Intergenerational differences in family language policy of Turkish families in the Netherlands

ABSTRACT

By using a quantitative approach, we examined the family language policy of Turkish parents in the Netherlands. Based on our earlier ethnographic work with 20 Turkish families and interviews with 35 parents, we designed and conducted a large-scale survey among 116 first-generation and 184 second-generation parents. The findings reveal that both first- and second-generation parents have bilingual language ideologies. Although second-generation parents are more positive about Dutch-oriented language practices of family members, both groups believe that Turkish should be a part of family life. Because parents establish a link between Turkish identity and culture, parents from both generations show strong attachment to Turkish language. Regarding language practices, both generations’ language preferences shift towards bilingualism although the first-generation choose to speak Turkish more. Second-generation parents and their third-generation children are more dominant in Dutch but they all report a preference for Turkish in the domestic domain. Regarding language management in the family, different generations of parents show different engagement strategies. With respect to the host society orientations to heritage language maintenance, the informants report unfavourable conditions in the Dutch media, politics and public institutions.

KEYWORDS:

• Dutch policy
• family language policy
• heritage language
• intergenerational differences
• language maintenance
• Turkish immigrants

Introduction

Within the conceptual framework of language maintenance and shift, we focus on the family language policy of Turkish families in the Netherlands. Using a survey, we explore language ideologies, practices, management of first- and second-generation Turkish parents.

The Turkish community has been of interest for scholars in the field of sociolinguistics for many years because they are the largest immigrant group in the Netherlands and have the highest heritage language maintenance (Backus 2013). Besides various research on bilingual skills (e.g. Leseman 2000), first language skills (e.g. Akoğlu and Yağmur 2016), language maintenance and shift (e.g. Yağmur 2009) and language change (e.g. Demirçay 2017), recently there has been more focus on the relationship between family dynamics and children’s bilingual upbringing in the home context (see, for instance, Aarts et al. 2016 and Scheele, Leseman, and Mayo 2010 for language use at home and for language proficiency of children). With the idea that language practices in an immigrant family cannot be divorced from the sociocultural context, we started our investigation of family language policy among Turkish families. Our initial aim was to understand linguistic interactions among second-generation families with parents who grew up mostly in monolingual Turkish family settings and became bilingual through their interactions in Dutch society. With this aim in mind, we initiated observations and in-depth interviews in 20 families. We initially observed linguistic interactions among family members, framing a general picture of families’ language practices and management, as well as micro and macro factors influencing them. Field notes, audio records and transcriptions helped us to explore recurrent themes and ideas systematically, and to design and conduct interviews with parents on their FLP and with teachers on their experiences and suggestions on working with children of immigrant families (see Bezcioğlu-Göktolga and Yağmur 2018a; 2018b for a detailed overview). Our findings showed that each family was unique and FLP dynamics of each family member was diverse. However, there were still shared patterns of language ideologies, practices and management. Observing these shared patterns during our qualitative research, we aimed to dig deeper by collecting large-scale data by means of a survey questionnaire.

There are a number of studies that have used quantitative research methodologies in their investigation of FLP among various bilingual and multilingual communities, including Korean families in the US (Kang 2015), Iranian families in the US (Kaveh 2018), and and Russian families in Israel (Schwartz 2008). However, there has not been a survey study on Turkish FLP in the Netherlands. With the aim of filling this gap, in the current paper, we present the findings of our survey among 300 first- and second-generation parents.

To contextualise our research better, we provide the theoretical framework for FLP studies and the sociolinguistics background of the Turkish community in the Netherlands in the following sections.

Family language policy

Family language policy is about implicit and explicit language planning in the family (Curdt-Christiansen 2009). It involves family members’ linguistic ideologies, practices and management in relation to language preferences and literacy practices (Curdt-Christiansen 2009; King, Fogle, and Logan-Terry 2008; Spolsky 2004, 2007).

Besides its intersection with multiple academic disciplines, including child language acquisition and language socialisation (Curdt-Christiansen 2018; King, Fogle, and Logan-Terry 2008; Schwartz and Verschik 2013), research on FLP is frequently informed by Fishman’s Reversing Language Shift Model (Fishman 1991), since much of the research on FLP examines the efforts of the families to maintain an ethnic minority language in the family as well as in the wider community (Schwartz and Verschik 2013). As De Houwer (2007) points out, children growing up bilingual speak the majority language one way or another, but if extra attention is not given to it, it becomes a big challenge for them to learn to speak the heritage language. The interaction between the primary caregiver and the child is in the centre of language maintenance and transmission (Fishman 1991). Such interaction mainly takes place within the family environment, so Fishman’s model has an undeniable place in FLP studies.

Keeping the contributions of language maintenance and shift theories and models mentioned above in mind, we draw on the language policy model of Spolsky (2004, 2007). Spolsky (2004, 2007) defines three components of language policy: language ideologies, language practices and language management. Language ideologies are about beliefs regarding languages and language use. Language practices are about what people choose and use among the varieties of their linguistic repertoire. Language management is any kind of effort and intervention to effect actual language practices. These components influence and are influenced by various intra-family and societal factors. Taking an immigrant family as an example, family members have their own beliefs and practices for home language use. These beliefs and practices are closely affected by their home environment, parents’ background and economic resources. Despite these certain norms and background for language use, family is not an isolated unit from society. Children go to school, interact in another language; parents go to work, they need to adapt themselves to socialise and to be competent in this outside world. This process of language socialisation brings the sociolinguistic, socioeconomic, sociocultural and sociopolitical environment into the home domain, having a powerful impact on language ideologies, practices and management of a family.

