Child Multilingualism
A case study of the influence of parental strategies and interlocutor
presence on children’s code-switching patterns in a multilingual
family living in the Netherlands
Anna Mikhienko
UvA-ID: 12733490
Master in General Linguistics, University of Amsterdam
Date: 25-06-2020
Supervisors: dr. Suzanne Aalberse
Second Reader: dr. Sible Andringa
Abstract
The present study investigates how Mexican-Ukrainian multilingual children living in the Netherlands use four languages (Spanish, Ukrainian, Dutch, and English as an international language) in conversations with different family members who are multilingual, bilingual and monolingual speakers. By analyzing children's speech, we aimed to examine how their language choice is influenced by the family's language ideology. As a result of this, how the children code-switch between all four languages.
The study provides evidence for the influence of interlocutor presence on children's language choice. By being present in the conversation and knowing the language, caregivers determine children's language use in a certain way. Since the children have four languages at their disposal, parental strategies employed in the family to stimulate children's language development are also described in detail.
Keywords: multilingualism, parental strategies, code-switching, interlocutor effect, bilingual mode hypothesis, trilingual language acquisition, language choice, early trilingualism
Table of content
1. Introduction ...3
2. Theoretical background ...4
2.1 The Interlocutor effect on Bilingual Mode activation...4
2.2 Code-switching...6
2.3 Parental strategies...9
2.4 English as an international lingua franca...14
2.5 Research questions and hypotheses...16
3. Method ...18
3.1 The participants and their linguistic environment ...18
3.2 Data collection...18 3.3 Data analysis ...19 4. Results ...20 4.1 Code-switching...20 4.2 Parental strategies...24 4.3 Interlocutor effect...31
4.4 English as an international lingua franca...34
5. Discussion...37
6.Conclusion...38
Acknowledgements...40 References
Introduction
Trilingualism is generally treated in the literature as another type of bilingualism (Wang, 2008). Recent studies do not provide answers on whether these two phenomena can be used interchangeably or not (Lanza, 2004; Cruz-Ferreira, 2006; Swisher, 2006; Wang, 2008; Chevalier, 2012). The difference between bilingual and trilingual subgroups of multilingual speakers is not well-defined in many linguistic disciplines. For example, the “One parent – one language” strategy is often chosen in bilingual families. It is interesting to know how this strategy would work in multilingual families where parents are multilingual themselves.
In this study, we focus on language choice and code-switching patterns in multilingual children exposed to four languages. This study investigates the naturalistic speech of Mexican-Ukrainian multilingual children when interacting with different interlocutors in their family setting in order to investigate how language choice and code-switching are influenced by the presence of multilingual and monolingual interlocutors. Would their code-switching patterns be different when talking to multilingual, bilingual, or monolingual family members? What are the main code-switching patterns? What is the influence of interlocutor presence on children’s code-switching patterns? In this study, it would be interesting to discover how children’s language choice and code-switching patterns would depend on the context and the interlocutors. Besides, we also look at the parental strategies to stimulate the children to speak and to direct their language choice patterns. A central question is whether there are any parental strategies specific to multilingual parents only.
To sum it up, in this case study, we investigate the influence of parental strategies and interlocutor presence on children’s language choice and code-switching patterns. This exploratory study on code-switching and interlocutor effect in specifically multilingual children, will shed some light on multilingualism as a separate field from bilingualism.
The next section provides an overview of previous research; Section 2.1 discusses the relevant literature on the Bilingual Mode Hypothesis and the Interlocutor effect. Section 2.2 then discusses the code-switching, and Section 2.3 provides a more detailed description of Parental Strategies. Section 2.4 gives an overview of background information about English as an international lingua franca. Finally, the research questions and hypotheses of the present study are presented in Section 2.5.
2 Theoretical background
2.1 The Interlocutor effect on Bilingual Mode activation
The concept of the Bilingual Mode Hypothesis was introduced by Grosjean (1989). The Hypothesis proposes that there is a continuum between two modes in which bilingual (and multilingual) speakers can be. One end of the continuum is the monolingual mode when bilinguals have to stick to one language when interacting with monolingual interlocutors; their other language is being suppressed for some time. The other end of the continuum is the bilingual mode, which means both languages are activated to a high degree.
In 2013, Grosjean included more information about multilingualism in his theory; he claims that trilinguals and multilinguals can be in bilingual mode (not ‘trilingual’ or ‘multilingual’ mode). In this case, “bilingual” mode would mean the state of the activation of more than one language but not the number of how many languages are activated at the moment. According to Grosjean, language mode in trilinguals or quadrilingual is like in bilinguals determined by the interlocutors who are present. Multilinguals are in a monolingual mode, when the people they are interacting with are monolingual in one of their three languages, or when they share only one language with another bilingual or multilingual speaker. However, they can easily switch to bilingual mode, if their interlocutors share the same second language, or languages in common and feel comfortable to speak both.
The movement along this continuum depends on several factors as the interlocutor effect, the situation, the content of discourse, and the function of the interaction (Dijkstra & G. van Hell, 2003). In this study, we look at monolingual and bilingual modes from the perspective of behavior, how the children behave in the presence of different interlocutors (multilingual mother, bilingual father and monolingual grandmothers), whether the children would tend to code-switch more in the presence of the multilingual speaker and whether they would be only in monolingual mode when talking to monolingual speakers without code-switches to other languages at their disposal.
It is hypothesized that bilinguals tend to code-mix more in bilingual mode. In this mode, both languages are activated; thus, the speaker is to choose a “language base,” which is the main language for interaction with a bilingual interlocutor. After defining the “base”, bilingual speakers intermix lexical or grammatical items of the second language; this process is called by Grosjean “language mixing”. Separate words, phrases, or sentences can be borrowed from the second language and integrated into the base language (Grosjean, 1989).
In the literature, theories and findings from studies of bilinguals are often assumed to apply to trilinguals by extension; thus, we can speculate that there is a possibility of trilingual and bilingual modes being different, or having differences which are not well-defined at the
moment. Bilingual individuals have been studied more thoroughly than trilingual ones. Based on this, we can assume that children who speak more than two languages, as trilingual or quadrilingual children, they can function in ‘trilingual’ or ‘quadrilingual’ mode, which might also have its peculiarities when compared to the bilingual mode.
Since multilingual families where parents speak more than one language tend to mix languages, we assume that their languages activated, and they tend to operate more in bilingual mode. Mixing languages and code-switching occur even if multilingual families have monolingual language ideology when one interlocutor speaks one language.
