Writing in L2: Norwegian Students’ Use of Translanguaging at the Draft Stage

This study explores Norwegian students’ use of translanguaging at the draft stage of writing in English (L2). 78 drafts were written under a translanguaging writing condition, that is when the students were prompted to alternate languages during the draft stage of writing as a scaffolding technique before producing a final product in English. Thirty-seven drafts containing language alternation were analyzed using the proposed integrated framework of translanguaging and written code-switching. The analysis showed that the students’ uses of translanguaging range from a strategic juxtaposition of linguistic elements fulfilling certain socio-pragmatic functions, to highly flexible language alternation with the aim of generating content or experimenting with one’s linguistic repertoire. The findings reported in this paper point to the fact that the integrated framework of translanguaging, pragmatic code-switching, and language mixing may be necessary to account for the students’ diverse and complex use of translanguaging in writing as it allows for a more differentiated approach to study the written form(s) of translanguaging. The present study aims to further translanguaging as an alternative pedagogy to contest the English-only approach to writing instruction in Norway because it provides the students with an opportunity to demonstrate a unique set of skills that remains obscured in the traditional monoglossic writing context.


Introduction
This paper reports on a study of Norwegian first-year upper secondary school students' use of translanguaging at the draft stage of writing an essay in English, their L2 1 . I collected 78 drafts written under a translanguaging writing condition, that is when the students were given a prompt (see Appendix) to alternate languages during the draft stage of writing as a scaffolding technique before producing a final product in English. This definition of translanguaging is based on its original conceptualization as a pedagogical language alternation practice used to promote students' literacy development in one or more languages (Williams, 1994(Williams, , 1996. Out of 78 drafts, 37 (47%) contained language alternation and were analyzed using the proposed integrated framework of translanguaging and written code-switching. Different ways of conceptualizing code-switching have led to the divergence of the code-switching research paradigm from the translanguaging paradigm. While acknowledging a certain degree of overlap between code-switching and translanguaging, many scholars are concerned with elucidating the differences between the two constructs (García, 2009;García & Li Wei, 2014;Slembrouck & Rosiers, 2017). More specifically, researchers often stress that code-switching is first and foremost a linguistic term invoking its structuralist heritage, whereas translanguaging is grounded in a broader sociolinguistic and ecological approach, not least as a pedagogy of language and an antidote to the monolingual norm and ideology in language classrooms and beyond (García & Kano, 2014;García & Wei, 2014;Jonsson, 2017;Lewis, Jones & Baker, 2012;Paulsrud, Rosén, Straszer, & Wedin, 2017).
So far, the pedagogical potential of the written form of translanguaging in mainstream multilingual classrooms has received scant attention in the literature (Canagarajah, 2011;Velasco & García, 2014). Among recent studies that adopt a translanguaging perspective on bilingual writing development is Velasco & García's (2014) investigation of young bilingual writers' use of translanguaging as a writing strategy in the planning, drafting, and production stages of writing. The authors qualitatively examined five writing samples produced by the 1 English is in the process of becoming a second language in Norway (Rindal & Piercy, 2013). However, English may be a foreign language (or L3) for some of the participants who did not report their proficiency in Norwegian (See Table 1). For simplicity, I use the "L2" abbreviation to refer to both FL and L2. K-4th grade Spanish-English and Korean-English bilingual students. The analysis showed that the young writers used translanguaging to organize ideas related to the topic and to engage the reader. In addition, translanguaging was employed as a vocabulary learning strategy (text annotations) and a discourse feature (internal speech). Canagarajah (2011) adopted an ethnographic, predominantly process-oriented approach to study a Saudi-Arabian graduate students' use of translanguaging as she was developing her academic writing skills in English. Canagarajah introduced the term codemeshing to refer to "the realization of translanguaging in texts" (p. 403) and reported on a positive effect of codemeshing on the student's writing process and products in terms of raised metalinguistic awareness, increased creativity, and ability to make conscious attempts to improve her writing. Other studies that address translanguaging in writing (García & Kano, 2014;Turnbull, 2019) involved participants engaging in oral translanguaging in the planning/preparation stage of writing in a target language, that is when participants could discuss the topic and/or plan their essays using L1 or a combination of languages before writing a text in English. However, to the best of my knowledge, no previous study offers a focused investigation of the discourse-related aspect of written translanguaging in L2 writing.
In this study, I propose to synthesize the translanguaging and code-switching research paradigms to examine the students' use of translanguaging in writing. This paper seeks to fill a knowledge gap in research on written translanguaging by arguing for the utility of the sociolinguistic perspective in research on code-switching in answering the call for more empirical research on translanguaging as a writing strategy and pedagogy in multilingual language classrooms. The main aim of the paper is (i) to lay the groundwork for an integrated framework of translanguaging and written code-switching and (ii) to demonstrate the utility of the integrated approach in enhancing our understanding of how Norwegian L2 learners may employ their linguistic resources at the draft stage of writing in English.
As stated earlier, this investigation focuses on the analysis of language alternation patters in the student's drafts. Drawing on the work by Alvarez-Caccamo (1998), I employ the term "language alternation" in a broad sense to denote the alternating use of two or more recognizable linguistic varieties that may or may not carry discourse meaning and thus function as an intended contextualisation strategy (Auer, 1999). In other words, the term 2020, 8 (2), 1-26 4 keeps "the notions of communicative code and linguistic variety separate" (Alvarez-Caccamo, 1998, p. 38) allowing for the possibility to switch the communicative intend (code) with or without a switch to another language (dialect, register, style, prosodic register). The opposite is also possible, i.e. the change in a linguistic form may not necessarily signify an intended discourse function.
I begin with a theoretical exploration of the origins of translanguaging and codeswitching as well as the relationship between the two concepts. The remaining part of the paper is concerned with the empirical investigation of the students' translingual writing. The analysis is based on the integrated framework of translanguaging and written code-switching proposed in the theory section of the paper. According to this framework, code-switching that serves as a contextualization cue (Auer, 1999) may constitute one manifestation of translingual practice (Canagarajah, 2013) under the umbrella term of translanguaging since the latter encompasses "a variety of discursive and pedagogical practices" (Cenoz, 2017) and can be understood as "the general communicative competence of multilinguals" (Canagarajah, 2011, p. 403).

