Young learners’ language attitudes in a multilingual school

A renewed interest has arisen on the importance of language attitudes in the acquisition of an additional language. Variables, such as age, gender, mother tongue, and the sociolinguistic context, have been taken into consideration when exploring language attitudes. However, the research on young learners’ attitudes towards languages in English-based schools is scarce. The present paper focuses on young learners’ language attitudes in a high English immersion context in the province of Castello de la Plana, in the Valencian Community, where two official languages coexist, Spanish as the majority language, and Valencian as the minority language. However, English is the leading language of instruction in this international school context. Many studies have been conducted within the framework of language attitudes across ages. However, as far as we know, no studies have accounted for young learners’ attitudes in an English immersion context, such as international schools. Therefore, the present study aims at filling in this research gap by analysing children’s language attitudes in an English immersion school context. The aims of this paper are twofold: in the first place, to analyse the language attitudes towards the three languages under investigation (Spanish, English and Valencian) of multilingual children attending an English immersion school. And, in the second place, to investigate the effect of age on language attitudes. The findings derived from the present study confirm the impact of age on young learners’ language attitudes. However, the school context does not present a determining role in young learners’ attitudes.


II. Introduction
When we talk about the acquisition of a second or an additional language, it is unarguably necessary to make reference to the factors that contribute to the process. One of the aspects that have always caught my attention is the way a language is acquired when we are not even able to stand on our own feet. In this process not only cognitive factors but also psychological factors that may be biased by outer and inner agents play a role. As noted by Portolés (2014, 2) «the affective side of languages is also paramount in order to better understand early language acquisition processes». One of these psychological factors that really caught my attention was the importance of attitudes towards languages and how the use of a first | Índex Fòrum de Recerca. Núm. 23/2018, p. 529-543 ISSN: 1139-5486. DOI: http://dx.doi.org/10.6035/ForumRecerca.2018 531 language may affect the way we acquire or learn an additional language. As part of human nature, somebody's acts, feelings and beliefs towards a language shape, both favourably and unfavourably, the way we see and, therefore, use and label a given language.
Many studies have been conducted within the framework of language attitudes across ages -children, adolescents, and adults. However, most of the research has analysed how native speakers of a language perceive, use and label their L1 or L2 within their speech community (Baker 1992). As far as we are aware, no studies have accounted for young learners' attitudes in an English immersion context, such as international schools. Therefore, the present study aims at filling in this research gap by analysing children's language attitudes in an English immersion school context. The aims of this paper are twofold: in the first place, to analyse the language attitudes towards the three languages under investigation (Spanish, English and Valencian) of multilingual children attending an English immersion school; and, in the second place, to investigate the effects of age on language attitudes.
We will first provide an explanation of language attitudes in a wider context, taking into account their definition as well as former studies conducted in relation to the topic of investigation. Then, the research questions will be posed together with the hypotheses. Afterwards, the setting, sample, instrument and procedure and data analysis will be carefully explained. Once the method has been presented, we will proceed with the analysis and discussion of the results obtained in the present study. The study will finish with a reflection on the results as a conclusion and the limitations that we found in the study as well as further research that can be done on the topic.

III. Aims and research questions
As far as we know, no studies have focused on multilingual children's language attitudes attending an English immersion setting. Hence, we aim at (1) analysing the language attitudes towards the three languages under investigation (Spanish, English and Valencian) of multilingual children attending an English-based school, (2) investigating the effect of the age on language attitudes. Taking into account the theoretical background and our aims, we have formulated the following research questions: First research question (henceforth RQ1): What are the attitudes of very young learners towards different languages?
Second research question (henceforth RQ2): What is the effect of the age on their language attitudes? 532 Considering the above research questions posed above, we have formulated the following hypotheses: Hypothesis I: Students in English-based schools will display the most positive attitudes towards English, followed by Spanish and finally, Valencian.
Hypothesis II: Older students will show more negative attitudes towards the minority and the foreign language than younger learners.