Family is ‘the most common and inescapable basis of mother tongue transmission’ (Fishman 1991, 94). How this transmission is realised and FLP is followed in the home domain can be explained through various micro factors. These factors are basically intra-family dynamics, including parents’ expectations, parents’ language experience, and parental knowledge of bilingualism. Micro dynamics within families are undeniable in forming the FLP (Curdt-Christiansen 2009, 2016), since parents’ perceptions and practices play an important role in determining the attitudes of children towards the immigrant language as well as their bilingual development (De Houwer 2007; King and Fogle 2006; Leung and Uchikoshi 2012; Liu and Lin 2019; Spolsky 2012). The outer world, or ‘macro forces’, is as significant as the micro family factors. They include political, sociocultural, economic aspects and sociolinguistic environment of the families. Curdt-Christiansen (2009) describes political factors as an influence on individual’s opportunities for education, civil activities as well as political decisions. Economic factors are about the connection between languages and economy, namely how economic power influences the practice of a specific language. Cultural factors are about the representation of symbolic values through language. Political decisions at macro-level regarding language policy, and especially decisions regarding education policy, can have a profound impact on the ideologies formed and decisions made by family members regarding their everyday language choices and practice (Spolsky 2004).

A quantitative look into FLP studies

Various research investigated language ideology, practice and management in the FLP framework. While the majority of the studies used qualitative methodologies (e.g. Curdt-Christiansen 2009; Kopeliovich 2010), some others benefitted from mixed-methods and quantitative approaches. For instance, Kaveh (2018) utilised a mixed-methods approach in her investigation of first-generation Iranian immigrant parents’ FLP and maintenance of Persian through Spolsky’ language tripartite model. In-depth interviews with 7 parents and a survey with 10 families revealed that parents have strong bilingual orientations for their children. Persian is the language of communication in their families during daily conversations and family gatherings, with occasional switches to English when needed. On the other hand, not all of the families have planned language management strategies apart from a few daily language strategies. Similarly, Kang (2015) examined the FLP of Korean American families in the US through a survey with 480 parents with children under 18 years old. The survey included questions on language practices during interaction with family members and discussions on various topics, language ideologies of parents on various aspects of bilingualism, language management for Korean and English, and parental assessment of children’s Korean language skills. The results indicated that parents believe parental use of Korean is essential for children’s bilingual development, while the use of English is not an essential part of English acquisition in the home settings. Besides, while Korean is the primary language of communication between adults, a mix of Korean and English becomes dominant when children are involved in the conversation, especially on school-related topics. Regarding language management, although more than half of the participating parents reported that they have Korean books for children at home, and they used both Korean and English while reading to their children, the majority of the children watch TV and read for pleasure in English. In the same line with this, Ó hIfearnáin (2013) benefitted from a mixed-methods approach in his research with Gaeltacht Irish speakers in Ireland. He triangulated a variety of data collection methods, including a survey, interviews and discussions, and demonstrated that no matter how devoted parents are for intergenerational transmission of Irish, the presence of English as the powerful language within and outside the family had a detrimental impact on Irish transmission.

In the second-generation immigrant families context, Schwartz (2008) examined the family policy factors affecting first language (L1) maintenance among 70 Russian–Hebrew-speaking children in Israel. She revealed that all of the participants supported the multilingualism of Israel as well as promoting intergenerational language transmission. Further, second-generation parents used mainly Russian in interaction with their spouses and parents, but inserted Hebrew in their conversations with children depending on linguistic circumstances. Language practices in the home domain included both Russian and Hebrew. Children’s reports also corresponded with parental practices. Children preferred to speak Russian more with their parents and grandparents while they favoured Hebrew over Russian in daily life.

All of these studies mentioned above show that there are inconsistencies between parental language ideologies, practices and management. Intended language ideologies of parents are not always realised in family language practices due to many reasons including different language preferences of each family member, children’s language socialisation, lack of context-related vocabulary, the power of the mainstream languages and the value given to the home language by the mainstream institutions. In order to understand how these are experienced by the target population of this study, in the following section, we provide an overview of the sociolinguistic characteristics of first- and second-generation parents in the Netherlands.

Sociolinguistic characteristics of first- and second-generation Turkish parents

The history of the Turkish immigrants in the Netherlands is similar to the other countries in Western Europe. The shortage of workforce after the Second World War resulted in the arrival of labour migrants in the Netherlands beginning from 1964 (Akgündüz 2007). Although it was initially a well-planned migration of a temporary nature, increasing economic opportunities caused a settlement of a big Turkish community in the host country. Currently, after Germany and France, the Netherlands hosts the third largest Turkish group (Yağmur 2015). The Turkish-origin population is around 400.000 (CBS 2017). Around 191.000 people in this population are first-generation, and 207.000 second-generation (but of these, around 45.000 have only one parent who was born abroad, while for 160.000, both parents were born abroad). There is also an emerging third-generation (with both parents born in the Netherlands), but CBS has no data on the size of this group. Besides, as Backus (2013) asserts, although there is a settled second and third-generation, the search for spouses in Turkey has until very recently brought about a continuous inflow of first-generation Turkish speaking immigrants.