Integrated Framework
The term translanguaging was conceived in an educational context by the Welsh educator Cen Williams (1994Williams ( , 1996 to denote a planned teacher-initiated pedagogical activity based on a purposeful concurrent use of two languages within a lesson or task. For Williams, translanguaging is a valuable strategy that should be encouraged and utilized in bilingual language classrooms to promote children's development in both languages. Since then the term has been further developed by a number of scholars (Creese & Blackledge, 2010;García, 2009;García & Kano, 2014;García & Li Wei, 2014;García & Otheguy, 2014;Jonsson, 2017;Lewis et al., 2012;Li Wei, 2018;Otheguy et al., 2015;Paulsrud et al., 2017) to refer to a wide range of complex language behavior in multilingual speakers in and outside of the educational context. Consequently, translanguaging evolved into an open-ended construct (Slembrouck & Rosiers (2017), which may accommodate diverse bilingual practices, including codeswitching (García, 2009).
Despite observable commonalities between pedagogical translanguaging and classroom code-switching (Lewis et al., 2012), code-switching studies have been criticized for supporting the dual competence perspective on bilingual proficiency since "codes" imply the separation of linguistic systems (García & Otheguy, 2014;Otheguy et al., 2015). In the next section, I show that in the sociolinguistic approach to code-switching the understanding of "codes" as separate linguistic systems has been reexamined by a number of scholars.
As a response to the critique of code-switching for supposedly promoting the dual competence model, MacSwan (2017) points to ample literature on bilingualism and codeswitching 2 that adopts the holistic perspective on bilingualism to account for high levels of complexity, systematicity, and creativity involved in language alternation. A comprehensive account of the dual versus so-called "unitary" view of the architecture of bilinguals' language system(s) is beyond the scope of the paper (for a debate of the issue in question see MacSwan (2017) and Otheguy, García & Reid (2018). Instead, building on the commonalities between the two research traditions, I propose to conceptualize pragmatic code-switching as one possible realization of translanguaging since both are rooted in the holistic view on bilingualism and, importantly, endorse and promote multilinguals' diverse language practices.
The analysis of the translingual writing presented in this paper shows that the students choose to employ their linguistic repertoire in a flexible fashion, i.e. as a continuum ranging from intended juxtaposition of linguistic codes to achieve a certain discourse effect (pragmatic code-switching) to hybrid language practices devoid of such juxtaposition. Therefore, an integrated framework offers a more differentiated, data-driven approach to the analysis of the students' use of their linguistic repertoire at the draft stage of writing in a target language.
In the remaining part of this section, I lay out the premises of the integrated framework of pragmatic code-switching and translanguaging in light of different interpretations of "codes".
Code-switching has been studied since the 1950s. Jakobson, Fant & Halle, (1952), Fano (1950), and Fries & Pike (1949) first introduced the notion of "switching codes" as a scientific phenomenon to refer to the coexistence of phonemic systems in the mind of a monolingual or bilingual speaker. A later definition of code-switching as an alternation 2 See MacSwan (2017) for the overview of studies. between languages (dialects, styles, etc.) stems from early studies on bilingualism (Weinreich, 1953;Haugen, 1956;Mackey, 1962) that focus on a systematic description of bilingual speech with the purpose of mapping linguistic choices of bilinguals. In order to predict when, how, and why the switch to another language would occur at least two distinct "codes" had to be identified. A "code" became synonymous with a language (and/or a language variety) and thus was reified by the analyst in actual speech data.
