The role of agency in the language choices of a trilingual two-year-old in conversation with monolingual grandparents

Abstract

This study examines the role of agency in a young trilingual child’s language choice in interaction with her Mandarin-speaking grandparents. The child was born in Japan to a Chinese mother and a Japanese father. English is used as a lingua franca in the family. The study demonstrates how the child asserts her agency to negotiate the decisions and efforts made by the grandparents. The bond of the heritage language and culture, and the value of child trilingualism are strongly desired and explicitly implemented in the grandparents’ monolingual or ‘bilingual-like’ discourse strategies in dealing with the child’s mixed codes. Meanwhile, the child’s flexible language use is not a passive response to the grandparents’ strategies but an exercise of her four significant senses and behaviour, which are: (1) resisting through no response; (2) moving on in a dual-lingual conversation; (3) assisting the grandparents to decode her non-Mandarin speech; and (4) modifying the language choices of herself and others. This study suggests that language choices of trilingual children are complex. It also provides empirical evidence that grandparents provide an important incentive in the planning of family language policy.

Keywords

Children’s agency discourse strategies family language policy grandparents language choice

Introduction

Family has come to be viewed as an independent domain in the study of language policy, and family policies themselves have come to be studied (King et al., 2008; Siiner et al., 2018; Spolsky, 2012). One emerging research area of family language policy is children’s agency. Luykx (2003) suggests that language socialization across generations is not ‘a one-way process’ but ‘a dynamic network of mutual family influence’ (p. 40). While adults teach children languages, children also affect adults’ language use (Gafaranga, 2010). This article focuses on the agentic role of a young trilingual, whose main home languages (English and Mandarin) differ from the community language (Japanese), in interactions with her Mandarin-speaking grandparents. Specifically, it will explore the child’s significant agentive senses and behaviour in response to the grandparents’ monolingual or ‘bilingual-like’ discourse strategies, and how these agency influence the grandparents’ language choices. Grandparents as core adult interlocutors in language practice have not been much investigated. Based on empirical data, this study addresses this gap by considering the incentive role of grandparents as a component of family language policy.

Children’s agency

Family language policy provides a framework for examining parental language ideologies, parent–child interactions, and child language development (King & Fogle, 2013, 2018). Studies of how family language planning impacts child language development have generally examined the role that parents play in response to the child’s mixed items (Curdt-Christiansen, 2016; Lanza, 2007). However, children ‘think about and reflect upon the nature and function of language’ (Pratt & Grieve, 1984, p. 2). This ‘socioculturally mediated capacity to act’ is referred to as agency (Ahearn, 2001, p. 112). Research on children’s agency extends the focus from a parent-dominated direction to a parent–child mutually-influencing bi-direction (Fogle, 2012; Fogle & King, 2013; Gyogi, 2015). Children’s agency appears in multiple forms, is flexible, and is contextual (Fogle, 2012). It is usually exercised in two dimensions: (1) sense of agency (a child’s agentic feelings for general or specific contexts), and (2) behaviour of agency (a child’s agentic behaviour as participation or action, or deliberate non-participation or non-action) (Mercer, 2012). Children always need to think about the relationship between a language and the context in which it is embedded (Lanza, 1997).

In her study of language mixing in English-Norwegian two-year-olds, Lanza (1997) describes how a monolingual or bilingual context can be promoted by a parent using any of five different discourse strategies. Lanza observes that while children sometimes follow the parent’s language choice, they may also use a different language than the one the parent has indicated to be most appropriate. Children’s agentic reactions reshape the conversational context and impact the parental strategies. The first strategy, ‘Minimal Grasp’, is where the adult pretends not to comprehend the child’s language and requires clarification from the child by using prompting words such as ‘what’ or ‘pardon’. This strategy is most likely to establish a monolingual context. In the second strategy, ‘Expressed Guess’, the adult seeks confirmation from the child through a ‘Yes’ or ‘No’ reply in the adult’s language. This also involves language clarification but less directly. In the third strategy, ‘Adult Repetition’, the adult uses the appropriate language to repeat the content of the child’s utterances. This method might work as a reminder to the child that a switch to the adult’s language is desirable. However, the child may also interpret it as a sign that he/she is understood and his/her language choice is acceptable. The last two strategies, ‘Move On’ in the conversation after noticing the ‘inappropriate’ language used by the child and, particularly, ‘Code-Switches’ to the child’s language, transmit a clear signal to the child that his/her language does not hinder the communication and that continuing to use the language is tolerable or even favoured. The adult who uses either of the last two strategies seems to have followed the lead of the child.

Fogle (2013) identifies three types of agency that are commonly exercised by children to resist adults’ language choices: not responding, interacting using ‘wh-’ questions of high frequency, and influencing the language choice of adults. In a case study, King and Logan-Terry (2008) show how an English-Spanish two-year-old’s responses to the mother’s passive strategies are acknowledged by the mother. The bilingual mother started with an open-ended question in Spanish and the child did not respond. The mother then moved to an either-or question. When the child eventually answered in English, the mother confirmed (‘okay’) and continued the conversation in English. The mother eventually switched to the child’s code following the child’s step-by-step strategic replies. In a different study, in which language shift from Rwandan Kinyarwanda to French is ‘talked into being’ in bilingual families in Belgium, Gafaranga (2010) discusses the strategies used by the parents to accommodate to their primary school children’s¹ constant medium requests. Gafaranga notes that when the parents do not conform to the children’s medium requests, a parallel conversation will be carried on in the preferred language of each.

