The Effect of Multicultural Family Structures on the Language Attitudes of Children and Adolescents

Attitudes toward languages of biand multilingual children and adolescents have not yet become a focus of research. Despite rapid growth in the number of biand multilingual children and adolescents across the globe, surprisingly few studies have been devoted to this significant phenomenon. Biand multilingualism is a major consequence of immigration and cross-cultural marriages. Regardless of whether cross-cultural marriage is involved, the immigration of families has considerable linguistic consequences on children and adolescents. This paper draws on five case studies in examining the complex factors determining the attitudinal patterns evinced by biand multilingual children and adolescents from immigrant families in Thailand. Five households agreed to participate in this research endeavor and data were also collected from surveys, interviews, observations and field notes acquired through the employment of ethnographic investigative methods. The data collected were analyzed through constant comparative method and content analysis. Findings showed consistent patterns for those biand multilingual children and adolescents whose Thaispeaking mother was linguistically dominant in a family with an immigrant father speaking a minority language. The results showed that these children were more likely to perceive Thai as more highly regarded language. By contrast, it was also found that Thai was not as highly regarded by the children of parents if both were minority-language speaking immigrants. It was concluded that the family structure of immigrant families is associated with the language attitudes of their children. By the same token, the type of marriage of immigrant families has long-reaching effects on the development of children and adolescents’ language attitudes. Moreover, data showed that a non-migrant mother’s dominant language played a more influential role in contrast to the minimal role-played by the migrant father’s minority language in the development of children and adolescents’ language attitudes.


INTRODUCTION THAILAND'S IMMIGRANT COMMUNITY
In broad terms, a cross-cultural marriage, otherwise known as a transnational marriage, an interracial marriage (Interracial Marriage, n.d.) and a mixed marriage (InterNations, n.d.), refers to marriages of couples from two different countries.However, in the context of Thailand, there are marriages, for instance, between Thai Muslims of Malay descent and Thai Buddhists, or between animistic hill tribe women and Buddhist Lana men in northern Thailand.In recent decades, cross-cultural marriage and subsequently bi-and multilingual children and adolescents in immigrant families have increased dramatically.
With a growing economy, Thailand has recently been recognized as a 'migration hub' (Huguet & Chamratrithirong, 2011) and the Thailand immigrant community (consisting of

PROBLEM STATEMENT
We found few studies focusing on the attitudes towards languages shown by bi-and multilingual children and adolescents.Such was the case of Thailand in particular, where only a handful of previous studies paid significant attention to language attitudes of young adults who were students at Thai universities (Saengboon, 2015;Sisamouth & Lah, 2015;Snodin & Young, 2015;Ploywattanawong & Trakulkasemsuk, 2014).However, in these papers, language attitudes of bi-and multilingual children and adolescents from immigrant families still remained under-researched.Moreover, most of the existing studies of bi-and multilingual children and adolescents primarily focused on one-parent-one-language families and the interactional strategies minority language-speaking parents adopted in dealing with their children.These strategies had significant consequences for the proficiency level of the minority language spoken by a bilingual child (Qiu & Winsler, 2017;Takeuchi, 2006;Venables et al., 2014).Still, in these studies, language attitudes of bi-and multilingual children and adolescents remained relatively unexplored.
The aim of this study was to shed light on five specific immigrant families in Thailand through exploring the attitudinal patterns of bi-and multilingual children and adolescents in these families.

THE CURRENT LANGUAGE SITUATION IN THAILAND
Thailand is ethnically a melting pot and linguistically a diverse country with an estimated 74 spoken languages (Ethnologue, 2005).In their respective regions, Northern Thai (Lanna or GEMA Online ® Journal of Language Studies Volume 18(1), February 2018February http://doi.org/10.17576/gema-2018February -1801-08 -08 eISSN: 2550-2131 ISSN: 1675-8021 124 Kam Meuang), Northeastern Thai (Isaan or Lao), Central Thai (Standard Thai) and Southern Thai (Pak Tai) play key roles as regional vernacular languages used in daily communication (Smalley, 1994).The linguistic situation in the southernmost border provinces and that of the numerous hill tribes in the north are essential for everyday language use.There are two distinctive Malay languages in the south adopted by Muslims.Moreover, there are various Khmer speakers.There are major dialect differences in the north east.There are numerous speakers of different Mynamrese languages, several speakers of Vietnamese and a large number of ethnic languages spoken in the north.Despite the wide linguistic range in Thailand, the Thai government has stipulated that Central Thai shall be the national standard and the only official language used in public domains, notably, government and education.This policy is justified on the grounds that it strengthens national unity.Influenced by the one nation-one language ideology (monoglossic ideology), Standard Thai, the lingua franca of Thais with different L1 backgrounds, is thus the most prestigious language.However, ethnic minority languages (e.g., in Isaan) with millions of speakers are not recognized across various public domains (Draper, 2010).
However, globalization has heightened interest in English on the part of those engaged in business, commerce, education, tourism, trade, medicine, scientific research, and journalism among other domains (Lee, 2015).English is taught as a compulsory first foreign language, as stipulated by the Ministry of Education's curriculum for public schools (Draper, 2012).Further, as China, Japan and South Korea deepen economic ties with Thailand, regionally hegemonic languages-Chinese, Japanese and Korean-have become increasingly popular.This is indicated by the fact that the teaching and study of the Chinese (Mandarin) language has been recently skyrocketing (Lee, 2015).