A different, interactional, approach in research on language alternation in speech was introduced and developed by Gumperz (1957Gumperz ( , 1964 who looked into the social functions of code-switching, i.e. code-switching was studied in relation to the enactment, maintenance, and deconstruction of communicative roles and social identities. The interactional approach to code-switching resulted in the framework of socio-pragmatic functions of oral code-switching (McClure, 1981;Gumperz, 1982;Poplack, 1980;Valdés-Fallis, 1976). A more detailed account of this framework and its application to the analysis of written code-switching is given in the "Analytical framework" section.
The interactional approach to code-switching marked an important shift from the structural to the interpretive perspective with the starting point in meaning rather than structure. The latter requires linguists to regard a speaker's and/or listener's interpretation of language alternation in conversation as a point of departure. In this connection, Alvarez-Cáccamo (1998) challenges the way code-switching research often presupposes the existence of at least two distinct codes in bilingual conversation, each used to fulfill a particular pragmatic function whereas numerous examples of bilingual data contain language alternation that is perceived by the participants and linguists as virtually functionless. Instances of socalled "unmarked" or "free switching" were labeled as such and have not been explicitly addressed in the literature. With this in mind, Auer (1998) points out that a more appropriate approach to the analysis of data containing language contact entails establishing function rather than codes, i.e. functionally meaningful transitions should be labeled "code-switching", whereas those not exhibiting any clear pragmatic intention on the part of the speaker, may be referred to as instances of "mixed code". When describing "mixed code", Auer (1998) emphasizes a frequent and seamless alternation between languages or language varieties that "does not carry meaning qua language choice for the bilingual participants" and often does not "receive neither discourse-nor participant-related interpretations" (p. 16, emphasis in the original). In his later work, Auer (1999Auer ( , 2014 employs the term "language mixing" to refer to the aforementioned type of language alternation. Auer (2014) maintains that "language mixing" should be distinguished from code-switching since, unlike code-switching, it lacks the discourse-functional aspect. The notion of "language mixing" arguably echoes some of the translanguaging definitions in the literature, for instance, "as bilingualism without diglossic functional separation" (García, 2007, p. xii), and later as a dynamic communicative practice where "languages are no longer assigned separate territories or even separate functions, but they may co-exist in the same space" (García, 2009, pp. 78-79).
Given what has been discussed so far, alternating use of languages by bilinguals appears to be much less homogeneous than a traditional understanding of code-switching entails. The last 40 years of code-switching research from the sociolinguistic perspective have shown that language alternation has properties of a continuum spanning from intentional pragmatic contrast of two codes (pragmatic code-switching) to a smooth and uninterrupted flow of language alternations in which none of the languages assumes any distinct discourse function ("language mixing") 3 .
Consequently, a more differentiated analytical approach is needed to grasp the diverse nature of language alternations in the students' drafts. In this approach, the designation codeswitching is reserved for the type of language alternation that carries a particular sociopragmatic function (e.g. clarification, lexical needs, quotes, parenthetical comments, etc.), while translanguaging is an overarching concept that includes all kinds of translingual practice, i.e. pragmatic code-switching, language mixing, translation, etc.
In the sections that follow, the integrated framework of translanguaging, pragmatic code-switching, and language mixing is applied to the analysis and interpretation of the students' translingual writing.