Apart from thinking about adults’ discourse strategies, children make planned decisions and give deliberate performances vis-a-vis adults’ language beliefs. In a study of two 12-year-old English-Japanese girls negotiating and constructing bilingual space at home in London, Gyogi (2015) found that one child resisted her mother’s monolingual Japanese use and flexibly adopted English and Japanese with the mother. Gyogi considers this girl’s ‘negative’ agency not to have been a complete contestation of the mother’s language beliefs since the child consistently used Japanese when addressing the Japanese researcher. In contrast, the researcher noticed that the other child used English to construct her social identity as a good English learner and negotiated the mother–child relationship taking into account the mother’s attitude toward English. Through this ‘positive’ agency, this second girl seems to have provided herself with a flexibility to use both languages. Gyogi (2015) thus recognizes that the responses of both children to their mothers’ policies are not controlled by the mothers’ beliefs but are ‘an exercise of children’s agency though contestation, negotiation and redefinition’ of the adults’ beliefs (p. 749).

Children also judge and act agentively with regard to the type of activities. Quay (2001) discusses how a trilingual one-year-old’s feelings about the language of a specific activity played a role in his language choice. The toddler used the community language, Japanese, to address either the English- or German-speaking parent whilst playing with toys or looking at books. These two activities were part of the boy’s routine in his daycare centre and therefore Japanese would be regarded as the ‘appropriate’ language in this situation (Quay, 2001).

The role of grandparents in family language policy

In addition to childcare, grandparents contribute greatly to the transmission of the heritage language and culture (Braun, 2011; Braun & Cline, 2010; Ruby, 2012). Thus, it is useful to look at the linguistic as well as cultural importance of the grandparents to core families in the research on family language policy. Early studies involving grandparents generally examine natural language learning and maintenance across three generations in immigrant families (King & Fogle, 2013). The loss of the intergenerational transmission that occurs in daily conversations within the family is seen as the key marker of language loss (Fishman, 1970; Spolsky, 2012). Language shift is often the norm that develops through practice, as parents and grandparents often find it difficult to establish rules and policies at home (Lanza, 2007; Zhu & Li, 2016).

Instead of looking only at the overall patterns of language transmission, recent studies attempt to seek the affective role of grandparents in different types of transnational and multilingual families (Braun, 2011, 2012; Braun & Cline, 2010; Chevalier, 2012; Smagulova, 2017). From a study of 70 trilingual families, half living in England and half in Germany, Braun and Cline (2010) regard the grandparents as having important sociocultural and linguistic influences on the parents’ language practices with their children.² In a different study of parental language choice with the same 70 families, Braun (2012) notices the residencies and language proficiencies of the grandparents to be affecting factors. However, in both studies, grandparents are not direct participants. The theme of grandparents is also touched on in other studies, where the families contained varying language combinations. Lee (2004) uses questionnaire-based methods to examine language use of young aborigines aged 16 and over in 195 families in Taiwan and discovers that the presence of grandparents supports the parents to maintain the home languages. In a study of two English-French-Swiss German children (aged between 2 and 4) carried out by Chevalier (2015), and also in a study of two Croatian-English-German children (aged between 1 and 3) conducted by Ivir-Ashworth (2011), it was found that language input by grandparents had beneficial effects on active trilingualism. In a survey of German-Chinese families in Hong Kong, Chong (2004) points out that grandparents’ positive ideas and attitudes toward bilingualism help with school-age³ children’s language acquisition.

The perception of the value of multilingualism across generations as well as its development have been extensively discussed. While some grandparents think multilingualism adversely affect family relationships and thus disapprove of their grandchildren being raised with more than one language (Braun & Cline, 2010), others consider the learning and use of the heritage languages and social languages to be valuable (Said, 2014). In a study of Chinese immigrant families in Quebec, Canada, Curdt-Christiansen (2009) emphasizes the influence of the sociopolitical and economic aspects on parents’ and grandparents’ expectations when they pursue policies for primary school children’s⁴ language and literacy education in Chinese, English, and French. Curdt-Christiansen argues that these aspects contribute to the adults’ opinions on the value of multilingualism, which in turn underpins their invisible language planning. In a different study of three diasporic Chinese families dwelling in the UK, Zhu and Li (2016) describe how the specific educational background and migration experiences caused two retired grandparents in one of the three families to feel a ‘loss’ of their mother tongue, Chinese. The authors discuss how the way in which the grandparents deal with the challenges of family multilingualism is affected by different sociocultural factors, as reflected in their perceptions of social relations and self-identities, and their construction of language strategies.

Curdt-Christiansen (2016) further discusses how family members translated their beliefs into policies in a trilingual Chinese family, a Malay, and an Indian bilingual family in Singapore, identifying inconsistencies between adults’ language ideologies and family linguistic practices. The trilingual Chinese family consisted of parents and two children (aged 7 and 10) who mainly spoke in Mandarin and English, and three extended family members, the grandmother, an aunt and an uncle, who mainly used Hokkien.⁵ In interviews with the grandmother and the aunt, both parties showed an appreciation of English’s instructional values and the mother tongues’ cultural functions, and valued both the bilingual policy, which recognized both English and the mother tongues as official languages, and the educational policy, which had established English as the medium of instruction (Curdt-Christiansen, 2016). The author also gives examples of parent–child exchanges in the Malay and Indian families,⁶ which show that the efforts made by the parents in communication were incongruent with what they professed to believe. Curdt-Christiansen explains that although maintaining the bond with the mother tongue and culture was strongly desired in the three families, the family members convincingly confirmed their affection for the instrumental values of language through their everyday practices.