LITERATURE REVIEW LANGUAGE ATTITUDES OF BI-AND MULTILINGUAL CHILDREN AND ADOLESCENTS
Drawing from perspectives of psychology and sociolinguistics (e.g., Edwards, 1999;Dörnyei & Csizer, 2002;Kinzler et al., 2012), specific theoretical inputs of Rosenthal (1974) and Garrett (2010) in particular provided the theoretical basis for the current study.
Attitudes toward languages were regarded as a complex phenomenon.As such, the macro-concept of language attitudes was viewed as a vast field in and of itself.The study of language attitudes has long been a central concern in social psychology, sociolinguistics and related fields (see Edwards, 1999, for a 'bridging' of the study of language attitudes in social psychology and sociolinguistics).In this paper, the author followed the definition of Garrett (2010), who offered an important insight into the essence of attitudes: "an attitude is an evaluative orientation to a social object of some sort, whether it is a language, or a new government policy, etc" (p.20).Similarly, this paper concurred with the contention of Myers-Scotton (2011) that "our views about language are scenarios we carry around in our minds that create attitudes towards speakers and the social communities to which they belong" (as cited in Mirshahidi, 2017, p. 146).Language attitude research has shown that our views towards any aspect of a language reflected our attitudes towards the speakers and users of that language.By the same token, our language attitudes were tantamount to languagebased social evaluations, as well as explicit preferences and judgments of a language and its speakers and users.
It was scarcely surprising that language attitudes have not become a major topic of research and a concern for public in Thailand and elsewhere.The cursory research so far conducted has shown that Thailand-based language attitude theorists have been interested in the attitudes of young adults and adults, who were university students, towards varieties of English.Some critics maintained that this body of literature (language attitudes toward English varieties in Thailand) was deemed not relevant to the current study, but it was not deniable that it was relevant to the previous studies of language attitudes.In other words, the current sub-section of literature review is relevant to provide theoretical inputs to previous studies in language attitudes.Nevertheless, it was found that they harbored little or no interest in non-English languages.Insofar as concerns Thai scholars, it was also found that they viewed idiolects, dialects, and native accents as grounded in ethnicity, country of origin, and group identity, especially socio-economic class and social status (Kinzler, Shutts and Spelke, 2012).For instance,1 Ploywattanawong and Trakulkasemsuk (2014) investigated Thai university students' attitudes-as measured by acceptance and understandability-toward ASEAN (Association of Southeast Asian Nations) English usage (e.g., Malaysian English and Singapore English or 'Singlish').Likewise,2 Saengboon (2015) surveyed Thai university students' attitudes toward varieties of English. 3 Snodin and Young (2015) investigated Thai university students' and Thai adults' attitudes towards a model or a target variety of English.In another study, 4 Sisamouth and Lah (2015) examined language attitudes of university students in southernmost Thailand.In summary,5 Thailand-based language attitude theorists focused mainly on language attitudes toward varieties of English among university students and adults.However, in current language attitude literature in Thailand, attitudes towards not only varieties of English but also non-English languages as expressed by children and adolescents was relatively under-researched.While attention has been given to Thai adults' and young adults' attitudes towards languages, the same attention has not been paid to preschool children and adolescents' attitudes toward varieties of English and non-English languages (e.g., the dominant language and the immigrant minority language) in Thailand.
Furthermore, beyond Thailand, we found a general lack of research on attitudes of younger children and adolescents toward their languages.It was in spite of a significant number of previous studies on child language acquisition and learning (e.g., see Woan & Tat, 2017).Rosenthal (1974) argued that the majority of previous studies focused on attitudes of adults and older children toward their languages.Another line of research has focused on parental attitudes toward their children's languages (e.g., see Zhang, 2010), without surveying language-related attitudinal patterns expressed by children themselves.Yet there was a different body of research, conducted almost exclusively by psychologists, on monolingual and bilingual children's social and friendship preferences for different types of monolingual or bilingual speakers (e.g., Kinzler et al., 2012;Byers-Heinlein et al., 2016).Approached from the developmental perspective, this body of psychological research on monolingual and bilingual children's social and friendship preferences generally showed little concern for children and adolescents' language attitudes as shown in evaluative comments towards languages and speakers of languages.The dearth of literature in this area was partly because of the widely accepted belief of a number of sociolinguists that 'children do not become aware of dialect differences until about the time of the onset of puberty' (Rosenthal, 1974, p. 55).In contrast, a large number of past research investigations focused on the need to support the development of minority languages as spoken by bilingual children through the so-called One-Parent-One-Language approach (e.g., see Logan-Terry, 2008;De Houwer, 2009), even though there was little or no discussion of bilingual children's language attitudes toward their languages.However, it should be acknowledged that children developed what Rosenthal (1973) termed 'social awareness of language differences' between the ages of three and six when children expressed their preferences for language varieties and social dialects (Kinzler et al., 2012).Despite the paucity of studies, recent research in children's language attitudes has provided a long list of factors that influenced children's language attitudes.These factors included parents, siblings, families, teachers, peers, classrooms, schools, playgrounds and communities, political and sociocultural changes, school instructions, familiarity with the languages, status of the languages in the children's circle of friends, foreign or host country of birth, the number of years spent in the host country, second language support in the host country, the trend of English as the world language, and so forth (Dörnyei & Csizer, 2002;Kinzler et al., 2012;Oliver & Purdie, 1998).
Although the focus of the current study was on children and adolescents, a review of literature with a focus on adults, to a large extent, provided theoretical inputs to the previous studies in language attitudes.There has been a growing awareness of the importance of adult attitudes toward languages, since adult language attitudes shaped their interactions with those who had different language backgrounds.These language attitude experiences were even believed to have long term consequences in respect to decision-making in planning and implementing language and educational policies (Lewis, 1981).Attitudes toward language also played a crucial role in efforts to maintain, alter, and revitalize endangered and minority languages (Fishman, 1980).Research in language attitudes in adulthood showed multiple factors (variables) accounting for varied language attitudes, including age, gender, geopolitics, socioeconomic status et cetera (Baker, 1992;Dörnyei & Csizer, 2002 inter al.).
Childhood bilingualism literature provided sufficient theoretical basis for the current study.One avenue of past research on childhood bilingualism partially focused on crosscultural marriage where there was a dominant language-speaking parent and a minority language-speaking parent with their offspring becoming bilingual (Qiu & Winsler, 2017;Takeuchi, 2006;Venables et al., 2014).In some cases, both parents were minority language speakers from the same language background, but not in a cross-cultural marriage (For example, a Filipino couple lived in Bangkok).Nevertheless, there were few studies of such families which were not in a cross-cultural marriage, but in a L2 dominant environment.
Taken together, the aforementioned bodies of research has not yet fully addressed language attitudes of bi-and multilingual children and adolescents toward their languages and how their attitudinal patterns emerged in the context of belonging to immigrant families.Given these gaps in the current literature, this study explored the following questions.(1) How do family structures (whether or not involving a cross-cultural marriage) influence language attitudes of bi-and multilingual children and adolescents?(2) What factors lead to varied language attitudes evinced by bi-and multilingual children and adolescents?

