Dynamics of learner affective development in early FLL

Affective learner factors were first considered as a cause of success in language learning. This was followed by a change in approach and recently authors (e.g., Edelenbos, Johnstone, & Kubanek, 2006) have considered them an important outcome, especially in early foreign language learning (FLL). Current research into affective learner factors in early FLL tries to catch the developmental aspects too, and studies are emerging that take a contextual view as well. This paper describes a study on affective characteristics of young FL learners that combines the developmental and contextual perspectives. Using the case study methodology the author analyses the affective profiles of three young learners of English as a foreign language who were followed for 4 years. The analyses are done taking into account their immediate language learning environment, home support, out-of-school exposure to English and language achievement. The findings suggest that affective learner factors contribute to the dynamic complexity of early FLL.

A closer analysis of how researchers have viewed the role of young learners' affective characteristics reveals several approaches to the issue.Until recently many (e.g., Burstall, 1975) considered affect primarily as a cause of success in early foreign language learning (FLL).Other researchers (e.g., Blondin et al., 1998;Edelenbos, Johnstone, & Kubanek, 2006), however, approached the issue from a different perspective and pointed out that it is precisely the development of learner affective characteristics (mostly positive attitudes and motivation) that is, in fact, the most relevant outcome or result to be looked out for in early FLL.Even more recently two more approaches have become quite salient.According to one, we should look for interactions of affective learner characteristics with other individual learner characteristics (e.g., language anxiety, learning strategies) or with contextual factors.The other focuses on the temporal dimension and investigates developmental aspects of various affective learner factors.

Major Insights into Affective Characteristics of Young FL Learners to Date
The traditional view of young learners as being so similar to one another that individual differences were not a real issue resulted in a scarcity of studies into their affective characteristics.Once this view was replaced with an awareness that young learners do differ among themselves, the need for investigations of learners' affect became obvious.In-depth research that involves young learners, however, requires triangulation of data because children often find it hard to articulate their perceptions and feelings (Lamb, 2004;Mihaljevi Djigunovi , 2009a).Since it has been observed that young learners' affect is not a set of stable factors but changes over time, the longitudinal approach is often applied.But longitudinal studies on young FL learners are still comparatively rare.
Some studies (e.g., Low, Brown, Johnstone, & Pirrie, 1995;Nikolov, 2002;Szpotowicz, Mihaljevi Djigunovi , & Enever, 2009) found that, generally speaking, young learners adopt the attitudes of their significant others (parents, siblings, relatives, friends, teachers).Nikolov (1999), however, stressed that their first experiences in FLL enable young learners to form their own attitudes, which are shaped by the FL classroom processes.Vilke (1993) found that the FL teacher plays an important formative role in attitude development.A closer look at the studies on the development of young learners' attitudes indicates that findings can be quite ambiguous.Some studies (Chambers, 2000;MacIntyre, Baker, Clément, & Donovan, 2002;Nikolov, 1999) found that earlier starters' attitudes are more positive than those of later starters but turn less positive over time.On the other hand, Cenoz (2003) and Mihaljevi Djigunovi (1998) showed that positive attitudes can be maintained over extended periods of time under favourable learning conditions.In some studies (e.g., Lasagabaster, 2003;Williams, Burden, & Lanvers, 2002) no significant age-related differences in young FL learners' attitudes to FLL were found.It is assumed that such contradictory findings should be attributed to different research designs used in the studies listed.Another possible reason may be that the discrapancies are due to the differences between the contexts in which the young learners were learning the FL.Contextual influences, both at the macro and micro levels, are now increasingly recognized as a significant factor in early FLL (Mihaljevi Djigunovi , 2009b).From the temporal perspective, an interesting recent observation of young FL learners' attitudinal development suggests that over the years young learners, especially the less successful ones, tend to become bored or disillusioned if they do not feel they have made expected progress in mastering the FL (Bolster, 2009).Such a turn of events may contribute to the young learner building up a negative self-concept.Driscoll and Frost (1999) suggested starting a different FL at certain points during primary education because a fresh start in language learning may lead to positive attitudes to FLL.The relationship of motivation and language achievement was found not to be consistent in all the studies.Some authors (e.g., Masgoret, Bernaus, & Gardner, 2001;Mihaljevi Djigunovi & Vilke, 2000) observed that it depended on what kind of measure of achievement was used: Stronger correlations were found with self-assessment, course grades and integrative tests than with discrete-point tests.Interestingly, it was also observed that the strong correlations between motivation and outcomes in younger learners decreased with age (Tragant & Muñoz, 2000).Young learners' self-concept was also found to correlate positively with learning outcomes (Mercer, 2011;Mihaljevi Djigunovi & Lopriore, 2011).Nikolov (2001) warned that initial attitudes, motivation and self-confidence of young learners should not be underestimated as they can have a long-lasting effect on future learning.

