TRANSLANGUAGING AS PLAYFUL SUBVERSION OF A MONOLINGUAL NORM IN THE CLASSROOM

Previous studies of educational translanguaging have described it as an instructional and inclusive practice supporting the active classroom participation of students from diverse linguistic backgrounds. This chapter demonstrates how in monolingually-oriented educational contexts, translanguaging can also constitute a form of subversive language play targeting the local monolingual norm. The data are video-recorded lessons from secondary-level CLIL (Content-and-Language-Integrated-Learning) classrooms in Finland. In CLIL classrooms L2 is often upheld as a normatively assigned medium of interaction, particularly in whole-class talk, and students use their shared L1 in peer interaction. This chapter offers a case study investigating how one student’s translanguaging, which takes place as a reaction to the teacher’s enforcement of the L2-only norm, is treated as a ‘language mix’ by other participants in the classroom. Drawing on conversation analytic (CA) methodology, we describe the sequential unfolding and the normative context of the focal student’s translanguaging, as well as practices of categorisation with which other students respond to him. We suggest that these kinds of situations can help to empirically tease apart some differences between translanguaging and code-switching. Further, we argue that the ‘meaning’ of translanguaging to participants cannot be established without considering its relation to locally upheld norms around language choice, which in the present case are employed as resources for the construction of language play and subversive identities.

Monolingualism and the separation of different languages have a long history in language education and its research. In many corners of the world, the prevailing attitude is to devote formal language arts classes to a specific language and, in the varied contexts of bilingual education, regulate how much each language is used in the teaching of subject matter. In this chapter we investigate classroom interaction with a focus on interactional translanguaging practices, their sometimes problematic nature for participants in a bilingual educational context in which the pedagogical rationale is to maintain a separation between the two languages, as well as the relationship between translanguaging and code-switching. Our aim is to shed light on an educational context and interactional situations in which the practice of translanguaging can be a contested affair, being something that is at odds with the established pedagogy.
From a practice-oriented perspective, translanguaging has been characterised as an umbrella term (Nikula and Moore, 2016) that refers to the flexible language practices of bi-and multilingual speakers. Such practices bring into contact various semiotic resources in a creative way (Blackledge and Creese, 2017). Consequently, research on translanguaging has tended to focus on bi-and multilingual interaction as a form of social action rather than on describing bilingual speakers' languages as compartmentalised sets of elements, rules and skills (Noguerón-Liu and Warriner, 2014, pp. 182-183). Such a distinction can be seen in terms of a contrast that Thibault (2011) makes between 'first-order languaging' and 'second-order language'. Whereas languaging is a fundamentally dialogical phenomenon taking place as people engage in social interaction, language refers to lexicogrammatical attractors that have been formed over time and that "guide and constrain first-order languaging" (Thibault, 2011, p. 216; see also Turner and Lin, 2017).
A running theme in educational studies of translanguaging has been how pedagogies could encourage fluid language practices and allow the use of minority language students' heteroglossic linguistic repertoires in the multilingual classroom (see e.g. Blackledge and Creese, 2014;García 2009;Wei 2011). As Cenoz and Gorter (2014) and Hélot (2014) advocate, a translanguaging pedagogy can highlight individual agency and counteract linguistic insecurity in the classroom by supporting student participation. In this line of work on translanguaging, it is possible to detect a socio-political motivation to legitimise the use of more than one language in 3 communication, which in the English-speaking countries has typically been problematic in the education of indigenous bilingual populations and immigrants (see García and Lin, forthcoming).
Instead, a translanguaging pedagogy encourages students to engage in a creative flow of semiotic resources, blurring the borders of different languages; it aims to make use of the full range of students' linguistic repertoires, histories and experiences (Blackledge and Creese, 2014;García, 2009;Wei, 2011). Currently, attempts are on-going to extend the notion of translanguaging to educational contexts involving majority students, too (Turner and Lin, 2017).
