“His Favorite Things Mahi Al-hadia ” Social Functions of Code Switching in Bilingual Children’s Conversations

,


Introduction
Code switching as an area of study is connected to linguistics, psychology, anthropology, and related disciplines. Each field studies the various aspects of code switching from different perspectives. In the field of linguistics, for instance, this phenomenon is investigated from syntactic, sociolinguistic, and psycholinguistic points of view, which has led to the development of a rich literature on this topic. Code switching is found among people from all ages, genders, societies, and languages. Many studies have analyzed adult-adult, child-child, and adult-child code switching in different language pairs around the world. However, to this researcher's knowledge, few studies have been conducted on Arabic-English bilingual children, with the main exceptions being Bader (1998), Bader and Minnis (2000), Alenezi (2006), Abugharsa (2013) and Al Omary (2020) who examined code switching mainly form a syntactic perspective. Due to the increasing number of Saudi transnational families who have settled to the U.S. and the growing number of Saudi children enrolled in American public schools, bilingualism has increased. Thus, there is a growing need to study code switching among Saudi bilingual children from a sociolinguistic perspective to better understand how kids of transnational families employ their linguistic repertoire in their conversations.
Sociolinguistic approaches have been applied to examine code switching among adults and children's' talks. Woolard (2005) stated that "a significant segment of sociolinguistic research since the mid-twentieth century has been devoted to understanding how bilingual and multilingual communities organize their multiple linguistic resources" (p. 73). Studies on code switching have been conducted at both the macro-level and the micro-level. Nguyen (2014) stated that "both approaches are concerned with the meaning of language but differ in the point where this meaning derives from" (p. 48). Macro-level approaches have led to language choice studies related to social factors, such as the ways in which alternations between codes or varieties occur with regard to the community (Ferguson, 1966) or the domains in which alternations occur (Fishman, 1966). Micro-level approaches have led to code switching studies related to the situational level, such as the pragmatic classification of types of code switching as Situational Switching, Metaphorical Switching, and Conversational Switching (Gumperz & Hymes, 1972;Gumperz, 1982), as well as to the conversational code-switching approach (Gumperz, 1982). However, other approaches, such as the Markedness Model of Myers-Scotton (1993), are an attempt to "incorporate the micro-and the macro-perspectives into CS [code-switching] research" (Boztepe, 2005). Therefore, this study adopted the markedness model as an inquiry tool to examine how Arabic-English bilingual children switch between codes and to provide a fuller picture of the macro-and the micro-level aspects impact children's employment of codes.

Code Switching and the Markedness Model
Myers- Scotton (1993) proposed the Markedness Model to explain the socio-psychological motivations behind code switching. According to Myers-Scotton, code switching in a multilingual community is based on the association of particular code choices with Rights and Obligations (RO) sets during social interactions. Each code indexes culturally determined RO sets, which are "derived from salient situational features and relevant cultural values" (p. 7). In this model, there is an unmarked code for any given situation, and use of this code indexes the persona and role relationship the speaker wishes to have.
Moreover, speakers' decisions about alternative choices are mainly based on assessing costs and rewards. According to the negotiation principle (Myers-Scotton, 1993, p. 113), speakers work out what code choice best suits their goals in a given conversation. Along with the role played by the negotiation principle in interpreting code switching functions, she also proposed three maxims for use in the analysis of code switching: the unmarked choice maxim, the marked choice maxim, and the exploratory choice maxim.
