COMPETITION OF JAVANESE AND INDONESIAN: A MIXED-METHODS APPROACH TO YOUNG MULTILINGUALS’ LANGUAGE CHOICE

Multilingualism differs from place to place, with different pressures and outcomes depending on each unique situation. In Indonesian context, an important factor is the status and over prestige afforded to the national language, posing a possible threat to local language vitality. This study reports the position of Javanese and Indonesian as parts of the language repertoire of young Yogyakartan multilinguals in three domains: home, school, and the street. A mixed-methods approach was used and the main data were collected through questionnaire and observations at ten participating high schools. The students’ survey was responded by 1,039 students. Their natural language use was recorded at school playground. Supporting data were collected from language teachers’ survey and interviews with school authorities. The findings show that Javanese and Indonesian still compete in the three domains but not in all sociolinguistic situations. The youths’ Javanese-Indonesian choice is mostly dependent on gender and parental level of education.


INTRODUCTION
Multilingual situations give rise to a number of interesting issues, such as the degree of multilingualism in speakers, code-switching and code-mixing, attitudes to language, and language use across domains. Other issues, particularly relevant to Yogyakartan youths, are the status of each language and language choice and preference. For example, Indonesian, as Indonesia's official language, is often connected to the shift away from Javanese. Smith-Hefner (2009) found that young Javanese people tend to perceive Indonesian being more "trendy", "cool", and "modern" than Javanese (p. 62). Indonesian is also perceived as easier than Javanese (Setiawan, 2013). Regarding the perceived importance of and high enthusiasm towards English especially in the educational context, some language activists and educationalists consider its negative impacts on younger generation's pride in their own language and culture (see e.g., Arafah, 2014;Hanna, 2012). Arabic is a crucial language due to its role in Islamic rituals and given that Islam is the religion of the majority of Indonesians. In Islamic education, Arabic is thus used more for religious purposes than daily communication.
Act No. 20 (2003) andAct No. 24 (2009) mandate the use of Indonesian as the main language of instruction and allow foreign languages such as English and Arabic to be used as instructional languages to support foreign language mastery. Javanese and other regional languages can be used in early stages of education (Act No. 20, 2003) and are still taught as a subject with minimum meeting hours in the school curriculum.
However, it was not tested in the high schools' highly valued National Examination, whereas Indonesian, English and Arabic (in Islamic schools) were. This has had a significant impact on language education. Indonesian, English and Arabic are taught in more meeting hours than regional languages. In the educational context, Javanese does not seem to have a good standing among students. It is commonly considered as complicated by young Javanese due to its speech levels (Smith-Hefner, 2009;Zentz, 2014), where choice affects politeness standard. Since multilingual education is important as, among others, a solid foundation for learning and multilingual literacy (UNESCO, 2014) and to promote educational policies that support linguistic, cultural, ethnic, and religious diversity and to strengthen the nation's identity (Cummins, 2001), research on students or youths' multilingualism is significant.
This current study of young Yogyakartan multilinguals' language choice is related to previous work that examines the shift from Javanese to Indonesian in Yogyakarta (Kurniasih, 2006;Nurani, 2015;Smith-Hefner, 2009) as well as in other regions (Musgrave, 2014;Purwoko, 2012;Setiawan, 2013;Untoro, 2011;Zentz, 2016). The language shift can be related to three issues: firstly the youths' language choice and preference; secondly views of the elder generation on their language use; and thirdly relationship between their language use and a number of factors such as gender and parents' education levels. This study addresses the use of not only Javanese and Indonesian but also English and Arabic, all of which are part of the youths' linguistic repertoire. The focus, however, is on the first two due to the examined shift.
In relation with language choice and preference, this study examines the use of youths' languages in three domains: home, school, and the street, in which all the youths are assumed to share similar opportunities of using their languages. The language domains refer to a range of activities or socio-cultural constructs or contexts representing a combination of settings or place, topic, and role relationship in which a language is used (Fishman, 1991;Romaine, 1995;Spolsky, 2003). In multilingual situations, different languages are commonly assigned to different domains (Spolsky, 2003), and multilinguals usually acquire and use their languages for either particular purposes, or depending on whom they are speaking with (Grosjean, 2006).