Method Participants and data
The original corpus of 78 drafts was obtained as part of a larger study of the effect of three writing instructions on the quality of essays written by Norwegian first-year upper secondary school students (Prilutskaya, Knoph & Hanssen, 2020). A total of 288 first-year upper secondary school students (age 15-16) from two mainstream schools in Norway received a task to write a fantasy narrative essay in English. The participants were assigned randomly (on a class-by-class basis) to one of three writing instruction groups: English-only, translation, and translanguaging. The students in the English-only group wrote their drafts and final essays in English; the students in the translation group were asked to write the drafts in Norwegian and then translate them into English. Finally, the students in the translanguaging group were prompted to alternate languages in their drafts before writing a final product in English. The language(s) of instruction including the writing prompts varied among the groups. English was used for the English-only condition, Norwegian for the translation condition, and a combination of Norwegian and English for the translanguaging condition (see Appendix). The drafts collected from the translanguaging group are the focus of the analysis in the present study. In addition to writing the essays, the participants in all three groups filled out an on-line questionnaire where they reported their language background and gender. The responses from the translanguaging group are summarized in Table 1:

Analytical framework: from oral to written code-switching
From the pragmatic perspective, oral code-switching was broadly divided into situational and metaphorical. The former is directly linked to the interactional context, for instance, when a speaker switches to another language to adjust to the needs of the interlocutor who might not speak a particular language. A metaphorical type of switches is a more complex phenomenon that is known to fulfill a pragmatic, emphatic, and creative function rather than being a direct response to a communicative situation. Language alternation patterns examined in the present study are of the metaphorical type since in written discourse the communicative context is established in advance, and remains uniform.
Most proposed taxonomies of metaphorical or socio-pragmatic functions of oral codeswitching include quotation, emphasis or clarification, triggered switches, stylistic switches, parenthetical comments, linguistic routines/idiomatic expressions, and lexical need switches (Valdés-Fallis, 1976;Poplack, 1980;McClure 1981;Gumperz, 1982;Zentella, 1997). Edelsky (1986) was the first researcher who consistently applied the functional categories for oral code-switching to code-switching patterns in bilingual writing. In her comprehensive study of writing by elementary students enrolled in an English-Spanish bilingual program, Edelsky found that the children code-switched for clarification, direct quotation, ethnic group identity, emphasis, lexical variety, and because they learned a word or a phrase in that language. More recent studies of socio-pragmatic functions of written code-switching in adult bilingual writing showed that bilingual writers utilized code-switching in a similar manner (Montes-Alcalá, 2005Losey, K. M., 2009).
As Sebba (2012) rightly points out, a number of researchers draw on the theories and analytical frameworks designed with oral code-switching in mind in their study of multilingual texts in written discourse. Use of written data to confirm hypotheses based on spoken data (mainly conversational code-switching) remains problematic since some written genres lack the interactional aspect of conversation. However, Sebba (2012) admits that Gumperz's (1982) fundamental contributions to the field of code-switching such as the notion of contextualization cues, a distinction between situational and metaphorical code-switching, and a taxonomy of discourse functions of code-switching, "are potentially applicable, and have been applied at some stage, to written language alternation" (p. 99). Further, similarly to the conceptualization of pragmatic code-switching given in the theoretical section of this paper, Sebba proposes to apply the term "code-switching" to a change in language that is meaningful and functioning as a contextualization cue for the reader, while other instances of language change should be classified as a different phenomenon. In this paper, I use Auer's notion of language mixing to describe such phenomena.
It follows then that previous empirical research on oral and written code-switching has provided the present study with a basic taxonomy of socio-pragmatic functions that is flexible, yet by no means exhaustive. It represents some of the most common functions of code-switching and thus offers a suitable starting point for examining language alternation in writing in the context that has not been explored before. The analytical procedures and the results obtained from them are described in the next section.