Although studies of multilingual families tend to look at families from the perspective of a broader language ecology, grandparents seem to have played only a peripheral role in the discussion of family language policy (Lanza, 2007). Grandparents as core adult interlocutors and their contributions to the making of family language policy have not been much explored. While the exchange of linguistic and cultural knowledge and skills through language, literacy and digital practices involving grandparents has been reported in two studies in Bengali-English homes in the UK, the attention in both is on the role of grandparents as home educators and supporters in early childhood education (Kenner et al., 2007; Parven, 2016). Smith-Christmas (2014) addresses this gap by providing data showing how parents and extended family members dynamically and dialogically tried to negotiate a strong Gaelic-centred policy with three children (aged 4, 8 and 12) on the Isle of Skye, Scotland. The policy was mainly reified by the mother and the grandmother, while the father and his extended family members (his brother and sister) occasionally interfered with this policy from a different perspective (Smith-Christmas, 2014). The father, uncle and aunt contributed to the low use of Gaelic among the children, socializing the children into a shift from Gaelic to English through negative practices, such as replying in English to the Gaelic speakers, keeping English as their peer group language, and disciplining the children in Gaelic as a ‘last resort’ strategy, which might have made children less inclined to use Gaelic in positive contexts (Smith-Christmas, 2014). The author argues that extended family members participating in family language policy may thus have negative as well as positive influences on the policy.

The study

This study investigates a trilingual child’s language choices in interactions with her Mandarin-speaking grandparents from the perspective of family language policy with regard to extended family members. Unlike many other studies that consider adults’ top-down policies alone, the present study emphasizes the child’s bottom-up agentive role vis-a-vis the grandparents’ explicit decisions and efforts. The following questions will be investigated: (1) What strategies do the grandparents use to deal with the child’s non-Mandarin utterances? (2) How does the child exercise her special agentic feelings and behaviour to negotiate the grandparents’ strategies and policies?

Participants

The trilingual girl (hereafter referred to as Y⁷) was born in 2014 in Japan to a Mandarin-speaking mother, who was also the present researcher, and a Japanese father. The mother had previous rich experience working in early childhood education in an English environment, and the father regularly read and wrote in English as a university lecturer. English was the lingua franca of the family. Since Y was born, she had been spoken to in three languages: Mandarin, by the mother; English, by the father and in family conversations; and Japanese, by both parents for reading and doing other activities related to that language. Y attended a small size nursery⁸ from the age of 1;11. In the total 12 hours of her time awake, Y’s daily language exposure was roughly as follows: English 4 hours (3 hours at home and 1 hour in the nursery), Japanese 5 hours (1 hour at home and 4 hours in the nursery), and Mandarin 3 hours (at home). Y was described as a friendly and active child who got along easily with both adults and children.

Y’s grandparents spoke Mandarin and an Eastern-Chinese dialect. Like the majority of Chinese grandparents of a similar age, even though they had received school education through junior high school, they knew almost no English or Japanese. Y and the grandparents had spent time together on two previous occasions: when the grandparents babysat her for 7 months after she was born (0–0;7) in Japan, and when Y visited them in China for a month (1;2–1;3). Y also communicated with them once a week through online video-calls. Y was spoken to in Mandarin by the grandparents and she addressed them predominantly in Mandarin. The grandparents paid a second visit to Japan from 24 November 2016 to 9 January 2017 (2;6–2;7), during which time this study was conducted. There were strong family bonds in Y’s family. The three of them played together every day. Y appeared to be closely attached to the grandparents.

Data

The data for this study were collected during the grandparents’ second visit to Japan over a period of 6 weeks and 3 days. Three methods of data collection were used: participant observation, diary entries and video-recording. The last method played the largest role. The aim of this study and the methods were explained to the grandparents, while the child was already familiar with them from her previous experiences. Both parties were aware that their interaction would sometimes be recorded. They were comfortable with the mother’s actions of observing and recording; thus the observer’s paradox did not have a noticeable impact on the speaker’s language use. The mother tried to keep an eye on the child when she had no household chores, making written note of what she had observed. Diary entries were made on a daily basis. A 25-minute video was taken once a week which recorded the child and her grandparents playing together in the living room. In total, six videos were made. In order to ensure natural communication, the camera was fixed in an inconspicuous corner and not adjusted to follow the child when she moved around. In three of the six videos, the child frequently moved away from the camera and her speech became either discontinuous or incomprehensible.

Because of the transcribing difficulties experienced, the three poor quality videos were not used in the study; only the three high quality videos were transcribed verbatim and used. Among these three transcribed videos: Video 1 showed the child and the grandparents playing cooking games, Video 2 showed them listening to a sound picture book⁹ and then drawing pictures, and Video 3 showed them playing Lego games and then the child leading the grandparents in a role-play, ‘calendar activity’, which was a daily routine in English¹⁰ in her nursery. Transcription of the videos was made in the CHAT format of CHILDES (MacWhinney, 2017) by the researcher. The speakers were coded as GrF (grandfather), GrM (grandmother), CHI (child), or MOT (mother). Utterances were marked as M (Mandarin), E (English), J (Japanese), or Mix (mixed items). Mixed items included the mixing of elements of two or three languages at the word or sentence level, or the switching from one language into another at a constituent or sentence boundary (Döpke, 1992). Unintelligible speech, onomatopoeia, and English loanwords in Japanese were not dealt with in the analysis. Forms of personal reference were associated with affection between family members and the use of a reference term from a different language was therefore not considered as a mixed utterance. Examples of coded transcripts are shown in Appendix. Three bilingual examiners, who were also parents of bi- or trilingual children, reviewed the transcripts. Minor corrections to transcriptions were made as a result of the examiners’ comments.