It is useful to acknowledge these two senses of translanguaging -practice and pedagogy -when considering how the term has been introduced in the linguistic literature alongside, and sometimes to replace, the concept of code-switching (CS). By now, there is an extensive literature that has approached code-switching from a micro-interactional, conversation analytic (CA) perspective, both in everyday and educational settings (for early work, see Auer, 1984;in classrooms, e.g., Bonacina and Gafaranga, 2011;He, 2013). While the epistemological assumptions behind the notion of code-switching may be somewhat different from the theoretical groudings of the emerging translanguaging research, the existing CS literature has provided a significant contribution to our understanding of the conversational structures and interactional functions of language alternation. It is not always easy to identify what the 'meaning' of an individual codeswitch is (Alvarez-Cáccamo, 1999), yet CS research has demonstrated that the use of diverse language resources within one situation does not signal a linguistic 'deficit' but can instead denote, for example, identity work or community-building (see e.g. de Fina, 2007).
Many interactional studies of code-switching inspired by conversation analysis ask the question 'Why this language now?', and search for evidence for the functions of a code-switch in the local sequential context. Recently, this has been problematised from the point of view that it contains the presumption that it is possible to distinguish a single clear 'code' (language) at all points in multilingual conversation (see also Auer, 2007). Languages 'leak' and are merged or crossed (Rampton, 2006) in many ways in interactional situations, which has given rise to criticism of the analytical orientation to identifying the base code as a practice that itself maintains the monolingual habitus (Gogolin, 1997) of language separation (e.g. Blackledge and Creese, 2014;Piller, 2016). These, and the observation that bi-and multilingual speakers do not always denote 4 an apparent meaning or function to a code-switch, are among main reasons why there is a need to rethink existing terms. The promise of translanguaging as an analytical concept is that it may help overcome this implication by paying attention to the speaker and their action instead of focusing on what the current code is (see also Blackledge and Creese, 2017).
Whether one subscribes to the notion of code-switching or translanguaging, many analysts of multilingual language practices are familiar with the difficulty of analysing highly fluid (firstorder) languaging practices without reifying (second-order) 'languages' by referring to named languages, or "Languages-with-capital-L" (Jaspers and Madsen, 2016, pp. 236-237). This is not only the analyst's dilemma, but 'languages' are undoubtedly very real, commodified entities in which participants themselves can invest in many settings (see also Jaspers and Madsen, 2016;Møller, 2016, p. 280), not least in education. In bilingual education, named languages become tangible through the distribution of content lessons between L1 and L2 in the curriculum, as well as teachers' classroom practice of maintaining language separation by way of sanctioning students for 'inappropriate' language choice (e.g. Amir and Musk, 2013;Copp Jinkerson, 2011;Jakonen, 2016). Moreover, in peer interaction, assessment, correction and criticism of language use can tap on a student's perceived (lack of) skills in a named language (e.g. Cekaite and Björk-Willén, 2013;Evaldsson and Cekaite, 2010). The aforementioned studies describe some ways in which students learn to reproduce discursive practices that establish hierarchies between more and less valorised languages (e.g. named languages vs. 'mixed' forms) and regulate their own and their peers' language choice. Such normative practices construct and maintain an educational regime of parallel monolingualism (Heller, 1999) in which only one language should be spoken at a time. From a translanguaging perspective, this can be seen to restrict student agency by not allowing them to mobilise the full range of their linguistic repertoires.
In this chapter we attempt to tease apart some differences between translanguaging and codeswitching as interactional practices. We analyse interaction in a bilingual educational context in which the teacher makes an explicit effort to keep a monolingual L2-only norm in the spirit of a pedagogical drive to push the students to use the L2, in this case English, as opposed to their shared L1 (Finnish). The analysis focuses on how one student does not align with such a normative prescription of a medium of classroom interaction (Bonacina and Gafaranga, 2011) 5 but instead does what we view as playful translanguaging: that is, he combines resources from different 'languages' in a highly fluid, unpredictable and creative manner, without marking the combination of these 'languages' as problematic. We show how the peers of the focal student respond to this kind of playful subversion of the local classroom norm by assessing and mocking his language competence as well as categorising his language as a 'mix'. Our argument is that in instances where students use more than one 'language' in a monolingual context, there can be a difference between whether one is indeed seen to mix languages or enact a short-lived codeswitch to L1. We suggest that these subtle differences can be socially consequential to students and index particular identities. Our overall aim is to broaden the range of existing translanguaging studies in education, which have to date focused on how translanguaging can be used as a constructive and emancipatory resource for instructional activities. In our case, translanguaging as an interactional practice comes in sharp conflict with the local pedagogical rationale. We suggest that translanguaging research needs to examine local contexts and locally upheld norms around language choice in close detail in order to understand what meaning and impact translanguaging practices have to classroom participants.