METHOD
This study investigated the complex interrelationships between the family structures of immigrant families whether a non-or cross-cultural marriage was involved in bi-and multilingual children and adolescents' language attitudes.The data were derived from a multiple-case and multiple-year study (more explanations are provided in the following subsections).A focus on routine interactions in the daily lives of five families in Thailand provided insights into the complex nature of the language attitudes of bi-and multilingual children and adolescents.

MULTI-SITED APPROACH: SAMPLE AND PARTICIPANTS
With the development of globalization, regionalization and other macro-trends (Lee, 2015), immigrant families-whether or not involved in a cross-cultural marriage-have increasingly become common in Thailand.Bangkok Metropolis, Thailand's capital, has the world's fast growing number of bi-and multilingual children and adolescents.A multiple-case study was carried out on five focal families in Bangkok.These families under study were acquaintances of the author.Children and adolescents were selected through the author's personal network, partly because it was claimed they were raised bilingually from birth, i.e., they were categorized under the rubric of 'bilingual first-language acquisition' children (as defined by De Houwer, 2009).The researcher's biases were controlled to a limited extent because the author consulted with a senior social scientist, parents of these children and secondary literature.In order to limit variability, all children and adolescents who participated in the current study were born in Thailand, thereby excluding foreign-born participants who were labeled as bi-and multilingual children and adolescents, and were between the ages of 4 and 17 (mean age = 8 years, median = 10 years) during the time periods at which data were collected.The age difference between participating children and adolescents was wide, because samples were chosen based on the criteria that either one of their parents or both parents were immigrants regardless of their family structure as in a non-or cross-cultural marriage.

CASE 1: FAMILY A
The married couple in family A is in a cross-cultural marriage.Family A consists of a Chinese (Mandarin) speaking father originally from Taiwan, a Thai-speaking mother from Thailand and their Thailand-born five-year-old daughter, Chimlin (pseudonym), who was surveyed and interviewed from the age of two to the age of five.Responses to a questionnaire were recorded and interviews were conducted by Chimlin's mother with the assistance of the author, because she knew better how to interact with her daughter and elicited more valid data from Chimlin than the author.Chimlin's mother is a trained researcher with a doctorate degree to carry out the semi-structured interview.She was given the list of questions by the author.The interview was conducted in the Thai language at home by Chimlin's mother.The reliability of the data was at an acceptable level, in spite of the young age of Chimlin.Some probing questions, for example, were the following: Which language do you use the most frequent?Which language do you like the most?Which language is the most useful?Which language is the least useful?

CASE 2: FAMILY B
The married couple in family B is not in a cross-cultural marriage.Family B is made up of a Filipino Tagalog-speaking father and a Filipino Tagalog-speaking mother from the Philippines and their three children, all of whom were born in Thailand.The oldest and the youngest children did not participate in the survey (because they were not present in family events when the survey was conducted), but their non-survey data were reported by their mother through interviews and were observed by the author during researcher-family interactions, for instance, the author's visits to the family B. The questionnaire completed by their middle daughter, now a 13-year-old adolescent, Jasmine, in 2016, when she participated in this study.She had interacted with and had been observed by the author from the age of 11 to the age of 13.

CASE 3: FAMILY C
The married couple in family C is not in a cross-cultural marriage.Family C includes a Bangla-or Bengali-speaking father and a Bangla-or Bengali-speaking mother from Bangladesh and their two Thailand-born children.Their younger child, a 1.5-year-old boy, was too young to participate in the questionnaire survey in 2016-2017.However, the questionnaire was completed by the mother in 2016-2017 to report about her older child, a four-year-old boy, Sumon (pseudonym).He also participated in this study from the age of one to the age of four, partly because he interacted with the author and was under observation by the author in the same neighborhood for four years.It should be recognized that the mother reported attitudinal data toward the languages of her older son during interviews with the author.With the assistance of Sumon's mother, Sumon was able to rate 1 (not very true), 2 (not true), 3 (may or may not be true), 4 (true) and 5 (very true) for 96-item questionnaire.Some items in the survey were the following: When asked my attitudes toward English, I have positive attitudes and am in favor of English.I see a need to learn English.When asked my attitudes toward Thai, I have positive attitudes and am in favor of Thai.