According to Myers-Scotton (1993), to use the unmarked choice maxim is to "make your code choice the unmarked index of the unmarked RO set in talk exchanges when you wish to establish or affirm that RO set" (p. 114). This maxim refers to the influence of societal norms in guiding the speaker to choose the expected codes in a medium of interaction. Under Myers-Scotton's maxim, code switching can be classified as sequential unmarked code switching and unmarked code switching. Sequential unmarked code switching occurs when there is a change in the situational factors during a conversation that lead the speaker to index the new unmarked RO set. An example of this type of code switching is when the speaker code switches as a result of a change of topic during a conversation. Here, the speaker would like to index a new unmarked RO set since the first unmarked RO set has been changed. Sequential unmarked code switching is mostly related to intra-sentential code switching. On the other hand, code switching as an unmarked choice occurs when "the speaker wishes to index two identities or attitudes toward the interaction (and therefore two rights and obligations sets) simultaneously" (p. 149). Unmarked code switching occurs frequently in certain communities.
Unmarked code switching is related to inter-sentential code switching (Myers-Scotton, 1993). According to Myers-Scotton, there are several conditions that should be met in unmarked code switching to let the switching occur. The first condition is that the speakers must be relatively equal in terms of bilingualism and socio-economic factors. The second condition is that the speakers should have the desire to embody the mutual identity. Third, the speakers must be capable of evaluating the codes that carry their identities in the interaction. Since code switching requires knowledge of the two languages, a fairly high degree of proficiency in both languages is necessary in this scenario.
The second maxim is the marked choice maxim, which is to "make a marked code choice which is not the unmarked index of the unmarked RO set in an interaction when you wish to establish a new RO set as unmarked for the current exchange" (Myers-Scotton, 1993, p. 131). In this maxim, conversational expectations are violated by the speakers and the marked choice is made to start a new RO set. According to Myers-Scotton, emotional factors and the negotiation of social distance are the motivations that push speakers to produce marked code switching. It should be mentioned that marked code switching is the most common type since it can occur within any community and under any conditions. The exploratory choice maxim occurs in the following type of situation: "When an unmarked choice is not clear, use CS to make alternate exploratory choices as candidates for an unmarked choice and thereby as an index of an RO set which you favor" (Myers-Scotton, 1993, p. 142). Thus, this type of code switching occurs because the speakers are not sure which RO set should be used; it is a result of vague situational values or conflicting norms. Therefore, in the process of exploratory code switching, the speaker introduces an RO set through a certain code to let the interaction occur. If a failure should occur, another code will be introduced. markedness model fit this study as it provided a clear picture of how bilingual children manage their communication and how social motivation played an important role in facilitating their communicative events.

Code-Switching and Bilingual Children
Several studies on children's language choice across languages have been conducted in view of the Markedness Model. Starting with Farris (1992) used Myers-Scotton's Markedness Model to investigate a type of register variation referred to as "babytalk" as well as adult speech in Mandarin at a Taiwanese preschool. In her study, the data were derived from naturally occurring conversations between preschool children and their teachers. This study was conducted in several preschools in Taiwan. The results revealed that there were two registers used in these preschool settings, the babytalk register and "the voice of authority," which refers to the language style used by teachers when speaking to preschool students. She claimed that both of these registers are motivated by Chinese cultural assumptions regarding children and childhood. Furthermore, in terms of talking to children, she found that teachers switched between the unmarked voice of authority and a babytalk register. In turn, when teachers used a babytalk register, children reacted differently, "the younger children in particular seeking [those] teacher[s] out for comfort or assistance" (p. 206). Thus, the children who participated in this study were aware of the variation between two registers and knew how to react to each register.
In another direction, Myers-Scotton (2002) conducted a study on a Malawain family containing two adults and two children to show that the unmarked choice can be identified through a quantitative analysis (a frequency-based criterion) and that frequency can also identify the marked choice. The data were collected from everyday conversations that were held between both parents and their two children. The participants were fluent speakers of both Chichewˆa, their first language, and English. The findings of this study revealed that when bilingual speakers did not use both languages equally in a given conversation, the more used language represented the unmarked choice. Furthermore, she found that when the participants in the same conversation were from different generations, there was no one language used as the unmarked choice for all participants. This meant that not all participants in a given conversation would have the same unmarked choice, which would vary. The study found Chichewˆa was the parents' unmarked choice, while English represented their children's unmarked choice.