A Study into Dynamics of Young Learners' Affective Characteristics
In this paper we would like to look into affect in early FLL using an approach that combines the temporal and contextual perspectives discussed above.We believed that such a complex aim could best be achieved using longitudinal case study methodology.Collecting detailed information about specific young learners as individuals over 4 years could provide us with rich, holistic, context-sensitive data that could serve as a good base for researching developmental aspects of affective learner factors.In our opinion such an approach has the potential of providing broader and, also, deeper insights into the complexity of early FLL.

Context of the Study
The qualitative study described below is part of the Early Language Learning in Europe (ELLiE) project 1 (www.ellieresearch.eu),a 3-year transna-tional longitudinal study that was carried out between 2007 and 2010.The study (Enever, 2011) was preceded by a scoping year 2 so that data were collected over a total of 4 years.Our focus here will be on the Croatian cohort.
The Croatian context is characterised by a long tradition in early FLL (Vilke, 2007).Since the first half of the twentieth century the starting age at which the FL was introduced as a compulsory school subject in primary school has been constantly lowered.In 2003 the national education authorities passed a law that requires all learners to start a FL at the very beginning of primary education, in grade one (age: 6-7).Primary education in Croatia includes eight grades and learners are 14 when they leave primary school.Although there is no regulation that stipulates which FL to start with most learners start with English, which is the most popular FL (Medved Krajnovi & Letica Krevelj, 2009).The role of English is enhanced by the requirement that those primary learners who start with a FL other than English must take English in grade four (age: 10) so that no learner ends primary education without having learned English.Following the recommendations of the Council of Europe, the Croatian National Curriculum (2010) allows for a second FL to be taken up at different points in the education cycle (in grades four or six of primary school or in secondary school).
FLL in both the primary and secondary sectors is centrally regulated and, currently, the main guidelines are included in two documents: the Croatian National Curriculum (Croatian Ministry of Science, Education & Sports, 2010) and the Croatian National Educational Standards (Croatian Ministry of Science, Education & Sports, 2005).It is proposed in the guidelines that early FL teaching should be communicatively oriented, holistic and multisensory.It is explicitly stressed that in the early years metalanguage and grammatical explanations should be avoided.The FL learning goals are laid out in terms of the Common European Framework levels (Council of Europe, 2001) aiming at the A1 level by end of grade four.
Except for schools in the areas near the border, most classes in Croatian schools are monolingual.Teachers of English to young learners can obtain their qualification either by getting a university degree in English Language and Literature (teaching stream), or a university degree in Early Education with a Minor in English.
Another characteristic of the Croatian context relevant for this study is a relatively high exposure to English.Foreign programmes in English shown on Croatian national channels have always been subtitled, popular music with lyrics in English can be heard via the electronic media all the time, most computer software that learners normally use is English, and contact with Englishspeaking foreigners (tourists, businessmen) is high too.With tourism as practically the main branch of national economy, Croatian policy makers have al- 2 The scoping year was partly sponsored by the British Council.
ways paid lip service to FLL and have claimed it is on their priority list, though national investments in this field do not substantiate this claim.Generally speaking, attitudes to learning FLs -English in particular -are very positive among most people.