If multilingual parents share the same languages with their children and the children are aware of this, then children’s bilingual mode will be activated when a multilingual parent is around. Multiple interlocutors activate the bilingual mode, and, as a result, children tend to code-switch more. In other words, constant code-switching and the ability to switch between monolingual and bilingual modes are key features of families where parents speak several languages to their children. It is claimed by Döpke (1992) that bilingual speech cannot but be mixed because all languages are activated at the same time. Döpke writes in her book that even among young children mixing has been found to result from a choice based on who the interlocutor was. In what follows, that mixed utterances are prevalent in multilingual families.
Language mixing is frequent among both the children and the parents in multilingual families (Mackey, 2000). When the parents are also bilingual, or highly proficient in two languages, they tend to make use of all language resources they have at their disposal, or ‘alternate use of two or more languages’. These linguistic switches cannot be equated with disfluency, and it is more evidence of linguistic advantages.
What Grosjean (2013) comments about language modes in multilinguals are that “trilinguals are in a monolingual mode when the people they are interacting with are monolingual in one of their three languages, or when they share only one language with another bilingual or multilingual”. Once again, the multilingual children are predisposed to code-switch more in the presence of multilingual interlocutors.
Additionally, it is highlighted that the naturalness in everyday communication is essential (Cruz-Ferreira, 2006), which in multilingual families, often involves the use of more than one language by the same person.
However, an important question would be how both multilingual parents regulate and control their switching, what holds them back from constant switching apart from an established language policy in the family? One factor might be the interlocutor effect. If a monolingual interlocutor is present is a critical factor or ‘a switch’ for moving along the modes to regulate when other languages should be blocked out.
Results of the studies (Genesee & Nicoladis, 1996) show that bilingual children show sensitivity to the language choice of their interlocutor and can reliably choose the language preferred by their interlocutor from at least two years of age, the same pattern can be observed among trilinguals as well (Quay, 2008). Thus, the interlocutor effect is a crucial factor in determining language choice, and the interlocutor predicts what language the children would speak back to him.
Cruz-Ferreira (2006) also claims that bilingual children alternate languages in order to communicate effectively with interlocutors that are identified with each of the languages. In her research, children demonstrate their awareness of the common knowledge that is tacitly assumed of their interlocutors: they know that they know what they mean.
Ultimately, the presence of different interlocutors activates bilingual mode and increases the number of code-switches in multilinguals. By taking into account what language interlocutors who are present in the conversation speak, multilingual children choose in what language to convey the information. Thus, the interlocutor presence determines children’s language choice.
2.2 Code-switching
As previously described, it is impossible not to mix languages in a multilingual environment when various languages are presented in the conversations (Cruz-Ferreira, 2006).
To characterize code-switching patterns in our data, I use Muysken’s framework (2007)1. He claims that there are three distinct, if not always clearly separable, strategies in code-switching among multilinguals: insertion, alternation, and congruent lexicalization. These three code-switching types are presented in terms of optimization strategies, which multilingual speakers use as a “shortcut” in conversations.
The first strategy is called insertion, the insertion of well-defined chunks of language B into a sentence that otherwise belongs to language A. Simply put, insertion is the incorporation of a word or words from one language into an utterance in another language. Usually, grammatical or lexical properties of the first language are used as the matrix or base language. To illustrate this strategy, Muysken (2013) gives the following example of inserting Spanish utterances into Quechua (1),
1 In 2013, Muysken added one more strategy in his framework called backflagging. However, in this study, we use the framework presented earlier.
(1) Q‘aya suya-wa-nki [las cuatro-ta] Qo-yku-sqa-sun-ña [bukis].
tomorrow wait-1OB-2SG at four-AC give-ASP-ASP-1PL.FU-con box “Tomorrow you wait for me at four. We’ll have a go at boxing.”
(Quechua/Spanish; Urioste, 1966, p. 7)
The second strategy is the alternation between stretches of words in different languages. In other words, alternation is the succession of fragments in language A and B in a sentence, which is overall not identifiable as belonging to either A or B. As in the Example (2)
(2) Ándale pues, and do come again. Go on well and do come again “That’s all right then, and do come again.”
(Spanish/English; Gumperz & Hernández Chavez, 1971, p. 118)
The third strategy is the congruent lexicalization of a shared language structure with words from different languages. The speaker uses the elements from either language in a structure that is wholly or partly shared by languages A and B. This strategy involves cases of code-switching where languages have substantial parts of their grammar in common, and the switching mostly involves inserting words from either language into the shared structure. For instance (3),
(3) Això a él a ell no li i(m)porta this to him to him not 3SG.CPRO matters “This he, he doesn’t care.” (Catalan/Spanish; Vila i Moreno, 1996, p. 393)
Furthermore, Muysken et al. (2007), elaborate more on extralinguistic factors which influence the code-switching patterns. Each of the three strategies (insertion, alternation, and
congruent lexicalization) has been influenced by extralinguistic factors favoring the pattern.
For example, stable bilingual communities with a tradition of language separation tend to use
alternations more than other code-switching patterns. Insertion is frequent in colonial settings
and recent migrant communities with a tradition of overt language separation; speakers display asymmetry in their proficiency in two languages. Congruent lexicalization pertains to the communities with two languages of roughly equal prestige and with no tradition of overt language separation.
According to this description, we hypothesize that the possible form of code-switching in our data will be an insertion, due to two reasons. First of all, both parents grew up in monolingual communities and later in life moved abroad, which can be considered “recent migrants”. Perhaps, this term “recent migrants” can be more specific, in our case, the family falls into the category of economic immigrants (skilled workers and business people). Second of all, as it is claimed by Muysken “asymmetry in speaker’s proficiency in two languages,” which may be the case for the family who participates in the research since the children have different environments for learning languages. They learn Dutch formally at school, and it is also the language to play with their peers, and Spanish and Ukrainian are home languages. This difference in language settings may play an important role in the “asymmetry” of children’s proficiency since formal education and daily conversations at home presume different vocabularies and language register. For more details, see Figure 1.
Codemixing pattern
Linguistic factors
favoring this pattern Extralinguistic factors favoring this pattern Insertion Typological distance
Colonial settings; recent migrant communities; asymmetry in speaker’s proficiency in two
languages
Alternation Typological distance Stable bilingual communities; tradition of language separation
Congruent lexicalization
Typologically similar languages
Two languages have roughly equal prestige; no tradition of overt language separation
Figure 1. Code-switching patterns
Importantly, when discussing what counts as a code-switch, we need to distinguish code-switches from loanwords. Loanwords are adopted from one language (the donor language) and incorporated into another language without translation. At the same time, code-switches occur when a speaker alternates between two or more languages in the context of a single conversation (Chesley, 2010). In the current data, we observed this strategy of borrowing words from another language.
