Constructing Playful Talk through Translanguaging in English Medium Instruction Mathematics Classrooms

Recent studies on English-Medium-Instruction (EMI) classroom interaction have begun to look at the role of translanguaging as a pedagogical practice in supporting participants to exploit multilingual and multimodal resources to facilitate content teaching and learning. The present study contributes to this growing body of literature by focusing on playful talk in multiple languages and modalities in EMI mathematics classrooms in a secondary school in Hong Kong. Based on the data collected from a linguistic ethnography, we analyze how the teacher constructs playful talk in order to achieve various pedagogical goals including building rapport, facilitating content explanation and promoting meaningful communication with students. The analysis demonstrates that translanguaging appears to be a critical resource and that several social factors, including the teacher’s personal belief, history, sociocultural, and pedagogical knowledge, play a role in constructing playful talk. The playful talk transforms the classroom into a translanguaging space, which in turn allows the teacher and students to perform a range of creative acts and experiment with a variety of voices to facilitate the meaning making and knowledge construction processes.


INTRODUCTION
In English-Medium-Instruction (EMI), English-as-a-Second/Foreign-Language students will learn all/some subjects through English.The 'multilingual turn' in education (May 2014), especially translanguaging as a pedagogical approach (e.g.Garc ıa et al. 2017), has recently attracted the attention of EMI researchers due to the need for a more nuanced understanding of the role of the learners' as well as the teacher's complex multilingual and multimodal repertoires in knowledge construction (e.g.Lin and Wu 2015;Lin and Lo 2017).Translanguaging challenges the monolingual pedagogical principle (i.e.English only) in EMI and encourages the learner and the teacher to draw on their familiar and available linguistic, semiotic, and multimodal resources to facilitate the processes of meaning making in the classroom.
The present study focuses on the role of translanguaging in constructing playful talk in an EMI classroom.Here, 'playful talk' refers to a range of verbal and multimodal activities and routines, including humour, parody, teasing, which can emerge in teacher's and students' talk (Lytra 2017).Previous research shows that playful talk can be a useful tool for motivating and facilitating second language (L2) learning in the classroom (e.g.Bell 2005;Waring 2013).There is, however, little empirical work on playful talk in EMI classrooms (e.g.Jakonen et al. 2018) or playful talk through translanguaging.Translanguaging scholars have emphasized the significance of the creative and playful dimensions of the practice as they challenge the power relations and hierarchical order in the process of knowledge construction (e.g.Wei 2011Wei , 2018)).Hence, studying the role of translanguaging in constructing playful talk in EMI classrooms can potentially allow researchers and teachers to understand translanguaging as a resource for enabling classroom participants to engage in diverse multiple meaning-making systems and subjectivities.This can create a classroom environment that promotes student participation and facilitates content learning.
To address this research gap, this study examines how translanguaging is employed by the teacher (male) to create playful talk in the EMI classroom in order to accomplish his pedagogical goals in the lessons.This study is 2-week focused classroom observations in a Hong Kong (HK) EMI secondary mathematics classroom.Observations with fieldnotes, ethnographic interviews with teachers and other stakeholders, and video recordings are collected.The classroom interactional data are analysed using Multimodal Conversation Analysis (MCA).The analyses of the classroom interactional data are triangulated with the video-stimulated recall-interview data, which are analysed using Interpretative Phenomenological Analysis (IPA) in order to analyse the teacher's reflections on his pedagogical and interactional strategies.Macaro (2018: 19) describes EMI as the 'use of the English language to teach academic subjects (other than English itself)' in countries where English is not usually spoken by a majority of the population.In HK, Chinese (spoken Cantonese and written standard Chinese) is the language of daily communication by the majority of the local population and English is considered a prestigious language due to HK's history as a British colony.In primary schools, Chinese is usually the medium-of-instruction (MoI) and English is taught as a separate subject.At tertiary level, all government-funded universities adopt EMI, mainly because of the need to align with international higher education and cater to a significant number of international students.It is in secondary schools where the MoI policies have undergone significant changes in recent years.

ENGLISH MEDIUM INSTRUCTION IN HONG KONG
Before the 1997 handover of sovereignty from Britain to China, the HK government adopted a non-interventionist policy which allowed schools to make their own choices in MoI.Parents and other stakeholders had a strong preference for English medium secondary schools because of the popular belief that learning English can provide stronger benefits in future job opportunities and potentially increase the students' social mobility.Over 90% of secondary schools claimed to be EMI schools although in actual classroom interaction, mixing Chinese and English were prevalent (Lin 2006).The HK government and education authorities regarded language mixing as a key reason affecting the students' English standards.A streaming policy was introduced in 1998, which mandated a clearcut division in secondary schools into either EMI or Chinese-medium-instruction (CMI) schools.One hundred fourteen secondary schools were granted an exception to adopting EMI in teaching content subjects while 307 schools deployed CMI.However, there was a strong demand from the general public for reinstating EMI in all secondary schools as they saw EMI as a potential enabler to improve the students' English proficiency.In 2010, the HK government offered CMI schools the autonomy to decide their MoI for content subjects if they met certain criteria (e.g. the students' learning ability, the teachers' language ability, requirements of individual subjects) (Education Bureau 2009).This policy has resulted in a diversified mode of MoI in schools, including CMI in all content subjects for all classes, CMI/EMI in different subjects in different classes, or EMI in all content subjects for all classes.
As Tollefson and Tsui (2014) argue, the debate of adopting EMI in secondary schools ignores the fact that such a monolingual rule offers limited opportunities for social interactions because teachers in EMI classes tend to adopt the lecture format to teach the content.Moreover, Lo (2014) has shown that L2 learning opportunities can vary in different EMI content subjects.For instance, mathematics and science lessons may not favour classroom discussion between the teacher and the students since these lessons are often treated as solving problems with set formulas and calculation procedures.Hence, students seldom have the opportunity to participate in a discussion with the teacher.In the article though, we aim to illustrate a different picture of a mathematics lesson and provide evidence that shows how a mathematics teacher and the students in an EMI class engage in playful interaction through translanguaging in order to create a space that facilitates content learning and promotes meaningful communication.