The early first generation, who arrived in the Netherlands in the mid-1960s lacked sufficient Dutch skills, and their social and linguistic contact was mainly within their own group (Smets and Kreuk 2008). Besides, they had low levels of education and limited access to Dutch classes. The temporary nature of their stay also impacted their devotion to learning Dutch and emphasising Dutch education for their children (Akgündüz 2007; Yılmaz and Schmid 2015). The characteristics of today’s first-generation, who also constitute the participant group in this study, are different from the firstcomers mentioned above. They moved to the Netherlands as a marriage partner to a someone of the second-generation, since most of the Turkish people in the Netherlands select their partners in Turkey (Lucassen and Laarman 2009; Extra and Yağmur 2010). The current first-generation have much higher levels of education compared to earlier generations, and start to follow Dutch classes as soon as they arrive in the Netherlands.

Second-generation parents are bilingual. They come mostly from low socio-economic backgrounds with a lack of rich and elaborate linguistic input from their parents (Leseman and van den Boom 1999). They grew up in disadvantageous conditions due to lower family income than European standards and unfavourable accommodations (Crul and Vermeulen 2003), and mostly attended schools where the majority of children had an immigration background. Mainstream education did little for the sociolinguistic integration of immigrant children until 1990s, which resulted in low levels of placement in higher secondary education, falling behind their peers (Aarts and Verhoeven 1999; Leseman 2000; Verhallen and Schoonen 1993). Still, as they grew up in the Netherlands and studied in Dutch schools, they are fluent speakers of Dutch (Doğruöz 2007). Turkish maintenance is also quite high (Backus 2013; Extra and Yağmur 2010). Many factors including Turkish dominant households until school age (Leseman 2000), close social contact with Turks in the Netherlands and in Turkey, presence of Turkish media in the home, and holidays in Turkey contribute to continuing language maintenance.

Various studies focus on the intergenerational developments in language practices of the Turkish community in the Netherlands. For example, Eversteijn (2011) showed that the younger generation tends to prefer Dutch when talking to each other while they use Turkish to address their parents and the older generation in general. Similarly, Yağmur’s (2016) data showed that although the second generation is more proficient in Dutch than the first, using Turkish is quite common for both generations. The second generation uses Turkish more in the domestic domain and Dutch more in the public domain. A study by Extra and Yağmur (2004) revealed that use of Turkish was considered the most significant indicator of ethnic identity, and the second and even the third generation show strong motivation to practice Turkish in their daily lives, although Dutch often becomes the dominant language after several years at school. Last but not least, Sevinç (2014, 2016) indicated that although first-generation participants in her study prefer Turkish, second and third-generation participants can express themselves better in Dutch than in Turkish and use mainly Dutch except in the domestic domain. All in all, regarding language practices, although the use of Turkish in various domains of daily life is very common in the second and third generations, younger generations have a preference for the majority language Dutch over Turkish (Extra and Yağmur 2010).

From an FLP perspective, our ethnographic study (Bezcioğlu-Göktolga and Yağmur 2018a, 2018b) revealed that in second-generation families, some of which include a first-generation spouse, Turkish maintenance is a very important part of family language ideologies. However, their language practices and management are quite diverse and complex. Language practices differ according to the interlocutors, the context of the conversations and language preferences of family members. Families have various language management strategies based on their needs and expectations. Besides, all of the families focus on the educational achievement of their children in school, so teachers have a significant agency on FLP. However, pro-Turkish parental ideologies are not always appreciated by teachers, since they believe immigrant families should put more emphasis on Dutch in the home context when children start school.

Such teachers’ ideologies are in line with the general language policy for immigrant languages and families in the Netherlands. Although home languages were included in the curriculum in the mainstream primary and secondary schools between 1974 and 2004, these classes were abolished when the integration policy changed in 2004. Indeed, the changes in the integration policy in the Netherlands have recently been described as ‘from diversity celebration to a color-blind approach’ by Stevens and his colleagues (2019). Like many other scholars (Driessen and Merry 2011; Extra and Yağmur 2006; Vasta 2007), they mean the shift in the mainstream policies that put emphasis on assimilation over multiculturalism. In the present study, we aim to increase our understanding of the Turkish families’ FLP as well as factors influencing these policies by adding a quantitative approach to the abovementioned studies.

Methodology

Given the sociolinguistic picture, the aim of the current study is to explore family language policy among first- and second-generation Turkish parents in the Netherlands. We present the findings of a survey designed based on data derived from ethnographic research with 20 Turkish families in the Netherlands by testing the following hypotheses:

1. First- and second-generation parents hold positive language ideologies towards Turkish maintenance in the family.

2. First-generation parents’ language practices are more Turkish oriented than second-generation parents.

3. First-generation parents are more involved in language management than second-generation parents.

Participants

Three hundred respondents filled in the survey. Among them, 116 (39%) were first-generation, and the remaining 184 (61%) were second-generation. We approached participants through posts on social media groups, contact with prominent Turks in different regions of the Netherlands, and contact with several Turkish organisations such as mosques and Turkish foundations. Respondents completed the questionnaire through a hard copy or online (using Qualtrics software offered by Tilburg University). It took approximately 15–20 minutes to fill out the questionnaire.

Table 1 presents the distribution of gender, marital status, age, and the level of education of the participants.