Data were analysed in speech turns with a single speech turn consisting of one or more utterances. The numbers and percentages of speech turns of the child and the grandparents were calculated. Based on the data, the grandparents’ ideologies and attitudes toward the child’s trilingual abilities and language use were revealed. In addition, excerpts containing instances of the child’s and the grandparents’ mixing and switching as related to the research questions were analysed. The theme (the description of a pattern or a phenomenon) of a discourse (an excerpt) was explained, and utterances in a discourse representing the theme were marked and discussed.

Results

Language choices of the grandparents

Grandparents are one of the important influences in language choices in trilingual families (Braun, 2012) and this is evidenced in the present study. The languages produced by the grandparents are summarized in Table 1.

Table 1.

Language choice of the grandparents.

Number of speech turns (%)
Video 1		Video 2		Video 3
GrF		GrF		GrF
M	25 (96%)	M	8 (89%)	M	21 (95%)
J	1 (4%)	Mix	1 (11%)	J	1 (5%)
Total	26 (100%)	Total	9 (100%)	Total	22 (100%)
GrM		GrM		GrM
M	67 (99%)	M	34 (74%)	M	78 (74%)
Mix	1 (1%)	J	18 (17%)	E	3 (6.5%)
		Mix	10 (9%)	J	6 (13%)
				Mix	3 (6.5%)
Total	68 (100%)	Total	46 (100%)	Total	106 (100%)

The data in Table 1 show that the grandparents’ native language, Mandarin, is a strong code in the family’s language practices. In Video 1, the grandparents and the child spoke predominantly in Mandarin: 96% of the grandfather’s and 99% of the grandmother’s utterances were made in Mandarin. In Video 2, the grandparents observed the child playing, and had some general conversations with the child related to but not specifically commenting on the picture book. Even though the use of non-Mandarin terms increased, the main code used by the grandparents was still Mandarin (89% of the grandfather’s and 74% of the grandmother’s utterances). In Video 3, non-Mandarin codes were sometimes mimicked by the talkative grandmother, but 74% of her utterances were in Mandarin. The grandfather observed the ‘calendar activity’ and mainly commented in Mandarin (95% of his utterances).

The grandparents sometimes naturally made comments on Y’s language and communication skills. It seemed that they felt the primary goal of communication was to encourage child participation. For example, when Y responded to their Mandarin in a different language, the grandmother once said, ‘yuàn yì shuō jiù hǎo’ (as long as she is willing to talk) and the grandfather said ‘huì shuō jiù hǎo’ (as long as she knows how to say it). When the mother heard these comments, she usually jotted them down. The grandparents also thought they had ‘learned’ some foreign words from the child. Although the grandparents sometimes mimicked Y’s pronunciation of an English or a Japanese expression, in most cases, they copied and used it in their own speech, rather than simply repeating it in the immediately following utterance after the child said it. They may not have pre-planned to use foreign terms in communication but were nevertheless not hesitant about using such terms. In Video 1, the grandfather included one Japanese expression among a total of 26 turns (= 4%) and the grandmother used one mixed item among 68 turns (= 1%). In Video 2, non-Mandarin speech increased when they listened to the audio picture book, and this accounted for 11% of the grandfather’s and 26% of the grandmother’s total utterances. In Video 3, the grandmother took part in the role-play and 26% of her utterances were made in one of the non-Mandarin languages. The grandfather mainly observed, but he also did adopt one Japanese item (= 5%).

The results depict Y’s grandmother as an active speaker while the grandfather talked relatively less. In their regular family life, the grandmother was a passionate conversationalist and the grandfather was an engaged, though less active, participant. The results thus spoke for the reliability of the video-recording methodology as they had reflected the grandparents’ natural behaviour.

Language choices of the child

The child’s ability to choose an appropriate language responding to the interlocutors is shown in Table 2.

Table 2.

Language choice of the child toward the grandparents.

Number of speech turns (%)
Video 1		Video 2		Video 3
M	62 (94%)	M	23 (37%)	M	24 (75%)
J	1 (1.5%)	E	17 (27%)	E	3 (9.5%)
Mix	3 (4.5%)	J	20 (32%)	J	3 (9.5%)
		Mix	3 (4%)	Mix	2 (6%)
Total	66 (100%)	Total	63 (100%)	Total	32 (100%)

Y’s Mandarin utterances in each video outnumbered any other single type of language when she addressed the grandparents (94%, 37% and 75% in Videos 1 to 3 respectively). Meanwhile, her mixing behaviour was also observed. It was noticed that when the grandparents spoke predominantly in Mandarin in Video 1, the child used very few non-Mandarin terms; whereas when the grandparents mixed more languages in Videos 2 and 3, the child’s non-Mandarin forms also increased. This parallels observations from studies conducted separately by De Houwer (2009) and Lanza (1997), which show how children tend to monitor and adjust to their parents’ modelled behaviour.

The association between an activity and the language in which it is acted out appeared to be one more reason why the child addressed the grandparents in more non-Mandarin forms in Videos 2 and 3. For example, Y preferred English in the ‘calendar activity’, a daily routine in her nursery, as shown in the following excerpt.¹¹

(1) excerpt from Video 3

@Situation: The child was role-playing the nursery’s ‘calendar activity’.

1 *CHI: What’s the weather like today?

2 *MOT: jīn tiān tiān qì zěn me yàng?

%eng: What’s the weather like today?

3 *CHI: wài pó.

%eng: Grandmother.