DATA AND METHOD
The empirical analysis draws on a corpus of 16 video-recorded English-medium lessons on British history that were taught to 14-15-year-old Finnish L1 students between December 2010 and February 2011. This took place at a lower secondary school in a bilingual (Finnish and Swedish) region of Finland. The lessons were part of what the school referred to as their English immersion programme, which is perhaps more aptly described as fairly small-scale content and language integrated (CLIL) instruction in that the amount and organisation of L2 instruction was relatively flexible and depended on the competences of the current staff (for a discussion on how these two bilingual approaches differ in the Finnish context, see Nikula and Mård-Miettinen, 2014). At the time of data collection , the two weekly history lessons represented the only academic subject that these Year 8 students attended in English. Their teacher, an experienced content specialist, was a Finnish L1 speaker fluent in English. She had previously worked abroad in English-speaking countries and was married to an Englishman.

6
The language policy of the classroom, both declared and practiced, was the exact opposite of what could be seen as a translanguaging pedagogy. A norm of L1 and L2 separation was promoted by the teacher: the students were to use English in all communication as soon as they entered the classroom. In actual practice, however, such a blanket requirement mainly extended to teacher-student interaction, so that the students routinely used Finnish in their peer interaction (see also Jakonen, 2016, on how such a normative conflict can be handled by the teacher with relatively 'soft' policing).
For the purposes of this chapter, we adopt a case study approach by focusing the analysis on how one student, Sakari, and his peers Inka and Susanna interact in the classroom (see also Waring, 2013, for a similar approach). During the two-month data collection, a general impression emerged of Sakari as a fairly talkative boy who quite frequently engaged in different kinds of off-task behaviour with his peers. In his communication, Sakari drew on a broad range of languages and semiotic resources, often for the sake of entertaining himself and others by way of language play and linguistic performance. Besides English and his L1 Finnish, a third 'named language' that he would sometimes use was Swedish, from which he inserted individual words or short phrases in his talk. As far as we are aware, Sakari was not from a bilingual family, and neither the other students nor the teacher seemed to orient to Sakari as a bilingual. This suggests to us that Sakari's Swedish usage may have been more an act of creative use of resources that still are in the local environment (both in the form of formal Swedish language lessons in school and in the local community) than something that could be traced to Sakari's family background.
Sakari was typically not treated by others as a student with a high level of English competence, quite the opposite in fact (and the following analysis will demonstrate some ways in which such a peer assessment of his skills would take place). What is remarkable is how such a negative social positioning did not prevent him from communicating actively in the classroom. To us, it seems that his fluid and creative use of unexpected language resources -that is, translanguaging -was exactly what enabled his participation, but did so at a certain cost. Therefore, our interest with the following empirical materials lies in how such language use on the one hand subverts the local 'normative climate' of monolingualism and, on the other hand, has quite tangible social consequences when it is evaluated, categorised and contested by both the teacher and Sakari's peers.
Methodologically, we draw on conversation analysis (for an introduction, see Sidnell and Stivers, 2013) in order to investigate how Sakari's translanguaging emerges and is received in a particular sequential context. CA is a data-driven and micro-analytical approach to studying social interaction, which has its intellectual roots in ethnomethodology and social sciences (see e.g. Maynard, 2013). In the following analysis, we present and analyse two extended sequences, each c. 1 minute 30 seconds long. These analyses show how Sakari uses a hybrid language form in response to the teacher's normative requirement to use the L2 only (extracts 1-2), and how other students tease him about it, categorising it as a language 'mix' (extracts 3-4).
The data have been transcribed following the Jefferson (2004) notation conventions. Turns that contain Finnish or Swedish language units have been translated into English, aiming at idiomatical equivalence. For selected turns-at-talk, we provide an approximation of their pronunciation and grammatical structure using IPA transcription and interlinear glossing (see Bickel, Comrie and Haspelmath, 2008, for the grammatical labels used). All names in transcripts are pseudonyms.

ANALYSIS
The first extract shows how three students, Sakari, Inka and Susanna, respond to the enforcement of an L2-only norm in a joking and playful manner, co-constructing a non-sensical performance.