Table 1. Gender, marital Status, age and level of education.

Characteristic	n	%
Gender
Female	182	60.7
Male	118	39.3
Marital Status
Married	263	87.7
Single (Parent)	37	12.3
Age (in years)
≤25	2	.7
26–35	68	22.7
36–45	158	52.7
46–55	63	21
≥56	9	3
Education – ISCED
Primary	10	3.3
Lower secondary	28	9.3
Upper secondary	111	37
University and/or higher	151	50.3

As for marital status, 37 of the respondents are currently single; however, they have children from previous marriages with a Turkish spouse. Regarding the age groups, the majority of the participants (n = 158, 52.7%) were between 36 and 45 years old. The mothers of all participants (N = 300, 100%) and the fathers of 296 of them (98.7%) were born in Turkey.

The first language of the majority of participants (n = 292, 97%) was Turkish.¹ Similarly, most (n = 213, 71%) state that their dominant language is also Turkish.² The majority of the participants (n = 189, 63%) had at least one child in primary school, 85 parents (28%) had children older than primary school age and 19 parents (6%) had children younger than primary school age.

Besides the personal characteristics presented above, participants were asked about the frequency of their visits to Turkey. They are in close contact with Turkey. More than half of them (n = 172, 57.3%) stated that they visit Turkey every year, 58 of them (19.3%) visited twice a year or more, 53 participants (17.7%) once every 2 years, 10 participants (3.3%) once every 3 years and only the remaining 7 participants (2.3%) paid a visit to Turkey once per 3 years or more.

Instrument

The resources for the design of the questionnaire were earlier observations with 20 families and interviews with 35 parents and relevant literature. The initial design of the survey was checked by two field experts, piloted with potential respondents, reviewed and revised again by intensive work with two professors in the field of sociolinguistics. The instrument was implemented in Turkish or Dutch, based on the preference of the respondent.

The demographics section consisted of 18 questions on participants’ gender, age, marital status, education, birth country of themselves, their partners and parents, the length of residence in the Netherlands, their dominant language, number of children and the frequency of visits to Turkey.

Following this, three main sections intended to elicit information on participants’ language ideology, language practice and language management.

Language ideology

The language ideology section sought answers on participants’ beliefs and attitudes towards the use of Turkish and Dutch within and outside the family. The initial sub-sections assessed attitudes towards language use among family members. In the section on the beliefs about language preferences of family members, participants were asked about the language they feel more comfortable speaking at home, the language they prefer their children to each other, inside and outside home, and the language they prefer their spouse speak to them and their children (6 items).

In the following sub-sections, participants were asked to state their level of agreement on pre-formulated statements. The sub-section on the beliefs about children’s use of Dutch asked if the participants feel comfortable if their children speak only Dutch to them, to each other and in the presence of Turkish friends and relatives, if they mix Turkish and Dutch and if they respond in Dutch to a Turkish statement (5 items). In the sub-section on beliefs about children’s use of Turkish, participants stated their level of agreement on the following statement: Children should speak only Turkish to each other at home, parents should speak only Turkish to their children before they start school, speaking Turkish in the family helps children’s cognitive growth, and Turkish language is a condition for the learning of Dutch (4 items). The following sub-section was on beliefs about language use in the family and consisted of statements on children’s use of Dutch to the participant, children’s use of Dutch before school age, and parents’ use of only Dutch in the presence of their children (4 items).

The beliefs on bilingualism sub-section asked participants to state their level of agreement on children’s bilingual achievement through the following statements: Children’s bilingual achievement depends on having sufficient exposure to both Turkish and Dutch, having sufficient interaction with the speakers of both languages, knowing a sufficient number of words in both languages, acquiring Turkish in a dominant manner before school age, and acquiring both languages from infancy (5 items). Similarly, the 7-item-scale on beliefs about school achievement asked participants’ level of agreement on parental support for children in mastering Dutch skills and homework, learning Dutch or acquiring Turkish before school age, the availability of Dutch books at home, children having better skills in Dutch than Turkish, and speaking only Dutch with children (7 items).

In the next sub-section, participants stated which language is more important to fulfil various tasks in the Netherlands: to make friends, to make money, to study, to find a job, to get better education, to live in the Netherlands, to have a say in mainstream society, to raise children, to be accepted in the Turkish community, to speak to Turkish friends, to be accepted by Dutch people and to speak to colleagues (12 items). Following this, participants stated their level of agreement on reasons for language maintenance, through statements including preserving Turkish identity, religion, mother tongue, culture and having better contact with the Turkish in the Netherlands and in Turkey (6 items).

The last 2 sub-sections were on schools and the society’s attitudes towards Turkish. The former one consisted of 5 statements including if children and parents are allowed to speak Turkish to each other in the schoolyard, if teachers value the bilingualism of Turkish children at school and if teachers appreciate the role of Turkish in learning Dutch. The latter consisted of 6 items including if participants agree that healthcare centres appreciate the use of Turkish in the family, common public opinion respects Turkish language, Dutch authorities value Turkish use at home and support Turkish maintenance, Dutch society respects Turkish culture, and the media supports the use of only Dutch at home.