Aim and Methodology
In this qualitative study we wanted to look into interactions of affective learner factors with contextual variables from the developmental point of view.The most suitable way to reach this complex aim was through the longitudinal case study approach.
We followed three young learners of EFL from the Croatian cohort of the ELLiE sample for 4 years, from grade one till end of grade four.They were between 6 and 7 years old at the start and between 9 and 10 at the end of the project.
Concerning the affective aspects we gathered data on the young learners' self-concept as learners of English, interest in FLs, attitudes to English as a school subject, motivational orientation in learning English, perception of difficulty, preferences for classroom activities and attitudes to teaching English.On the contextual side, we collected information on the immediate learning context (the school environment, principals' attitudes to early language learning, characteristics of classroom teaching, teachers' attitudes), on the home environment (parents' education, knowledge of English, using English for professional purposes, attitudes to early FLL, family involvement in the child's learning of English) and on exposure to English out of school.Finally, information on the young learners' achievement in English was also documented over the 4 years.
The instruments used to collect the data included oral interviews, questionnaires and classroom observation schedules that were used each year; vocabulary recall tasks (grades one, two and three), listening comprehension tasks (all grades), reading comprehension tasks (only grade four) and oral tasks (grades two, three, four) and teacher assigned end-of-year grades in English.Data on learner attitudes and motivation were elicited through oral interview questions about the favourite school subject, preferred classroom activities, perception of difficulty of English as a school subject and about why it might be good to learn it.To elicit data on attitudes to teaching English the young learners were asked during the oral interview to look at four pictures: One presented a traditional classroom with the teacher pointing to some writing on the board and pairs of learners sitting in three rows of desks and all facing the teacher and the blackboard; one presenting a classroom where group work was going on; one where both the teacher and her young learners were sitting on the classroom carpet in a circle and doing an activity with flash-cards; one with young learners jumping around and the teacher just standing at a loss.The learners were asked to choose the classroom in which they thought they would learn English best and to explain why.In order to provide information on their self-concept the learners were asked in the oral interview to compare themselves to their classmates and say whether they thought they were the same, better or worse at English than their peers, and to explain how they could tell.Data on the immediate learning environment were elicited from the school principals and English teachers by means of questionnaires and oral interviews.Home support and out-of-school exposure to English were described on the basis of information collected from the young learners (oral interviews) and their parents (questionnaires).Parents' questionnaires also provided information on the parents' education levels, background in learning and using English and the learners' expressions of attitudes about learning and using English out of school.Measures of language achievement were endof-year English grades, and vocabulary, listening and reading comprehension as well as oral production tasks used in different grades as specified above.Using both end-of-year grades and specifically designed language tasks made it possible to use both discrete-point and integrative measures of language achievement.Triangulation was secured by collecting information about the same aspect from different participants (e.g., the child and the parents) and using different instruments (e.g., oral learner interview and classroom observation).All the instruments used were designed jointly by the ELLiE team.More information about the instruments can be found in Enever (2011).

Results
In this section we will provide descriptions of each young learner based on the data gathered.Each description will comprise (a) the young learner's affective profile as it developed throughout the four years, (b) summary of the characteristics of the immediate learning environment each learner was exposed to, (c) summary of the characteristics of the home environment, (d) summary of the learner's out-of-class exposure to English, and (e) summary of the learner's language achievement.
Case study 1: Ana Affective profile.Ana's language self-concept was consistently low during the 4 years.Every year she said that, compared to others in class, she performed badly.She explained her claim by referring to the low end-of-year grades she kept getting.Ana believed it was good to learn English because it would be very useful if she went to England one day.She wanted to learn German as well because her grandmother lived in Germany.In grade one, English was among her favourite school subjects and she particularly liked colouring and drawing; in grade two it was just an 'ok' subject and she reported liking reading, playing and learning new songs; in grade three she liked some activities such as singing, but disliked tests because she found them difficult; in grade four she particularly enjoyed acting and, through acting, learning new words but still disliked tests because they were simply too difficult for her.From grade two on Ana felt English was always more difficult than the previous year.She consistently preferred the traditional classroom arrangement explaining that it involved seeing things written up on the board and hearing the teacher well.While in grades one and two she thought her parents were happy about her learning English, from grade three on she felt her parents were not happy anymore because of her low grades in English.
During classroom observation it was noticed that, in spite of high interest, Ana's engagement and performance ranged from average to low.She would consistently split her attention during classes between the teacher and other learners, was very active and loud in whole-class activities, but was often restless and got distracted easily.On many occasions she would raise her hand and quickly put it down after she heard the full question.
Ana's parents reported that she was not keen on learning English, did not enjoy speaking English at home, felt insecure about her knowledge and found learning English rather hard.