Secondly, since young people are observed to experience pressure and shaming from the elder ones related to their use of Javanese Krama, for example through heavy corrections (Zentz, 2014), it is important to know the elder generation's view about the youths' language use. As Gafaranga (2010) claims, the young people might not realise or aware of the ongoing language shift. However, elder community members have conscious knowledge of the changing language and the shift (Karan, 2011).
Related to the third issue, factors associated with language choice: gender and parents' educational levels are scrutinised, with following hypotheses. 1) Female research participants tend to choose Indonesian, rather than Javanese, because it has higher prestige and is a language of wider communication. This hypothesis is based on scholars' claims that there is positive association between gender and the use of prestigious versus less prestigious languages (e.g., Eckert & McConnell-Ginet, 2003;Labov, 1990;Tannen, 2010;Trudgill, 1972). Relevant research includes Kurniasih (2006), Smith-Hefner (2009) andBissoonauth (2011). 2) Participants with a higher parental education level tend to choose Indonesian, rather than Javanese, compared to those with a lower parental education level. This hypothesis is built on both traditional and contemporary works (e.g., Bernstein, 1960;Bissoonauth, 2011;Kurniasih, 2006;Labov, 1963Labov, , 1972Labov, , 1990 confirming that a higher socio-economic status marks the more frequent use of language in terms of prestige.
To sum up, this study has the objectives of explaining the use of Yogyakartan youths' languages in the three domains, the elder generation's perspectives, and relationship between their language use and gender as well as parents' education levels.

METHOD
No research method, quantitative, qualitative, or mixed-methods, is intrinsically better than the other because the choice depends on the research problem (Creswell & Clark, 2011;Riazi & Candlin, 2014;Teddlie & Tashakkori, 2010). To achieve the objectives, this present study applied both quantitative and qualitative approaches through survey, interview, and observation. Both types of data are complementary and are not necessarily mutually exclusive (see Dornyei, 2007) and used to investigate the research object which is multidimensional (Riazi & Candlin, 2014).
All types of schools in the population were included for accurate coverage (Lohr, 2018). 'Yogyakarta' where the schools are located refers to the city, not the province of Special Region of Yogyakarta. Based on principals' and students' consents, ten schools that have both male and female students participated as representatives of all school types: state and private general schools, state and private vocational schools, and state and private religionbased schools (Table 1). A student survey was conducted and responded by 1,039 student participants, aged between 12 and 18 years old. The SPSS was used to measure the frequency of data occurrence and find correlations between language use and gender as well as parental education level. With regard to the hypotheses, the language use is claimed to be dependent on the two independent variables if the p-value/value of the Pearson Chi-Square has less than the significance level of .05.
To triangulate these student data, 34 language teachers who gave consent from the ten schools were also surveyed. The majority that participated were teachers of the compulsory language subjects at their school: 8 Javanese teachers, 10 Indonesian teachers, 9 English teachers, and 3 Arabic teachers. Two French teachers from different schools, 1 German teacher, and 1 Japanese teacher also participated.
Relevant emergent data from interviews with the school authorities were used to corroborate the statistical findings. Four junior high and five senior high school principals from the ten participating schools were interviewed because one was sick. Two principals considered that being accompanied by a staff member was complementary and could provide a more complete contribution, making 11 interviewees in total: the nine principals, one vice principal of student affairs, and one language coordinator. The choice of school principals is relevant to qualitative sampling which is mostly to select participants purposefully due to their rich knowledge, information, or experience.
It was also important to establish actual language use by observing and recording verbal interaction in real communications. To get real examples of natural language exchanges, a number of short, overt non-participative observations in playgrounds were done.

RESULTS AND DISCUSSION Results
With regard to demographic information on gender, the number of female and male students were 632 (60.8%) and 373 (35.9%) each; 34 (3.3%) students did not respond to gender identification. Parental level of education is shown in Table 2. Following are relevant data related to Javanese youths' languages choice.

Home
The number of young people reported using Javanese to their family is slightly larger than that using Indonesian. Table 3 indicates that a few more number of participants reported speaking Javanese than that reported using Indonesian with their mothers and fathers. More than twice as many reported speaking LJ Ngoko over HJ Krama to both their parents. With their siblings, more respondents reported speaking Javanese than speaking Indonesian.