Language practices

Language practices in and outside the family were measured through questions using a 5-point Likert scale. First, participants were asked which language they generally use with their parents, siblings, spouse, children, Turkish friends and relatives, on the phone and in Turkish workplaces (11 items). In the second sub-section, similar questions were asked about the language use of their children (6 items: language use with the respondent, the spouse, siblings, grandparents, relatives and Turkish friends). The third sub-section looked into the language use of children, both before they started school and after they had started school (2 items).

The following 2 sub-sections measured the language preferences of participants in various aspects of their daily lives. In one section, comprising 8 items, participants were asked which language they generally chose for daily activities, including thinking, dreaming, counting and calculating, using social media, texting messages, reading on the internet, watching TV and listening to music. The final sub-section asked more detailed questions about language choice on social media such as which language participants use on the social media pages or groups they follow, and when writing comments on posts, contacting friends and relatives and looking for information (5 items).

Language management

In this section, participants were asked what kind of language management activities they are involved in to improve the Turkish and Dutch skills of their children inside and outside the home context. In a 29-item scale, participants stated how frequently they are involved in the following language management activities in Turkish and in Dutch: reading books, reading picture books, listening to songs and singing songs, watching a movie, watching TV, teaching children how to read and write, correcting children when they mispronounce a word, looking for online resources, helping children with schoolwork, teaching them vocabulary, correcting children when they mix languages, extra-curricular language schools, and involvement in children’s school related activities.

Results

Before the main analysis, we checked each scale to ensure reliability (Table 2). The internal consistencies of three scales were above .90, 9 of them were above .80, two of them were above .70, and the remaining three of them were above .60. (Note that the internal consistency of the scale on the school’s perceived attitudes towards the use of Turkish was initially .58 for 6 items, but one item was found to be problematic, so it was deleted. The reliability score for the remaining 5 items was acceptable.)

Table 2. Scale characteristics.

Scale	Number of items	M (SD)	Α
Language choice of the participant	11	3.95 (.62)	.87
Language choice of the children	6	3.25 (.89)	.86
Language use before/after school	2	3.49 (1.01)	.77
Language dominance	8	3.46 (.82)	.90
Social media language use	5	3.39 (.77)	.82
Beliefs about language preferences of family members	6	3.39 (.78)	.86
Beliefs about children’s use of Dutch	5	2.81 (1.08)	.86
Beliefs about children’s use of Turkish	4	3.24 (.96)	.74
Beliefs about language use in the family	4	2.73 (.73)	.62
Beliefs about bilingualism	5	3.93 (.75)	.81
Beliefs about school achievement of children	7	2.57 (.52)	.68
Language importance	12	2.31 (.42)	.82
Reasons for language maintenance	6	4.04 (.84)	.89
School’s attitude towards Turkish	5	3.21 (.72)	.64
Society’s attitudes towards Turkish	6	2.50 (.80)	.89
Language management Turkish and Dutch	29	3.28 (.62)	.91
Language management (mother vs father)	10	2.52 (.78)	.91

Note: M refers to the mean value in a scale of 5, in which 1 refers to only Dutch, 2 more Dutch than Turkish, 3 both Dutch and Turkish are equal, 4 more Turkish than Dutch, 5 only Turkish. The values below 3 indicate higher Dutch use while values above 3 indicated more Turkish use, values, practices and so on.

Besides checking the descriptive statistics for each scale, we conducted a t-test analysis to see if there are statistically significant differences between groups.

Language ideology

In this section, we first provide a short description of the participants’ beliefs on home language use, children’s language practices in the family, and their beliefs on the language ideologies of school and the mainstream society. Parents have equally positive attitudes towards the use of Turkish and Dutch, which might be an indication of bilingual ideologies and a belief that the presence of both Turkish and Dutch in the family is a good thing, yet first-generation participants feel more comfortable speaking Turkish at home while the second-generation participants feel comfortable speaking both Turkish and Dutch to a similar extent (Table 3).

Table 3. Language Ideologies.

		N	Mean	SD	t	p
Beliefs about language preferences of family members	FG	116	3.59	.89	3.64	.000*
	SG	184	3.26	.67
Beliefs about children’s use of Dutch	FG	116	2.86	1.16	0.73	.463
	SG	184	2.78	1.03
Beliefs about children’ use of Turkish	FG	116	3.29	1.08	0.66	.504
	SG	184	3.21	0.87
Beliefs about language use in the family	FG	116	3.19	0.79	−1.33	.184
	SG	184	3.30	0.68
Beliefs on bilingualism	FG	116	3.97	0.76	0.89	.372
	SG	184	3.89	0.75
Beliefs on school achievement – reversed	FG	116	2.54	0.57	−0.88	.377
	SG	184	2.59	0.49
Language Importance	FG	116	2.25	0.38	−1.84	.065
	SG	184	2.34	0.44
Reason for language maintenance	FG	116	4.05	0.92	0.79	.937
	SG	184	4.04	0.78
School’s attitude towards Turkish	FG	116	3.17	0.79	−0.77	.437
	SG	184	3.23	0.67
Society’s attitudes towards Turkish	FG	116	2.47	.81	−0.59	.554
	SG	184	2.53	.80

Note: FG refers to first-generation while SG refers to second generation. M refers to the mean value in a scale of 5, in which 1 refers to only Dutch, 2 more Dutch than Turkish, 3 both Dutch and Turkish are equal, 4 more Turkish than Dutch, 5 only Turkish. The values below 3 indicate higher Dutch use while values above 3 indicated more Turkish use, values, practices and so on.