Contextual factors
Immediate learning environment.Ana attended a village school with about 150 pupils.The school offered English and German as FLs.Pupils did not have a chance to use English with native speakers but they could use it with the Dutch (the school was involved in a number of international projects with the Netherlands) and Danish people (visits connected with international help to war-stricken parts of the country).Both the school principal and teaching staff had highly positive attitudes to starting with English in grade one but were not as enthusiastic about starting a second FL during the primary years: They believed it was too much for children.Most parents were happy and proud about their children learning English.The school was moderately equipped with digital media (CD player, video, computers) but only the CD player was used in English classes.In the school library there were only two English dictionaries and no books for children in English.
Ana had the same teacher of English during the 4 years.Her teacher had a college degree in English and was qualified to teach English in primary school.Her competence in English was at the C1 level.She had visited England and the US several times.She believed in early FLL and thought the age of 5 was the optimal age to start.In her opinion, FLL was easy and natural for children, especially with regards to pronunciation.Her knowledge about her learners' lives outside school was rather poor.Her teaching was dynamic and included a lot of games with flashcards, role-plays and songs.She said that her learners liked most playing games and writing tests, while reading and doing textbook work were the least popular activities.Her classroom management skills were well developed and there were no major behavioural problems in class.On occasions she would be a little impatient with her learners.She used English about 75% of class time and resorted to L1 when she thought her learners could not understand what she was saying.Children's seating arrangement alternated between the U-shaped format and the desks in three rows with the learners sitting in pairs and all of them facing the blackboard.
Home environment.At home Ana had one English dictionary but did not use it.She would tell her family about what she had learned in her English classes and would sometimes ask for help with homework.In grade one her mother would help.Since her parents never leared English at school, in later grades it was her brother who could help her with homework more than the parents.
Ana's parents had primary school education only.Her mother was a housewife and her father a factory worker.The father did not need English at work.They believed that it was useful to learn English, that it contributed to a child's openness to other cultures and that in the future knowing English would be as important as knowing L1.However, they thought that early FLL was not really fun for children and that it could cause difficulties in learning other school subjects.
Out-of-class exposure to English.According to her parents, Ana spent about 2 hr per week watching programmes in English and about 1 hr listening to music in English.She did not have access to the Internet at home.During summer holidays she had opportunities to use English with foreigners but never used them.
Language achievement.Ana's final grade in English was a 3 3 in grade one and from grade two it was a 2. Her vocabulary knowledge was average in grades one and two and turned below average in grade three.Still, although her lexical diversity was low, her oral production showed consistent progress.Listening comprehension varied over the years: It was above average in grades one and three but below average in grades two and four.Ana's reading comprehension scores (grade four) were among the lowest third in the Croatian cohort.
Case study 2: Ivan Affective profile.Throughout the 4 years Ivan thought very highly of himself as a language learner.He attributed this to the extra private classes he attended.He also pointed out that he could recognise many English words and stressed that his parents often praised him.In Ivan's opinion, learning English was a good thing because one could talk to many people.Ivan wanted to learn German and French as well.English was a favourite subject only in grade one.By grade four German became his favourite subject because, he said, he was good at it; but he pointed out that he also enjoyed speaking English.His favourite classroom activities changed from learning new words (grade one) to reading, playing and arguing (!) in English (grade two) to reading (grade three) and, finally, to speaking (grade four).He disliked his classmates fighting in class and, in grade four, he did not like singing because of 'silly' words such as 'twinkle, twinkle little star.'His classroom teaching preferences varied: In grade one he claimed he liked the chaotic classroom best, in grade three he preferred the sitting on the floor format, while in grades two and four he opted for the traditional classroom arrangement.He felt his parents were satisfied with his knowledge of English and he himself felt proud when they would ask him about some words and he would know them.
Ivan looked very interested in all classroom activities and his engagement was generally very high.He often took the role of a leader during wholeclass and group-work activities.When he found an activity too easy, he would quickly get bored, start entertaining his classmates and would become quite disruptive.He seemed very aware of his language abilities and often tried to show off.Sometimes he would misbehave to attract the teacher's attention.
According to parents' reports Ivan was enthusiastic about learning English, proud of his knowledge, liked speaking English, was not embarrassed to use it and did not perceive learning English as hard.