CASE 4: FAMILY D
The married couple in family D is in a cross-cultural marriage.Family D consists of a Bangla-or Bengali-speaking father from Bangladesh and a Thai-speaking mother from Thailand and their two children.Their younger child, a 1.5-year-old girl, was too young to take part in the questionnaire survey in 2016-2017.Their oldest child, Barsha (pseudonym), a five-year-old girl, was a participant of the study from the age of one to the age of five.The questionnaire was filled out by Barsha's mother in 2017.It should be acknowledged that the mother reported her two daughters' attitudinal data toward languages during interviews with the author.Data collected by the author was also based on observations of researcher-family interactions, because the family lived in the same community with the author for five years.

CASE 5: FAMILY E
The marriage of family E is not in a cross-cultural marriage.Family E, at the present time, consists of a Nigerian-accented English-speaking mother and three Thailand-born children.Only her middle son, a 17-year-old boy, Emmanuel (pseudonym), provided answers in the 2016 questionnaire survey by himself, while his mother reported his data (during formal interviews and informal chats) from the age of 15 to the age of 17.Using interview methods, the mother reported attitudinal data toward languages of her three children, highlighting the experiences of Emmanuel.Data collection was also conducted on the bus and at the workplace, especially since the author shared rides on the bus and worked in the same organization with the mother for five years.

DATA COLLECTION, INSTRUMENTS AND ANALYSIS
In examining the language attitudes of bi-and multilingual children and adolescents, the author took into consideration languages to which the subjects of investigation were exposed: Bengali, Chinese (Mandarin), International Standard Varieties of English (principally American English and British English) referred to as English, Nigerian-accented English, Tagalog and Thai.These languages were chosen in view of the fact that they had a strong presence in the daily interactions of the five focal families under study.In selecting these languages, it was assumed that potentially significant comparison and contrast results would be generated in respect to bi-and multilingual children and adolescents' attitudes, which in themselves were indicative of the dominating influences of ethnic group affiliation and the effort to achieve greater economic power.
The same guidelines and questions were adopted for questionnaire surveys and field interviewing.Data for this multiple-case study were gathered by semi-structured interviews: researcher-to-parent or face-to-face interviews, follow-up parent interviews, parents' selfreported responses to a 96-item questionnaire survey, direct assessments and field-notes from naturalistic (e.g., parent-child interactions and sibling-peer interactions) observations by the author.For sample items in the questionnaire survey, please refer to the preceding subsection of case 3 -family C. The survey and the interview questions were developed from comparable questionnaires and interviews employed in previous studies (e.g., Draper, 2012;Lai et al., 2015;Snodin & Young, 2015).They consisted of demographic information (section I) and open-ended language attitude questions (section II).More specifically, the questionnaire and the interviews asked parents to elicit answers from their children in regards to their attitudes towards international standard varieties of English (i.e., American English or British English) and languages spoken by minority language-speaking parents in addition to perceptions towards regionally hegemonic languages (i.e., Chinese, Japanese and Korean) and the dominant language-the official state language (i.e., Standard Thai).Some items were deemed not applicable to some participants whose parents were not from these countries, thereby modifications of questionnaire and interviews were made to these items.The participants were given at least six months to complete the questionnaire.Family visits, field interviews and researcher-family interactions and observations occurred before and after conducting the survey phase of investigation.
During the initial phase of data analysis, data were analyzed by adopting the constant comparative method (Glaser & Strauss, 1967), open coding and the content analysis convention for categorizing large themes in multiple data.Coding schemes (coded patterns) were derived from recurring patterns rather than past studies.During the second phase of data analysis, the author inferred the presence of specific attitudinal patterns derived from large themes after triangulation with in-depth analyses of survey responses, transcribed interviews, interactions and field notes.
Through the employment of factor analysis, some underlying reasons for these attitudes were additionally revealed.The complex interplay between language attitudes and types of immigrant family structures (those in and those not in cross-cultural marriages) were classified into three broad patterns as main attitudinal findings (see Table 1), which can in turn be divided into six sub-patterns (see Table 2).Further intriguing patterns (as combinations of sub-patterns) can be obtained if a comparative analysis was conducted across the two subgroups (i.e., families with parents who are not in cross-cultural marriages and those that are).

RESULTS
The findings addressed how family structures (non-or cross-cultural marriages) influenced the extent of language attitudes of bi-and multilingual children and adolescents (answers to research question 1).It was found that there were differences between the families in view of the two types of family structures shown by the families, viz., those which were crosscultural marriages and those which were not.The results revealed the consequences of immigrant family structures (i.e., non-and cross-cultural marriages) that largely accounted for the language attitudes of bi-and multilingual children and adolescents.Table 1 summarized overall patterns as categorized by non-and cross-cultural marriages, the dominant language, the minority language and English.
Three broad patterns can be observed.First of all, a potentially important but unanticipated result was that the relative exposure to the dominant and the non-dominant language (i.e., language use frequency) was not predictive of language attitudes (the broad pattern 1).For those bi-and multilingual children and adolescents in families in which parents were in non-cross-cultural marriages more often displayed use of the minority language-thereby being most exposed to the minority language-as compared with those with their parent in a cross-cultural marriage.However, the relative exposure of bi-and multilingual children and adolescents to first and second languages (as reflected in the frequency of use of each language) may not adequately account for variances in their language attitudes.
In addition, an unexpected but presumably crucial result emerged (as the broad pattern 2) when a set of analyses was conducted to compare the dominant language (i.e., Thai) and English.English was more positively regarded than Thai in families in which parents were in non-cross-cultural marriages (e.g., both parents came from Bangladesh, Nigeria and Philippines) in comparison to those in cross-cultural marriages, whereas Thai was more positively regarded than English in these families in which the parents were in cross-cultural marriages as compared with those who were not in cross-cultural marriages.
Finally, the minority language as spoken by migrant parents in contrast to nonmigrant parents was less positively regarded across these two types of families.