According to Stavans and Swisher (2006), if there are more than two languages spoken at a time, it is worth noting that switches mostly occur between two of the three languages at the sentence level. In this case, it is interesting to look at our data and see when code-switches mostly occur and what are the most common code-switching patterns.
To conclude, Muysken (2007) characterized the phenomenon of codeswitching in terms of bilingual speech strategies (insertion, alternation, and congruent lexicalization), which we
described in this section; further analysis of our data through this framework will be presented in the results section.
2.3 Parental strategies
After describing key code-switching types from Muysken’s framework, in this section, we focus on language ideology and parental strategies more in detail. Two main frameworks we look through at parental strategies are Lanza's (2004) and Döpke's (1992) frameworks. These two frameworks have some similarities and differences. In terms of differences, Lanza's parental strategies are used more as a reaction towards child language mixing; they comprise a range of caregiver responses to children's use of a language which is different to that of the caregiver, while Döpke's strategies are teaching techniques of proactive language teaching, stimulating children to know more words and use them actively. In terms of similarities, both the teaching techniques described by Döpke (1992) and the parental discourse strategies identified by Lanza (2004), have been shown to play a role in a young child's motivation to speak another language (Chevalier, 2012). Thus, these two strategies, in combination, influence children's language choice and code-switching patterns.
Besides practical strategies that parents use, language ideology plays an important role in children's motivation to speak different languages. Palviainen & Bergroth (2018) highlight the importance of beliefs and attitudes about languages and language development in mixed-language families. Ideology is seen as 'a set of beliefs' and as 'the underlying thought system that is expressed through discourse' (Määttä & Pietikäinen, 2014). In her research Curdt-Christiansen (2013) demonstrates the following, it is worth quoting at length: "when tracing the processes of bi/multilingual development and minority language maintenance, researchers have shown that language ideology is often the underlying force in family language choice and decisions on what language to practice and what measures to employ in order to influence family members' language behaviors". Thus, language ideology and parental strategies play an essential role in children's multilingualism, language choice, and code-switching patterns. Based on language ideology and language believes, parents can use different approaches toward language teaching, whether to separate the languages or not, and whether code-switching is acceptable or not. In this study, we look at parental strategies through Lanza's (2004) and Döpke's (1992) frameworks, described in detail below.
Lanza's framework represents the idea of two language modes previously described by Grosjean (1989, 2013): monolingual and bilingual modes. If in monolingual mode, primarily one language is spoken at a time, two or three languages are activated in bilingual mode, and speakers tend to code-switch and bring lexical items of one language into the other language
base. Both modes are located along a monolingual-bilingual continuum. Lanza presents five basic types of parental discourse strategies in response to code-switching by children. These strategies are explained thoroughly and with examples below. Additionally, Figure 2 provides a visual overview of these strategies described by Lanza.
Figure 2. Lanza (2004): parental discourse strategies
The framework starts at a monolingual mode when a child switches to another language. Her parent explicitly shows that it is crucial to stay in a monolingual mode by requesting clarification. With the Minimal Grasp Strategy, in other words, request for clarification, a parent shows the necessity to keep speaking the same language and stay in a monolingual context, it is a signal for a child that she is speaking the wrong language from the perspective of the caregiver. For example (1) given by Chevalier (2015), the mother indicates that she cannot understand the child's answer, so she keeps asking the same question.
(1) Mother: You're gonna eat some food?
Thomas: og Mama and Mama
Mother: What am I gonna do? <- minimal grasp
Thomas: spis eat
Mother: What am I gonna do? <- minimal grasp
Minimal grasp responses are usually the following words 'what?', 'huh?', 'I do not understand', 'say that again' and Wh-questions.
The next strategy is the Expressed Guess Strategy, where the parent reformulates or questions the child's utterance using the target language after having indicated comprehension
of the child's use of non-target language. These responses are constraining because the adult queries the child's utterance and requires them to react in some way. In the example (2) below, the child speaks Swiss German to her English-speaking aunt, and the child says 'Ich Schriibe öppis' ('I am writing something'). As a reaction to that, her aunt responds in English with an expressed guess by asking whether the child is going to write or not, and what exactly she is going to write, thereby moving back to English as the language of discourse.
(2) Daughter: Ich Schriibe öppis I write something “I am writing something”
Aunt: Are you gonna write? What are you writing? (Chevalier, 2015)
The following strategy is the Repetition Strategy, which is used when the parent repeats the child's meaning using the other language in a non-question form occur. It is a response in which the caregiver translates at least one lexical item of the child's utterance. The following example (3) illustrates this strategy. The same child from the example above addresses her aunt in Swiss German by saying 'chueche' ('cake'), her aunt reiterates the same word but translating it into English 'cake'.
(3) Daughter: Chu:eche “cake”
Aunt: That is a cake! (Chevalier, 2015)
In the Move on Strategy, children are stimulated to be in a bilingual mode; the parent continues the conversation in the language A, the language spoken before the child switched to the language B, exhibiting comprehension of the child's use of the other language. Lanza (2004) gives this example (4), excerpted from Ex.29 (from Lindholm and Padilla, 1978b)
(4) E: Cómo va? 'Where does it go?'
C: In the back, it goes in back E: Qué son estos?
C: Parrots. E: Qué hacen? 'What do they do?'
As we can see from the given example (4), one interlocutor continuously asks Spanish questions in response to the second speaker's answers in English.
Lanza (2004) claims that Move On strategy reveals a bilingual identity of speakers. Both roles of a monolingual and bilingual are on the front stage, with one producing the one language and the other comprehending the language.
The outer end of the continuum between monolingual versus bilingual mode strategies is the Code-Switching Strategy, in which two languages are activated so that the speaker can have more code-switches in his speech. This strategy consists of both inter-sentential and intra-sentential code-switching. The intra-intra-sentential code-switching is when a parent incorporates the child's use of non-target language into his utterance. An occurrence of inter-sentential code-switching is when a parent switches into another language at the sentential level. For example (5), Lanza (2004) describes when Siri and her mother just finished lunch.
(5) Siri: jeg mett “I full”
Mother: You're mett? What does mama say? Lanza (2004): 268
In example (5), the mother integrates the child's response in Norwegian in her clarification question in English. Lanza asserts that ‘what Mama says’ and ‘what Papa says’ strategies are common in families adhering to the one person, one language strategy. This type of code-switching is widely used by the mother, who participates in our research (See the Results section).
To sum it up, all five strategies can be used by parents to show language policy at home, whether the child should speak one language and stay in a monolingual mode or switching to other languages is acceptable.