Table 3 reveals that except for the language ideologies of participants regarding language preferences of family members, there are no statistically significant differences between the first- and second-generation participants on their language ideologies. For the first sub-scale, looking at the mean values, we see that the first-generation participants are closer to ‘mostly Turkish’ language ideologies compared to the second-generation. To provide more details on the rest of the sub-scales, the table shows that participants from both generations believe neither only Dutch use nor only Turkish use is common in the family. Indeed, looking at the mean values of their beliefs about children’s use of Dutch or Turkish at home, they have more negative attitudes towards the use of only Dutch than only Turkish. Furthermore, both groups have similar beliefs on the bilingual achievement of their children. They agree that children need to acquire Turkish in a dominant manner before school age (FG: M = 3.48, SD = 1.28; SG: M = 3.64, SD = 1.12), they should have sufficient interaction with speakers of both Turkish and Dutch (FG: M = 4.08, SD = .97; SG: M = 4.08, SD = .85), and know a sufficient number of words in both languages (FG: M = 4.16, SD = .98; SG: M = 4.05; SD = .88).

Regarding the importance of languages, although the majority of the participants believe that both Turkish and Dutch are equally important for making friends (FG: M = 2.87, SD = .50; SG: M = 2.74, SD = .56), raising children (FG: M = 2.86, SD = .57; SG: M = 2.82, SD = .58), being accepted by the Turkish community (FG: M = 3.27, SD = .72; SG: M = 3.18, SD = .54), and speaking to Turkish friends (FG: M = 3.58, SD = .68; SG: M = 3.18, SD = .52), they consider Dutch to be more important for other variables such as to have a say in the mainstream society (FG: M = 1.71, SD = .73; SG: M = 2.11, SD = .76), to earn a living (FG: M = 1.99, SD = .67; SG: M = 2.21, SD = .69) and to get a good education (FG: M = 1.57, SD = .70; SG: M = 1.84, SD.79).

The highest mean values in the language ideologies scale are the items in the sub-scale on the reasons for language maintenance. Participants place higher values on Turkish maintenance to preserve their Turkish identity, culture, and religion, to protect the status of Turkish as a mother tongue, as well as for a better connection with Turkey and with other Turkish people in the Netherlands (Table 4).

Table 4. Reasons for language maintenance.

It is important to maintain Turkish because …		Total		Descriptives
	Gen	N	f	M	SD
We can preserve Turkish identity.	FG	116	100	4.07	1.14
	SG	184	100	4.02	1.06
We can preserve our mother tongue.	FG	116	100	4.32	.96
	SG	184	100	4.25	.85
We can establish better contact with Turkey.	FG	116	100	4.15	1.06
	SG	184	100	4.24	.84
We can contact other Turkish speakers in the NL.	FG	116	100	4.02	1.06
	SG	184	100	3.84	1.06
We can preserve Turkish culture.	FG	116	100	4.16	1.05
	SG	184	100	4.19	.83
We can preserve our religion.	FG	116	100	3.61	1.40
	SG	184	100	3.73	1.21

Note: M refers to the mean value in a scale of 5, in which 1 refers to ‘do not agree at all’, while 5 refers to ‘fully agree’.

In our earlier ethnographic study, we saw that school and mainstream society exert influence on the construction of the language ideologies of the parents. The findings of the present study show that both first- and second-generation parents have similar ideologies regarding the attitudes of schools (Table 5) and the mainstream society (Table 6) towards Turkish. Around half of the participants state that children are not allowed to speak Turkish among each other in the school yard, but parents are permitted to use Turkish. Besides, participants are undecided about whether teachers value the bilingualism of Turkish children at school and whether they appreciate the role of knowing Turkish in learning Dutch. On the other hand, participants believe that the members of the mainstream society in general have negative attitudes towards the use of Turkish. They do not agree with the statements that common public opinion respects Turkish language and culture, and that Dutch authorities and media value and support the maintenance and use of Turkish at home.

Table 5. School’s attitudes towards Turkish.

		Total		Descriptives
	Gen	n	f	M	SD
Children are allowed to speak Turkish among each other in the schoolyard.	FG	116	100	2.67	1.20
	SG	184	100	2.95	1.17
Parents are allowed to speak Turkish among each other in the schoolyard.	FG	116	100	3.44	1.17
	SG	184	100	3.52	1.08
Parents are allowed to speak Turkish to their children in the schoolyard.	FG	116	100	3.42	1.23
	SG	184	100	3.43	1.09
Teachers value the bilingualism of Turkish children at school.	FG	116	100	3.11	1.06
	SG	184	100	3.22	1.01
Teachers appreciate the role of Turkish in learning Dutch.	FG	116	100	3.22	1.18
	SG	184	100	3.07	1.00

Note: M refers to the mean value in a scale of 5, in which 1 refers to ‘do not agree at all’, while 5 refers to ‘fully agree’.

Table 6. Society’s attitudes towards Turkish.