Contextual factors
Immediate learning environment.Ivan attended an elite city school that offered, besides the regular Croatian national curriculum, international programmes in English for foreign children, as well as the Croatian national curriculum in English.This allowed Croatian children opportunities to communicate in English with foreign pupils in the school.The school was involved in many national and international projects, and promoted multicultural communication, holistic learning, critical thinking, life skills, and development of self-confidence.The FLs on offer were English, French and German.By Croatian standards it was extremely well equipped: The teachers had at their disposal video players, CD players, interactive whiteboards, and each classroom had at least one computer with English software.Ivan's class could enjoy an English corner display area, a self-access area with English readers and games, and a class library with books in English.The school library contained authentic books for children and children's dictionaries.
During the 4 years Ivan had classes with three teachers of English.His first grade teacher had a university degree in Early Education with a Minor in English.Her communicative language competence was at the B2 level.She believed that even an earlier age may be the optimal time to start learning FLs.She saw the main benefit of the early start in easier acquisition of pronunciation and development of writing skills.She insisted on oracy but introduced reading and writing quite early using age-appropriate teaching strategies.She knew her learners well and could point to individual differences among them.She was aware of what they were exposed to outside school and used the English they acquired elsewhere in classes.Her use of English amounted to 50% of class time: She often code-switched in the middle of the sentence, strategically using L1 to make her teaching point more salient.The second grade teacher had the same qualification and the same level of competence in English as the first one.She considered starting in kindergarten as optimal, believing that an early start helped young learners to avoid language anxiety and to prepare well for later language learning.Her classes included a wide range of activities because she knew children could not focus on one thing for a long time.Her use of English varied between 25 to 50% of class time.The third teacher taught Ivan's class in grades three and four.She had a university degree in English Language and Literature and was not trained specifically to teach young learners.However, she believed that early learning was a good idea because it was easier to fight language anxiety later.Her competence in English was at the C1 level.In her teaching she put special emphasis on listening comprehension because she found that this was a problem for her learners.Her use of L1 was excessive: She kept translating practically everything she said into L1.She stressed that in grade four individual differences in motivation and self-confidence became prominent.
The learners were occasionally seated in groups of four with two desks put together to form a bigger working surface.However, most of the time the children sat in pairs with desks arranged in rows and all the learners facing the blackboard.
Home environment.Ivan had picture books in English at home, and often watched cartoons, children's programmes and movies with his parents.He also had access to the Internet, played games and, by grade four, participated in group networks.
Ivan claimed he did not ask anyone for help with English because he did not need any.However, his mother reported that the family often helped by watching movies with him and checking homework.They sometimes sang and read books together, too.
Both parents had learned English at school, obtained university education and often used English at work.They believed early learning of English was useful, fun, not hard at all, had a positive effect on later learning and did not stand in the way of learning other school subjects.
Out-of-class exposure to English.According to his parents, Ivan spent on average 3 hr per week watching films and other programmes in English, 2 hr playing video or computer games, 2 hr listening to music in English on the radio, 1 hr reading and 1 hr speaking in English.He met English-speaking people fairly often and communicated with them: Besides speaking English to foreign children attending his school, during the summer he would also meet children from other countries, family friends and relatives who lived abroad, and had fun using English.