For a long time, the parents' role as language teachers for their children has been downplayed (Döpke,1992). To the best of my knowledge, one of the first works related to parental language teaching strategies was the book "One Parent – One Language," written by Döpke in 1992. The concept of "teaching technique" has been put forward; she illustrates how parents' behavior and approaches to "teach" a language to their child influence a child's fluency
in a minority language. Additionally, parents' awareness of their function as a language teacher plays a vital role in child second language acquisition. In her book "One Parent – One Language", Döpke presents a framework of teaching techniques that comprises modeling, rehearsing, patterning, and eliciting techniques. Each of these strategies is explained below.
Eliciting strategies describe in what way children respond to questions by using various
techniques, such as asking choice questions ("or" - type questions, when a child is supposed to choose one); encouraging children to speak more; requesting for translation (asking children to translate a specific term or a whole utterance); asking wh-questions (real questions which require a single element answer, as in "What is this?"); asking yes-no questions (contracting polar questions) and giving prompts (or request for insertion, when parents provide slots for the child to fill in, as in the example (6).
(6) P: that is the …. C: wolf
Rehearsing strategies include various techniques as well, such as incorporation (parents
incorporate parts of child's utterance into their speech); repeating (repeating child's utterances); modeling; elaboration (when the function of the object is explained, for example, "the car is to ride in"); paraphrasing (general paraphrase which conveys the same meaning with different words) and semantic correction (correction of the meaning, for example (7)),
(7) C: This is a mouse P: This is a cat
The last strategy is called feedback, which can be either positive (when a parent shows to the child that the child's last utterance was linguistically right) or negative (when a parent shows that the child's utterance was unacceptable).
To sum it up, after presenting two relevant theories to this research framework, we may conclude that Döpke's and Lanza's strategies complement each other when it comes to children's motivation to speak different languages. Lanza's approach elaborates more on parental approach towards code-switching, and Döpke's strategies presuppose teaching techniques that parents can use to stimulate children's language learning.
2.4 English as an international lingua franca
Apart from analyzing code-switching patterns in general, in this study, we specifically look at code-switches in English since this language is spoken neither at school nor at home. In this section, we look at the English language as "a world language" widely spoken among expatriates all over the world and, notably, as a second national language in the Netherlands.
Firstly, in expatriate communities all over the world, English is being used as a lingua franca among multilingual speakers. After analyzing trilingual language acquisition in children in several multilingual families, Chevalier (2015) claims that the status of English as the global language has resulted in positive attitudes towards learning this language, as well as fluency in speaking it among non-native speakers. Due to the fact that English has a high status, children may have more motivation to speak it. If it is common and natural to speak English among adults who live abroad, then it is interesting to discover what influence it would have on their children's language acquisition.
"The spread of English as a lingua franca threatens not only the traditional ownership of English as a property of its native speakers" but the status of the native speaker (Davies 2003). Thus, English is owned by native speakers of other languages. Furthermore, Seidlhofer (2005) pointed out that by definition, nobody speaks ELF (English as lingua franca) natively and advocates that speakers of ELF should be considered as language users in their own right. Meierkord (2005) stresses that English does not have a stable community of language users but is in constant flux. This implies that English is less 'foreign' to non-native speakers (McArthur 1996)
Secondly, as a consequence of internalization, Dutch society is increasingly penetrated by the English language. Taking into consideration that the Netherlands is "one of the most advanced countries in Europe concerning the integration of instruction of English in the national education system," English is considered as a national second language (Ammon & McConnell, 2002). Nortier (2011) claims that the number of Dutch people with little to no knowledge of English has decreased over the years due to the position of English in the Dutch educational system. Due to these changes in education, 'there is hardly any chance to complete a university degree program without demonstrating a high level of linguistic competence in English' (Dybalska 2010)
What is more, English is not only a language of instruction in a bilingual classroom setting, but it is often used as a working language or second language next to Dutch in commerce (Booij, 2001). As follows, it is also a necessary skill in business communications and job search. It is ostensibly seen as an essential skill in job requirements in the Netherlands;
candidates are assumed to speak English by default, even if sometimes proficiency in Dutch is not required (Berns et al. 2007).
In terms of the media and culture, "young people are plugged into digital media as well as global culture and trends" (Edwards, 2016). Additionally, the Netherlands follows the practice of subtitling movies rather than dubbing, which has a significant impact on English proficiency.
Some previous studies have also noted the influence of mass media on language acquisition. For example, preschoolers, who viewed Sesame Street, showed vocabulary growth (Vandewater & Bickham, 2004). Maier et al. (2012) highlight the benefits of contact with English which occurs primarily through the media such as television when shows and films are broadcasted in the original version with subtitles. This accessibility of input and its high status stimulate naturalistic learning.
As Edwards (2016) resumes, English is widely used in the Netherlands in several domains, including education and the media. Thus, it has a high status in society, but it is also widely accessible to children in daily life. Given that, we may claim that children in the Netherlands learn English due to two main reasons. First, the language has a high status in society. Falk and Bardel (2010) claim that language status is an essential factor that comes into play when a multilingual acquisition occurs. Second, children have enough input in English; thus, they can acquire all necessary grammatical and lexical components of the language to reach a certain level of proficiency.
In multilingual contexts, many languages are learned in the natural environment; thus, languages may intertwine, and it becomes harder to define the notion of "first, second, and third language". Given that, we may count Spanish and Ukrainian as one first language in children's life, Dutch as a second school language, and English as a third language. In this case, many studies that look at multilingualism and third language acquisition can be applicable. Jessner (2006) claims that English is the third language for many school children who are speakers of other heritage languages. The sociolinguistic or cultural surroundings play a decisive role in the structure and specifications of multilinguality and multilingualism (Jessner, 2006).
After all, language acquisition of school and home languages and English as an international language is a complex issue that requires not only language knowledge but also extra information about the social aspects of the linguistic environment. Bonnet & Siemund (2018) claim that the typical third language learning situation is increasingly difficult to find since the globalization and international migration have changed the distribution and use of languages, especially in younger generations. As an aside, it is worth noting that English language acquisition is a specific topic in multilingual studies and requires more attention when
studying expatriate families. Research specifically concerning a third language acquisition is a relatively new branch of the study of language acquisition, and it is expanding substantially.
2.5 Research questions and hypotheses
In this study on child multilingualism, two multilingual Mexican-Ukrainian children living in a third language country (the Netherlands) were observed in conversations with different family members: multilingual mother, bilingual father, and monolingual grandmothers. In order to investigate how the children use all four languages in monolingual and bilingual settings, parents' parental strategies have been thoroughly studied. Code-switching patterns affected by the presence of different interlocutors are taken into consideration.