TRANSLANGUAGING
The term translanguaging was originally used to describe a pedagogical practice of moving flexibly between different input and output languages in Welsh revitalization classrooms (Williams 1994).L iWei (2014) further develops the concept as a process of knowledge construction, which involves going beyond different linguistic structures and systems (i.e.not only different languages and dialects, but also styles, registers and other variations in language use) and different modalities (e.g.switching between speaking and writing, or coordinating gestures, body movements, facial expressions, visual images).Li Wei (2014) emphasizes the transformative nature of translanguaging practices as they create a translanguaging space for multilinguals by bringing together different sociocultural dimensions, including the speakers' social identities, life histories, beliefs, and their knowledge of the wider institutional environment, as resources in the process of negotiation of meaning (see also LiWei 2011). Li Wei (2018) further argues that the concept of 'translanguaging space' includes two notions that are essential to bilingual education: creativity, which refers to the ability to 'push and break boundaries between named language and between language varieties and to flout norms of behaviour' (p.15), and criticality, which refers to the ability to use 'available evidence insightfully to inform different perspectives of cultural, social and linguistic phenomena and to challenge and express ideas through reasoned responses to situations' (p.23).A number of translanguaging studies have demonstrated that bi/multilinguals are provided with agency to employ various linguistic and semiotic resources creatively and critically to challenge the traditional configurations, categories, and power structures, and construct new meanings and new configurations of language practices through everyday interactions (Choi 2019;Li Wei 2016;Zhu et al. 2020).However, few studies have looked at the transgressive dimension of translanguaging in classroom interactions . Li Wei (2014) examines classroom interactions between the children and their teachers in the UK Chinese heritage language schools.The findings indicate that the Cantonese-English-speaking students sometimes employ Cantonese characters in their schoolwork in order to approximate the Mandarin expression.Li Wei argues that the students' creative and critical expressions of meanings in their schoolwork indicate their agency in constructing their sociocultural identities, attitudes and values, and challenge the dominance of Mandarin as the Chinese lingua franca.