Language achievement.
Ivan obtained the top end-of-year grade in English in all the 4 years.He performed above average on listening and reading comprehension tasks.His listening skills showed constant progress from year to year.His vocabulary size increased as well and by grade three it was the largest in class.His oral production was characterised by high lexical diversity from the start, and it developed considerably over the years.

Case study 3: Vesna
Affective profile.In grade one Vesna thought she was better at English than her classmates.She changed her mind in grade two and for the rest of the lower primary years she thought she learned at the same rate as others.She believed it was good to learn English because she could use it to speak to many people.She wished she could also learn French and Italian.In grade one all school subjects were her favourite, in grade two English was one of her four favourites, in grade three one of three, while in grade four it was her single favourite subject.Vesna liked everything about her English classes but her favourite classroom activities were playing (grade one), singing (grade two), listening to stories (grade three) and learning new words (grade four).English was not getting difficult for her over the years, she claimed.But in grade three she said some things were easy and some were not that easy.She consistently preferred the traditional classroom arrangement explaining her choice by saying there were things written up on the board; in grade two she said she liked it because both English and Croatian words could be put on the board, while in grade four she pointed out that she preferred it because the teacher could show new words on the board.Vesna knew her parents were happy about her progress in English because she would ask them about it and sometimes she would read and write something for them to show what she could do in English.
Vesna's parents claimed she was very keen to use English, enjoyed speaking English a lot and felt proud of her language knowledge.She did not find it difficult and experienced no language anxiety.

Contextual factors
Immediate learning environment.Vesna attended a small town school with around 800 pupils.It offered English and German as FLs.The school administration advocated early FLL and had offered it long before it became mandatory in 2003, with the local community covering the costs.The attitudes to English were very positive but German was highly popular too since many families had someone who had lived as a guest worker in German-speaking countries.The school principal was a very ambitious person and had won support of local authorities for many extra school activities.The school was very modern and well equipped, as well as generally well staffed.Each classroom had a computer with FL learning software and a video and a CD player, there was an interactive whiteboard in the IT classroom, while the school library contained authentic English books for children.The school was visited by pupils from Canada several times, so the young learners in the school had opportunities to interact with native speakers.
The EFL teacher had a university degree in Early Education with a Minor in English.She was very active, took part in many professional events and was a regional INSETT leader.Her communicative language competence in English was at the B2 level.She used the communicative approach and her classes were highly interactive.She often based her classes on playing, singing and story-telling.Classroom observation showed she used English 50% of the time, switching to Croatian when she thought it was more efficient to use L1.She believed that grade one was the optimal starting age because at that age FLL came more naturally, was less stressful and made acquisition of pronunciation easy.An earlier start might be good for some children, she claimed, but not for all.The teacher enjoyed teaching young learners, thought it was a highly appealing job, but felt it was difficult to keep her English at a high level without teaching more advanced learners, too.Her classroom management skills were very good.In grades one and two the children sat at desks arranged in a circle.From grade three on the desks were arranged in three rows with the children sitting in pairs and all facing the blackboard.
Home environment.At home Vesna had access to picture books in English and read them frequently.She often listened to music in English on the radio and played computer games.She never asked any member of the family for help with English.
Both of Vesna's parents had secondary education.They had not learned English at school and did not need the knowledge of English at work.Their attitudes to early FLL were very positive: They believed it was very useful and fun, it helped children to be more open and to entertain positive attitudes to other cultures.They thought it was not hard to learn a FL at an early age and it did not interfere with learning other school subjects.
Out-of-school exposure.On average, Vesna spent up to 2 hr per week watching TV programmes in English, about 1 hr playing computer games and the same amounts of time listening to music in English, reading and speaking in English.Although she would never ask her family for help, she would regulary inform them about what she had learned in her English classes.She would have opportunities to speak English with foreign tourists during summer holidays and when meeting some relatives and friends who lived abroad.She reported enjoying speaking English.
Language achievement.Vesna had top grades during all 4 years.Her command of vocabulary increased steadily from average to above average to second highest in class.Her oral production was characterised by average lexical diversity.In the first two years her listening comprehension was below average but increased in the last two years.Her scores on reading comprehension tasks were, however, among the lowest in class.