Relatively similar figures reflect the more number of participants that claimed speaking Javanese than that reported using Indonesian to their peer and older neighbours. Far larger number of young people reported using LJ Ngoko than that claimed using HJ Krama to peers, but contradictorily much more number of respondents claimed to speak HJ Krama with older neighbours.
A different trend appears when speaking to relatives and guests. The number of participants claimed the use of Javanese to relatives is lower compared to that of Indonesian. The use of Javanese and Indonesian with guests shows a large disparity.
This contrasts with the results of the principals' interview data (Table 4).  The results from the student survey and interview data are probably best described in a comment given by one of the school principals during an interview: "Moreover, many of them do not use Krama in their neighbourhood. They only speak Ngoko at home. Some only use Indonesian. These days, they use more and more Indonesian for their daily conversations at home." [P of JH4] The Chi-square tests of relationships between the young people's gender and their use of Javanese and Indonesian in the home mostly result in p < .05, indicating a significant association, except in communication with their fathers. The tests show that χ 2 = 7.756 and p = .021 for language use with mothers; χ 2 = 4.573 and p = .102 with fathers; χ 2 = 17.750 and p = .000 with siblings; χ 2 = 67.156 and p = .000 with relatives; χ 2 = 97.380 and p = .000 with peer neighbours; χ 2 = 10.471 and p = .005 with older neighbours; and χ 2 = 19.578 and p = .000 with guests. The data show that female participants tended to speak more Indonesian than male participants did and the sharpest contrast was shown in their report of communicating with peer neighbours.
The other Chi-square analyses suggest a significant association between their parents' levels of education and their home languages (Table 5).

Schools
Among the participants, the reported use of Indonesian is extremely dominant in the school environment.
However, Tables 6 and 7 indicate that the students prefer to speak Indonesian with all school members except their peers. Competition between the use of Javanese and Indonesian appears in communications among peers, both inside and outside class. For communication in class, more respondents claimed to speak LJ Ngoko than Indonesian whereas students who claimed to speak LJ Ngoko were smaller in number than those who stated using Indonesian outside class. Only a few number of participants stated they used Javanese in interactions with non-academic school members like school janitors, parking attendants and canteen assistants. Of these interactions, more reported choosing LJ Ngoko than HJ Krama as Table 7 shows.
The students' data on their language use at school were triangulated with the teachers' (Tables 8 and 9). Most teachers admitted the dominant use of Indonesian when students speak to them both in and outside class and few teachers recognized the use of English to teacher in class. Most teachers also observed that Indonesian dominates student interactions with principals, administrative staff, school janitors, parking attendants, and canteen assistants. These findings confirm the students' self-reported data. However, there is some discrepancy between the data from students and from teachers regarding interaction with schoolmates. More teachers considered that Indonesian dominates language use inside class and Javanese is dominant outside class. However, more students admitted they use Javanese more frequently than Indonesian inside class and they use more Indonesian than Javanese outside class.
Almost all school principals recognised the dominant use of Indonesian among students at school (Table 10). The principal of JH2, for example, stated that "Indonesian is used for daily conversations. Nowadays, children tend to speak Indonesian very often." A closer investigation of peer interactions shows that there is some gender difference in the choice of LJ Ngoko and Indonesian (Table 11). The number of female students who said they spoke Indonesian with their peers at school is higher than those who said they spoke Javanese. In inclass interactions, the number of female students who said they used Indonesian is nearly double that of students who said they used LJ Ngoko and in outsideclass communications, the number is more than three times.
By contrast, more male students said they used Javanese more than Indonesian. The number of male students who said they used Javanese with classmates in class is more than four times as many as those who said they used Indonesian and related to outside-class exchanges the number is more than triple. The results of the Chisquare analyses suggest that there is a significant association between gender and the use of LJ Ngoko or Indonesian with their friends -with χ 2 = 185.693 and p = .000 for communication in class, and χ 2 = 248.939 and p = .000 for communication outside class.