Fuller, Elsman and Self (2007) examined the linguistic performance of Spanish-English bilingual children who live in the U.S. The participants consisted of seven children in the 4th and 5th grades who were enrolled in a school where Spanish and English were both languages of instruction. English was the language of instruction for morning classes while Spanish was used for the afternoon classes. One of the authors attended classes with these children for about one academic year, twice a week, to collect the data during classroom activities. In order to create a comprehensive analysis of the data, two models were employed, the Markedness Model (Myers-Scotton, 1993;Myers-Scotton & Bolonyai, 2001) and the Sequential Approach (Auer, 1988(Auer, , 1995Li, 1988). Although both models provided a socio-pragmatic and structural analysis for code switching in this study, the authors favored the Markedness Model as it could provide an explanation for each level of linguistic performance and could stand for social identities.
Both qualitative and quantitative analysis showed that children used the Spanish code to communicate with each other, i.e., Spanish was the peer code. On the other hand, English was used in exchanges with teachers or during teacher-fronted activities. Particularly, in view of the Markedness Model, most instances of code switching can be seen as a strategy to index and negotiate roles within an exchange. On the other hand, the Sequential Approach introduces language choice as a contextualization cue and sequential conversational structure, which means there is interactional meaning for using one code over another. Thus, the Sequential Approach provided an explanation for conversations where the overall pattern of interaction consisted of switching back and forth between languages. Bolonyai (2005) investigated the linguistic choices among Hungarian-American bilingual girls in the ages of six and seven. The aim of this study was to examine how these Hungarian-American girls, who were English-dominant, used code choices to construct relations of dominance and subordination during school activities, which consisted of competitive conversation. This study's data were derived from a corpus of 18 hours of six American-Hungarian families' conversations that were tape-recorded in the United States. Out of the entire corpus, only 90 minutes were analyzed for the purpose of this study. Bolonyai's (2005) results showed that Hungarian-American bilingual pre-adolescent girls used different linguistic strategies in order to manage power asymmetries in their talk. They switched between two codes, Hungarian and English, according to their personal reasons and desires. For instance, one of the girls asked her mother during their playtime about the best player in English, the unmarked code for her, multiple times even though the rule was to communicate only in Hungarian. Bolonyai argued that young bilingual children are rational and social actors who can employ code switching to serve their personal desires and to have optimal outcomes.
In general, these studies have shown the Markedness Model to be a helpful theory when analyzing bilingual children's language use. In particular, it provides a better understanding of the social aspects of language alternation. However, it can be said the Markedness Model as a theory has not been applied to analyze the social aspects of code switching among Arabic-English bilingual children, which help us understand the social functions for code switching among Arabic-English bilingual children. These empirical studies mainly elicited data from bilingual children in different age groups, which would indicate that this theory could be used to analyze the present study's data. Furthermore, these studies have focused on different ages and kinds of bilingual children, preschool Mandarin-speaking children (Farris, 1992); Chichewˆa-English bilingual children (Myers-Scotton, 2002), 4th and 5th grade Spanish-English bilingual children (Fuller, Elsman, & Self, 2007), and six-and seven-year-old Hungarian-English bilingual girls (Bolonyai, 2005). The present study was thus intended to supplement these previous studies as it targeted a group Saudi Arabic-English bilingual child, which to the best of my knowledge had never been studied before for the purpose of exploring the role of social motivation in language choice and code switching.