		Total		Descriptives
	Gen	n	f	M	SD
Healthcare centres appreciate the use of Turkish in the family.	FG	116	100	3.02	1.15
	SG	184	100	3.16	1.05
Common public opinion respects Turkish language.	FG	116	100	2.35	.90
	SG	184	100	2.44	1.02
Dutch authorities value the use of Turkish at home.	FG	116	100	2.49	1.06
	SG	184	100	2.49	.98
Dutch authorities support Turkish maintenance.	FG	116	100	2.19	1.00
	SG	184	100	2.25	.96
Dutch society respects Turkish culture.	FG	116	100	2.73	1.06
	SG	184	100	2.67	1.04
The media supports the use of Turkish at home.	FG	116	100	2.06	.87
	SG	184	100	2.17	.93

Note: M refers to the mean value in a scale of 5, in which 1 refers to ‘do not agree at all’, while 5 refers to ‘fully agree’.

Language practices

Concerning the language choice of the participants with their parents, children, siblings and spouses, both first- and second-generation participants are inclined to always use Turkish with their mothers (FG: n = 114, M = 4.94, SD = .27; SG: n = 184, M = 4.71, SD = .62) and fathers (FG: n = 109, M = 4.84, SD = .59; SG: n = 180, M = 4.66, SD = .61). However, based on the descriptive statistics, it seems that the second-generation participants more often choose both Turkish and Dutch in interaction with their siblings (n = 182, M = 3.10, SD = 1.03) and spouses (n = 112, M = 3.58, SD = 1.17) than the first-generation participants (siblings: n = 108, M = 4.74, SD = .57; spouses n = 106, M = 4.11, SD = 1.06). Both groups are inclined to choose both Turkish and Dutch equally in interaction with their oldest (FG: n = 112, M = 3.58, SD = 1.17; SG: n = 176, M = 2.79, SD = .94) and youngest children (FG: n = 100, M = 3.53, SD = 1.20; SG: n = 164, M = 2.95, SD = 1.04). Children, on the other hand, have different language practices. The majority of the children of the first-generation (n = 105, M = 4.57, SD = .79) and the second-generation (n = 178, M = 4.29, SD = .91) are reported to speak only Turkish to their grandparents, but they prefer to speak mostly Dutch with their siblings (FG: n = 105, M = 2.83, SD = 1.30, SG: n = 167, M = 2.59, SD = 1.15). Besides, while the first- (n = 113, M = 4.23, SD = 1.04) and the second- generation parents (n = 177, M = 375, SD = 1.21) state that their children always speak Turkish to them before school age, children are more oriented to speaking Turkish and Dutch after they started school (FG: n = 110, M = 3.30, SD = 1.03; SG: n = 176, M = 2.90, SD = 1.05).

As Table 7 indicates, there are statistically significant differences between the first- and second-generation parents regarding their own language use and the language choices of their children. Although the language choice of the first-generation parents is more towards the use of Turkish, bilingual language use is more common among the second generation. One other significant finding concerns the language preferences of children before and after school age. The table indicates that children use more Turkish before school age; it also signifies that they are submerged more in Dutch after starting school. This finding can give us significant insights regarding the language management strategies at home. Parents continue using Turkish in the home context, which balances extensive Dutch input received in the school context. In many cases, leading to Dutch-Turkish bilingualism.

Table 7. Language practices.

		N	Mean	SD	t	p
Language choice of the parents	FG	116	4.33	0.53	9.69	.000*
	SG	184	3.70	0.55
Language choice of the children	FG	114	3.43	0.93	2.70	.007*
	SG	178	3.14	0.84
Language choice of the children before/after school	FG	113	3.76	0.94	3.67	.000*
	SG	177	3.32	1.02
Language dominance	FG	116	3.93	0.72	8.76	.000*
	SG	184	3.16	0.74
Language use – social media	FG	116	3.83	0.78	8.74	.000*
	SG	184	3.11	0.63

Note: FG refers to first-generation while SG refers to second generation. M refers to the mean value in a scale of 5, in which 1 refers to only Dutch, 2 more Dutch than Turkish, 3 both Dutch and Turkish are equal, 4 more Turkish than Dutch, 5 only Turkish. The values below 3 indicate higher Dutch choice and dominance while values above 3 indicated more Turkish choice and dominance.

*p ≤ 0.05.

Language management

Parents are involved in various language management practices, in which the mean values closer to 1 indicate ‘never’ whereas the mean values closer to 5 indicate ‘always’. The most frequently employed language management strategies in Turkish are correcting children when they mispronounce a word (FG: M = 4.00, SD = 1.11; SG: M = 3.53, SD = 1.11), listening to and singing songs in Turkish (FG: M = 3.78, SD = 1.12; SG: M = 3.42, SD = 1.06), and watching Turkish films together (FG: M = 3.61, SD = 1.17; SG: M = 3.22, SD = 1.05). The most frequently used language management strategies in Dutch are helping children with their Dutch homework (FG: M = 3.87, SD = 1.13; SG: M = 3.94, SD = 1.00), correcting them when they mispronounce a Dutch word (FG: M = 3.54, SD = 1.21; SG: M = 3.97, SD = .98), and teaching them how to read (FG: M = 3.65, SD = 1.29; SG: M = 3.74, SD = 1.10) and write (FG: M = 3.59, SD = 1.37; SG: M = 3.76, SD = 1.12).

An analysis of the language management activities for Turkish and Dutch reveals interesting findings. There is a statistically significant difference between the first- and second-generation parents regarding their Turkish language management practices while we do not find such a difference for Dutch language management activities (Table 8). Second-generation parents are less frequently involved in Turkish language management compared to first generation.

Table 8. Language management Turkish – Dutch.