4 *GrM: qíng tiān.

%eng: Sunny.

5 *MOT: jīn tiān tiān qì zěn me yàng?

%eng: What’s the weather like today?

6 *CHI: jīn tiān shì sunny and cloudy.

%eng: Today is sunny and cloudy.

Although Y was capable of using the Mandarin equivalents for the English terms appearing in the above exchange, she called on the grandmother to answer the mother’s translated question (Line 3). More interestingly, even after the grandmother gave a model answer in the expected language, the girl switched back to English in the middle of processing a Mandarin utterance (Line 6).

Y’s agency

Y generally used her three languages flexibly rather than passively following the grandparents’ lead. She demonstrated special senses and actions in reply to the strategies developed by the grandparents to deal with the non-Mandarin utterances that she produced in conversations. Y’s agentic responses included: resisting through no response, moving on in a dual-lingual conversation, assisting the grandparents to decode her non-Mandarin speech, and modifying the language choices of herself and others.

Resisting through no response

The first agentic action taken by Y was resisting by failing to respond when language clarification was required by the grandparents. In the following excerpt, the grandmother tried to negotiate a Mandarin context and insisted that the child’s reply be clarified in Mandarin. However, it seemed that the grandmother was unsuccessful in her attempt to draw Mandarin out of Y.

(2) excerpt from Video 3

@Situation: The child was role-playing the nursery’s ‘calendar activity’.

1 *CHI: What’s the weather like today?

2 *GrM: tīng bù dǒng.

%eng: I don’t understand.

3 *CHI: What’s the weather like today?

4 *GrM: yòng zhōng wén shuō yī biàn.

%eng: Say it in Mandarin.

5 *CHI: (0).

%com: No response.

In the above exchange, Y refused to make a language correction from English to Mandarin and moved on in English (Line 3) even when the grandmother asked for either a content repair, ‘I do not understand’ (Line 2), or a language choice repair, ‘say it in Mandarin’ (Line 4). When asked more than once, Y simply ignored the adult’s request (Line 5). The grandmother’s approach to eliciting Mandarin was similar to the first strategy, ‘Minimal Grasp’, employed by the bilingual mother of a two-year-old in Lanza’s (1997) study. That mother successfully feigned the role of an English monolingual and insisted on a rephrasing of her child’s Norwegian utterances (Lanza, 1997). However, in the above excerpt, it seemed impossible to achieve alignment as Y did not switch to the grandmother’s medium choice regardless of the efforts that the grandmother had made.

Moving on in a dual-lingual conversation

The fact that the grandparents guessed the meaning of Y’s non-Mandarin utterances and responded appropriately to the content were positive evidence of their comprehension of the child’s speech even if they responded in Mandarin. Y thus perceived that she was permitted to move on in a different code in parallel to the grandparents, who attempted to decode her non-Mandarin utterances but continued in Mandarin. The grandparents’ behaviour coincided with the parental strategy ‘Move On’ described in Lanza’s (1997) study. Through this practice, a multilingual context is negotiated and a dual-lingual paradigm is adopted in which speech partners mutually understand each other’s codes, but where one partner consistently uses one code, while the other partner another (Smith-Christmas, 2014). A dual-lingual conversation between Y and her grandparents is shown as follows:

(3) excerpt from Video 2

@Situation: The child is showing her drawings to the grandmother.

1 *CHI: oekaki, oekaki.

%eng: Draw pictures.

2 *GrM: oekaki shì shén me?

%eng: What is oekaki?

3 *CHI: Anpanman.

%com: Anpanman is a Japanese cartoon character.

4 *GrM: ā! Anpanman.

%eng: Ah-ah! It’s Anpanman.

%com: The grandmother knew the character.

5 *CHI: mite.

%eng: Look!

6 *GrM: mite. kàn yī kàn.

%eng: Look.

%com: The grandmother translated the Japanese term mite into its

Mandarin equivalent kàn yī kàn.

7 *CHI: oyama.

%eng: A hill.

%com: The child drew a hill.

8 *GrM: zài shàng miàn.

%eng: (Anpanman is) on the top of the hill.

In the above exchange, the grandmother first asked Y the meaning of the Japanese word oekaki (draw pictures) (Line 2). As said earlier, this insisting strategy did not work well. Once the grandmother recognized the character Anpanman, she immediately showed her understanding using a Mandarin exclamation ā (ah-ah) and said the character’s name in Japanese (Line 4). Interestingly, the child accepted the grandmother’s Mandarin expression zài shàng miàn (on the top) (Line 8) as an appropriate reply to her Japanese word oyama (a hill) (Line 7). A medium negotiation began with a disagreement between the two participants about which language to use, and it ended with the grandmother giving in to Y’s preferred language. The grandmother had noticed the child using another language, but still responded to the child by attending to the content of the conversation. The desire to continue a conversation, regardless of the language form, may be enough of a reason to move on. Whether or not a code-switching into the child’s language is done, a bilingual context is created and a parallel conversation in the two languages ensues.

Assisting the grandparents to decode her non-Mandarin speech

Being aware of the grandparents’ attempts to guess the meaning of her non-Mandarin speech, Y sometimes assisted them in decoding her English or Japanese message. In the following excerpt, with the aid of the child’s nonverbal cues the grandparents successfully guessed the meaning of an English expression and modelled its Mandarin equivalent.

(4) excerpt from Video 2

1 *CHI: Pee-pee.

%com: The child pointed to the toilet.

2 *GrF: ‘Pee-pee’ shì shén me?

%eng: What is ‘pee-pee’?