Participants
This study employed ethnographic techniques to examine how bilingual children use their linguistic repertoire and shift between codes fur social motivation purposes. Six bilingual children were chosen as participants for the purpose of this study. All were boys between the ages of 9 and 11. At the time of the study, all of the participants were living in Carbondale, Illinois. In the current study, the participants were assigned the following pseudonyms: Ali, Mido, Sari, Nori, Misho and Fadi. They were enrolled in American public schools in which the language of instruction and education was English. At the time of this study, two of the children were in third grade, three in fourth grade, and only one in fifth grade. Additionally, every participant also regularly attended the Sunday school at the Islamic Center of Carbondale that was established in 2013. All of the participants were originally from Saudi Arabia and had lived there for some years. They were selected for several reasons: 1) they were born in Saudi Arabia and thus had experience speaking Arabic for at least two years or more; 2) they had spent a minimum of four years in the USA; 3) they were in the third grade or higher and thus had acclimated to the American culture and had adopted English as their second language; 4) their parents tended to communicate with them in their native language, Arabic; and 5) they were attending the Islamic Center Sunday school in order to retain and continue learning Arabic. An additional reason for the selection of these children was that at least one of the parents of every child was a monolingual Arabic speaker-therefore Arabic was expected to have been reinforced in the participants' home on a regular basis, thereby balancing their language use. The participants were familiar with the researcher beforehand because he was a teacher for this group of children between September 15 and December 15, 2014, and he was a friend of their parents.

Research Setting
This research was conducted in a private Sunday school in Carbondale, Illinois. This institute is located in the Islamic Center of Carbondale, a non-profit organization funded by the Saudi Student Association in Southern Illinois University. The aim of this school is to teach, enhance, and help Arabic competency in Arabic-speaking children. It educates individuals of four to ten years-of-age, who are separated into three levels: pre-school children, first to second grade, and third grade and above. A requirement for children to be admitted into this school is to have enough knowledge of Arabic to communicate in that language. At the time of conducting this study, each of these three grade levels contained six to ten individuals.
Because this school was only open for two hours a week, every level had its own schedule. In the highest level-which was the tier most relevant to this research-the school period was separated into three classes. The initial 25 minutes was for reciting and reading the Qur'an, followed by a five-minute break. The second class consisted of an hour of activities for the children to practice their Arabic language skills. The final half hour was for purely recreational activities.

Data Collection and Analysis
The data obtained in this study were part of 'telling a story' activity, a series of six images-each addressing a different subject matter-were shown to the participants as a course activity to increase their productive language skills in Arabic. Each participant was asked to look at an image and tell a story about what he saw. This exercise was undertaken in six one-hour sessions across a six-week timeframe. Time was allocated equally between the participants, each being given approximately five minutes. Because it would have been challenging to allow all participants to speak at the same time about a single image in a single session, a rotational scheme was employed to sidestep any such potential issue. The six images were employed in every session. Throughout the initial three sessions, the participants were required to tell a story in Arabic. It should be mentioned that the participants were divided into two groups because after conducting the first session, I found that it was difficult to analyze the speech of six children in each session and with each story. In addition, it was not easy to recognize the speakers when I did the transcription later on. The participants were divided according to age: Group A included three that were 11 years old, while Group B included other participants who were 9 and 10 years old.
The schedule in Table 1 shows the allocation of the sessions and topics. This schedule helped avoid repetition and prevented participants from affecting one another's responses. Additionally, the activity was designed to boost the participant's enthusiasm for participating. The first four images were adopted from Terveen (2013). The initial image is of fire fighters rescuing a cat, the second depicts a boy learning to ride a bike, the third depicts a boy washing laundry, and the fourth a child receiving a gift and then playing with it. The fifth image shows a boy who fell down when he was playing and was retrieved from the website ESL Printables (Zeynep), and the sixth shows a boy shopping at a grocery store, which was retrieved from Turtlediary.com. The final two images were added to Terveen's pictures to address more generalized subject matter and to boost participant interest in the activity. The researcher, who was the teacher for this group for two consecutive years, led the discussion in these sessions and I tried to make field notes beside using two different recordings to record the children's talk.