		N	Mean	SD	t	P
Turkish Language Management	FG	116	3.14	0.88	3.52	.000*
	SG	184	2.80	0.78
Dutch Language Management	FG	114	3.42	0.83	3.90	.697
	SG	178	3.39	0.70

*p ≤ 0.05.

Overall, parents who took part in the survey have maintenance-oriented language ideologies besides their aspirations for bilingual development of their children. On the other hand, children prefer Dutch especially after they start school. Parents make use of various language management strategies both for Turkish and Dutch.

Discussion

We examined the FLP of first- and second-generation parents in the Netherlands in this article. The findings of the survey point out similarities with the results of the observations and interviews as well as other quantitative studies in FLP (e.g. Kang 2015; Kaveh 2018; Ó hIfearnáin 2013). In all of these studies, parents have strong bilingual ideologies, and language maintenance orientations. However, their language practices and management do not always show strong consistency with the language ideologies. Although Turkish language use has a significant place in families’ everyday life for reasons including the maintenance of identity, culture and contact with the same group, rather than parental language ideologies, actual family language practices are influenced by other resources, including children’s language practices, different linguistic domains and contexts, and the expectations of the mainstream society. The same is valid for language management.

An interesting difference between the first- and second-generation parents is that first-generation parents are more involved in language management activities that the second-generation. Indeed, they put extra effort for bilingual skills of their children. The findings are in line with our earlier ethnographic research (Bezcioğlu-Göktolga and Yağmur 2018a). The first generation moved to the Netherlands as grown-ups, and they had to master Dutch to succeed in society. They also brought their identity and cultural backgrounds, and the language they are most comfortable with is Turkish. They are well aware from their first-hand experience that Dutch is a must for upward mobility and Turkish is indispensable for one’s own roots. The challenges they had themselves cause them to be more proactive for their children. That is why they provide their children with the best language learning opportunities.

The survey also gave an opportunity to compare the FLP of parents with different migration histories. The language use and acculturation of Turkish people in the Netherlands have been widely researched (see Yağmur 2016 for intergenerational language choice and the acculturation patterns of Turkish speakers in four immigration contexts). Yağmur (2016) found that both the first and the second generations have pluralistic ideologies: they show strong attachment to their Turkish identity and culture, their social contact is mostly with the Turkish group, and they relate Turkish maintenance to the maintenance of identity, but the second generation claims they have a better understanding of the Dutch language, culture and mentality, and they choose Turkish less often than the first generation in their interactions with their surroundings. The findings of the current study are in line with these results. We can infer that there is a shift in the FLP in second-generation Turkish families towards more bilingualism regarding language ideologies and practices. They are strongly in favour of their own Turkish identity and culture, but they also acknowledge and embrace Dutch language and mentality. As long as they also pay attention to Turkish language management activities at home, and as long as they can transmit these values to their children as well, it is possible to predict that the next generation will have a similar orientation. On the other hand, this is not a one-sided affair, and it cannot be fulfilled through the efforts of only the minority group. There should be more understanding between the minority and majority cultures to strengthen mutual relationships.

The results show that first-generation informants choose to speak Turkish more, compared to the second generation. However, both generations’ language preferences indicate bilingualism. Indeed, in line with the three-generation language shift model of Fishman and considering the bilingual language preferences of the children of the second generation, it is clear that they have bilingual language practices. Although the third-generation children easily understand and use Turkish, as also asserted by Eversteijn (2011), they tend to favour Dutch in their daily interactions. On the other hand, it should also be noted that it is difficult to make any generalisations regarding generational differences, since due to marriages from Turkey, there is not a real third generation yet. Both linguistically and culturally, children get input from mainstream society as they reach school age, and this influence gets more powerful as children get older. This effect may cause them to support the mainstream ideology as valid and necessary, and cause more distance to the heritage language.

The contradictive ideologies of the families and the mainstream society complicate the realisation of FLP for Turkish families. Despite the bilingual ideologies of parents, their Turkish-maintenance orientations are not favoured by educators and mainstream institutions. Interviews with teachers also show that teachers have monolingual expectations from immigrant families (Bezcioglu-Göktolga and Yagmur 2018b). These findings are in line with other quantitative studies on FLP (Kang 2015; Schwartz 2008). Mainstream ideologies tend to neglect the importance of home language maintenance for bilingualism of immigrant families. A successful FLP can be realised only if there is a partnership between families and the mainstream society (Schwartz and Verschik 2013). Blaming the mainstream institutions for their lack of involvement in the FLP of immigrant families would not be a solution in this case. What makes them collaborate would be to raise awareness among all the relevant stakeholders, and provide them with rich resources. Such resources can be provided in various ways. Eisenchlas and Schalley (2019) illustrated a very significant example of how immigrant and refugee families in Australia can be assisted in bilingual practices. They organised workshops with various groups including minority families of different generations, mainstream teachers and authorities and in the languages of the audience, to raise awareness of the value of bilingualism, to empower parents to make better decisions, and to give families practical suggestions for home language maintenance. There are individual efforts in the Dutch context. For instance, there are language consultants offering workshops for day cares, primary schools, and healthcare centres (see, for example, Ute’s International Lounge 2019).

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1 Circassian, English and Kurdish were three other languages mentioned.

2 76 informants report Dutch as their dominant language while 9 informants report balanced dominance in Dutch and Turkish. Kurdish and English were two other languages reported by two informants.