Setting
The study was carried out at an international school. The school is placed in the city of Vila-real, located in the province of Castelló, in the Valencian Community. This centre is a private educational centre which covers academic years at all levels; from nursery, infant, and primary, to secondary education and sixth form. The school offers students the opportunity to study the English National Curriculum. Additionally, the school offers updated facilities that benefit students learning and provide them with a high-quality educational experience.
In order to understand the setting as a high English immersion context, several important aspects as regards the school context have been deeply detailed hereunder according to Cenoz's (2009) continua of multilingual education, which are organised as follows: Concerning the language of instruction used among the teachers, English is the main language used for communication in the school. As aforementioned, almost all content subjects are taught in | Índex Fòrum de Recerca. Núm. 23/2018, p. 529-543 ISSN: 1139-5486. DOI: http://dx.doi.org/10.6035/ForumRecerca.2018 533 English, except for those regarding language courses, which are taught in the target language. Table 3 shows the percentages of languages instruction in the school. As shown in Table 2, most content subjects are taught in English, while Spanish is only used in primary and sixth form to teach content subjects such as social science or business. Additionally, Valencian is introduced in primary and taught through Key Stage 2 and 3 (i.e. primary and secondary). Also, a third language is introduced in secondary education. However, the rest of the subjects are taught in English such as Maths, Geography, History, Arts and Crafts or PE. Hence, all teachers in the school know English as well as personal staff. However, not all native speakers of English can speak Spanish, for instance, there is a number of teachers who do not understand Spanish language. In addition to teachers and staff, students are all multilingual. For example, there are Chinese, African, and English students who have to deal with more than three languages everyday, as well as those students who are Spanish or Valencian native speakers. In fact, those children who use Spanish in the classroom are always told to use only English both to speak with the teacher and with the classmates.
When referring to the language of instruction, there are three main languages -English, Spanish, and Valencian. On the one hand, if we focus on the classroom context, English is the only language used with the exception of the language used in language subjects, such as Valencian, Spanish, French, German, or Chinese, where the target language is the language used. As regards teachers' meetings, English is also the language of communication. However, there are some Spanish native speakers who use Spanish to express personal opinions or talk about important circumstances concerning students, as it is the case of the teacher of Valencian in primary education. On the other hand, the language used to address parents is usually Spanish since most of them do not speak nor understand English. Still, parents are able to access a website where they can be updated on their children's exams, homework, or field trips. In this way, parents are able to choose between English or Spanish as the school website offers both versions. However, if parents want to arrange a | Índex Andrea Rubert Albiol. Young learners' language attitudes in a multilingual school 534 tutorial with an English native speaker and this particular professor is not able to carry out a conversation in Spanish, there is always another professor who interprets the tutorial. Despite the English language being the majority language in formal contexts, the school offers several extracurricular activities which are not necessarily taught in English such as robotics or ICT.
Finally, we observed the sociolinguistic context, which comprises the way languages are used in terms of majority and minority languages or heritage. As mentioned, Valencian is hardly used in the school aside from native children who talk in Valencian during break time or Valencian teachers who may refer to other Spanish teachers using the language in hand. Nonetheless, the language is not used if the interlocutor is not sure whether the other person involved in the conversation is a Valencian native speaker. Hence, Spanish would be the language chosen for communication in these situations. In primary education, as previously stated, students use the English language to communicate with each other and with the teacher. Nonetheless, students in secondary education hardly use English to talk with their classmates but only to address the professor. Additionally, children would mainly use Spanish with each other outside the classroom. As the vast majority of professors are English native speakers, English would be the language used for communication with each other. Therefore, the English language is the majority language at the school followed by Spanish, whereas Valencian is the minority language. Foreign languages (i.e. Chinese, German, French), however, are only present within the language courses as included in the school's curriculum.
On the whole, the British School in which the present study was conducted should be considered a multilingual school and, thus, it holds a high English immersion context.

Participants
All the participants in this study (N=70) were primary students from a single private school in an industrial town in the Valencian Community. In order to carry out the study, two classes from the first grade (mean age: 5.5) as well as two classes from the fifth grade of primary education (mean age: 8.5) have been examined for the study. Nonetheless, they all share a similar socio-economic status (high-class). As Table 3 presents, more than a half of the students come from fifth grade (n=42; 60%) whereas the rest of the participants are currently enrolled in the first grade of primary education (n=28; 40%). Moreover, 57% of the students in Year 1 are female and 43% are male, while the majority of participants from Year 5 are male students (55%) and the rest are female (45%). The respondents in this study have different languages as a mother tongue, although the majority of them speak Spanish or Valencian as their first language. In order to better illustrate these data, Figure 1 displays the different L1 of students. Figure 1 highlights a great difference between those students who noted to have Spanish as their mother tongue (n=28; 39%) and those who stated to use Valencian as their first language (n=11; 16%). However, 17% (n=12) of the students considered both Valencian and Spanish as their mother tongues. When students were asked the reason why they considered to have both languages as a first language (L1) they all remarked that they use one language with one parent and the other language with the other parent. As regards the mother tongue of the other 45% of the students, almost 13% (n=9) spoke Chinese as a first language, either together with Spanish or not.