The audio recordings of the data were transcribed and coded, following conversation analytic conventions based on Atkinson and Heritage (1984). Then, the data were qualitatively examined through the use of content analysis techniques. After annotating the instances of code switching and measuring the frequency, the types of code switching that occurred in the data were identified. Furthermore, each instance of code switching was classified in according to Poplack's (1980Poplack's ( , 2001 classification of code switching (inter-sentential, intra-sentential, and tag switching). Additionally, every instance of code switching was assessed in a qualitative analysis in terms of the Markedness Model (Myers-Scotton, 1993).
It should be mentioned that the English translation was inserted when an instance of Arabic code switching transpired. Any nouns borrowed from another language were exempted from the research. Therefore, all proper nouns, including the titles of fast-food items, TV shows, and television channels were exempted from the analysis to maintain its objectivity.

Code Switching and Social Function
In this study, I argued that Arabic was the unmarked code choice while English was the marked choice. This is because of the nature of the setting, which required participants to tell a story using Arabic. Thus, Arabic was the expected code as it was the language of the task. When a participant switched to English, he switched to the unanticipated code. Switching to English in this study was considered a marked code choice most of the time when the motivation behind switching was to increase or decrease social distance or to show emotions through one's speech (Myers-Scotton, 1993, p. 111). However, I have found some instances where switching to English was still unmarked when the switch carried no indexicality or change in the RO set. Furthermore, any mid-utterance (intra-sentential) code switching was seen as unmarked since it took place within a single phrase with no impact on the functions of the conversation or the RO set. Following is a discussion of some examples in which switching to English was seen as a marked choice to fulfill social functions such as forging alliances and bonds, changing the role of relationship, and invoking authority. Then, I provided other examples where switching to English was seen as an unmarked code to fill language gaps, finish the conversation in a positive way, or indicate the bilingual identity.

Switching to English as the Marked Code Choice
In Example 1, Mido told a story about a boy who received sports clothing as a gift. He started the story in Arabic following the researcher's instructions to relate the story in Arabic. He successfully completed the task and described the picture in Arabic (lines 1−5). At the end of his story, he produced the last utterance in a high pitch to indicate that he had finished. There was a pause for about two seconds. I tried to encourage the other children to engage in a discussion about their friend's story by asking (line 8) one of the children for his opinion about the story. There was another pause of about two seconds during which time the boy was looking at the picture and clearly thinking. At this point, Mido spoke up again to encourage the others to participate. He switched to English as seen in lines 10 and 11. He asked the others, "you got any comments, come on" and then followed this with "don't be scared" to encourage them to talk. Ali then started the conversation by asking Mido about why the child in the picture had a number one. It seems that this question was rather spontaneous because many pauses took place during Ali's turn. Interestingly, Ali chose to respond in Arabic to Mido's English turn. Mido answered in Arabic that it was a gift. Ali was not satisfied with Mido's answer, so he asked again "why number one". He clearly meant to ask why the boy had received a shirt with number one on it. Also, Ali produced his turn in Arabic (line 14). Mido was angry so he answered in a high pitch in Arabic that he did not know. Then Sari engaged in the conversation and chose to defend his friend, Mido. He answered in Arabic that number one was for the goalkeeper, which he repeated (lines 16 and 17).
Mido and Sari did not appear to understand Ali's question; they thought that he was saying something against them and that he had asked the question at random. Ali asked in Arabic again "why? why? why?" Sari took the argument personally and started attacking Ali verbally, claiming that Ali did not like soccer. Ali answered in Arabic that they already knew he did not like soccer. Sari added that "you like basketball". Ali was trying to ijel.ccsenet.org Vol. 12, No. 1; interrupt (as in line 23), but Sari did not let him talk. Sari resumed his turn by explaining the distribution of numbers in soccer. During Sari's turns, from lines 19−25, although he used Arabic turns, he inserted some English words within clause boundaries in lines 21 and 24. As a result, Ali asked why the boy's friends had given him a gift (line 26). Sari tried to answer but failed (line 27). Mido answered the question repeatedly to signal the end of the conversation. All the turns taken between the three children from line 25 to the end were in Arabic.