It takes place as the class has been writing a short summary of their essays. They have been doing this while drinking tea, which was a routine way to begin the lesson in this class, a teacher's way to familiarize the students with what she viewed as the British culture. As the extract begins, the three students have just returned their task papers to the teacher, who continues to circulate in the classroom collecting student work. This means that they now have some extra time before the teacher assigns a next task. During this transitory period, the students have been joking in Finnish about drinking various beverages through the nose, perhaps inspired by the tea routine in the class. The teacher's reproach (line 9) targets this L1 talk and brings about a shift in language and activity.  The teacher enforces a monolingual classroom norm (see also Amir and Musk, 2013;Jakonen, 2016) by labeling Finnish as the 'wrong' language (line 9) and by issuing a directive to the students to use English instead (line 12). Her reproach is double-edged in that it addresses two kinds of violations of the local classroom order: the absence of institutionally-assigned L2 and off-task talk, or 'gossiping'. Her turn thus enacts a language policy of 'strict' L1 and L2 separation, but does so in a relatively soft and humorous manner by permitting the other violation, 'gossiping', as long as it takes place in the L2. As we see in their responses, the reproach is initially complied with by all three students: Inka acknowledges a 'slip' (line 11), Susanna continues talk on the same topic but in English (line 13) and Sakari begins a new sequence in English (line 15).
Sakari's sequence-initiation at line 15 is a smilingly produced 'confession' that he added energy drink (brand name 'ED') in his tea. The turn brings about a shift in activity from gossiping about friends to doing translingual language play, a non-serious performance that fulfils the L2 norm pro forma. In constructing this turn, Sakari points at his cup and pinches his index finger and thumb together to gesture the amount, which he produces in Finnish (vähäse, 'a bit' At line 29, Sakari recites in English a well-known Finnish nursery rhyme that is used when making sand castles (or 'sand cakes' in Finnish) in the sandpit while he taps a teabag with his spoon, in the same manner as children would pat on a 'sand cake' before removing the bucket. In this sequential context, such a recital works to ward off the two girls' requests and insisting pursuits (lines 23, 26) for sharing the energy drink. A nursery rhyme also underlines the playful and nonsensical character of the students' language use in response to the teacher's normative reproach: the students orient to the topic of talk being of secondary importance as long as talk takes place in English.
Extract 2 shows how, as the situation unfolds, the students' normative orientations to the medium of interaction come in conflict. In the extract, Sakari shifts the topic to tell a story about what happened to him in the morning. His telling is constructed by drawing on and combining a broad range of language resources, which the story recipients orient to not so much as creative expression as 'bad English'.

Extract 2. Translanguaging in storytelling.
(Finnish = bold, Swedish = bold and italics) Another kind of normative orientation to language use is visible in the turn design of Susanna's responses to Sakari's story (lines 41, 51-52). In addition to a simplified syntax, she pronounces her turns with an exaggerated and stereotypical 'Finnish' pronunciation of English. In Finnish public discourse, such pronunciation features are sometimes referred to as 'rally English', a derogatory term that finds inspiration in Finnish (male) motor racers' and athletes English skills and indexically points to poor linguistic competence (Kivistö, 2016;Tergujeff, 2014; see also "Swenglish" in Sweden, as described by Kontio, 2017 be a means to display language competence and enforce power hierarchies in multilingual peer interactions. While Inka's teasing is quite a bit more subtle than the imitation described by Evaldsson and Cekaite (2010) -which may be related to differences in age -what seems to combine them is that in both contexts a monolingual ideology is being maintained in and through these actions.
In summary, extracts 1 and 2 show how the context (sequential, activity and normative 'climate') can be significant aspects in establishing the local sense of translanguaging to participants. These extracts take place after a teacher reproach for L1 use, in moments of transition between pedagogical tasks, and by doing 'being non-serious' in response to that kind of action the students subvert the local institutional norms (extract 1). However, as Sakari begins to tell a

story, Inka and Susanna's orientation toward monolingual interaction comes in conflict with
Sakari's translanguaging. Sakari's story is a complex and unpredictable fusion of different kinds of language resources while Inka and Susanna do interactional work to keep the two languages apart by translatory practices and by conveying to Sakari that his language form is problematic.