The findings from the observations on out-of-class peer interactions support the statistical findings. There were 11 sets of recordings of naturally-occurring conversational data. The use of only Javanese appeared in five male-to-male conversations, and only Indonesian was used in two female-to-female conversations; an example of each is in the next paragraphs. Indonesian with a little Javanese was used in three femaleto-female conversations and mixed Javanese-Indonesian was used in one male-female conversation, with the male participant using Javanese while the females used Indonesian.
Following is a dialogue between two male students after a class in School SH1. Student 1 asked why Student 2 did not reply to a message he had sent via online chat. Both spoke LJ Ngoko only.
'Was there any notification on your phone?' The following is an Indonesian conversation between four female students in School SH2. The conversation is characterised by a style that imitates the Jakartan dialect, which in many ways is considered 'trendy and modern' (see Smith-Hefner, 2009), and the Javanese influence of particles -e and -pa, which characterised locality.
'No idea. Just leave her.' The principals of JH4, JH5 and SH2 confirmed the influence of the sociable style on the youths' Indonesian, as exemplified in the following comment.
"Nowadays, there are many students who use Indonesian, but in the youth's sociable style. They use that language style to send short messages." [P of SH2] The common use of nonstandard Indonesian is not only heard in their daily speech but also in written communication, like SMS. This phenomenon worries the older generation, including principals and teachers, who would prefer young people to use the standard language regardless of the formality of the settings.
Relevant to Javanese-Indonesian peer interactions, a set of Chi-square tests indicates significant relationships between the parent's level of education and the use of these local and national languages with friends in class and outside class -fathers': χ 2 = 31.344 and p = .000 inside class, and χ 2 = 16.374 and p = .000 outside class; mothers': χ 2 = 14.262 and p = .001 inside class. However, there is no significant association between mothers' level of education and the use of languages in outside class peer interactions, with χ 2 = 3.951 and p = .139.

Street
Language use on the street is typically informal, consisting of greetings that serve the phatic function and small talk among neighbours, friends or acquaintances as stated in the questionnaire. Table 12 indicates that Indonesian is used slightly more frequently than Javanese in greetings with peers and with older neighbours. The reported frequency of using Indonesian between school mates is also higher than Javanese. The number of young people who claimed to use Indonesian to greet acquaintances is much higher than those who claimed to use Javanese. Only a small number of young people reported using Arabic greetings but this number is higher than that claimed to use English. The Chi-square analyses show that there is significant relationship between gender and the use of Javanese and Indonesian on the street, with the following details. To greet peer neighbours (χ 2 = 125.041 and p = .000); older neighbours (χ 2 = 11.135 and p = .004); school mates (χ 2 = 237.575 and p = .000); acquaintances (χ 2 = 37.951 and p = .000). The Chi-square statistics in Table  13 reveal the significant association between parental education level and language use in five out of eight cases.
The case of Javanese use is similar with small talk, as Table 14 shows. The only difference is that more young people reported speaking Javanese, especially the low variety, with their peer neighbours. With older neighbours, small talk occurs more frequently in HJ Krama than LJ Ngoko, as is the case with greetings. The Chi-square tests result in significant associations between gender and the use of Javanese and Indonesian except for communication with older neighbours. The details are as follows: Interactions with peer neighbour (χ 2 = 99.005 and p = .000); older neighbours (χ 2 = 8.061 and p = .018); school mates (χ 2 = 194.497 and p = .000); and acquaintances (χ 2 = 29.257 and p = .000). As Table 15 depicts, no significant association is shown only by relationship between language use for communications with acquaintances with mother's education level.

Discussion
Young Yogyakartan multilinguals' higher use of Javanese than Indonesian appears in three domains: home, school, and street, depending on particular interlocutors. More specifically, the more frequently use of Javanese is due to the contribution of the used L variety.
In the home domain, a larger number of participants claimed to speak LJ Ngoko than HJ Krama to their mothers, fathers, sibling, relatives, and peer neighbours because LJ Ngoko is the mother tongue (see also Purwoko, 2011). In the case of the young people's Javanese communications with their parents, the rules of the Javanese speech levels are not applied (Hudson, 2002). The young people's intimacy with their parents might make them disregard an interlocutor's age as a determinant for using HJ Krama, and to some extent make the child-parent informal relationship stronger.