In this example, as stated previously, Arabic was the unmarked code choice because the participants were following the instruction to talk about the picture in Arabic. Mido successfully told the story in Arabic. However, when he wanted to encourage his peers to comment, he did so in English. Here, Mido switched to the language, English, that he knew would decrease the social distance with his peers as he was aware that this language was the one preferred by the group members and because they were used to using it as a medium of instruction. Furthermore, choosing English redefined the existing relationship roles, from directing the speech to the researcher and peers to directing the speech solely to the peers by relying on the peer code. For these reasons, switching from Arabic to English in a setting where the expected language was Arabic, was considered marked. However, Ali responded to this English turn with Arabic, the unmarked code choice. It seems that Ali understood that Mido was trying to encourage them to interact so he engaged in the conversation by directing a question to Mido about his story. Considering that Ali was the most fluent child in this group in both languages would explain why he responded in Arabic instead of English. It appeared he was aware that the language of instruction for this task was Arabic. Finally, the changing of the code in this example was observed to fulfill the function changing the addressee.
The intra -sentential code switching in lines 21 and 24 were unmarked because they were used to compensate for the lack of Arabic linguistic ability in recalling a single word or phrase with no obvious change in the RO or situation. According to Myers-Scotton (1993), intra-sentential code is one of the features of unmarked code switching.
Another example of switching to English to change the role relationship is given below. Example 2 is a conversation that took place before the participant started telling the story. I gave a picture to Fadi to talk about. Fadi was happy even before looking at the picture because he thought it was about soccer, which was his favorite topic. Nori, his classmate, commented in Arabic, calling Fadi "the king of soccer" (line 3). Fadi responded with an unclear word and then asked in Arabic at what point in the story he should begin telling it. There was a pause for about two seconds, so I tried to keep the conversation going by answering his question, telling him that he could start at any point. Nori repeated my statement to help his friend. Fadi was still confused and needed time to think, so he switched to English in line 9. Nori again wanted to help his friend, so he suggested that Fadi start from a place Nori pointed out to him. Nori used a mixed utterance (line 10). Fadi responded in English about the order of the events in the picture.
Then a pause occurred for about two seconds. Misho added a comment, but it was not apparent to me what it was when I did the transcription. Nori used Arabic to show Fadi the order. Fadi felt comfortable with Nori's suggestion and produced the agreement in English (line 15). Misho was unsatisfied with Nori's suggested order, so he suggested he start from a different point. Fadi first asked Misho about the order and then agreed to follow it after he made a statement in Arabic: "yeah, yeah, right". All turns from line 16 until the end were produced in Arabic. It should be mentioned that Fadi then told the entire story in Arabic. In this example, Fadi switched to English, a peer code, as a sign of his bilingual identity and to seek help from his friends who carried the same bilingual identity. Before this code switch, he directed a question in Arabic asking for help, and he did not find the researcher's answer sufficient. As a result, he switched to English when he decided he needed his friends to interact with him. This action meant that he switched to an English code in order to negotiate different RO sets. He also code switched to increase the social distance between speakers. Therefore, English in this example was marked and used for the purpose of changing the role relationship of the setting.
Example 3 is a story told by Ali about a boy helping his mother clean his clothes. It was followed by a conversation between Ali, Mido, and Sari about Ali's story. The setting was informal and Ali narrated his story in a relax tone. He used Arabic as he was asked to do. While narrating this story, Ali added some imaginary events, such as saying that it was a girl instead of a boy in the picture and that her name was "Cinderella". Also, he went beyond the picture to explain why she was cleaning her clothes and working at home. He referred to the harsh life she had with her grandmother. Ali narrated all these events in Arabic from line 1 to 17.