Analysis and Results
To begin with, I grouped the drafts written under the translanguaging condition according to the students' choice of language(s). Here the term "translanguaging" is narrowed down to refer to language alternation within a single draft to separate it from translation. Table 2 shows the students' choices of language(s) in the drafts: cases, it was difficult to determine where one case of language mixing ends and another one begins. In order to avoid a situation when the proportion of language mixing would be artificially inflated due to the density of alternations within one text, I treated these dense language alternations as evidence of a single case of translanguaging, namely, language mixing.
The distribution of the categories (both raw frequencies and %) that were identified in the drafts is illustrated in Figure 1:

Distribution of Categories
In the following subsections, I provide a brief description of the outlined categories followed by the examples from the corpus.

Emphasis
This type stands for around 4% of the total number of switches. As the name suggests, these switches are introduced to emphasize a certain idea by repeating the same or similar message in two languages. Specifically, in Example 1, the author emphasizes their fear of being alone in a big city by adding a sentence in Norwegian that further explains why the author feels intimidated by Tokyo. In Example 2, the author reinforces the image of spending three

Direct speech, quotation
This is a productive category that accounts for 27% of the switches. There is a variety of languages associated with the switches for quotation, such as, English to Spanish (Example 5), Swedish to Norwegian (Example 6), French to Norwegian (Example 7), and others.
Interestingly, the students make use of the opportunity to alternate not only between the languages but also between the standard and local varieties of Norwegian as in Example 8 where the author alternates between standard Norwegian and the local dialect.

Linguistic routines
Around 9% of the total number of switches falls under this category. Linguistic routines imply use of a range of brand names, words, and phrases that are routinely used by Norwegian speakers in the original language, which is primarily English. Compared to the lexical need switches for stylistic purposed described above, linguistic routines switches are employed for convenience and thus lack a creative aspect of stylistic switches.
(27) "Hvorfor bruker du ikke bare Google translate?" ["Why don't you just use Google translate?"] (28) Plutselig kommer vi på at vi må bestille flybillettene hjem, men når jeg skal betale, kommer det "error". [Suddenly we realize that we have to book our flight tickets home, but as I am about to pay, there is an "error"] (29) Etter at jeg var ferdig gikk jeg ut og brukte google maps får finne en shopping mall får å kjøpe noen klaer […]. [After I was done I went out and used google maps to find a shopping mall to buy some clothes]

Parenthetical comments
This least productive category includes three cases and amounts to around 2% of the total number of switches. The first two examples (30 and 31) come from the same essay. Here the author inserts additional information in Norwegian to clarify and expend on the message expressed in the English sentence. The parenthetical comment in (32) serves a different purpose, i.e. it functions as a mental note to remind the author to add more information to the text at a later point of time.
(30) Carpets are usually soft and nice to walk on, which is why you find them in so many places, one of those places may be Japan (ae veit ikke, har ikke vaert der). [I don't know, I haven't been there] (31) Seeing how I can buy practically anything that I want (som e lovlig i hvert fall).