Discussion
Our results paint a different picture about young learners' affective development from the one that quantitative studies have painted (e.g., Mihaljevi Djigunovi & Lopriore, 2011).The wealth of data about the three young learners described above offers a very dynamic and complex view of this development.It confirms conclusions of some other case study research (e.g., Cekaite, 2007) that individual learner's FLL is not a linear process, and suggests that the development of the affective aspects of early FLL is cyclical too.
Contrary to popular belief as well as to results of many quantitative studies, affective profiles of our three young learners show that not all children necessarily start FLL with a positive outlook nor that such an outlook is maintained.In our study two (Vesna and Ivan) out of three learners began learning English as "learning optimists" (Pinter, 2011), assessing their language learning abilities high, but only one (Ivan) maintained such a positive selfconcept throughout the whole lower primary education.It is very likely that this was the result of a very strong home support and a high exposure to English outside school.This corroborates findings of Muñoz and Lindgren (2011) about the role of home support and language exposure.Ana, on the other hand, began her FLL by seeing herself as less able than her peers.She based her self-assessment of language learning ability on her grades.Ana started with a grade 3, which is considered very low for the first grade (her class grade point average in English being 4.7), and continued getting a grade 2 for the rest of the lower primary years.A comparison with her classmates' English grades apparently suggested to her they were doing better than her.Each year she perceived English as more difficult than the previous one.Although she was moderately exposed to English through watching TV programmes and listening to the radio, she had no experience with using English in real life communicative situations, which possibly could have boosted her linguistic self-confidence, as suggested by Marschollek (2002).Her vocabulary acquisition, listening and reading comprehension were characterised by ups and downs over the years and, though consistent progress of oral production was recorded, her speaking skill was lower than her peers'.Taking into account that Ana's parents did not speak English, Ana's overall learning achievement seems to confirm the findings of Muñoz and Lindgren's (2011) study that stress the impact of parents' knowledge and professional use of the FL on young learners' achievement.The lower English grades, coupled with her parents' growing dissatisfaction with her performance in English, probably enhanced her negative self-concept.In contrast, Ivan developed a very positive self-concept from the very start and maintained it throughout the 4 years.The evidence he relied on were not only his top grades and parents' acknowledgement of his high performance, but also the facts that he had extra, private tuition in English and ample opportunities to use English in real life communication with English-speaking people, all of which he reported as successful experiences.Vesna, on the other hand, started with a high self-rating of her language learning ability but soon perceived herself to be not really better than others but just at the same level as her classmates.Her exposure to English out of school was higher than Ana's and she had positive experiences with using English in real life situations.Vesna's teacher-assigned grades were high during all 4 years and her parents praised her consistently.However, her perfo rm anc e o n E LLiE tasks did no t alw ays reflec t her teac her-assi gned to p grades.From her reports on the preferred classroom arrangement it is clear that Vesna valued teacher written prompts on the blackboard a lot and had not yet developed much learner autonomy.In contrast to Ana, neither Vesna nor Ivan perceived English as getting difficult in each subsequent grade.
All three learners seemed very motivated to learn English and wished they could learn other FLs too.Although their motivational orientation was of the communicative type, which was found to be connected with high achievement among Croatian learners of English (Mihaljevi Djigunovi , 1998), in Ana's case motivation was not positively related to achievement, probably due to the interfering effect of her negative self-concept.Preferences for classroom activities showed different developments among the three learners, too.Ana and Vesna seemed to prefer playful activities (playing, singing) during the first three years and mentioned learning new words as favourute activities only in the last year.Ivan, however, preferred learning new words from the start, and reported liking typical FLL activities, such as reading and speaking, from grade two on.It seems that he developed a high meta-learning awareness very early.As far as attitudes to teaching English are concerned, Ana and Vesna seemed to like traditional teacher-controlled classrooms, which implied peace and quiet, hearing the teacher well and seeing the language material written on the board.Ivan, who had more experience with different teaching formats, changed his preferences every year but the traditional teaching prevailed as his choice.Although all three learners experienced different types of classroom arrangements during the 4 years, the general prevalence of the traditional format suggests that at a young age learners need structure and clear guidance in their FLL.This is in line with the findings of the quantitative study on young learners in the ELLiE project (Mihaljevi Djigunovi & Lopriore, 2011).
As stressed elsewhere (e.g., Pinter, 2011), discrepancies are often observed between what young learners claim they think and feel and what they do in the classroom.Therefore, observation of the learner's language learning behaviour is very valuable as a source of information, although it can be rather ambiguous, too.Thus, Ana seemed very interested in all classroom activities but it was not reflected in her engagement in class, which sometimes turned to be quite low.Although her teacher described her as an average language ability pupil, already in grade one she perceived English difficult and even found it hard to fully understand the questions the teacher asked.She was also having difficulties with focusing attention on learning: She seemed to be trying to catch the teacher's and peers' attention and secure her place in the group.Such efforts on her part proved to be overwhelming and caused a lot of distraction.Ivan seemed to manage his classroom behaviour better.He very early took the leading role in the class, both in terms of learning activities and socially.Being more advanc ed in English than mo st o f his peers, the classroom ac tivities o ften seemed not challenging enough for him.This was especially evident in grades three and four, when his teacher kept translating all she said into Croatian.As none of his three teachers adjusted their teaching to his cognitive and affective needs, he frequently behaved disruptively.Vesna, who showed the same high level of interest as Ana and Ivan in everything that was going on in class, engaged very actively in all classroom activities.Although at times she behaved in a somewhat shy way, probably thanks to her positive self-concept, this did not prevent her from being continually participative and active.