After Ali finished his story, Sari made a negative comment in English about it. Ali returned to the story and made a concluding statement in Arabic. After that, there was a pause for about three seconds, so I directed a question in Arabic to the other children about the story (line 18). Sari again started attacking Ali's story in a high tone in English. Here, Mido interacted with the group but used an Arabic word first and then translated it into English. Then he made an unclear comment. Mido and Sari switched to English to criticize Ali's story as seen in lines 23−32. In lines 31 and 32, Mido asked Ali two questions about the story in English. Ali responded to the second question first and his tone indicated seriousness, so Mido again paraphrased his first question in line 35 and lightened his tone by saying "Mr. Boss" at the end. Ali did not answer the question but rather directed a question to Mido, imitating Mido's English accent. Sari and Mido answered at the same time (lines 37 and 38) In the above example, as Ali started with the unmarked code, Arabic, Sari chose to switch to English to increase the social distance with Ali. The same occurred with Mido, who chose to challenge Ali by using English instead of Arabic. As a result, Ali responded to them in English, which meant he accepted the change in the conversation pattern. It should be considered that Ali started his story by using Arabic in a relax atmosphere, but when Sari and Mido switched to English, the atmosphere had become more serious, and as a result the code changed. For these reasons, English was a marked code in this conversation. Furthermore, it could be claimed that the choice of codes in this example was to invoke authority, show seriousness, and deal with social distance. Myers-Scotton (1993) has stated that the motivation for making a marked choice is "to indicate a range of emotions from anger to affection and to negotiate outcomes ranging from demonstrations of authority or of superior educational status to assertions of ethnic identity" (p. 132).
In summary, choosing English in examples 1−3 was observed as a marked code in the data because it met the conditions for marked code choice proposed by Myers-Scotton (1993) in the Markedness Model framework. Myers-Scotton stated that marked code choice plays an important role in establishing a change to increase or decrease the social distance among the participants (p. 132). One of the features of marked code choice is its use ijel.ccsenet.org International Journal of English Linguistics Vol. 12, No. 1; to show seriousness, invoke authority, and to express personal feelings ranging from anger to affectation. It can therefore be seen throughout the previous examples that English was inserted by the speakers as a means of playing with social distance. Moreover, English as a marked code in the current study helped redefine the relationship between the interlocutors. This goes hand-in-hand with Myers-Scotton's (2006a) claim that marked choice "always calls for a move from the expected social relationship of the participants, a readjustment of the expected social distance which would hold between them" (p. 216). Finally, the code switching in the previous examples served the following functions: change the addressee, make alignment, expand, and invoke authority.

English as the Unmarked Code Choice
The following examples illustrate how English was considered an unmarked code choice. In Example 4, Sari was given a picture to talk about, and he initially produced unknown words, causing all of the children to laugh. Ali laughed too and followed up with an Arabic utterance that made all of them laugh again. Sari produced an English sentence (line 5), although the preceding turn was in Arabic and the instructions were in Arabic as well.
Ali again took an Arabic turn, asking Sari about whether the person in the picture was a man or a woman. Sari ignored him and started narrating the story in Arabic (lines 8−10). In example 4, Sari switched to English to express his personal feeling toward the picture. This switch was preceded and followed by Arabic turns produced by Ali. The setting was relaxed and informal while the children were laughing, joking, and talking to each other. Although Sari's first turn in this example was unclear, he produced this turn in English less clearly. It appeared that he was not trying to change the role relationship because he was already in a peer-role relationship. In other words, there was no change of RO set or situation because of Sari's choosing another code. Therefore, the switch to English in line 5 could be seen as an unmarked choice that occurred to express personal thoughts. As for the intra-sentential code switching in line 14, it could be explained as unmarked code switching that occurred to substitute a word from the other language when the speaker was having difficulty recalling the target word, a change that left no effect on the conversation pattern or even changed the code of the next person to speak.