Language mixing
This is the most prolific category in the corpus with around 32% of coverage. Language mixing includes cases where the alternation between languages seems to be unmotivated in terms of the socio-pragmatic functions mentioned earlier. The patterns of the switches in this category vary substantially, i.e. from rather dense multiple alternations within a single phrase or a sentence (Examples 33-38) to more orderly structured patterns where the author sticks to one language per paragraph, such as in Example 41 with one paragraph written in a nonstandard French and the next one in Norwegian. Mixing more than two languages is another distinct characteristic feature of language mixing, which is illustrated in Examples 39 and 40.
In Example 40, the author shuttles between German (in italics), Norwegian, and English (in bold), whereas Example 40 contains English, Norwegian, Spanish, and Greek. In the following pages I discuss the empirical findings in light of the proposed integrated framework.

Discussion and concluding remarks
As was pointed out in the introduction, the aim of this paper was to develop an integrated framework of translanguaging and code-switching in order to explore the students' use of translanguaging at the draft stage of writing a text in English. Drawing on the theoretical discussion of this paper, I have proposed a more nuanced approach to the conceptualization of code-switching as an intended contextualisation strategy (Auer, 1999) that fulfills one or more of the socio-pragmatic functions described in the adopted analytical framework. The data analysis showed that the students utilized code-switching for reasons similar to those identified in previous research, i.e. they switched for a quote or direct speech, for emphasis, for stylistic purposes, lexical needs, and for linguistic routines. A large proportion of the switches for quotation found in the present corpus (27%) is not surprising as switching to the original language of speech or thought helps achieve authenticity in written texts.
Furthermore, when seen through the lens of Bakhtin's notion of heteroglossia, the insertion of another language/code for emphasis and, in particular, for direct speech and quotation adds a new dimension to the discussion of the students' use of translanguaging as a discourse strategy. The result of "incorporating intratextual discourses into the narrator's text" (Tjupa, 2009, p. 125) is the effect of multiple voices belonging to the narrator and characters and coexisting within the fabric of a single text. The use of a "hybrid construction" (Bakhtin, 1981) or "utterance within utterance" (Vološinov, 1929(Vološinov, /1973 allows the narrator to distance themselves from the characters by giving the characters their own voice. As Tjupa (2009) points out, "the direct speech of a character often serves to express that character's linguistic view of the world, which can differ to a greater or lesser extent from the view of the world on which the narration is based" (p.126). The effect of multiple voices (heteroglossia) is not necessarily due to the use of different national languages, i.e. in some cases (see Example 8) the students turn to their local Norwegian dialect with the same purpose of creating another voice for themselves. It is possible that this voice is a better match to their sense of identity than the standard variety of Norwegian employed in the main body of the text. Strengthening the author's identity through the use of dialects alongside the standard variety may benefit Norwegian learners who reported lacking confidence in their English writing skills (Horverak, 2015).
Next, lexical need switches, emphasis, and parenthetical comments as functional categories of code-switching in writing show the students' ability to use their linguistic repertoire both as a scaffolding strategy (word retrieval or creating a mental note to remind the writer to edit a piece of text at a later point in time) and a discourse strategy to enhance the communicative ability of their writing (emphasizing an idea or using other languages for stylistic purposes).
The categories discussed so far have a clear functional profile, and thus are indicative of pragmatic code-switching. However, the category of language mixing manifests a different phenomenon since it was not associated with any particular socio-pragmatic function.