Instrument
The research instrument used to gather the data was based on Portolés (2014) and adapted for our purposes. Therefore, in order to analyse children's language attitudes towards the three languages (Spanish, English and Valencian), we employed a matched-guise technique. First of all, participants were asked to write personal information such as their names, age and first language. Considering the age of students, we made use of red and green stickers to report | Índex Andrea Rubert Albiol. Young learners' language attitudes in a multilingual school 536 young learners' language attitudes. As stated in Portolés (2014), these colours hold positive and negative connotations as they are directly linked with the colours of traffic lights. However, we omitted the use of the yellow colour, that was used by Portolés (2014) to portray neutral choices and, therefore, only positive and negative language attitudes will be analysed within our study. The questionnaire included three pictures of the two main characters of the story with a blue, a pink and a green background. These three colours respectively made reference to the language used to tell the scripts of the story; the blue picture made reference to Spanish, the pink picture to English, and the green picture to Valencian (see Appendix A).
The script was adapted from a current Disney film that became very popular among the younger population in the recent years -Frozen. The script (see Appendix B) was adapted and translated into the minority language (Valencian), the majority language (Spanish) and the foreign language (English). The translation was carefully done in order to have the final result as similar as possible in the three languages.
According to the criteria of the matched-guise technique, only one person should be representing the different speeches to avoid external variables affecting the participants' responses. In this line, the script was read aloud by the main researcher of the present study, native speaker of Valencian and Spanish, and proficient in English. The tone, intonation, and emphasis were thoughtfully taking into consideration.

Procedure and data analysis
The data collection procedure took place in an English-based school set in the province of Castelló de la Plana, where this study was conducted. First of all, the deputy director of the school was informed about the purposes of the study. After that, I explained the aims of the thesis and the time that would be needed to conduct the activity to the different teachers. Once I received their consent and we agreed on a date for the implementation of the task, we started with the data collection.
To start with, we printed out two questionnaires in each sheet of paper and then we cut it out in a half. As a result, a hundred A3-sizequestionnaires were printed, from which seventy-one were used. Additionally, the three pictures used in the matched-guise questionnaire were printed in A4 format. As aforementioned, questionnaires were distributed to each of the students together with three green stickers (positive attitudes) and three red stickers (negative attitudes) that they had to put next to the three different pictures. These questionnaires were identical for Year 1 and Year 5 students.
| Índex Before starting with the activity, students were provided with the explanation of the task and the instructions to follow. Moreover, they were explicitly told not to copy their classmates' answers, while they were supervised by their tutor and the main researcher during the whole activity to make sure that their answers were individual choices. Once the activity was explained and understood, students proceeded to write their names, ages and mother tongue. Then, the researcher showed the A4 picture that corresponded to the audio track that was about to be played. In this way, students could follow the activity and know where they had to put the sticker. Finally, a pen-drive with three different audio tracks was used to play the recordings in the class computer. The implementation of the activity was conducted in the classrooms and it lasted a total of four days; the first two days we collected the data in both of the classes in Year 5 and the last two days in those of Year 1. The length of the activity varied depending on the class, with a higher demand of time when dealing with the younger students. Hence, around thirty minutes were needed to finish the activity in Year 1 and fifteen minutes in Year 5 (see Table  4).