The more frequent choice to use HJ Krama over LJ Ngoko with older neighbours shows that these young speakers do have some knowledge of the socially determined speech levels. They are aware that non-intimacy plus an interlocutor's age becomes a determinant for the choice of the H variety over the L variety. Slightly more young people reported they preferred Indonesian to HJ Krama when speaking to older people, which is probably, as G. Poedjosoedarmo (2006) and Zentz (2014; describe, a safer choice to speak politely rather than using LJ Ngoko (see also Ravindranath & Cohn, 2014).
More young people reported using Indonesian to their relatives and guests than Javanese. This trend is increasing, which shows that the greater the social distance between interlocutors, the more likely it is that they will use Indonesian. Again, the respondents showed that Indonesian was safer than either LJ Ngoko or HJ Krama. 'Being younger or older' than relatives in the Javanese context does not necessarily mean only considering the relative biological ages of the interlocutors but also requires consideration of the age of their parents. Therefore, HJ Krama might be used with relatives because either the speakers are biologically younger than the interlocutor or their parents are younger than the interlocutor's parents. The tendency to choose Indonesian with guests indicates that spatial distance is also likely to be perceived as social distance or non-intimacy. Indonesian is considered neutral because speakers can use it regardless whether they are familiar or not with their interlocutors and whether they are older or younger than them. For many Yogyakartan youths, Indonesian has become 'a language escape' when they find themselves in Javanese situations that normally require them to use the formal or polite form, that is the H form, from the speech levels.
The finding that more speakers of Javanese reported their use of LJ Ngoko rather than HJ Krama with family members, relatives and peer neighbours indicates that the young multilinguals consider intimacy and familiarity first regarding these interlocutors. The finding that more young people claimed to speak HJ Krama than LJ Ngoko to older neighbours and guests indicates that age is a determinant in their choice between these varieties (see Koentjaraningrat, 1985;S. Poedjosoedarmo, 1968).
The Javanese-Indonesian language patterns in their home domain are illustrated in Table 16.  These findings relate to gender and parental education level. This present study confirms the traditional view on different language use by gender in the literature (e.g., Eckert & McConnell-Ginet, 2003;Labov, 1990;Tannen, 2010;Trudgill, 1972). They also confirm the trend strongly claimed in a number of bi-/multilingualism studies that females tend to choose the more widely used language (e.g., Bissoonauth, 2011;Kurniasih, 2006;Smith-Hefner, 2009). With their mothers, siblings, relatives, peer neighbours and older neighbours, girls are more likely to speak Indonesian, while boys are more likely to speak Javanese. The relationship to parental education level is illustrated in Table  17, with the arrows' directions showing the larger numbers of speakers.
The difference in language use at home by parental level of education is in line with a number of studies on language prestige and socialeconomic status (e.g., Bernstein, 1960;Bissoonauth, 2011;Kurniasih, 2006;Labov, 1963Labov, , 1972Labov, , 1990. Labov (1990) claims that the use of either objective parameters: education, occupation and income; or subjective measures of social stratification, similarly refer to "the hierarchical organization of the speech community". In this present study, the higher the level of the parents' education, the more frequently the young people use Indonesian, and vice versa.
In the school domain, young people reported using Indonesian dominantly to almost all interlocutors, and using Javanese irrespective of relative age or level of formality. For example, in the informal situation when pupils interact with non-academic and non-administrative school staff, the majority of young people prefer to use > > > Indonesian. If they do choose to use Javanese, LJ Ngoko is preferred over HJ Krama.