The children's speech was recorded with four different interlocutors. The mother usually addresses her children in Ukrainian when she is alone with them, but as they know that she is a multilingual speaker, they may speak English, Spanish, and Dutch with her. The father speaks Spanish to the children, and the children know that he speaks English with the others. Thus, the daughter inserts English code-switches when talking to her father. In this study, we investigate whether children's code-switching patterns differ when talking to multilingual mother, bilingual father, or monolingual grandparents.
As we already discussed in section 2.1, Grosjean (1989) developed the general notion of a 'language mode', referring to the relative state of activation of a bilingual's two (or more) languages and language-processing mechanisms. One language (the base language) may be active, while the state of the other language may vary from deactivated (monolingual language mode) to relatively active (bilingual language mode). Therefore, the hypothesis is that the children are in bilingual mode when multiple interlocutors are around or when interacting with other multilinguals (for example, their mother). We expect more code-mixing in bilingual settings than in monolingual settings, due to greater activation of the other language. For example, the children would insert more foreign words when conversing with their mother rather than their monolingual grandmother, who is not able to understand the meaning of their code-switches.
According to that, the children are expected to be in monolingual mode when talking to monolingual speakers; thus, they have fewer code-switches. As the older sister grew up in an expatriate community in Nigeria, she is assumed to be more proficient in English; we expect her to show a higher preference for English in bilingual contexts, than in her younger brother, who was born in the Netherlands. Baker (2011) highlights the importance of the language choices of siblings, environmental influences such as neighbors and friends, mass media, and nursery school. Since the younger son went to a Dutch playschool, and according to his mother,
Dutch is his dominant language; thus, we expect him to show a general preference for Dutch and produce more Dutch code-switches in a monolingual Spanish or Ukrainian context.
Beliefs and attitudes about language ideologies in the family and language development play a significant role in mixed-language families (Palviainen & Bergroth (2018). Given that, it would be interesting to analyze what strategies the parents use to support multilingualism in the family by using languages separately and what attitude parents have towards code-switching to other languages.
Code-switching can be viewed in terms of optimization strategies, which multilingual speakers use as a "shortcut" in conversations (Muysken, 2013). In this research, we want to investigate the main code-switching patterns and parental strategies in multilingual families. The overarching question, 'How do Mexican-Ukrainian multilingual children code-switch with
multilingual, bilingual and monolingual interlocutors, and what are the parental strategies used in the family?', is addressed by answering the five sub-questions that are summarized
below.
1. What languages do the children speak to what interlocutors? 2. With interlocutor X, what language do the children switch to?
3. What types of code-switches do the children display in their speech the most?
3. Method
3.1 The participants and their linguistic environment
The children that this study focuses on are Mexican-Ukrainian trilingual girl of 6;5 years old, who will be referred to as 'Victoria' and her younger brother (4 years old), who will be referred to as 'Daniel'. The names of the two children are pseudonyms. The children currently live in the Netherlands, together with her parents.
Victoria was born in Ukraine and grew up in Nigeria in an English-speaking expatriate community since the family lived in Nigeria for several years. English was her first school language. After moving to the Netherlands at the age of 2 years old, she experienced difficulties in learning Dutch, however, after 2-3 years, Dutch became her dominant language, and now, she confidently speaks this language to her peers and friends. Daniel was born in the Netherlands and was exposed to the Dutch language since birth. Dutch is also his language to converse with his friends in the playgroup. Victoria and Daniel prefer speaking Dutch to each other when playing since it is their dominant school language. Spanish and Ukrainian have always been children's home languages.
The mother is a native speaker of Ukrainian and Russian, and she is multilingual; she additionally speaks English, German, Dutch and Spanish. She is majoring in Translation Studies and has rich experience in teaching languages; she has designed her active language learning methodology. Her professional expertise may influence the way she uses languages with her children. The father is a native Spanish speaker from Mexico who understands Dutch and Ukrainian at a beginner's level. The mother typically speaks Ukrainian to her children when she is alone with them, and she teaches them other languages (Spanish, English, and Dutch). The father only speaks Spanish to the children, and when the whole family is together, the main language of conversation is Spanish. English is a language spoken between parents. The children also study English at school. During the recording period, the maternal grandmother, who is a monolingual Ukrainian speaker, was visiting. The family has regular trips to Mexico and Ukraine to visit their grandparents and extended family.
3.2 Data collection
The data for this study consist of audio recordings of spontaneous speech of two children in interaction with their parents and grandparents in the home. These recordings were mainly opportunities to analyze the children's language use and provided support for the analyses of the children's language use in parent-child interactions. Their mother made audio recordings of their speech during everyday activities. They were instructed not to pay attention to the voice
recorder and act as naturally as possible. The parents were not given any instructions on how to act, and were allowed to determine the activity for their taping sessions (e.g., playing with toys or reading a book). After the recordings, an informal interview with the mother was done in which we discussed the child's development and language-related activities. The parents were collaborators in the research endeavor.
The research project was approved by the Ethics committee of the Humanities Faculty of the University of Amsterdam, and informed consent was obtained from the parents.
3.3 Data analysis
A total of 18 recordings, adding up to 2 hours 44 minutes of speech, were analyzed. The recordings were transcribed and analyzed in ELAN (https://archive.mpi.nl/tla/elan).
Languages were coded 'Dutch', 'English', 'Ukrainian', 'Spanish' and 'code-switching', based on the lexical content of a sentence. Tiers' names include both the interlocutor and the language, for instance, 'Mother Ukrainian', 'Father Spanish'. When decoding, language settings were also mentioned in the data file. To answer our research questions, we calculated proportions of English, Dutch, Spanish Ukrainian, and 'code-switching' utterances in each of the four languages. Proportions of each of these languages were calculated for each interlocutor in the corresponding context. Interlocutors were tagged as 'mother', 'father', 'daughter', 'son', 'maternal grandmother', and 'paternal grandmother'. Furthermore, words like 'papa' and 'mommy' were treated as proper names, so that they were never interpreted as instances of code-switching. Established loans which were conventionalized in the family were not coded as code-switches. Figures 3 and 4 provide a visual overview of the coding methods.
Figure 3. Example of the coding schemes in ELAN
In conclusion, the study illustrates how interlocutor effect and language mode influence the child's language choice and code-switching at the conversational level. For example, how does the code-switching pattern change in the presence of different interlocutors? This part of the analysis allows for a more explorative examination of the data. Additionally, the parent's reaction to code-switching was studied and analyzed, and the main parent teaching strategies were discussed in the research.