PLAYFUL TALK IN SECOND LANGUAGE INTERACTION
The importance of playful talk in language learning and development has been discussed by a number of scholars (e.g.Cook 2000; Bell 2005).Playful talk is an interactional practice whereby linguistic resources are being manipulated to achieve ludic effects (e.g.Cook 2000).Waring (2013: 192) builds on Cook's definition of language play and conceptualises 'doing playful talk' as 'stepping into an alternative world unfettered by the roles and the setting of the classroom and doing so lightheartedly'.According to Tarone (2000), language play aims to entertain, lower the affective filter, stretch a speaker's sociolinguistic competence and destabilize the interlanguage system.Davies (2003) studies playful talk in peer interactions between first language (L1) and L2 English speakers.The analysis demonstrates that L1 speakers assisted L2 speakers in learning how to engage in playful talk, 'but also to experience its social meaning in American society' (p.1382).Warner (2004) discovers occurrences of play with the form, the concept and the frame during computermediated communication in two German online courses.Bell (2005) analyses how L2 verbal humour is constructed by L2 English speakers as they interact with L1 English speakers.The findings suggest that playful talk can be an indication of language proficiency as more advanced speakers employ L2 linguistic resources in more creative ways.Moreover, the findings also reveal that playful talk could potentially lead to a deeper processing of lexical items, making the meanings of the lexical items more memorable.Some L2 classroom interaction studies have identified the social functions of playful talk as a face-saving device (e.g.van Dam 2002) and as a strategy to create new selves and new social relations (e.g.Belz 2002).Cekaite and Aronsson (2005) explore young children's L2 playful talk in immersion classrooms and the findings illustrate that through the use of various verbal resources, including codeswitching, artful variations in pitch, playful talk generates opportunities for the learner to learn the accurate L2 lexical items and grammar.Broner and Tarone (2001) analyse young learners' playful talk in L2 during a Spanish immersion classroom and demonstrate that it allows them to deploy various linguistic resources in constructing classroom jokes and creating worlds that do not exist.
One of the first attempts to provide a conversation-analytic account of how playful talk is constructed in adult English-as-a-Second-Language classrooms, where students may not share a common L1 with the teacher and other students is that of Waring (2013).She finds that participants mobilise identity as a resource for doing being playful and argues that playful talk can allow classroom participants to perform a range of subversive acts and experiment with a wide range of voices, including as teachers, parent, child, pop culture expert.Tai and Brandt (2018) demonstrate how a learner employs both multimodal resources and her limited English repertoire to construct an embodied enactment in a humorous manner in order to display her understanding of a target lexical item in a beginner-level adult English-for-Speakers-of-Other-Languages lesson.As shown, playful talk can be seen as useful in facilitating meaning-making, creating a jocular environment, negotiating relationships, promoting student engagement and expressing students' identities (Waring 2013;Lytra 2017).
To date, there is little empirical work that explores the construction of playful talk in EMI classrooms.Although EMI classrooms are in a sense also L2 classrooms, they focus on subject contents and have pedagogical goals and agendas that are different from language classrooms.Jakonen et al.'s (2018) study analyses how a student's translanguaging practices subvert the English-only norm in a junior secondary Content and Language Integrated Learning (CLIL) history classroom in Finland and is treated as 'language mixing' by other peers.Although the definition of CLIL is different from EMI, the CLIL classroom in Jakonen et al.'s study employs L2 (English) as the MoI for teaching history.The analysis illustrates that the student's translanguaging practices involve deploying a wide range of linguistic resources, through combining lexical items and grammar of English and Finnish and uttering English words with a stereotypical Finnish accent, to challenge the institutional norm of using only English in the classroom.
Based on the review on existing literature on language play, EMI and translanguaging, this study aims to bring together the concepts of translanguaging and language play in order to extend our understanding of translanguaging practices in subject teaching and EMI.In particular, we hope to achieve a finegrained understanding of how playful talk is constructed through teacher's translanguaging practices and what pedagogical goals do playful talk accomplishes in situ.Moreover, translanguaging practices are complex in nature since different sociocultural factors, such as personal history, life experience, identity, can potentially play a role in influencing an individual's use of meaning-making resources in the process of constructing knowledge.We also hope to uncover how the teacher understands his own pedagogical practices at specific moments in the interaction and how the classroom interactions are shaped by multiple sociocultural factors.In this study, we aim to demonstrate that translanguaging can serve as a resource of creativity and language play and it can be deployed to create a translanguaging space in the EMI classroom to promote content learning and students' participation.

DATA AND METHOD
This study aims to address the following research questions (RQs): 1 How does the HK EMI mathematics teacher use translanguaging in constructing playful talk? 2 How does the HK EMI mathematics teacher perceive his use of translanguaging in constructing playful talk? 3 How do the findings of this study provide implications for EMI policy?

Participants and data collection
The participating school is a prestigious secondary school in the New Territories of HK, and it is the first EMI school in the local district.The school is subsidised by the HK government and provides education from secondary one to six based on the curriculum guides set by the HK Education Bureau.The school uses EMI to deliver most of the lessons (except Chinese, Mandarin, and liberal studies classes), and the school examinations are assessed through English.Although the school's mission statement is explicit that it aims to develop students to be bi/ multilinguals, the school language policy places heavy emphasis on the use of English on the school campus.All morning assemblies and staff meetings are conducted in English.All teachers and students are explicitly informed that English has to be used during the content lessons.Moreover, English-for-all-days are held on every Monday when everyone (all teaching staff and students) in school must use English for communication.However, in practice, the actual implementation of English-for-all-day could vary as not all students are willing to speak English to their peers and teachers outside the classrooms.Chinese week is also held to promote Chinese acquisition, but these events are only held annually.Hence, it can be seen that the school's language policy is biased in favour of English over other named languages (Cantonese and Mandarin in this case).
The mathematics teacher, who agreed to take part in this study, has at least eight years' experience in teaching mathematics in English.He is a Cantonese L1 speaker and previously attended an EMI school for his secondary education.His bachelor's degree in mathematics and IT education and MSc in Mathematics were obtained from two top-ranked universities in HK.These universities also use EMI.During his undergraduate studies, he occasionally taught drama at several HK secondary schools.He did not receive any EMI teacher training when he was pursuing his education degree.
The first author carried out intensive fieldwork in the school.A one-hour presemi-structured interview was conducted with the teacher in order to understand his professional training, his linguistic knowledge, his perceptions of the best practices and his attitudes towards using multiple languages in the EMI mathematics classrooms.During the fieldwork period, the first author observed a year 9 class.There were 18 students in the class and this class was classified as an enhancement class.Students, who ranked below average among their cohort in the internal mathematics examination, were enrolled in this class.All students in the class were 15-year-old and they spoke Cantonese as their L1s except two students in the class.These two students spoke Mandarin as their L1s, and they were migrants from the mainland China.All students have received at least 6 years of primary education, where Cantonese was employed as the MoI and English was taught as an L2.Based on the first author's initial conversations with the teacher, most of the students passed the internal school English examinations which involved reading, writing, speaking and listening.
Eleven 40-min lessons were observed and video-recorded.Five informal interviews were conducted with the teacher and students during the twoweek observational period in order to gain detailed information about the observed lessons.These informal interviews lasted for 5-15 min and they can be considered as ethnographic interviews (Spradley 1979) since they took place spontaneously rather than being scheduled with participants in advance.A 1-h post-video-stimulated recall interview was conducted with the teacher in order to compare his actual translanguaging practices and his interpretations of his practices.All interviews were carried out in Cantonese.