Broad patterns of language attitudes
Pattern 1.The relative exposure to the dominant and the non-dominant language (i.e., language use frequency) was not predictive of language attitudes.Pattern 2. English was more positively regarded than the dominant language (i.e., Thai) in families in which parents were in non-cross-cultural marriages than those in cross-cultural marriages, whereas the dominant language (i.e., Thai) was more positively regarded than English in families where parents in cross-cultural marriages as compared with those who were not in cross-cultural marriages.(A non-migrant mother's dominant language played a more influential role than a migrant father's minority language in the development of language attitudes of bi-and multilingual children and adolescents.)Pattern 3.The minority language (spoken by migrant parents as opposed to non-migrant parents) was less positively regarded across these two types of families.
Three straightforward sub-patterns were found in regard to families with parents in non-cross-cultural marriages.Despite a minority language was being reported as being used the most often (with the most exposure to the minority language), bi-and multilingual children and adolescents with parents in non-interracial marriages (i.e., both parents were minority language-speaking migrants) expressed that the dominant language (i.e., Thai) and English were more positively regarded than the minority language.Furthermore, they exhibited that English was more positively regarded than the dominant language (i.e., Thai).
In addition, another three sub-patterns were consistent from families with parents in cross-cultural marriages.While the dominant language (i.e., Thai) was reported as being the most frequent, positive attitudes were held towards both the dominant language (i.e., Thai) and English.Another attitudinal difference was also observed.They displayed that the dominant language (i.e., Thai), as their non-migrant mothers' language, was more positively regarded than English.Moreover, the minority language, as their migrant fathers' language, was less positively regarded.Parents in non-cross-cultural marriages Sub-pattern 1.Although the minority language used the most frequent, it was less positively regarded.Sub-pattern 2.English was more positively regarded than the dominant language.
Sub-pattern 3.The dominant language and English were more positively regarded than the minority language.
Parents in Cross-cultural Marriages (No available data from the present study)

Mother Non-Migrant -Dominant Language Speaking Parent
Parents in Cross-cultural Marriages Sub-pattern 4.The dominant language used the most frequent, partly leading to more positive attitudes toward it.
Sub-pattern 5.The dominant language was more positively regarded than English.
Sub-pattern 6.The minority language was less positively regarded.
Parents in Non-cross-cultural Marriages (No available data from the present study) Accordingly, the one exception was found in case 5. Data derived from Emmanuel, a 17-year-old adolescent from family E, was difficult to analyze partly because he showed that he most liked the dominant language (Thai) even though he perceived the dominant language (i.e., Thai) as the least useful.He perceived the dominant language (Thai) as the most positively regarded (due to his familiarity with the Thai language), but Thai was the least useful language to help him apply for university admissions abroad and foreign university scholarships.
Moreover, results showed how factors leading to varied language attitudes emerged from bi-and multilingual children and adolescents (answers to research question 2).There were several possible interpretations of these results, as well as a number of factors affecting language attitudes of bi-and multilingual children and adolescents toward their first and second languages.The three broad patterns (Table 1) and six sub-patterns (Table 2) can be explained in the following fashion.