In the context of traditional Javanese speech, relative age, with addressee being older, predicates the choice of HJ Krama, but the findings from this present study show that this is not the case for these Yogyakartan youths. The seemingly inappropriate choice of languages by these young multilinguals can be explained according to their age, their level of competence in Javanese, and their prioritising of the intimacy dimension. First, some young people avoid using their hierarchal ethnic language because of their age, which will leave them with HJ Krama as the right choice. Since this is inappropriate for them, they switch to Indonesian. Second, other young people do not know how to use Javanese speech levels, so for them, the choice is LJ Ngoko, and this is what they use. Third, some young people know the speech levels but value the intimacy dimension more highly than other social dimensions. This also leads them to choose the low variety. The small number of young people who do speak HJ Krama probably do so because they have a good understanding of the speech levels and their social functions, and are adept at using the high variety appropriately. The shift from HJ Krama to Indonesian by most Javanese young people has been observed for more than three decades, for example, in Surakarta, another centre of Javanese culture (see Errington, 1988). Javanese, represented by LJ Ngoko, competes with Indonesian only in the situation of peer interactions, as shown by the student survey data and the observations conducted inside and outside classes. Actual use and adult expectations differ. The surveyed teachers reported that they thought Javanese should have been the dominant language in outside-class peer interactions, while the interviewed principals assumed that generally more students would use Indonesian than Javanese. These different opinions probably stem from their own personal associations with particular situations. As language teachers, the surveyed teachers logically considered sociolinguistic factors, such as the formality or informality of the setting or relative age. They would predict that more students would choose LJ Ngoko over the other languages. The principals just expressed their opinions based on what they observed: in general, students' poor use of Javanese and extensive use of Indonesian in the playground.
The young people's language choice tends to be highly associated with gender and parental levels of education, as the case of Javanese-Indonesian rivalry in the home domain. That there is relationship between gender and parents' levels of education and the use of languages at home and at school is comparable to Kurniasih's (2006) findings.
In researching university-aged young people in Yogyakarta, Smith-Hefner (2009) claims that young Javanese women's preference for Indonesian over Javanese relates to their gender expectations on marriage. The issues of marriage and gender equality are not so relevant to this present study's female participants, given they are much younger than Smith-Hefner's (2009) subjects. However, for both groups the young women's language choice fits with the perceived prestige of Indonesian being "current", and especially the youth's everyday language, which is valued as "trendy", "cool" and "modern" (p. 62). The young women's use of Indonesian over Javanese might also be a signal that the use of Indonesian is an incoming norm that is being led by young women.
The data on language in the street domain shows Javanese-Indonesian competition for greetings and small talk with neighbouring peers, older neighbours and schoolmates. It also shows that within the context of speaking to older neighbours, more young people chose Indonesian than Javanese, which is different from the findings on language choice within the home domain. The significantly higher proportion of young people choosing Indonesian to interact with acquaintances both in greetings and small talk most likely indicates that social distance is an overriding factor influencing the young people's choice of Indonesian instead of Javanese, as also shown in the home domain.
The young people's use of languages to all interlocutors in the street domain has significant association with gender. Their language use to most interlocutors also has significant correlation with parental level of education. With regard to interactions through greetings and small talk with acquaintances, there is no correlation between the used languages with mother's level of education. This can be related to the majority of respondents, around 90%, who admitted to use Indonesian to less intimate people.

CONCLUSIONS
Javanese and Indonesian still compete in home, school, and street domains, but not in all sociolinguistic situations. The findings show that Yogyakartan youths are most likely to use the L form of Javanese in informal situations with peers or with older people, like parents, with whom they feel close. The competition between ethnic and national languages in the three domains suggests that even though Javanese is slightly more prevalent than Indonesian at home, its vitality is weakening. In the context of Javanese urban youth under this study, Javanese is losing ground. The evidence is that Indonesian has been replacing Javanese in various domains and functions.
If we consider that Javanese shift to Indonesian has occurred within the Javanese community itself, the concerns that many parties have about the further use of Javanese maybe well founded (e.g., Bahasa Jawa Mulai Ditinggalkan 'Javanese is Becoming Obsolete Errington, 1992;2003;Hanna, 2012; Motivasi Memakai Bahasa Jawa Makin Tiada 'Motivation of Using Javanese is Fading Away', 2009). The shift away from Javanese in Yogyakarta needs serious attention because this city is the centre of Javanese language and culture (Errington, 1998(Errington, , 1998Smith-Hefner, 2009). The call for a close look at the vitality of Javanese (Ravindranath & Cohn, 2014) and any better future support systems to Javanese (Vander Klok, 2019) needs follow-ups from the government, educational institutions, communities and other relevant parties.