Having described how participants can largely implicitly orient to different language use norms in classroom peer interaction, we now move on to discuss how Sakari's hybrid language form is more or less explicitly evaluated and categorized. The next data extract, which is for the sake of convenience divided into extracts 3 and 4, shows how such evaluation can find its source in the monolingual classroom policy. It takes place as the teacher is finishing instructing a group task.
In the course of doing so, she has just reminded the class to use only English in order to get 'points' which will contribute towards earning a reward (see also Amir and Musk, 2013), joking with the possibility to check English use with the help of the researchers' audio recordings. As the teacher reiterates the ultimatum (lines 7-10), Susanna capitalizes on the joke and teasingly complains that Sakari's English competence prevents the two girls from keeping the Englishonly norm in conversation with him. The teacher's threat at line 10 operates with the classic logic of instrumental conditioning whereby desirable behaviour is reinforced with a reward. Susanna's 'yes but' prefaced turn at line 12, which is produced with a smiley voice, aligns only partially with the teacher's action (see also Steensig and Asmuß, 2005): while it displays compliance with the L2-only norm, it also complains about a problem in doing so. Thus, in this sequential context, treating Sakari's presence in their group as a problem can be taken as criticism towards his English language competence, albeit done in disguise of continuing the teacher's joke. Notice how Susanna's complaint is aligned with by the teacher and Inka by suppressed chuckles (line 13) and outright laughter particles (line 14) in overlap with its production.

Extract 3. Evaluating another student's language competence
Given the delicate nature of Susanna's joking complaint, line 15 represents a sequential location in which the teacher faces the practical task of displaying a stance towards the criticism and its acceptability in the classroom. Rather than providing a response that, for example, would admonish Susanna for negatively evaluating a peer or counter the cricitism, the teacher turns the sequential relevance upside down by initiating a canonical three-part (Initiation-Response-Evaluation) instructional sequence to invite Susanna to list Sakari's 'options'. Such a management of the situation avoids taking an explicit stance towards Sakari's language proficiency, which Susanna has questioned, but instead reinforces the L2-only language norm.
By asking Susanna to produce the options, the teacher is very much treating them as already 'known answers' (Mehan, 1979). At line 28, Inka finally provides a response that reiterates the monolingual policy and is accepted, albeit with minimal linguistic correction (line 29). At lines 30-31, Sakari asks for help with beginning the task with a turn that is left unanswered. His turn thus offers a concrete display of English competence in the pragmatic context of task work, contrary to what Susanna's earlier tease implied.
What in extract 3 was somewhat implied criticism of language competence is made quite explicit and socially indexical in the next extract, which shows how the situation continues as the teacher gives the groups a go-ahead with the task (line 35) and approaches the focal group. Susanna continues to tease Sakari about his English skills (line 36) but does so in a manner that allows Sakari to even the score, at least temporarily, before Inka and Susanna categorise Sakari's talk as a 'mix' of languages and evoke the social categories which the 'mix' indexes.
Extract 4. Categorising talk as a mix of languages. Susanna's smilingly produced playful tease that she addresses to Inka (line 36) uses passive voice, which treats the non-understanding of Sakari's 'English' as a shared concern and conveys that the problem is not really in the recipient but rather in the speaker. However, using Finnish to produce the tease leaves it vulnerable to actions that uphold the L2-only norm, and that is exactly what Sakari does at line 39. His normative policing by way of 'snitching' on Susanna's 'slip' is recognized early on (lines 40-41) and appreciated (line 42) as an artful retort.
The teacher's response at line 44 is interesting insofar as she abstains from normative sanctioning of Susanna's language choice but instead turns Sakari's comment into a vocabulary teaching activity. She does this by constructing a 'we-group' of competent English users who 'say' things. The lack of any uptake during the silence (line 45) that follows the teacher's instructing action shows how such a proposed shift in activity is not aligned by the three students. This suggests that, as in earlier extracts, Sakari orients to there being no need to incorporate the L2 word provided by the teacher ('fault') in his talk by doing any retrospective correction of his prior turn. As the teacher therefore has nobody to teach, she shifts the topic to the classroom language norm (lines 46-47). Even if the teacher prefaces her turn with a display of empathy ('it's horrible Sakari') she nevertheless 'sides' with the girls ('they force you to speak English') and repeats the requirement to use the L2.