Interestingly, language mixing was the most productive category, which points to the fact that the students seem to have good command of both pragmatic code-switching and language mixing. A feasible explanation of the ample use of language mixing in the drafts was provided by the students in their answers to the on-line survey about their attitudes towards the translanguaging writing mode . The students explained that they alternated languages in their drafts as they tried to generate as many ideas as possible irrespective of the language of thought. Perhaps, allowing the thoughts to be expressed in whatever language available at a given moment helped reduce cognitive load and facilitate access to content schemas in long-term memory. The content schemas deliver information about the topic and content organization and thus are important in terms of coping with cognitive demands of writing in a target language. The students also noted that they incorporated other languages into their drafts to enhance stylistic and rhetorical features of the texts. According to the students' responses, being able to integrate multiple languages in their writing helped engage the reader and convey desired atmosphere. However, not every case of language mixing indicated its use as a writing strategy as three of the participants reported mixing languages due to the nature of the instruction and writing prompt as well as for the sake of experimenting with languages when given a chance to do so. With this in mind, Examples 33 and 39 may represent the task-induced artifacts rather than cases of naturally occurring language alternation since the switches often do not comply with syntactically definable constituent boundaries. The latter stands in contrast to a more commonly observed "constituent-by-constituent" pattern of switch in bilingual as well as monolingual speech production and processing (Azuma, 1996). Despite being possible artifacts of the elicitation process, these examples of language mixing may attest to the students' willingness to engage in translanguaging for the purposes of exploration, which they expressed in the survey. This demonstrates the students' openness to experiment with crosslingual writing practices in classroom settings, which in turn can be a potentially effective tool for creating engaging and stimulating learning activities that may facilitate students' ability to employ their linguistic repertoire in more innovative and learner-oriented ways.
Finally, the category of triggered switches represents a curious phenomenon, i.e. on the one hand, the switches of this nature mark clear boundaries between the languages and in this regard are unlike the dense and multiple alternations of language mixing. On the other hand, the switches are not driven by any specific socio-pragmatic function. Montes-Alcalá (2000) suggested the following explanation of triggered switches that she identified in her data: It could well be that the sentence is already constructed in an abstract way in the mind of the speaker, before he/she knows in what language it will come out, so if a word is going to be switched later on, that could trigger a switch […].
(p. 207) The cognitive processes underlying triggered switches are likely to occur at the stage of translation that in Flower & Hayes's (1980) cognitive model of writing refers to the process of converting thoughts/ideas into symbols (written language). In this respect, triggered switches are more analogous to language mixing than pragmatic code-switching as they manifest the flexible and interconnected use of languages. However, since one can identify a trigger word or a phrase, triggered switches may occupy the middle ground between intentional switching and free language alternation.
Overall, this study confirms that the students utilize their linguistic resources in diverse ways. Forty-one students (53%) chose to stick to one language per draft (either English or Norwegian), even when they were prompted to draw on their background languages. The participants' monolingual preferences are compatible with the monolingually oriented context of teaching English in Norway, i.e. in addition to the English-only format of nationally administered exams, the purpose of introducing other languages to ELT is limited to fostering metalinguistic awareness between English and L1 (The Norwegian Directorate for Education and Training, 2013). Nonetheless, 37 students (47%) demonstrated that they are in possession of an array of diverse translingual writing strategies that range from a skillful and strategic use of contrasting linguistic elements fulfilling certain socio-pragmatic functions, to highly flexible language alternations that may occur for the purposes of generating content or experimenting with one's linguistic repertoire. The findings reported in this paper point to the fact that the integrated framework of translanguaging, pragmatic code-switching, and language mixing may be necessary to account for the students' diverse and complex use of translanguaging in writing as it allows for a more differentiated approach to study the written form(s) of translanguaging. The present study aims to further translanguaging as an alternative pedagogy to contest the English-only approach to writing instruction in Norway because it provides the students with an opportunity to demonstrate a unique set of skills that remains obscured in the traditional monoglossic writing context.