3 *GrM: ‘Pee-pee’ jiù shì xiǎo biàn.

%com: The grandmother translated the English word ‘pee’ into Mandarin.

4 *CHI: Pee-pee.

%com: The child nodded.

The child told the grandparents, in English, that she wanted to pee and explained the meaning of the word by pointing to the toilet (Line 1). The grandmother confidently translated the English term ‘pee’ into its Mandarin equivalent (Line 3). The child might have thought a switch into Mandarin was no longer necessary since her English message had been successfully transmitted, so she confirmed the grandmother’s translation in English (Line 4).

Y also used nonverbal communication cues to support the grandparents’ efforts to comprehend and translate her Japanese utterances. An example of this is shown in Excerpt 3. The child pointed at the hill that she drew and asked the grandmother to have a look. Assisted by the child’s body language, the grandmother successfully guessed the meaning of the Japanese term mite (look) (Line 6) and offered its Mandarin equivalent kàn yī kàn (Line 6). Another example is shown in the following exchange, where the grandmother reacted to the child’s Japanese expression by contradicting the child’s claim.

(5) excerpt from Video 3

@Situation: The child and the grandmother were assembling puzzles.

1 *CHI: wài pó, dekinai.

%eng: Grandma, you can’t do it.

%com: The child offered the grandmother a piece of puzzle and pointed at a

spot on the puzzle board where the piece should be put.

2 *GrM: wǒ huì de.

%eng: I can.

%com: The grandmother did not accept the help.

3 *CHI: dekinai. muzukashiyo.

%eng: You can’t do it. It’s difficult.

%com: The child tried to offer the grandmother another piece of the puzzle and

tapped her index finger on the puzzle board.

4 *GrM: wǒ huì de.

%eng: I can.

%com: The grandmother refused again.

It is possible that the grandmother had heard the Japanese term dekinai (can’t) (Line 1) used by the child in her conversations with others. The grandmother contradicted Y’s Japanese statement by continuing the conversation in Mandarin. The grandmother’s reply might have functioned as a clue to the child that her Japanese had been understood, so the child went on to emphasize her opinion in Japanese (Line 3). Confirming a tentative interpretation of the co-participant’s talk is referred to as an ‘understanding check’ and is used by children as a strategy for medium request in parent–child interactions; however, an understanding check does not necessarily lead to alignment at the level of language choice (Gafaranga, 2010).

Modifying the language choices of herself and others

Aspirations for smooth communication across generations even prompted the grandparents to imitate the child’s utterances and deliberately ‘switch’ to her language. Y always monitored the grandparents’ ‘code-switching’ behaviour and modified her own and others’ language choices. Unlike the code-switching strategy used by the bilingual parents in Lanza’s (1997) study, two languages were not available as alternatives in this bilingual-like context, as shown in the following excerpt.

(6) excerpt from Video 3

@Situation: The child and the grandmother were playing Lego games.

1 *CHI: kore.

%eng: This one.

%com: The child picked a Lego piece and gave it to the grandmother.

2 *GrM: kore.

%exp: The grandmother did not accept the child’s help, found one by herself,

and imitated the child’s pronunciation of kore.

3 *CHI: iyada.

%eng: That’s not right.

%com: The child moved her hands back and forth seeing the grandmother

putting the piece in a wrong place.

4 *GrM: ieja.

%com: The grandmother insisted on putting the piece in the place she wanted, and tried to imitate the child’s pronunciation of the Japanese word, iyada.

5 *CHI: dame.

%eng: Don’t do it.

%com: The child tried to grab another piece from the grandmother but failed.

6 *GrM: tama.

%com: The grandmother continued to put the piece in another wrong place while trying to imitate the child’s pronunciation of the Japanese word, dame.

7 *CHI: bù shì. bù shì zhè lǐ. %eng: No, not here.

%com: The child switched to Mandarin and told the grandmother that she was wrong.

In the above exchange, Y tried twice to stop the grandmother from putting the Lego pieces in the wrong places while speaking in Japanese accompanied by body language (Lines 3 & 5); however, the grandmother repeatedly refused her help (Lines 2, 4 & 6). Y eventually switched to Mandarin after her protest in Japanese had failed once again. She did not continue in Japanese after the grandmother switched to Japanese. Her response to the adult’s code-switching act was different from what is usually seen in bilingual (Döpke, 1992; Lanza, 1997) and trilingual children (Chevalier, 2012; Quay, 2001). For example, the trilingual child Freddy, in Quay’s case study, recognized the parents’ attempts to use the community language, Japanese, and indirectly socialized them into using it by responding in Japanese to their Japanese utterances.

Discussion

This study shows how a trilingual two-year-old exercised her agency to resist, negotiate and modify her monolingual grandparents’ language choices. Rather than looking at the adults’ top-down policies alone, the present study highlights four types of bottom-up agentive senses and acts displayed by the child, as responses to the explicit decisions and efforts made by the grandparents, including the discourse strategies that they employed in dealing with the child’s mixed languages.