FAMILY STRUCTURES: GENDER ROLES OF PARENTS (EXPLANATIONS FOR THE BROAD PATTERN 1)
To further explicate the results of the broad pattern 1, a combination of sub-pattern 1 and 4 (i.e., relative exposure to the dominant versus non-dominant language was not predictive of language attitudes), the author more closely analyzed the data across case 1-5.When asked which language was most frequently used, it was not surprising that bi-and multilingual children and adolescents with parents in non-cross-cultural marriages (i.e., where both were minority language-speaking migrants) spoke the minority language (i.e., their parents' first language) more frequently and obviously concomitant with greater exposure than those with parents in cross-cultural marriages.When examining the interplay between language use frequency and language attitudes, data revealed that the more positive attitudes were held towards the dominant language (i.e., Thai) and English than the minority language for families in which both parents were migrants (from the same countries of origin) in noncross-cultural marriages.This was the case in spite of the fact that the minority language was reported to being used most frequently in families in which both parents were migrant parents (non-cross-cultural marriages) in stark contrast to families with one migrant parent and one non-migrant parent (cross-cultural marriages).
Case 1 attributed the higher use frequency of the dominant language (i.e., Thai) and the more positive attitudes towards the dominant language (i.e., Thai) than the minority language (i.e., Chinese) due to the roles of parents for Chimlin, a five-year-old Chinese-Thai bilingual girl.In family A, Chimlin's dominant language-speaking Thai mother was largely responsible for taking care of her with minimal assistance from her minority languagespeaking migrant father, a Chinese speaker originally from Taiwan (Chimlin's father was preoccupied with duties at his workplace).A fairly similar child-care arrangement can also be observed in family D in which the dominant language-speaking (i.e., Thai) mother was the main caregiver for two daughters including Barsha, partly because the minority languagespeaking (i.e., Bengali) migrant father was absent from daily family routines.
Case 2, 3 and 5 demonstrated that bi-and multilingual children and adolescents spoke their respective minority languages more frequently than the dominant language (i.e., Thai) and English in the absence of any dominant language-speaking parents, partly because both minority language-speaking parents were migrants from the Philippines in family B, Bangladeshi in family C and Nigeria in family E, respectively.For instance, Jasmine, a female English-Thai-Tagalog multilingual adolescent from family B (both parents were migrants from the Philippines), reported that she most frequently spoke Tagalog.A similar pattern was observed in case 3. Sumon, a four-year-old Bengal-English-Thai multilingual boy from family C (both parents were migrants from Bangladesh), reported that he most frequently spoke Bengali.Even though they reported using their respective minority languages most frequently, thereby incurring greater exposure, Jasmine and Sumon still exhibited more positive attitudes towards English than Tagalog and Bengali, respectively.Their more positive attitude towards English was due to the prestigious status of English in Thailand.Their less positive attitude towards their minority languages were associated with the relatively less prestigious status of their minority languages.
Let us turn to Emmanuel, a 17-year-old male adolescent, for an analysis of the broad pattern in 1.Unlike Jasmine and Sumon who reported speaking the minority language most frequently, Emmanuel reported using American English more frequently than the minority language (i.e., Nigerian-accented English) and the dominant language (i.e., Thai), despite the fact that both of his parents were minority language (i.e., Nigerian-accented English)speaking migrants from Nigeria.Further, Emmanuel reported that Thai was the language he liked the most in comparison to American English, the language he most frequently used.It has thus become apparent that Nigerian-accented English was less positively regarded, due to the stigma attached to the Nigerian-accented English.
The overall picture, nonetheless, was that these bi-and multilingual children and adolescents (across family B, C and E) held more positive attitude towards the dominant language (i.e., Thai) and English than the minority language (namely, Tagalog, Bengali, and Nigerian-accented English, respectively).Not surprisingly, Jasmine regarded English more positively, since it was the language she liked the most and thought most useful in comparison to Tagalog, Thai, Chinese (Mandarin), Japanese and Korean.Since Sumon thought English was more useful, he exhibited more positive attitude towards English in invidious comparison to Bengali, Thai, Chinese (Mandarin), Japanese and Korean.
This suggested that language attitudes of bi-and multilingual children and adolescents were more complex and nuanced than expected.On the one hand, bi-and multilingual children and adolescents did not merely hold more positive attitudes towards the language to which they were most exposed and used most frequently.On the other hand, bi-and multilingual children and adolescents did not simply view less-frequently used languages less positively.One interpretation of this result was that some significant sociocultural variables (e.g., the status of the dominant vis-à-vis the non-dominant language) largely affected language attitudes rather than the highest degree of exposure and the most frequently used language.

SOCIAL ISOLATION AND INTRA-ETHNIC NETWORK
The consistent pattern of minority languages being more frequently used than English and the dominant language (i.e., Thai) can be partially explained by the relative social isolation of immigrant families with both parents as minority language-speaking migrants (non-crosscultural marriages).Jasmine from family B reported most frequently speaking Tagalog with her immediate family (both parents were minority language-speaking migrants) and members of the Filipino community rather than other languages.Due to dense intra-ethnic networks, Sumon from family C showed that he most frequently spoke Bengali with his Bangladeshi family (both parents were minority language-speaking migrants), relatives and family friends in the Bangladeshi community in lieu of other languages.However, a different pattern was observed in case 5. Emmanuel from family E contended that he spoke Nigerian-accented English less frequently than English and Thai with his Nigerian family (Emmanuel's mother spoke Nigerian-accented English to him), family friends from his mother's workplace, church members (family E attended a Nigerian church).In contrast, families in cross-cultural marriages (A and D) with one dominant-speaking parent (i.e., Thai-speaking non-migrant mothers) from the mainstream Thai society had children who spoke the dominant language (i.e., Thai) more frequently than any other language.

FAMILY STRUCTURES: GENDER ROLES OF PARENTS (EXPLANATIONS FOR BROAD PATTERNS 2 AND 3)
In this case, sociological factors were explanatory of the broad pattern 2, a combination of sub-patterns 2 and 5, i.e., English was more positively regarded than the dominant language in families in which parents were in non-cross-cultural marriages than those in cross-cultural marriages (e.g., case 1 where a Taiwanese-Chinese migrant husband married a Thai wife), where the dominant language (i.e., Thai) was more positively regarded than English.
Bi-and multilingual children and adolescents differed in their attitudes towards the dominant language (i.e., Thai) and English.Chimlin, a five-year-old girl, from family A, and Barsha, a five-year-old girl from family D, reported the dominant language (i.e., Thai) as the most useful (more useful than English), whereas Jasmine, a 13-year-old female adolescent from family B, and Sumon, a four-year-old boy from family C, claimed English to be the most useful language.To justify this claim, case 1-5 attributed the varied attitudes towards the dominant language (i.e., Thai) and English to family structures and traditional gender roles, particularly in regards to the mother's role (across families in non-or cross-cultural marriages).Both Chimlin from family A and Barsha from family D (in which a dominant language-speaking non-migrant mother from Thailand married a minority-speaking migrant father in a cross-cultural marriage) exhibited more positive attitudes towards the dominant language (i.e., Thai) rather than English.In contrast, Jasmine from family B and Sumon from family C (where both minority language-speaking parents were migrants) expressed more positive attitudes towards English than their counterparts from families A and D. In general, the mother was the caregiver in case 1-5 (families A-D).