At lines 49 and 51, Susanna pushes the teasing further by smilingly claiming that the girls cannot understand when Sakari 'tries' to speak English. Such a statement can be heard as an insult, and indeed both Inka (line 52) and the teacher (line 55) treat the teasing as having gone too far.
Inka's forceful and overlapping rejection of the claim does not prevent Susanna from bringing her turn into completion with a categorisation of Sakari's talk (line 54) once in the clear. The categorisation is provided as the reason behind the unintelligibility of Sakari's language, namely that it is a 'mix' of different languages. The mix contains not only languages that are familiar in the local bilingual municipality (Swedish and Finnish) and the educational context (English), but also languages that are offered as an extreme case formulation (Pomerantz, 1986) ('every other Muslim language'). 2 This way the categorisation also constructs a hierarchy of languages between those that are named and those that are provided by way of an ethnicised description.
Such a description is treated as problematic by Susanna herself, who produces it hedgingly and with a smiley voice, and Inka, whose responding high-pitched shriek (line 59) treats the final category as sensitive and bordering on insult.
Susanna's second, more explicit tease puts the teacher again in the position where managing the acceptability of jocular teasing and the border between it and bullying is a relevant concern. The teacher handles the situation in a way that balances between this and the task of maintaining and reinforcing the L2-only norm. By producing an emotionally laden ('aww') rejection of Susanna's problem with Sakari's English' (line 55), she conveys that Susanna's criticism is unfounded and out of place. On the other hand, the teacher's proposal for Susanna to understand 'only the English bits' (line 60) in Sakari's 'mix' of languages can be heard not so much as a request to manage with whatever partial understanding Susanna can assemble from Sakari's talk, but more like a permission to 'ignore' what meaning Sakari draws on from other languages. Therefore, the turn reinforces a pedagogy that conceptualises multilingual turns-at-talk by way of codeswitching in the sense that the 'meaning' established by resources of one language can be 'picked out' and made sense of. It also valorises languages differently as resources for constructing meaning. Such a pedagogical stance is further underlined by the follow-up directive to Sakari, which pleads him to 'try to remember' the words, even by looking them up in a dictionary if need be (lines 62-63, 65). Thus, by orienting to L2 use as an effort that nevertheless needs to be made (such as by picking up and using the word 'fault'), the teacher is treating translanguaging as a problematic phenomenon.
As the teacher is already moving away from the group, Inka does a caricaturised performance of a person using a basic English expression (lines 64, 66). She ties this to the notion of a 'muslim language' by way of providing a typical Turkish male name (Erdi), which can be seen as a continuation to Susanna's ethnicised and racialized joke that evokes the Other. In this case, the exoticized and caricaturised Other and Sakari are grouped together since Inka claims that Sakari's language practices belong to unnamed (and perhaps even non-legitimate) languages labelled as Muslim languages.

DISCUSSION
In this chapter, we have examined situations where hybrid language practices become in normative conflict with locally established language policies and pedagogies in a bilingual classroom. By zooming in on how one student's language use both subverts such norms that regulate language choice, and how his talk is negatively treated as a 'mix of languages', we have shown an educational context and some interactional situations in which translanguaging does not necessarily constitute a desirable form of communication in the eyes of participants.
Secondly, our rationale behind relying on the participants' perspective, as displayed in their observable interactional orientations in situ, in identifying what counts as translanguaging has been to suggest new ways to conceptualise the relationship between translanguaging and codeswitching as concepts used in research on multilingualism. We will elaborate on these matters below.
Previous educational translanguaging literature tends to view translanguaging practices and pedagogies in a distinctly positive light, as something that support the agency and meaningmaking of multilingual students by allowing them to use their full linguistic repertoires.