Combining Multimodal Conversation Analysis with Interpretative Phenomenological Analysis
This study integrates MCA with IPA to study playful talk and translanguaging practices in EMI mathematics classrooms.Such methodological approach falls under the umbrella of Linguistic Ethnography.Linguistic Ethnography affords the capacity of a linguistic-oriented analysis to 'tie ethnography down' and 'open up' linguistic analysis (Rampton 2006: 395) without excluding ethnographic data so that the strengths of each complement the weaknesses of the other.In order to capture the complexities of translanguaging practices and the sociocultural factors that affect individual's meaning-making resources, it is essential to utilize a flexible framework that can combine different methodologies.MCA 'focuses on how social order is co-constructed by the members of a social group' (Brouwer and Wagner 2004: 30) through fine-grained analysis of the social interaction.It takes an emic/participant-relevant approach in order to explicate the detailed process of how social actions, such as learning, are co-organized and achieved through talk-in-interaction.The data are transcribed using Jefferson's (2004) and Mondada's (2018) transcription conventions.As Waring (2013) suggests, in order to identify playful sequences in classroom talk, it is important to explore extracts where talk is treated as playful by the participants themselves (e.g.laughter).
This study also draws on the analytical approach of IPA to investigate how the mathematics teacher perceives his own translanguaging practices at specific moments in the interaction.IPA focuses on the in-depth exploration of personal experience and how individuals understand and make sense of their experiences.In addition, IPA acknowledges the investigation of the meanings of the participants' experiences as an interpretative enterprise on the part of both researchers and participants.Thus, in order for researchers to understand how participants make sense of their world, a dual interpretation process called 'double hermeneutic' is adopted.Such a process requires researchers to try to make sense of the participants trying to make sense of their world (Smith et al. 2013).By doing so, it allows researchers to take an emic approach in order to understand the participants' personal experience case-by-case.After conducting the IPA analysis, we design a table with four columns in order to help readers to understand how the analyst makes sense of the teacher trying to make sense of his own teaching.From left to right, the first column presents the classroom interaction transcripts.The second column includes the stimulated-recall interview transcripts.The third column illustrates the teacher's perspectives of his own pedagogical practices.Finally, the fourth column documents the analyst's interpretations of the teacher's perspectives, which aligns with IPA's interpretation process.

ANALYSIS
We now analyse examples of playful talk for facilitating content learning (Extracts 1 and 2) and promoting meaningful communication (Extracts 3 and 4).

Constructing playful talk to facilitating content learning
In the dataset, four instances were identified which illustrate the occurrence of playful talk in the main instructional sequences for promoting content learning.This can allow classroom participants to engage in humorous talk while the teacher is teaching the content.Extracts 1 and 2 are typical examples that reveal this interactional phenomenon and illustrate the role of translanguaging in creating the playful talk.