Lanza (1997) considers ‘Minimal Grasp’ the most successful strategy for drawing a language spoken by a parent out of a child when that language differs from the one used in the larger social context. However, this highly-constrained practice did not always function well in the present family practices. Y refused to conform to the grandparents’ code and provided no appropriate responses. Y’s responses suggested that the adults’ attempts at a Mandarin-centred policy might not in fact result in the child’s use of that language. Smith-Christmas (2014), in her study of an extended bilingual Gaelic-English family, also discovered that the reification of a strong pro-minority language policy is not necessarily correlated with a high degree of minority language use. One possible explanation for why a strict elicitation strategy may result in divergent language production in children is that, as described by De Houwer (2009), child caretakers vary in their impact beliefs – that is, the degree to which they see themselves as duty-bound and skilful in shaping a child’s language choices. Both Barnes (2011) and Chevalier (2012), in separate studies of trilingual families, found that children tend to be more active in each of the parents’ languages when they are required to speak the language of each parent with that parent, while children are less active when they are not required to do so. Another explanation may be related to Döpke’s (1992) recognition that adults differ in their use of interactional techniques as a part of their language practices, including whether or not they are consistent. Takeuchi (2006), for example, discovered that Australian-born Japanese children, whose mothers consistently spoke in Japanese with them at home, appeared to accept the role of this language in their daily lives and the expectations placed on them to use it.

Y’s grandparents might have considered the primary goal of an interaction to be to encourage language behaviour irrespective of its form. In monolingual families, this motivation is shown by parents’ tolerance of a variety of linguistic errors (Kasuya, 1998, p. 342). Despite the fact that the grandparents knew only one language, a strong desire to strengthen family bonds by means of smooth intergenerational communication caused them to inadvertently demonstrate some ‘bilingual’ features. For example, similar to bilingual parents examined in Lanza’s (1997) study, they moved on, and decoded and translated Y’s non-Mandarin utterances. It is true that the discourse strategies discussed in this article were originally described as those employed by bilingual speakers; however, it appears that even functionally monolingual speakers – such as the grandparents in the present study – may at times appropriate some of them. To a certain extent, a bilingual-like context was more or less successfully negotiated in the grandparent–child conversations. Decoding and translating from one of a multilingual speaker’s languages to another appears natural and requires little effort for multilinguals; however, it was a more difficult linguistic as well as cognitive task for the monolingual grandparents.

The grandparents’ ‘bilingual’ strategies led to some bilingual practices at home, such as dual-lingual conversations. Dual-lingualism has also been reported in Smith-Christmas’s (2014) study of the English-Gaelic family, where the mother and the grandmother persistently used Gaelic while the children normally answered in English. The Gaelic-dominant grandmother occasionally felt it difficult to continue in Gaelic in a dual-lingual conversation with the grandchildren who constantly replied to her in English, which the grandmother said was ‘demoralising’ (Smith-Christmas, 2014). However, Saville-Troike (1987) argues that these types of dual-lingual conversations are qualitatively different from truly bilingual conversations.

From the diary data of the present study, one can see that Y showed a general awareness of the grandparents’ language abilities. She once commented ‘they can only speak Mandarin’ (recorded in the mother’s diary entries). This indicates that Y might not have associated the grandparents’ ‘bilingual’ behaviour with their real language abilities. It is also reasonable to assume that Y’s language decisions and actions were made not only based on the grandparents’ language choices but as a part of an overall consideration of the adults’ flexible policies and practices in respect to various linguistic and contextual factors. Although it is not clearly shown in the current data, the exercise of this significant agency sometimes influenced the grandparents to follow the lead of the child in the conversations.

Conclusion

This study contributes to the emerging discussion of children’s agency within the scope of family language policy. It involves grandparents as the core interlocutors, which broadens the perspective of family language policy in connection with the family language ecology. Children’s agency is a persistent and developing notion that needs to be understood as it is enacted in specific ways to suit each family. Children’s growing agentic senses and behaviour are closely related to their age-appropriate underlying metacognitive development. Language policies and language choices at home often become more complicated after a child starts school. Different challenges may arise when there is an interplay among multiple factors, such as the socioeconomic privilege connected with the community and school languages, family bonds, conﬂict over cultural values, and children’s own identities. There is still much to be learned about children’s agency in relation to how children experience and engage in family language policy over time. Additional interactional data analysed from a turn-by-turn perspective are needed in order to understand the various issues. In further research, there should also be continued attention to the role of grandparents as well as other extended family members in children’s acquisition of heritage language and culture through diversified intergenerational practices. This would include an examination of their contributions to children’s acquisition of literacy, as well as social and scientific knowledge, and their role as home educators and supporters in early education.

Footnotes

Appendix

Funding

The author received no financial support for the research, authorship, and/or publication of this article.

ORCID iD

Ying Zhan

Notes

References

Ahearn

(2001). Language and agency. Annual Review of Anthropology, 30(1), 109–137.

Barnes

(2011). The influence of child-directed speech in early trilingualism. International Journal of Multilingualism, 8(1), 42–62.

Braun

(2011). The role of education in the language practices of trilingual families. In Varcasia

(Ed.), Becoming multilingual: Language learning and language policy between attitudes and identities (pp. 113–134). Peter Lang.

Braun

(2012). Language maintenance in trilingual families: A focus on grandparents. International Journal of Multilingualism, 9(4), 423–436.

Braun

Cline

(2010). Trilingual families in monolingual societies: Working towards a typology. International Journal of Multilingualism, 7(2), 110–127.

Chevalier

(2012). Active trilingualism in early childhood: The motivating role of caregivers in interaction. International Journal of Multilingualism, 9(4), 437–454.

Chevalier

(2015). Trilingual language acquisition: Contextual factors influencing active trilingualism in early childhood. John Benjamins.

Chong

(2004). Bilingualer spracherwerb in deutsch-chinesischen familien in einem englischchinesischen umfeld - Eine erhebung in Hongkong [Bilingual language acquisition in German-Chinese families in an Anglo-Chinese environment: A survey in Hong Kong] (Unpublished doctoral dissertation). Ludwig-Maximilians-Universita.