However, in the present research context of English-medium history teaching (CLIL) to Finnishspeaking secondary school students, English is a foreign language that has been normatively assigned by the teacher as the language of communication instead of the local majority language (Finnish). In such a context, language separation, together with teacher practices for encouraging and enforcing students' L2 use can be seen to serve a distinct pedagogical purpose and the 19 rationale of communicative language teaching (see e.g. Richards, 2006). In other words, ensuring that students do make the 'effort' to use L2 (extract 4) embodies a pedagogical stance that one learns L2 by using it meaningful situations, which is at the heart of the bilingual approach that CLIL is. It is against this kind of pedagogical and ideological background that Sakari's translanguaging (extracts 1-2) takes place, as a sequential response to it. We argue that whatever meaning or function his translanguaging takes cannot be satisfactorily established without considering its relation to the local language choice norms and the pedagogical rationale behind those norms. In the context of French immersion in Canada, Ballinger et al. (2017) have recently questioned the pedagogical usefulness of allowing increased use of students' L1 (English), which is also the local majority language. We agree with Ballinger et al. (2017) in that there is a need to consider how best to adapt crosslinguistic pedagogies such as translanguaging to the different contexts of bi-and multilingual education, particularly in settings where there may be a conflict between the goals of translanguaging and the existing pedagogical practices.
Given the sequential context, Sakari's translanguaging can be seen as a means for displaying a certain kind of identity and, perhaps, for showing non-investment in education. In this regard, it is therefore not so much a 'sincere' attempt to draw on all linguistic resources in order to learn subject content than it is an instance of verbal play during a 'boring class' (Lin, 2005). The hybrid language form serves as a resource for alignment and disalignment with the institutional norm concerning language choice in the classroom: it can be used to play a game of 'doing being a good/bad student' (Emura, 2006), which between these students revolves around the notions of language purity and separation. For the teacher, Sakari's translanguaging is thus a 'problem' that might not disappear even if she relaxed the monolingual mindset because there is a potentially endless reservoir of resources for subverting the institutional work that is supposed to take place in classrooms.
Our approach to identify translanguaging by way of relying on participants' observable interactional orientations can also shed new light on the relationship between translanguaging and code-switching. Teasing these concepts apart has proven to be tricky in much of the existing literature; for example, García (2009) and Nikula and Moore (2016) suggest that translanguaging goes beyond code-switching to include phenomena such as translation and morphological 20 derivation whereas Lewis et al. (2012, p. 659) maintain that the "distinction…is ideological". As Nikula and Moore (2016, p. 3) further suggest, in much of recent research one can easily get the impression that translanguaging is treated as a "re-branding of code-switching", in the sense that data extracts of translanguaging can in terms of their conversational organisation look very much like instances of code-switching. We propose that one analytical way to deal with the apparent confusion is to investigate how participants orient to the normativity of language use when they code-switch or translate language items. The present analysis suggests that both interactional activities can serve the maintenance of language separation: cases in point are Inka's codedoubling (He, 2013) in L2 of an utterance that she first produces in a shared L1 (extract 2) and the teacher's translation of Sakari's L1 word as part of his turn into the L2 (extract 4). In these kinds of cases, an utterance that is either partly or fully produced with the help of L1 resources is treated to be in need of translation to the L2; the participants orient to the L1 and L2 as separate codes. In the present data, such language practices are in stark contrast with Sakari's talk that displays an orientation to L2 translation as unnecessary. Besides these largely implicit participant orientations to what the medium of classroom interaction is and what it should be (see also Bonacina and Gafaranga, 2011), normativities could fruitfully be investigated in situations where participants produce categorisations of each other as language users, for example by talking about 'mixing languages' and language resources, and thereby convey that such resources should not belong together (extract 4).
In describing an educational setting in which participants treat languages as bounded entities in which to invest, and from which departures can be sanctioned as a 'mix', the observations made in this chapter find resonance in recent sociolinguistic studies that have begun to reconsider and synthetise the plethora of concepts used for multilingual practices and pedagogies in a 'languagised' world (see e.g. Jaspers and Madsen, 2016;Møller, 2016, p. 280). Thus, even if from a sociolinguistic perspective the reification of (second-order) languages (Thibault, 2011) can be problematic, it is evident that there are social situations in which people treat "Languageswith-capital-L" (Jaspers and Madsen 2016, pp. 236-237) as distinguishable and socially consequential entities. In this chapter, we have attempted to argue that this observation should be taken seriously in researching both translanguaging practices and pedagogies. The promise of a careful microanalysis of interaction is that it can shed light on participants' local definitions for 21 and nuances in translanguaging, its overlap and frictions with notions such as code-switching, and the social consequences that language alternation has for participants in educational interaction.