Extract
1: Constructing a mnemonic to facilitate students' memorisation Prior to the extract, the teacher (T) was teaching the concept of a slope using English.T explained to students that when the straight line goes upward to the right, then the slope is positive.If the straight line goes downward, then the slope is negative.If it is a horizontal line, then the value of the slope is zero.During the T's explanation, T was drawing the slanting lines (going upward and downward) and a horizontal line on the blackboard which formed a triangle (see Figure 1).In Extract 1, T's translanguaging practices can be observed through his use of Cantonese rhyming words, repetition, stress and an English technical term 'slope' to create a mnemonic (line 80) and reinforce the value of the slope of the horizontal line.Concurrently, T's translanguaging practices also involve his deployment of gestural and semiotic resources (e.g.drawings on the blackboard) in order to display the flatness of the horizontal line that mathematically represents the value of zero.
In line 76, T first utters a Cantonese particle, 嗱 ('hey'), to draw students' attention.T then switches back to English to utter 'memorise' twice and enunciates another Cantonese particle, '吓' ('okay?'), in order to emphasise the need for students to remember something.After a 0.4-s pause, T switches back to Cantonese to mention the mnemonic, that is created by T, to the students, '呢 個口訣 (1.1) 我發明嘅 (this mnemonic (1.1)I invented it)' (lines 76-78).In line 80, T suddenly speaks slowly when uttering '條線 (the line).Simultaneously, T's hand movement visually indicates to students that the horizontal line on the blackboard is the line that T is referring to (Figure 1).When T utters the word '平 (flat)', T moves his right-hand along the horizontal line (Figure 2) in order to visually illustrate the flatness of the horizontal line to students.T continues to construct the second part of the mnemonic by uttering '個(the) slope (0.9) 係零 (is zero)'.It is important to also notice that the word '係 (is)' is repeated twice.Second, the English word 'slope' is used here to reinforce the technical term in the mathematical discourse.Third, the words '平' [ping4] and '零' [ling4] are rhyming words in Cantonese and coincidentally the meanings of these two words (i.e.flat and zero) reinforce the mathematical concept that the slope of the horizontal line must be zero.This message is also further emphasized as T points at the '0' on the blackboard (Figure 3) when he is uttering '係零' with stress.After T's introduction of the mnemonic, the students are clapping (line 81) in order to express their enjoyment of listening to T's mnemonic.S3 acknowledges the funniness of T's mnemonic by saying, '又幾好笑個喎 (it's quite funny)' (line 85).
In this extract, the construction of the mnemonic is considered as playful as signalled by the teacher's and student's reactions (e.g. a verbal acknowledgement in line 85 and the teacher's laughter in line 78).During the post-videostimulated-recall-interview, T comments that this mnemonic was created by him when he was a secondary school student.The researcher is interested to understand the T's reasons for using rhyme in creating this mnemonic (see Table 1).
T explains that he personally enjoys integrating rhyming words in his talk in order to create a doggerel effect.Particularly, he likes positioning the rhyming words at the end of the sentence in order to draw students' attention to the rhyming words.Towards the end of line 16 in the interview, T shifts the footing by voicing out his students' reactions as they hear the mnemonic: '咦, 有 啲, 又啱音喎 (oh, that's, that rhymes)' and '佢講嘢一陣咬字又會唔會啱音嘅呢 (oh, that teacher is talking now and will he use any rhyming words in his utterances?)'.Here, it can be suggested that T displays his expectation that the mnemonic will draw students' attention.In the MCA analysis, it is evidenced that T's introduction of the mnemonic is received with applause from students (line 81 of the interaction) and a verbal endorsement (line 85 of the interaction).This suggests that T's use of mnemonic is considered as playful and funny by the students.T also mentions that using these rhyming words can prevent boredom in the classroom and facilitate students' memorisation of the mnemonic due to the rhyming effect.Therefore, it can be argued that T's personal interest in using Cantonese rhyming words and his pedagogical goal (i.e.assisting students in memorising the mathematical content) shape his translanguaging practices in constructing the mnemonic and creating a humorous context in the classroom interaction.
Extract 2: Creating an imaginary context to facilitate students' memorisation Extract 2 is the immediate continuation of Extract 1. T aims to introduce the imaginary context of going hiking to facilitate students' memorization.T picks up a red-colour chalk from the tray in line 108 and in the subsequent interaction, everything that T writes on the blackboard is in redcolour, as opposed to the typical white-colour.In this extract, it can be shown that the triangle on the blackboard, which was constructed before the commencement of Extract 1, momentarily represents a hill and the hand-drawn person on the blackboard is often referred to as the students in the class and occasionally as T himself.In particular, T adopts a character viewpoint by imagining himself who goes hiking.By doing so, T translanguages through switching his intonations and displaying his facial expressions to enact the feelings of going up the hill, which is laminated with a tone of nonseriousness.This is treated as playful and laughable by the class.blackboard, which was previously created prior to Extract 1, figuratively represents a hill.T then adopts a smiley voice and asks whether students will feel happy when they go hiking (line 120).By adopting the smiley voice and enacting a running gesture (Figure 6), T is conveying a sense of happiness in walking up to the hill.However, student 3 (S3) provides a negative response in Cantonese and T displays his disappointed facial expression by narrowing his lips while S3 is speaking (line 122).Several students are laughing in line 124 since S3 challenges T's prior assumption regarding the students' reactions.In line 126, T repairs S3's negative comment by reiterating the happiness of hiking, '我好開心嘅 (I am very happy).In line 128, T switches footing from instructional frame to hypothetical frame by imagining himself as the person and voicing aloud his own feelings in English: 'I am very happy I am very (0.3) positive'.Note that when T utters the word 'positive', he moves the finger along the slanting line (upward) and this allows students to realize that T's jocular and positive feeling is associated with the mathematical meaning of positive (i.e.above zero).After a 2.6-s pause, several students are laughing (line 130) and clapping their hands (line 131) to applaud T's performance.In line 132, T switches back from the hypothetical frame to an instructional frame in order to provide explicit explanation to students by stating, 'that's why the slope is positive'.
In this extract, T translanguages through utilizing various multilingual and multimodal resources, including gestures, intonations, smiley voice, facial expressions, use of Cantonese, the drawings (hand-drawn person and a triangle which represents a hill), to adopt a character viewpoint and create a congenial scenario where he walks up to the hill.This allows T to connect the mathematical idea of the positive value of sloping upward with a delightful feeling.In the post-videostimulated-recall-interview, T offers his opinion regarding his use of drawings to facilitate the construction of the imaginary context (see Table 2).
The researcher first draws T's attention to his drawings on the blackboard.In the interview, the researcher is wondering whether T's illustration of an everyday life example can assist students in understanding the mathematical concepts.T then points out that since mathematical numbers dominate the mathematical discourse, using pictures can visualise the mathematical concept to the students.In particular, T shifts the footing by imagining himself as his students and voicing out their reactions when they look at T's drawings: '所以 佢哋就會, 嘩, 又係留心你畫啲乜 (So, they would be like: wow.They would pay attention to what you were drawing.)'.This illustrates his expected reaction that he will receive from his students.In the MCA analysis, it is evidenced that several students are laughing while T is drawing (e.g.lines 116 and 144).This indicates that the students are paying attention to his drawings and they treat it as humorous.Therefore, it can be argued that T's pedagogical goals (i.e.drawing students' attention and visualising the mathematical concept) motivate T in drawing images on the blackboard which contributes to T's translanguaging practices and the creation of a humorous context in the classroom interaction.

