially in regards to decentring (Dolan, 2014, p. 9). The Soccer Fence illustrated excellently the cultural situation that existed in South Africa and helped them understand a world that is changing (Dolan, 2014, p. 17). The topic of apartheid can be difficult to teach at lower levels, but the picture book proved useful to describe the situation from a child's angle, making it more relatable to the participants.

Limitations
It is not possible to generalise the findings of this study to a larger population, simply due to the low sample size (N=21). Also important is the weakness of human error in scoring, e.g. errors in planning, execution, scoring or researcher bias, in addition to the fact that a control group was not used to compare results. Without a control group, one might question how one can be sure that it is the intervention that influences the experimental group and not something else, such as the environment or the procedure of testing. Another interesting aspect is that of ethnicity. Some participants had one parent of Norwegian descent and some had parents who were both from other countries. All the participants had regular contact with other cultures.
This raises a question about whether it is a possibility that the participants were already more decentred than the average 5th grader.
Limitations also exist regarding the method, meaning formulation of the questions and the choice of pictures for the tests. The pictures are from different parts of the story, and display different levels of emotions. Some may evaluate the posttest picture to be more emotionally clear than the pretest picture. However, to ensure that participants had not simply copied their pretest answers, the decision was still made to include two different test pictures.
Regarding the intervention activities, one can discuss whether the chosen tasks were the best possible and whether different activities would have provided even higher results. One can also question whether the participants' English proficiency had increased over the weeks of the study in such a way that it could have affected the participants' ability to decentre.
However, the short time of the study (three weeks) makes this explanation unlikely.
In regards to reliability, one must stress the possibility of bias or over-emphasis on the part of the researcher. However, an important strength is that the questions chosen have no right or wrong answers, thus challenging the participant to express their thoughts. To check reliability in coding, two other master's students were trained and asked to score each answer.
There is, however, always the possibility of misunderstandings or misinterpretations, although the process was handled with great care and precision.

Conclusion
During the three weeks, a literary context had been created in which the participants were to interpret and take positions on situations different from the ones they were used to. The decentring ability was not forced upon them, but a context was introduced that would help them take the characters' positions, and make them reflect on the thoughts, values, and norms of the different characters. The participants questioned the context and perspectives continuously as they tried to understand the mentalities that governed during apartheid. To the majority of the participants, this process bore fruit, and increased their ability to decentre significantly. It was noticeable that most participants developed their ability to interpret a situation based on the context in which it happened and to imagine the thoughts and feelings of fictional characters, even though they might have collided with their own opinions. It was apparent that the participants came with their own predispositions, but to a large degree were able to temporarily show openness to alternative ways of thinking, and reflect on how the society they live in affect their thoughts and values. The main findings present a piece of research that shows how a narrative has the potential to develop intercultural competence in the foreign language classroom, as well as providing a hands-on description of how intercultural competence could be taught.