4. Results
Section 4.1 first provides the results of the analysis of language code-switching patterns of both children. Subsequently, Section 4.2 discusses parental strategies used in the family, and Section 4.3 presents the influence of the presence of different interlocutors in a conversation. Section 4.4 gives more information on English language acquisition in an international context.
4.1 Code-switching
The parents recorded 2 hours 44 minutes of children’s speech at home, which consisted of 1 hour 23 minutes of recordings in Ukrainian, 34 minutes in Dutch, 37 minutes in Spanish, and 8 minutes in English with different family members (mother, father, paternal and maternal grandmothers). The home recordings included dinners, breakfasts, playtime, cooking, reading, and drawing time. In this section, code-switching patterns to these four languages will be analyzed. See Table 1 and Table 2 below.
The children in this study are exposed to four languages in daily life. In total, Ukrainian is spoken the most in the family (1h 23 min 58 sec). Surprisingly, home language Spanish (37 min 50 sec) and children’s school language Dutch (34 min 50 sec) have almost even distribution. English is spoken less than the other languages (8 min 15 sec). The mother speaks all four languages in the recordings; thus, it is interesting to see how exactly she alternates between all these languages.
There are 63 code-switching utterances in the recordings, among which 22 utterances are code-switches to English (34%) and 41 Dutch code-switches (66%). It is worth noting that the children do not have Ukrainian or Spanish code-switches even though their father understands a bit of Ukrainian, and their mother speaks Spanish.
Interlocutor/Language Ukrainian Dutch Spanish English Total Mother 0:32:36 0:02:24 0:04:04 0:01:18 0:40:22 Father 0:00:10 - 0:09:32 - 0:09:42 Daughter 0:33:17 0:17:10 0:13:08 0:06:22 1:09:57 Son 0:05:55 0:15:16 0:08:16 0:00:35 0:30:02 Maternal grandmother 0:12:00 - - - 0:12:00 Paternal grandmother - - 0:02:50 - 0:02:50
Table 1. Distribution of languages per speaker
Interlocutor/Language Ukrainian Dutch Spanish English
Mother 0:32:36 0:02:24 0:04:04 0:01:18 Father 0:00:10 - 0:09:32 - Daughter 0:33:17 0:17:10 0:13:08 0:06:22 Son 0:05:55 0:15:16 0:08:16 0:00:35 Maternal grandmother 0:12:00 - - - Paternal grandmother - - 0:02:50 -
Amount of time per
language, total 1:23:58 0:34:50 0:37:50 0:08:15
Table 2. Distribution of time per language
In terms of code-switching patterns, based on Muysken’s framework (2007), the most common code-switching pattern used in the family is insertion. Two common code-switching patterns found in the data are one-word insertion and one phrase insertion, which are incorporated word or words from one language into an utterance in another language. We can characterize these insertion code-switches concerning the lengths. In total, there are 35 one-word code-switches (55%) and 28 one-phrase code-switches (45%). Other strategies from the Muysken’s framework (2007), such as alternation and congruent lexicalization, were not found.
Regarding language choice, since Dutch is an official language of school, it is one of the main languages for communication between the brother and the sister. The son prefers speaking Dutch more than other languages since he was born and raised in the Netherlands, while the daughter actively speaks other languages. Moreover, the children often borrow Dutch words related to school into other languages, for example, zomer (‘summer’), ik ben dood (‘I am dead’), de muis (‘mouse’). In Example (1), the mother asked her daughter questions about her class at school, how many new students they have, and who they are. She says a Dutch word because she could not remember the Ukrainian equivalent.
(1) Mother: Кто еще? Who else? “Who else?”
Daughter: и один jongetje and one boy “and one boy”
Mother: хлопчик boy
“a boy” (N15, 6:58)
One category of words that were not counted as code-switches and has been coded as “family words”. It is a category of words has been conventionalized in the family vernacular, they are used by some family members but understood by all family. These words are frequently used by different family members by inserting them in other language utterances. The reason for these insertions might be that speakers lack translational equivalents, and feel more comfortable about the word-meaning when the word-form is borrowed. The children do not switch to Ukrainian, apart from one word ‘мультик’ (cartoon) which they use in four utterances, it has not been counted as code-switches. For example (2a), the son inserts a Ukrainian word “мультик” (‘cartoon’) into his sentence in Spanish when addressing his father. He also uses the same word in his question to other family members (2b).
(2a) Yo quiero мультик I want cartoon
“I want a cartoon” (N10, 05:08)
(2b) Donde el мультик? Where a cartoon
“Where is the cartoon?” (N13, 1:00)
The same word, “мультик” is used by the daughter as well. She mixed even more languages in one utterance, and she uses Ukrainian, English and Spanish words in the same sentence when talking to her parents (3)
(3) En un мультик Muy funny мультик. Em a cartoon very funny cartoon
‘This is a cartoon, a very funny cartoon’ (N18, 10:28)
Another example (4) of family words which are unique words related to Dutch realities, such as “slagroom” (dairy product) and “bakfiets”(a type of bicycle for parents and children). The word “slagroom,” which is a milk product, is challenging to translate because it requires background knowledge of the process of how exactly the product had been made. “Bakfiets” exist only in the Netherlands, that is why it does not have an equivalent in other languages. Instead of defining the word in other languages, the family uses an optimization strategy and borrows it. As in Example (4),
(4) Mother: на чем поедем? on what ride ? “How will we get there?”
Daughter: на «bakfiets» on bakfiets “by bakfiets”
Mother: bakfiets у нас на ремонте bakfiets we under repair
“It is under repair” (№15, 11:50)
Our results suggest that the code-switches occur mostly when speaking to multilingual speakers, such as the mother who speaks all four languages presented in the study and the father who speaks Spanish and English. Since the father is proficient in English, the daughter feels comfortable to insert English words in her speech. For example (5), the daughter was telling her father that she had already used the spray.
(5) Daughter:
cuando va a tener comezón, va a estar así, un spray... when go have itch, go to be here a spray
“when it's going to be itchy, um ... We can do like this, a spray” (N5, 7:43)
Since English has become an international language, often the children use these most common international worlds, such as “baby”, “Christmas (N7, 0:42)”, “muy funny/ muy dangerous” (“very funny/ very dangerous”) (N9 00:49), “маленький cookies”(“little cookies”) (N15, 08:20)
As we can see in the recordings, when the daughter does not know the correct word, her mother would guide her, ask clarification questions, and help her if she still cannot remember. Those words which the girl does not know would be the trigger for switching. This code-switching is used as a teaching tool to make sure the children practice lexical retrieval of the same item in Spanish, Dutch, and Ukrainian. Their mother’s reaction is pedagogical; deliberately asks them to say the same word or phrase in other languages.