Playful talk for promoting meaningful communication
In the dataset, three instances were identified which involves the teacher and students engaging in extended discussions, which may not have direct relevance to the content subject.The playful talk in Extracts 3 and 4 are typical examples which reflect this feature.These extracts are different from the playful talk, which were analysed in the previous extracts since the interactions in the previous extracts have the pedagogical goals of promoting content-related learning.Rather, the pedagogical goals in the following examples aim to promote communication with the students so that the interaction values the students' ideas and expressions of their life experiences.
Extract 3: Drawing on the limited linguistic knowledge of Mandarin Prior to the extract, T read out the mathematical question that students needed to solve.After reading aloud the question, T initiated a question by using rhyming words at the end of each sentence to create a rhyming effect.However, when he uttered the last sentence, he could not think of an appropriate rhyming word/phrase that could be used at the end of the sentence.This led to students' laugher in the classroom.Student 12 (S12) then asked T whether he was able to pronounce '我覺得不行 (I don't think so)' in Mandarin.In Extract 3, it is evidenced that T and students are engaging in discussions about T's ability in pronouncing a Mandarin phrase.

Table 2 Continued
In response to S12's question, T attempts to use Mandarin by saying '我 (w o) (i.e.I)' (line 54).However, T discontinues uttering his responses and he then switches back to Cantonese to explain to the students that he needs time to process the Mandarin pronunciation.Although T attempts to utter in Mandarin in line 57 to respond to student 1's (S1) comment, T fails to construct a proper sentence as evidenced in the repetition of '我 (w o)' and the short pauses in between the utterances, '我覺 (w o jue) (i.e.I think) (0.3) 我 (i.e. w o) (I) (0.2) 我 (w o) (i.e.I)' (line 57).As shown in line 58, T's truncated Mandarin utterances are received with laugher from students.
However, T has not given up on using Mandarin in the classroom.T specifically selects S3 as the next speaker by announcing her name in Mandarin in line 60 and pointing at S3 (figure #9, line 61).It is important to note that S3's first language is Mandarin, and she and her family are migrants from mainland China (T's pre-interview).T then makes a request to S3 in line 64 by saying '告訴我 (tell me)' in Mandarin.By asking S3 to inform him the correct way of uttering '我覺得不行' in Mandarin, T is treating S3 as the linguistic expert who has the ability for repairing his Mandarin pronunciation.In line 66, S3 responds to T's request by offering the correct Mandarin pronunciation of the phrase, '我覺得不行 啊 (w o ju e de b u x ıng a)'.T takes the next turn and attempts to repeat S3's pronunciation in order to display his understanding in lines 68 and 71.
Notice that T points to the top and raises his right-hand upward to his face when he utters '我 (w o)' in line 71 (Figure 10) in order to visually illustrate the high intonation of this word.Instantaneously, S3 repeats the correct pronunciations again while T is speaking in order to provide corrective feedback to T (line 72).T continues to mispronounce '我' and coincidentally employ an iconic gesture to represent the first tone in Mandarin (Figure 11) (i.e. a horizontal line above the vowel).After a 0.3-s pause, T slants his index finger down towards his left (Figure 12) when he utters '覺得 (ju e de)'.This iconic gesture is possibly referring to the fourth Mandarin tone (also known as a falling tone) but T does not enunciate the words, '覺得 (ju e de)', in the fourth tone.T's attempt in using Mandarin is immediately received with laughter from the students.Simultaneously, T recognizes his failure in enunciating the correct Mandarin pronunciations through smiles and bending over the desk (figure #13).
Based on T's self-reflection during the pre-interview, T considers that his Mandarin proficiency is below average.In this extract, it can be seen that T is translanguaging as he draws on his limited linguistic knowledge of Mandarin, accompanied by various gestures, to construct a humorous atmosphere in the classroom.He takes this opportunity to invite S3, who has linguistic expertise in Mandarin, to participate in the classroom interaction.Typically, the linguistic codes (Cantonese and English) are mostly employed in the classroom.Hence, allowing S3 to translanguage (i.e.drawing on her familiar language, Mandarin) in the classroom makes the meaning-making process much more inclusive and honours the diverse communicative resources available in the classroom.During the post-video-stimulated-recall-interview, T comments on his use of Mandarin in the classroom (see Table 3).