ETHNIC MAJORITY GROUPS VERSUS ETHNIC MINORITY GROUPS
An alternative explanation emerged when considering the ethnic and cultural groups to which these bi-and multilingual children and adolescents belonged.It has become even more apparent that a dominant language (i.e., Thai)-speaking mother accounted largely for Chimlin (family A) and Barsha (family D) to consider themselves as members of the mainstream Thai group (as native Thais), even while in the presence of a minority language-speaking migrant father.In contrast, the apparent lack of a dominant language-speaking parent largely accounted for Jasmine (family B), Sumon (family C) and Emmanuel (family E) to consider themselves as outsiders in the host country as Filipino, Bengali and Nigerian, respectively, in the presence of both migrant parents from Philippines, Bangladeshi and Nigeria.Perhaps another possible explanation was the clear evidence derived from data that Jasmine, Sumon and Emmanuel turned to English for a sense of security in the absence of a dominant language-speaking parent.Yet, it became apparent that cross-cultural marriages-see the cases of family A and D, respectively and jointly, where one minority language-speaking migrant father married to a non-migrant majority language-speaking Thai mother-played a key role in influencing the extent to which the dominant language (i.e., Thai) was more highly regarded than English.However, the status of non-cross-cultural marriages in which both parents were minority language-speaking migrants affected the extent to which English was more higher regarded than the dominant language (i.e., Thai) by their children.
Nonetheless, family structures in relation to traditional gender roles in immigrant families cannot alone adequately account for language attitudes.Some complementary interpretations of the results involved the effects of the medium of instruction in formal language training.

EFFECTS OF MEDIUM OF INSTRUCTION AT SCHOOLS
In addition to family structures, considerable attention was paid to taking cognizance and analyzing the school experiences of bi-and multilingual children and adolescents.As active learners of English in formal school settings, Jasmine, Sumon and Emmanuel were enrolled in predominantly English-medium international schools (largely due to that they were foreigners with both migrant parents and were not allowed to enroll in Thai public schools) and displayed higher positive attitudes toward English than those who were non-active learners of English (i.e., Chimlin was enrolled in a predominantly Thai-medium school while Barsha was enrolled in a Thai-English bilingual school, because they were Thai citizens with non-migrant Thai mothers and migrant non-Thai fathers).The medium of instruction in formal schooling settings was a powerful force that did more than merely language use (communication) in the classroom setting, since it shaped the experiences of bi-and multilingual children and adolescents with English and Thai both inside and outside of the classroom.Those who were educated in English (i.e., Jasmine, Sumon and Emmanuel) were more likely to hold positive attitudes towards English outside the classroom than those who were educated in Thai or Thai-English (i.e., Chimlin and Barsha).In order to further build an alternative perspective and to develop an argument concerning the effects of the medium of formal language instruction, let us turn to family A for some clues.Despite the fact that Chimlin's migrant Chinese-speaking father communicated with her non-migrant Thaispeaking mother in English, case 1 demonstrated that she displayed only to a limited extent of understanding of English.Her Thai-medium schooling experiences in conjunction with her majority language-speaking (non-migrant) Thai mother were more pertinent to Chimlin's situation and for the present analysis.
Furthermore, considerable attention was paid to identifying and examining factors that contributed to relatively negative language attitudes towards the minority language adopted by bi-and multilingual children and adolescents.As a complement to the abovementioned factors such as the gender roles of parents, the present study attributed the less positive attitudes towards a migrant father's minority language to the development of bi-and multilingual children and adolescents' social relationships.

EFFECTS OF SOCIAL RELATIONSHIPS (EXPLANATIONS FOR THE BROAD PATTERN 3)
Case 1 and 4 revealed that there was a tendency for relatively less positive attitudes towards the minority language to be associated with bi-and multilingual children and adolescents with a minority language-speaking father and a dominant language-speaking mother from Thailand rather than by not having a dominant language-speaking mother.An important consideration was that the presence of a minority language-speaking migrant father in a family may not result in more positive attitudes towards the minority language by his child or children, due largely to the importance of the motherhood for child-caregiver relationship.In other words, children in these young ages were more attached to their mothers, hereby their mothers' language (i.e., Thai) was more highly regarded than their fathers' respective minority language.
The present study attributed this complex phenomenon to the development of social relationships (as explanations for the broad pattern 3, a combination of sub-patterns 3 and 6) of bi-and multilingual children and adolescents.Consider the participant Chimlin (family A) in case 1, for example.It has become apparent that the benefits of developing personal social relationships (i.e., the teacher-child relationship and peer relationships during the first two years of school) were based on Chimlin's Thai speaking ability rather than being based on her father's minority language.By the same token, the dominant language (i.e., Thai) enabled her to participate socially both inside and outside the classroom by virtue of her familial affiliation.She, nonetheless, claimed that Chinese (Mandarin) was not useful, partly because she only communicated in Chinese (Mandarin) with her minority language-speaking migrant father from Taiwan.Something similar occurred in regards to Barsha (family D).In contrast, there was a general lack of social support, participation and affiliation with Thai communities on the part of Sumon (family C) and Emmanuel (family E), both of whom having non-Thai migrant parents (non-cross-cultural marriages).