Conclusions
In the qualitative research described above we collected data through 4year longitudinal tracking of affective development of three young learners of English as a FL.Building our approach on insights from existing quantitative studies as well as the rare qualitative studies on young FL learners, we looked into young learners' attitudes, motivation and self-concept, contextual factors and language achievement.Data collected about the three learners were analysed with the aim to look into how affective learner characteristics developed over time as well as how they interacted with contextual variables and language achievement.This required a mixed-method approach and triangulation of data.
Our findings suggest that the development of affective learner variables in early FLL is both dynamic and complex.We found that attitudes and motivation, as well as learner self-concept, can be unstable and still developing at this early age.At the same time they enter into complex interactions with each other, and with contextual factors.Our case studies reveal that, consequently, the relationship of affective learner characteristics with language achievement is not linear.Thus, for example, in our study interest was not necessarily reflected in the young learner's engagement during FL classes, and was not always related to language achievement, either.One thing that emerged from the three case studies is that learner self-concept seems to be a good predictor of motivated behaviour, just as Henry (2009) pointed out.It seems that a host of variables, both individual and contextual, interact between and among themselves resulting in very complex relationships.These relationships change over time as any of the elements develops.The affective learner development during the first 4 years seems to be making a considerable contribution to the dynamism and complexity of early FLL.We could say that early FLL is a dynamic and complex phenomenon that is hard to fully grasp without including everything that can potentially have an impact.This in itself is next to impossible at the present level of knowledge.It seems to us that the qualitative research paradigm has a lot to offer and will one day, when an increasing number of case studies have accumulated enough insights, get us closer to what now seems next to impossible.

Implications for Further Study
Like any study about FLL, this one has its limitations too.Three case studies are, of course, not enough to make any sort of generalisations about affective development of young FL learners.Thus, in future studies it would be useful to increase the number of cases included.As contextual factors may be crucial to understanding early FLL, it would be enlightening to study learners from different macro and micro contexts.Extending the number of affective learner factors in future research may also contribute to better understanding of affect development in early FLL.New insights may be obtained also if different methodological approaches are used.Following up the young learners till at least the end of the 8-year primary education may also prove to be very informative.