T comments that he is having fun with his students and suggests that his pedagogical goal is to provide a break time for students.It is noticeable that T often uses phrases such as 'take過break (take a break)' and '休息 (rest)' in Cantonese to reinforce the need for the students to take a break.This is possibly because the students in this class are low performers in mathematics.T seems to understand his students' ability as he is aware that the students will not be able to concentrate during the mathematics double lesson which lasts for 90 min.Hence, taking a break can allow students momentarily move away from mathematics.Additionally, T acknowledges the fact that students enjoy teasing his Mandarin pronunciations.Particularly, T's words, '咁所以我咪比佢 哋笑下, 開心下 (So, I allowed them to laugh at me and enjoyed the laugh)', further reiterate his casual attitude towards the students' laughter.Hence, it is possible that T's motivation to befriend his students contributes to the creation of a translanguaging space, which allows students to engage in translanguaging and promotes a jocular classroom environment for students to relax.
Extract 4: Raising the issue of linguistic discrimination Extract 4 is the immediate continuation of Extract 3.After the students' laugher in line 76, T switches back to English and draws students' attention back to part b of the mathematical question (line 78).In this extract, T notably translanguages by drawing on his full linguistic repertoire (i.e.imitating a foreigner's Cantonese accent, using his limited Mandarin proficiency and L1 Cantonese) to construct a humorous classroom environment which does not only promote genuine communication with students, but also promote the examination of social issues including linguistic discrimination.
While T is specifying the sub-question 'b', S1 speaks concurrently and criticizes T's Mandarin as '馬來西亞嗰啲普通話 (Malaysian-like Mandarin)'.S1 continues to criticize T's Mandarin in line 83 by switching his speech to Mandarin and uttering the words with high intonation, '對"啊" (yes) (0.5) 不" 行"啊" (not okay)', possibly in order to imitate T's flawed Mandarin pronunciation.In line 85, T justifies his Mandarin pronunciation by explaining that '香 港人講普通話就係咁㗎啦 (this is how Hong Kong people speak Mandarin)'.Subsequently, T makes a comment in Cantonese, '你哋唔會笑啲外國人講中文 㗎喎 (you guys won't laugh at the foreigners speaking Chinese)' (line 87).By criticizing the students' views, T switches the focus of the discussion (i.e.T's Mandarin discussion) to students' perceptions about the way foreigners speak Chinese.
Interestingly, T creatively imitates a foreigner's Cantonese accent by altering his Cantonese intonations, '我哋都唔知 (we do not know) (Yale Cantonese Romanization: [o1] [dei6] [do1] [m4]/dZi+/)' (line 89), in order to portray himself as a foreigner who does not speak Cantonese.Such appropriation of a foreigner's accent is obviously different from the way T normally speaks Cantonese in the lessons.T continues his talk in line 91 by voicing aloud the students' perception, '佢哋你哋覺得好可愛個喎 (you guys think that the foreigner's Cantonese accent is cute)'.Through creating a performance of an 'acceptable' Cantonese accent in line 89, T aims to allege accent discrimination.In line 93, T initiates a rhetorical question to prompt students to reflect on their perceptions of different accents, '點解你哋唔可以用另外一個方法你覺 得好可愛呢?(why can't you try to use another way to perceive that as cute?)'.After a 0.2-s pause, T utters the Mandarin phrase, '我覺得不行', again and he mispronounces 我(w o) in this instance.By uttering the Mandarin phrase in this way, T is possibly affirming his HK Mandarin accent.However, S3 uses Cantonese to criticize T for not being 'cute' from her perspective (line 101).T immediately turns his body to face at S3 and asks '為什麼啊 (why)' in Mandarin.T's utterance is also marked with a loud voice as well as the exaggerated non-verbal gesture of dropping his hand (figure #14) in order to playfully enact his frustration towards S3's criticism.T's reply is treated by the class as a laughable, as shown by the laugher in the next turn (line 103).
As demonstrated in Extract 4, the playful talk momentarily becomes a translanguaging space which encourages open discussions between T and the students to identify their own biases and critically reflect on them.During the post-video-stimulated-recall-interview, the researcher is wondering what motivates T to have the classroom discussion about linguistic discrimination (see Table 4).
In order to make sense of T's pedagogical goals, the researcher asks T to explain whether he has another pedagogical goal in mind for having a class discussion about his 'non-standard' Mandarin pronunciation.T then states that he wants to prompt students to reflect upon their own biases.Such reflection confirms the MCA analysis that T creates playful talk in order to allow students to reflect on the issue of accent discrimination.T also briefly recounts his experience of being teased because of his poor Mandarin proficiency.Such incident motivates him to think about the issue of linguistic discrimination.It is noticeable that T shifts the footing by imagining himself as his students and voicing out their reflections: '咁樣笑我都好似唔係好合適喎 (Oh.Laughing at me in this manner.It might not be the most appropriate action)'.This shift of footing illustrates T's expectations that his students will reflect upon their behaviour.In the interview, T shifts to an unknown person's voice when he says: '我朋友好有趣呀講, 學人講廣東話 (My friend is so interesting.He is trying to speak Cantonese)'.By imitating an unknown person's voice, T attempts to portray the perception that people typically hold about foreigners speaking Cantonese.Hence, it is argued that the translanguaging space, which is created by T, is shaped by T's personal experience and his own reflection regarding this social issue.