Curdt-Christiansen

X. L.

(2009). Invisible and visible language planning: Ideological factors in the family language policy of Chinese immigrant families in Quebec. Language Policy, 8(4), 351–375.

10.

Curdt-Christiansen

X. L.

(2016). Conﬂicting language ideologies and contradictory language practices in Singaporean multilingual families. Journal of Multilingual and Multicultural Development, 37(7), 694–709.

11.

Döpke

(1992). One parent-one language: An interactional approach. John Benjamins.

12.

De Houwer

. (2009). Bilingual first language acquisition. Multilingual Matters.

13.

Fishman

(1970). Sociolinguistics: A brief introduction. Newbury.

14.

Fogle

(2012). Second language socialization and learner agency: Adoptive family talk. Multilingual Matters.

15.

Fogle

(2013). Parental ethnotheories and family language policy in transnational adoptive families. Language Policy, 12(1), 83–102.

16.

Fogle

King

(2013). Child agency and language policy in transnational families. Issues in Applied Linguistics, 19, 1–25.

17.

Gafaranga

(2010). Medium request: Talking language shift into being. Language in Society, 39(2), 241–270.

18.

Gyogi

(2015). Children’s agency in language choice: A case study of two Japanese-English bilingual children in London. International Journal of Bilingual Education and Bilingualism, 18(6), 749–764.

19.

Ivir-Ashworth

K. C.

(2011). The nature of two trilingual children’s utterances: Growing up with Croatian, English and German (Unpublished doctoral dissertation). University of East Anglia.

20.

Kasuya

(1998). Determinants of language choice in bilingual children: The role of input. International Journal of Bilingualism, 2(3), 327–346.

21.

Kenner

Ruby

Gregory

Jessel

Arju

(2007). International learning between children and grandparents in East London. Journal of Early Childhood Research, 5(3), 219–243.

22.

King

Fogle

(2013). Family language policy and bilingual parenting. Language Teaching, 46(2), 172–194.

23.

King

Fogle

(2018). Family language policy. In MaCarty

May

(Eds.), Language policy and political issues in education (pp. 315–327). Springer.

24.

King

Fogle

Logan-Terry

(2008). Family language policy. Language and Linguistics Compass, 2(5), 907–922.

25.

King

Logan-Terry

(2008). Additive bilingualism through family language policy: Ideologies, strategies and interactional outcomes. Calidoscópio, 6(1), 5–19.

26.

Lanza

(1997). Language mixing in infant bilingualism: A sociolinguistic perspective. Clarendon Press.

27.

Lanza

(2007). Multilingualism and the family. In Auer

(Eds.), Handbook of applied linguistics, Vol. 5: Multilingualism (pp. 45–67). Mouton de Gruyter.

28.

Lee

(2004). A survey of language ability, language use and language attitudes of young aborigines in Taiwan. In Hoffmann

Ytsma

(Eds.), Trilingualism in family, school and community (pp. 101–117). Multilingual Matters.

29.

Luykx

(2003). Weaving languages together: Family language policy and gender socialization in bilingual Aymara households. In Bayley

Schecter

(Eds.), Language socialization in bilingual and multilingual societies (pp. 25–43). Multilingual Matters.

30.

MacWhinney

(2017). The CHILDES project: Tools for analyzing talk—Electronic edition Part 1: The CHAT transcription format. http://talkbank.org/manuals/CHAT.pdf

31.

Mercer

(2012). The complexity of learner agency. Apples – Journal of Applied Language Studies, 6(2), 41–59.

32.

Parven

(2016). A young child’s intergenerational practices through the use of visual screen-based multimodal communication to acquire Qur’anic literacy. Language and Education, 30(6), 500–518.

33.

Pratt

Grieve

(1984). The development of metalinguistic awareness: An introduction. In Tunmer

W. E.

, et al. (Eds.), Metalinguistic awareness in children: Theory, research, and implications (pp. 2–11). Springer-Verlag.

34.

Quay

(2001). Managing linguistic boundaries in early trilingual development. In Cenoz

Genesee

(Eds.), Trends in bilingual acquisition (pp. 149–199). John Benjamins.

35.

Ruby

(2012). The role of a grandmother in maintaining Bangla with her granddaughter in East London. Journal of Multilingual and Multicultural Development, 33(1), 67–83.

36.

Said

(2014). A sociolinguistic study of multilingual talk at mealtimes: The case of an Arab family in London (Unpublished doctoral dissertation). Birkbeck, University of London.

37.

Saville-Troike

(1987). Dilingual discourse: The negotiation of meaning without a common code. Linguistics, 25(1), 81–106.

38.

Siiner

Hult

Kupisch

(2018). Situating language acquisition planning. In Siiner

Hult

Kupisch

(Eds.), Language policy and language acquisition planning (pp. 1–10). Springer.

39.

Smagulova

(2017). Ideologies of language revival: Kazakh as school talk. International Journal of Bilingualism, 23(3), 740–756.

40.

Smith-Christmas

(2014). Being socialised into language shift: The impact of extended family members on family language policy. Journal of Multilingual and Multicultural Development, 35(5), 511–526.

41.

Spolsky

(2012). Family language policy – The critical domain. Journal of Multilingual and Multicultural Development, 33(1), 3–11.

42.

Takeuchi

(2006). The Japanese language development of children through the ‘one parent-one language’ approach in Melbourne. Journal of Multilingual and Multicultural Development, 27(4), 319–331.

43.

Zhu

(2016). Transnational experience, aspiration and family language policy. Journal of Multilingual and Multicultural Development, 37(7), 655–666.