The use of English in example 4 can be explained based on the overall pattern of code switching, not only in terms of language choice with particular utterances. If we look at the overall pattern in this example, it can be argued that switching to English in both instances is an unmarked choice because the norm was for the participants to switch back and forth. Moreover, according to Myers-Scotton, one of the features of code switching as an unmarked code choice is that it is not necessarily made with the purpose of indexicality; instead, in general it has a communicative motivation (p. 117). It can be postulated that Sari switched to English more than once during a single interaction to trigger the group's dual identity as bilingual speakers. Fuller (2012) stated that "switching back and forth is not significant in the details but in the larger picture, and this is a picture of hybrid language and identities which challenge essentialist social categories" (p. 77).
Example 16 is a short description by Misho of a picture of a boy who received a gift. Misho briefly talked about ijel.ccsenet.org International Journal of English Linguistics Vol. 12, No. 1; the picture and inserted English sentences during his turn (lines 3 and 4). I provided the following turns to see if there was any impact from this code switching on the other speakers and to determine the motivation behind it. Finally, it could be said that code switching occurred 1) for the purpose of elaboration and explanation and 2) to finish the conversation in a positive light.
In summary, switching to English in examples 4 to 6 was classified as an unmarked code choice in the current study because the code switching occurred smoothly with no noticeable change in or negotiation of the current RO set in those interactions. Also, the speakers chose the unmarked code to save the utterance function and to maintain the flow of their social relationship. Daniel-Wayman (2016) stated that "speakers choose to speak in unmarked codes to maintain the status quo within a conversation" (p. 5). Furthermore, it can be seen that these instances of code switching occurred unconsciously, which goes hand-in-hand with Myers-Scotton's (1998) claim that the unmarked choice usually occurs unconsciously (p. 27). Therefore, these findings were in line with what Myers-Scotton claimed about the unmarked code choice in the Markedness Model. Finally, the code switching in these examples served certain functions, such as expressing a personal thought and word-finding difficulty.

Conclusion
The Markedness Model by Myers-Scotton (1993) was applied in order to determine the speakers' social motivation for employing code switching. In this study, I argued that the language of the instruction, which was Arabic, was the unmarked code as it was the code expected to be used and was also the unmarked index of the unmarked RO set for the given interaction. Therefore, using a code other than the unmarked code could be seen as a marked code when it violated the anticipated norms, thus carrying extra social meaning and was chosen to establish a new unmarked RO set (Myers-Scotton, 1993). This inserted code was seen as an unmarked choice if it was a pattern of interaction with no apparent intention to change or affirm the existing RO set, such as was the case with intra-sentential code switching most of the time.
Since Arabic was the expected code and therefore the unmarked code choice, most of the time, participants switched to English to change the RO set or the role relationship for interaction. In addition, the application of this approach in the current data revealed that the participants switched to a marked code to serve certain functions, such as to change the addressee, engage in interaction, make alignment, ask for translation, expand, invoke authority, and finish the conversation. More interestingly, the analysis of the data according to the Markedness Model showed some instances where switching to English fell under the unmarked choice maxim. That is, when the switch to another code did not carry a particular social meaning, it would not indicate a change in the RO set. This time, switching to English as the unmarked code choice was found to fulfill the function of expressing personal thought and word-finding difficulty.
Overall, this study contributes to the current research on the Markedness Model among bilingual children. Evidence for Myers-Scotton's (1993) marked and unmarked code switching was observed in Arabic-English Saudi bilingual children. This study also agreed with the findings of previous studies (e.g., Bolonyai, 2005;Fuller, Elsman, & Self, 2007;Klapicová, 2017;Myers-Scotton, 2002) arguing that bilingual children are rational and social actors who chose a given code intentionally to achieve certain social goals in an interaction.
As with previous studies, this study had certain limitations, such as the presence of the researcher during data collection, the small sample size of participants, and most importantly, the small number of pictures used to elicit and generate conversations. Thus, further studies could benefit from collecting data on linguistic performance during playtime or family conversations. It is also recommended for future studies to have more participants and gather more data to obtain more accurate results. Future studies on transnational bilingual children are encouraged to investigate the individual's language choice during different settings and contexts.