Code-switching to other languages is triggered when the children do not know how to say one phrase, or a word in the language when the interlocutor is her mother. For instance, their mother asks them to name the same object (or the same phrase) in other languages, and she would help the children to name all objects correctly, keeping all languages separately.
To conclude, the children switch to English and Dutch; the majority of code-switches are one-word insertion to the ‘base’ languages (Spanish and Ukrainian). Since Dutch is a school language of the children, there are many school-related vocabulary code-switches. Regarding English, the children do not speak the language, but they insert word or one-phrase utterances when speaking other languages. There is a group of words that have been conventionalized among family members, called ‘family words’; these words are loanwords borrowed from one language and adopted in another language.
4.2 Parental strategies
The most common way to establish boundaries between the two languages is by separating the languages by parents, the so-called "one person-one language" (Ronjat, 1913). In general, "one person – one language" principle is successfully employed in the family. In the family, the father and paternal grandparents speak Spanish to children, the mother and maternal grandmother speak Ukrainian. The ideology of overt language separation (Muysken, 2007) plays an important role in the family. Each language functions as a complete system without constant code-mixing; languages should be kept apart at all times. Within this approach, switching between languages is not considered helpful, since this might disturb the distinct separation of the languages. Thus, the monolingual approach is used in the family when one language should be spoken at a time.
Furthermore, code-mixing does not have the communal support, by which we mean that it is not common to switch between Dutch, English, Spanish, and Ukrainian in the society where the family lives. What is more, it is not common to switch between these four languages among adults in the family either. In the light of the observations in the family, a high level of adherence to the one person, one language strategy has been observed when the interlocutors are paternal and maternal grandmothers and the father. However, when the mother in this study, who is multilingual herself, is engaged in the conversation, other types of parental strategies are used.
Two approaches by Döpke and Lanza are described in the current study. Looking at the data through Lanza's framework, we might see one interesting peculiarity. Strategies that are located at the monolingual end of the continuum are mostly used by monolingual speakers, such as children's maternal and paternal grandmothers. They use monolingual strategies such as the Minimal Grasp and the Expressed Guess. When Victoria does the painting and naming game with her maternal grandmother, she suddenly says "watermelon" in English. When looking at the corresponding picture, the grandmother shows surprise and asks Victoria with a raising intonation in Ukrainian "кавун?" ('watermelon' in Ukrainian). This example illustrates the Minimal Grasp strategy.
An example of the Expressed guess is presented below (1), the maternal grandmother is asking where she can go with this hairstyle and her granddaughter answers "в музео", the grandmother hears this grammatically wrong case marker and guesses that her granddaughter meant "в музей". Thus, by guessing the utterance, she displays the Expressed Guess strategy.
(1) Grandmother: И куда я пойду с такой прической? and where I go-FUT with this hairstyle “Where can I go with his hairstyle afterward?”
Daughter: В музео? In museum
“In a museum” (mispronouncing)
Maternal grandmother: В музей? In museum
The Adult repetition strategy is actively used by their mother. When her daughter counts
in Ukrainian, she experiences difficulties with the word 'four', after having said this word, the mother repeats it out loud for her, once again, see in the example (2).
(2) Mother: Порахуй Count- IMP “Count”
Daughter: один, два, три, чотири One, two, three, four “One, two, three, four”
Mother: Да, чотири штучки Yes, four pieces
“Yes, you've got four pieces” (N1, 00:27)
At the same time, children feel free to code-switch in bilingual mode with their mother, who is a multilingual speaker for them. Interestingly, the mother's language teaching strategies are positioned in bilingual mode. Most of the time, she deals with children's code-switches, she educates them what languages they speak and requires them to switch back to the target language when a code-switch occur. For the mother, code-switching is a technique of teaching children other languages/another language. For example, when Victoria could not remember the word 'count' in Ukrainian, her mother first gave her prompt and, then, did a follow-up in the end, by saying that a Ukrainian word 'Порухувать' (‘to count’) would be 'optellen" in Dutch. (N1, 00:27).
The following example illustrates that code-switching is one of the parental language teaching strategies for the mother. Sometimes, when playing a game, children switch to Dutch instead of answering questions in Ukrainian. In this case, the mother patiently explains to them what the language setting is. For instance (3),
(3) Mother: blauw это на голландском blue this in Dutch
In this way, she highlights that the language she was expecting them to answer and the language they used is different; it is a form of negative feedback previously described by Döpke.
In the current study, the mother actively uses the intra-sentential code-switching
strategy described by Lanza (2004). This type of code-switching was already studied by
Saunders (1982) in his studies on multilingualism. He coined this type of code-switching as "quotational switching" when leaving quotations in their original language while speaking another language. The mother inserts both Dutch and Spanish words in her questions when addressing her daughter, as in Examples (4a) and (4b).
(4a) Mother : Расскажи нам кто такие ‘diertjes’ так они называются? Tell-IMP us who this animals this they called? “Can you tell us what ‘diertjes’ are?” (N9, 00:07)
(4b) Mother (to daughter): А что значит ‘valiente’? What mean brave?
“What does ‘valiente’ mean?” (N9, 01:03)
When the mother hears a code-switch into other languages, she incorporates the child's word said in another language into her question in the language of setting, which is, most of the time, Ukrainian. As an example (5) illustrates, Victoria inserts the Dutch word “jongetje” into her story when she could not remember the Ukrainian word ‘хлопчик’ (‘boy’).
(5) Daughter: И один jongetje and one boy “And one boy”
Mother: ‘jongetje’ як сказати по українськи? boy how say in Ukrainian “How would you say 'jongetje' in Ukrainian?”
Daughter: хлопчик boy
Döpke's strategies for language stimulation described in chapter 2.3 are used by the family members as well. For example, eliciting strategies, such as encouragement. The children are often encouraged to speak more, their grandmother and their mother tend to ask many questions which help children to convey more complex information, for example, asking what the children did the other day with their grandmother (6a) and what places they can go to with a new hairstyle (6b).
(6a) а что ты вчера с бабушкой робыла? and what you yesterday with grandmother do-PST “what did you do with your grandmother yesterday?” (N10, 2:52)
(6b) Куда я пойду с такою прическою? where I go-FUT with this hairstyle? “where can I go to with this hairstyle?” (N11, 0:59).
In the next example (7), the mother is requesting the translation, by saying ‘Translate what you just said’.
(7) Переводи что ты сказала Translate-IMP what you say ''Translate what you say” (N17, 00:50)
Rehearsing strategies described by Döpke, such as incorporation and repeating overlap with Lanza's interpretation of the intra-sentential code-switching strategy described by Lanza (2004). See the examples (3) and (4).