Academic year Duration
Year 5 A 15 Year 5 B 15 Year 1 A 30 Year 1 B 30 Total 1h30' The answers collected were codified in an Excel page in order to consider the different variables for analysis. In the process of codifying the data, we found out that one student used half of a green sticker and half of a red sticker to illustrate a «neutral» answer. Despite the fact that they were told not to use both, this student did not consent to use either a red or a green sticker to determine his attitudes towards Spanish. Therefore, this data had to be removed from the data analysed. This is the reason why we only considered seventy questionnaires instead of seventy-one that were distributed in the very beginning. This being solved and all the valid data codified, percentages were employed in order to perform the qualitative analysis, which is presented and discussion in the next section.

| Índex
Andrea Rubert Albiol. Young learners' language attitudes in a multilingual school 538

V. Results and discussion
Considering the research questions presented in the method section, we will start by analysing the results in order to answer RQ1 in the first place. To start with, global results will be analysed taking into account both academic years as a whole. In this vein, language attitudes in an English-based school will be determined. Then, in the RQ2, we will continue by considering each academic year individually in order to examine to what extent the variable of the age may affect the way children perceive different languages.

Results and discussion related to hypothesis I
With the main concern of analysing the language attitudes of school children in an English-based school context, we have posited Hypothesis I, which claims that pre-schoolers and primary students would display more favourable attitudes towards Spanish, followed by English and finally, Valencian. In line with this assumption, we examined the global percentage of positive language attitudes towards the three languages under investigation. As it can be observed in Figure 2, these findings suggest that the participants' overall attitudes towards languages are highly positive as 71.9% of the responses reported favourable attitudes. The overall results of positive language attitudes do not seem to provide a significant difference between Spanish (n=54; 35.8%) and Valencian (n=58; 38.4%). In fact, the Figure shows a preference towards Valencian rather than towards Spanish. With reference to English, more unfavourable attitudes are portrayed in the findings as only 25.8% (n=39) of the students noted to have favourable attitudes. Contrary to our previous expectations, the minority language reported the most positive attitudes, followed by Spanish, and finally English. Our study contradicts Portolés' (2014, 144) findings in which participants appeared to have higher favourable | Índex Fòrum de Recerca. Núm. 23/2018, p. 529-543 ISSN: 1139-5486. DOI: http://dx.doi.org/10.6035/ForumRecerca.2018 539 attitudes towards Spanish than towards the other two languages under investigation, that is Valencian and English. However, our findings are in line with Nightingale's (2012) study on adolescents' language attitudes, who reported to have more positive language attitudes towards Valencian in the first place, followed by English and finally Spanish. Still, the foreign language (i.e. English) is placed in the last place in the present study, even though the context is an English-based school.
As aforementioned, teachers at the school were also impressed with the findings since students' everyday attitudes towards the minority language are far more negative that what we found in our results. In fact, one of the teachers claimed that even Valencian native speakers are reluctant to use Valencian in an educational context and they all resort to the use of Spanish or English to address teachers or other students. Moreover, we believe that other variables which have not been taken into consideration in the present study might have played a role in students' choices. The order in which the recordings were played may have also influenced the way children perceived each recording in the different languages. As Spanish was the first audio track to be played, this might have caused the better understanding of the story in Valencian, that is, the third audio track.
It is important to highlight that according to our findings, children's behaviour do not match the language attitudes' affective component. While their behaviour shows negative attitudes towards Valencian, the affective part shows the contrary. That is to say that respondents displayed very positive attitudes towards Valencian but then they do not use it, probably due to the low prestige that Valencian has in the present school. The same happened in O'Rourke and Ramallo's (2015) study about Galician speakers' language attitudes, who displayed commitment and very positive attitudes towards the language (i.e. cognitively and affectively) but very low language practices (i.e. the behavioural component). In line with these findings, it can be stated that the behavioural and the affective component of the tripartite attitudes' construct can be analysed individually.
Taking into account our findings, the hypothesis is not confirmed and, therefore, our previous expectations have been contradicted since our participants displayed more favourable attitudes towards Valencian, followed by Spanish and finally, English.