DISCUSSION
Overall, this paper was among the first to investigate the extent to which bi-and multilingual children and adolescents' language attitudes were influenced by family structures, particularly with regards to the parental type of non-and cross-cultural marriages.It might well be that the data did not replicate the data of previous studies.
In general, the result reported as the broad pattern 2 provided a more complex view of variables that contributed to language attitudes of bi-and multilingual children and adolescents, thereby indicating that previous studies may or may not be replicated.Across previous studies, the author found no evidence that factors such as family structures (i.e., non-and cross-cultural marriages) affected bi-and multilingual children and adolescents' language attitudes towards English and the dominant language.
Nevertheless, some previous results were replicated.The broad pattern 3 further confirmed numerous existing studies that have shown that minority languages have undergone language shifts and endangerment in immigrant communities across the globe (Fishman, 1980;Lee, 2014).Moreover, the present study confirmed previous studies (e.g., Dörnyei and Csizer, 2002) in regards to the effects of language instruction (i.e., positive attitudes were held by children towards languages used as the medium-of-instruction in their educational settings).Nevertheless, there remained questions regarding parental roles and social preferences versus language attitudes.The findings did not totally disconfirm the frequent observation made in the body of literature concerning parental roles in children's language learning and language use.When exploring late English-Korean bilingual adolescents in New Zealand, Kim and Starks (2010) argued that "the language use of Korean mothers is not associated with their children's patterns of language use or their children's L1 proficiency" (p.285).They went on to claim that "parental language use plays a minimal role in the L2 acquisition" (p.285).The present study was in agreement with Kim's data by virtue of having shown that just because bi-and multilingual children and adolescents had a longer exposure to their parental languages, this did not entail that they hold positive attitudes towards the languages used at home.On the one hand, the present study agreed in part with Kim's analysis of migrant fathers who played a minimal role in respect to the language practices of their children.On the other hand and contrary to Kim's assertions, the current study revealed that a non-migrant mother's dominant language played a more influential role than a migrant father's minority language in the development of language attitudes of bi-and multilingual children and adolescents.

SOCIAL PREFERENCES VERSUS LANGUAGE ATTITUDES
This might be a strong assentation to state that the literature dealing with social preferences no longer suffices in addressing the attitudinal patterns of bi-and multilingual children and adolescents.It might be due to familiarity that bi-and multilingual children and adolescents preferred to associate with minority language speakers who spoke one of the minority languages to which these children were exposed (e.g., Kinzler et al., 2012;Byers-Heinlein et al., 2016).Regardless of social preferences, the present study warned us that bi-and multilingual children and adolescents may not evince prohibitive attitudes toward these minority languages.

CONCLUSION AND IMPLICATIONS
In sum, language attitudes of bi-and multilingual children and adolescents were influenced to a large extent by familial structures (i.e., those with or those without cross-cultural marriages), and, to a lesser extent, relative exposure, i.e., the frequency of use of each language.The factor of familial structure in relation to the gender roles of parents carried high explanatory value, since it exerted a much greater degree of influences on children and adolescents in contrast to communities and societies.
Simply to believe that bi-and multilingual children and adolescents who had greater exposure to their minority language-speaking fathers were likely to develop positive attitudes towards their fathers' minority language would be mistaken.Across existing studies, the effects of minority-language-speaking migrant fathers on language attitudes of bi-and multilingual children and adolescents received little recognition.To provide recommendations for future researchers, a minority language-speaking parent's role (particularly a migrant father's role) on language attitudes of bi-and multilingual children and adolescents deserves a further investigation.
What implications for the fields of bi-and multilingualism and language attitudes emerged from the voices of bi-and multilingual children and adolescents in the current study?In recent years, studies on bi-and multilingual children and adolescents and bi-and multilingualism at both the family and the country level have increased and continue to attract the attention of policy makers, researchers and parents.Despite these efforts, the success of bi-and multilingual children and adolescents remained a crucial challenge.The findings of the current study can inform policy makers, researchers, teachers, educators, practitioners and parents in regards to how to better planning, policing, researching and teaching bi-and multilingual children and adolescents.
These findings were strong indicators that issues that were identified by immigrant families in both non-and cross-cultural marriages regarding children's language attitudes should not be treated separately.This was partly because the variables that were addressed were inextricably linked with one other.
It has become apparent that the complex interplay and interaction of factors among types of immigrant family, types of cross-cultural marriage, parents' first languages, caregiver (mother)-child relationships, the dominant versus non-dominant parental language in the face of English, the effects of formal language instruction at schools, and largely monolingual or largely bilingual communities were factors of paramount concern for bi-and multilingual children and adolescents when their language attitudes emerged.
For bi-and multilingual children and adolescents who were raised in predominantly monolingual minority language families with both migrant parents (non-cross-cultural marriages), they were in need of language-support programs for learning the dominant (majority) language.However, for bi-and multilingual children and adolescents who were raised in predominantly bi-and multilingual families whether or not they were typical oneparent-one-language families (one migrant parent versus one non-migrant parent in a crosscultural marriage), they were in need of language-support programs to learn and maintain the minority language of their migrant parents.In a predominantly L2 monolingual environment (such as Thailand), minority language-speaking parents' attempt to forge a minority language environment at home remained a crucial challenge.Yet, increasing positive attitudes towards the minority language was preferred and supported by minority language-speaking parents (e.g., Chimlin's father) for family language planning and practice in the long term.

TABLE 1 .
Broad patterns (combinations of sub-patterns) of language attitudes across families in and not in cross-cultural marriages

TABLE 2 .
Sub-patterns of language attitudes of bi-/multilingual children and adolescents when comparing different types of family structures (parents in non-or cross-cultural marriages)