The existing frameworks by Döpke and Lanza fit a bilingual setting, but they are incomplete for a multilingual setting with interlocutors who operate in more than one language. Thus, in this article, we propose two more specific strategies for multilinguals: the Language Naming strategy and the Educational Language Game, which we will discuss below.
Since their mother speaks multiple languages, she uses special techniques available only to multilingual parents. We coin this 'Language Naming Strategy' and 'Educational Language
Game', and I will explain them below. The Educational Language Game is one of the key
language activities held by children's mother throughout the day. It is code-switching with an educational purpose to stimulate children’s language development. Usually, the mother plays a language game around one topic using, one by one, all four languages as a medium of
instruction. The mother asks her children to name the same object in multiple languages; if the children do not know, she will help them by naming the object herself. This strategy has a positive impact on children's language awareness; it highlights and raises the importance of each language spoken in the family and at school. The mother initiates this language activity stimulating the children to practice lexical retrieval in all languages they have at their disposal. As it was claimed before, parent's multilingualism can be a determining factor for children's degree of active multilingualism (Moerk, 1976). We can see the example (8) of the children who are asked to produce words related to the same concept in three languages. First, the children start counting in Ukrainian after their mother’s request in Ukrainian. Then, their mother repeats her request to count in Dutch, and the children continue in Dutch. The last language is Spanish, the mother and the children both switch to Spanish.
(8) Mother: Порахуй Count-IMP “Count”
Children counting in Ukrainian
Mother: Rekenen in het Nederlands Count-IMP in the Dutch “Count in Dutch”
Children counting in Dutch
Mother: En español como contar? In Spanish how count? “How would you count in Spanish?”
Children counting in Spanish (N4, 1:10)
Educational Language Game requires children to switch into several languages,
whereas Language Naming strategy requires children to identify what languages they speak. Languages in the family are often used separately, and code-switching is not appreciated by the parents. This may be related to language ideology when one language is spoken at the time.
Overall, we see this tendency in the family. Examples (7) and (8) provide evidence of monolingual language ideology (one language at one time) and metalinguistic awareness when it comes to distinguishing between languages. If children are asked to speak Ukrainian, but occasionally they switch to Dutch or English words, they will be informed about their switch
to another language. What is more, the children will be immediately corrected by their mother because their language policy at home presupposes that languages are separate and should be used as such. An example of metalinguistic awareness practice is provided in (9), children are told what languages they respond in.
An example of this educational language game can be observed when the mother talks about various colors and asks her children to name colors in Spanish and Ukrainian. In the following example (9), the mother asks how the children would say ‘yellow’ in these two languages.
(9) Mother: а як по-іспанськи? and how in Spanish? “And in Spanish?”
Daughter: Amarillo Yellow “Yellow”
Mother: а як по українськи? And how in Ukrainian? “and in Ukranian?”
Daughter: Жовтий Yellow
“Yellow” (N12,16:20)
In the example (10), the mother continues asking the children to name colors in other languages, such as English. When the son replies in English instead of Dutch, the mother confirms with him that he was speaking English but not Dutch.
(10) Mother: це який колір? This what color “What is the color?” Son: Red in English
Mother: Це на якій мові? це ти англійською сказав This in what language, this you in English say-PST “What language is it? You are saying this in English” (N12,16:20)
This type of Educational Language Games and Language Naming strategy leads us to the following speculation. Traditionally, parents in monolingual societies teach their children only one language. However, with changing realities of the contemporary world, language learning has become an irreplaceable part of life; thus, more and more people become active multilinguals later in adulthood; thus, they can educate their children in one more language. Moving away from the traditional approach of teaching children only one language, we might have come across a new paradigm of teaching children multiple languages by their parents at home. In our study, the mother is multilingual, which has an impact on children's multilingualism.
To sum it up, various parental strategies are used in the family to stimulate children's language development; most of them have been previously described in the literature by Lanza and Döpke. However, additional strategies available to only multilingual parents have been discovered in the research, namely 'Language Naming Strategy' and 'Educational language Game'.
4.3. Interlocutor effect
In terms of interlocutor presence, several interlocutors are presented in the recordings: a bilingual father who speaks Spanish to his children and English to the mother and family friends, paternal grandmother who speaks Spanish, a Ukrainian-speaking maternal grandmother and multilingual mother who speaks mostly Ukrainian to children daily, sometimes Dutch when their conversation related to school topics, the language of interaction between the mother and the father is Spanish and English. In the current study, we can see the influence of the interlocutor effect, which will be described further.
Their mother provided a detailed description of their language situation at home; thus, she showed a keen awareness of language strategies they use at home to grow children multilingually. Linguistic behavior and strategies of parents are defined by established language ideology at home. There is a monolingual ideology where one language is spoken at a time.
There are typical patterns in code-switching when different interlocutors are presented. First, although the basis of the conversation is in Spanish, the daughter switches from Spanish to English when talking to her father. It is worth noting that the father does not speak English to his children; there were no English utterances produced by the father in the recordings. As
we can see in the recordings, the father had some Ukrainian utterances in his speech (00:10 sec), and he did not have any English utterances. However, the father is still associated with English, and the children insert English words assuming that he understands. When the father is around, the daughter starts inserting English words. Thus, the interlocutor’s presence triggers the activation of more languages resulting in a bilingual mode and code-switching.
Second, the daughter switches from Ukrainian to Dutch when talking to her mother. Dutch is never chosen by the daughter when the interlocutor is her father; thus, we see an effect of different interlocutors, how their presence influences the language choice for code-switching.
An exciting and unique language teaching strategy is chosen by the mother when children start speaking in the non-target language. This strategy has been partly described by Lanza (2004). The mother facilitates a switch to the target language by introducing a corresponding interlocutor, which we call the Interlocutor Strategy. The following examples demonstrate this technique.
In Example (1), the daughter started speaking Ukrainian to her Spanish-speaking father, who does not understand Ukrainian. Her mother introduced the interlocutor for the daughter to switch to Spanish, the language which is usually spoken between the daughter and her father.
(1) Mother: Скажи папе Tell-IMP papa-DAT
"No, no tell this to your father". (N12, 45:45)
The same technique is used with other interlocutors. When the daughter could not remember the word in Ukrainian, the mother asked her how she would say the same word to her grandmother (2).
(2) Mother: А что ты бабушке скажешь?
And what you grandmother-DAT say-FUT
“How would you say this to your grandmother?” ( N12, 18:51)
When talking on Skype to paternal grandmother, the mother initiates the conversation by asking the daughter various questions (3), when she started answering in Ukrainian instead of Spanish, the mother highlights who the interlocutor is.