Results and discussion related to hypothesis II
The second research question seeks to determine whether the age variable may have an impact on young learners' language attitudes (Baker 1992, Nikolov 1999, Mihaljević-Djigunović and Krevelj 2009, Nightingale 2012, Portolés 2014 in an English-based | Índex Andrea Rubert Albiol. Young learners' language attitudes in a multilingual school 540 school context. Our hypothesis II claims that younger learners would display more positive attitudes towards the three languages than those in older learners. In order to test the second hypothesis, Figure  3 illustrates the overall positive attitudes from Year 1 and Year 5 students. As it can be observed in Figure 3, there seems to be a significant difference between the amount of positive language attitudes shown in Year 1 and Year 5. At first sight, first grade school children have the most positive attitudes towards the three languages. As opposed to the overall positive attitudes analysed in the previous subsection, this figure presents that younger learners show higher positive attitudes in Spanish (93%) than in Valencian (89%). However, English follows Valencian really close as positive attitudes refers. Yet, it is still presented in the last place (85.7%).
Considering the results regarding school children in the fifth grade, the reported attitudes are quite lower than those presented in first grade. Year 5 students show a clear preference for Valencian (78.5%), followed by Spanish (66.7%) and finally English (35.7%). It is important to highlight that the younger learners, those whose ages range from five to six years old, always speak in English with each other as well as with the teacher. However, as they grow older they start using Spanish to talk with their classmates and only resort to English when they address to the teacher, this can be clearly observed in secondary or sixth form. Self-awareness and self-esteem might be some of the key factors that influence language use within puberty. Therefore, students' attitudes towards the foreign language seem to be affected by the age variable since we can observe a great different between younger learners' (85.7%) and older learners' (35.7%) attitudes towards English.
| Índex These results, thus, are in line with other studies' findings (Baker 1992, Nightingale 2012, Portolés 2014. Portolés (2014, 147) also found out that preschoolers show a preference for the minority and the foreign language when compared to primary students results. She also claimed that «young children lack awareness of the low status and vitality of the minority language» and, thus, this is reflected in the present results since younger learners' displayed high positive attitudes towards the minority and the foreign language compared to the older learners. As stated by Portolés (2014, 150), «Baker (1992) reported that favourable attitudes towards the minority language decrease with age». However, this is not the case in the present study since fifth grade students showed more favourable attitudes towards the minority language (78.5%) than to the majority language (66.7%). In this vein, it cannot be stated that the majority language has got the higher positive attitudes as regards older learners. Thus, these findings contradict previous research (Hoffmann and Ytsma 2004, Lasagabaster and Huguet 2007, Rindler Schjerve and Vetter 2012, Portolés 2014 in which respondents reported to have more favourable attitudes towards the majority language (i.e. Spanish) as they grew older, than to the minority or the foreign language. These authors also claimed that Spanish is considered to have a higher prestige in shared bilingual communities and, therefore, it is the language used in formal situations or places considered to be of high social status. In line with these assumptions, if we consider our respondents' language practices, Spanish is the language mostly used among them in the school, followed by English. We probably agree on the fact that this might due to the high prestige that Spanish has got in the present school context.
Taking into account these results, the hypothesis II is not confirmed since older students' showed more negative attitudes towards the majority and the foreign language than younger learners. Therefore, as discussed in the previous section, the affective and behavioural components do not match either since older learners seem to have positive feelings towards Valencian (i.e. positive language attitudes) but they rather use Spanish or English to communicate at the school.

VI. Conclusions
Language attitudes may play an important role in the process of learning an additional language and, therefore, they may affect the degree to which we are able to master it. Existing research (Safont 2007, Nightingale 2012, Portolés 2014) has examined learners' language attitudes. However, no studies have been conducted in English-based schools. Thus, the present study aimed at covering that research gap by examining young learners' language attitudes in an English immersion context. Variables, such as age have been taken into consideration. Our results, as regards the age variable, show that positive language attitudes drop over time and, thus, these findings are in line with previous studies (Baker 1992, Nikolov 1999, Cenoz 2009, Lefever 2009, Nightingale 2012, Portolés 2014. Taking these findings into consideration, it cannot be stated that the sociolinguistic context (i.e. the English immersion school context) necessarily affects the development of language attitudes across the years.
However, considering our results, the minority language gains favourable attitudes as regards older learners while majority and foreign languages' attitudes decrease over time. Considering the school context as a paramount element affecting language attitudes and leaving the age variable out of the equation, it can be noted that it may affect language attitudes in a certain way especially on a daily basis. For example, students do not use Valencian in class and are quite reluctant to receive instruction in this language as stated by several teachers in the school. Nonetheless, this is not reflected on our results since the minority language reported the most favourable attitudes, followed by Spanish and, finally, English.
Our findings suggest that, although the language attitudes' tripartite construct can be analysed by considering its three components to be interrelated, they can also be analysed individually since sometime what we feel or believe does not match with actual responses or behaviours.
In our view, these results have provided valuable insights that should be further investigated by applying statistical methods in order to get determining results. Baker, Colin. 1992