DISCUSSION
The analysis of the extracts has revealed that translanguaging can be used as a critical resource for constructing playful talk, which allows the teacher to achieve his pedagogical goals.In response to the first RQ, the sequential analysis of the interactions has demonstrated that playful talk can be constructed for facilitating content learning (Extracts 1 and 2) and promoting meaningful communication between the teacher and students (Extracts 3 and 4).In all cases, playful talk is oriented to by the classroom participants as humorous (Waring 2013).A range of specific pedagogical goals can be achieved through playful talk, including establishing an imaginary context, circumventing possible limitations in comprehending abstract or complex explanations and discussing social issues.Similar to prior studies (e.g.Broner and Tarone 2001;Warner 2004;Waring 2013;Tai and Brandt 2018), interactional features in Table 4 Continued playful talk are identified in this paper, which includes adopting an informal register, exploiting unusual lexical items, laughers and playfully initiating uninvited responses.In Extract 1, the teacher creatively employs linguistic resources, including Cantonese rhyming words, repetition and an English technical term, to form a mnemonic and deploys gestural resources to visually illustrate the meaning of the mnemonic to the students.In Extract 2, the teacher skilfully shifts footing in order to enact a discourse identity (i.e. the imagined person going hiking) which creates an imaginary context to facilitate students' understanding of the mathematical concepts.In Extract 3, the teacher deploys his limited Mandarin repertoire to respond to the student-initiated playful comments as a way offer a space for students to take a break from doing mathematical questions.Extract 4 also demonstrates how T and students engage in extended discussions about accent discrimination.Particularly, T shifts footing to facilitate his portrayal of a discourse identity as a foreigner who cannot speak Cantonese in order to promote the examination of this social issue.
With regard to the second RQ about how does the teacher make sense of his use of translanguaging in creating playful talk, the analysis of the videostimulated-recall-interview demonstrated that translanguaging does not only enable the teacher to bring together multiple linguistic and multimodal resources to construct meaning.It enables the teacher to bring his prior life experience as a student (Extracts 1 and 2), his personal interest in adopting particular linguistic features (Extract 1) and his prior experience of being teased (Extract 4) into the playful talk which contributes to the creation of translanguaging spaces in the classrooms (Li Wei 20112018).In addition to bringing along the teachers' personal interests and his prior life experience to the classroom interactions, the findings further highlight that the teacher brings his various pedagogical knowledge and beliefs (e.g.knowledge of students' academic and linguistic backgrounds, knowledge of scaffolding strategies and understanding of students' personality traits) into his teaching.These are crucial factors to be considered in order to understand how the teacher creates a translanguaging space to achieve a range of pedagogical goals in playful talk.
The findings of this study have pedagogical implications for both EMI teachers and students.Throughout the analysis section, we have demonstrated that playful talk in EMI classroom helps to create a translanguaging space, which allows classroom participants to bring in a range of linguistic and multimodal resources and different kinds of knowledge into the lessons.It moves away from the typical view to EMI mathematics classrooms which provide limited opportunities for students to interact with the teacher (Lo 2014;Tollefson and Tsui 2014).Through playful talk in EMI classrooms, the participants transform the traditionally teacher-fronted interaction to negotiate a space for voicing their thoughts and create a more dynamic and contingent environment to facilitate students' participation.

CONCLUSION
The findings of this study reveal the potential of constructing playful talk through translanguaging in a multilingual but nominally EMI mathematics classroom.This study has implications for implementing EMI policy at the classroom level.EMI has raised important concerns regarding the complicated interrelationships between language use and content learning (Lin 2006;Macaro 2018).In this study, we have demonstrated that translanguaging can serve as a source of creativity and language play which allows classroom participants to bring in a range of linguistic resources, including composition (e.g.smiley voice, laughter, the volume of voice, word choice), multimodal resources (e.g.gesture and drawings), various pedagogical knowledge and skills, personal experience and interests into the lessons.In this way, translanguaging helps to uncover the misconception of the monopoly of English as the norm in EMI classrooms.This prompts the policymakers to recognize translanguaging as an empowering tool for promoting linguistic diversity in the EMI classrooms and maximizing language users' full linguistic and semiotic resources in knowledge construction.Such perspective treats the multilinguals' ability to speak multiple languages and deploy various semiotic and sociocultural resources as an asset instead of a hindrance affecting their learning processes (Li Wei 2018).The findings provide insights into the need for HK to develop a robust and socially responsive plurilingual model which can offer discursive spaces for various multilingual and multimodal resources along with the target L2 (English).
One limitation of this study is that it is restricted to one EMI teacher and one content subject from a HK secondary school.A longitudinal case study examining the role of translanguaging in creating playful talk by different teachers in different EMI classrooms can enrich our understanding of how using translanguaging in playful talk can lead to positive outcomes on students' content acquisition and English language development.Although this study has framed playful talk in EMI classroom interaction in a positive light, it is possible that playful talk may not be always understood by all students within the classroom (Waring 2013).Hence, it is worth investigating how translanguaging can exclude those who find themselves unable to participate in playful interaction for different reasons.