Translanguaging in the development of EFL learners’ foreign language skills in Turkish context

ABSTRACT

This study investigates the potential impact of translanguaging pedagogy on EFL learners’ four language skills, as well as their perceptions towards its in-class implementation as a pedagogy. The study was conducted for 10 weeks with 60 pre-intermediate and 60 upper-intermediate students at a high school in the north of Turkey. In each group, half of the students were assigned as the experimental group, while the other half as the control group. The experimental groups were subject to a 10- week (40 h of teaching) of translanguaging pedagogy, whereas one of the control groups was taught through grammar-translation methods, while the other, through communicative language approach. In the study, the quasi-experimental mixed methods design was employed. The analysis of a paired-sample t-test and one-way analysis of ANCOVA indicated that translanguaging played an effective role in improving students’ four English language skills. Semi-structured interviews revealed that the students reported constructive, cognitive, interactive, and affective benefits through translanguaging pedagogy since they were able to draw upon all the linguistic resources for meaning-making and negotiation, a sense of comfort, and a sense of motivation to use English. We offer implications for teachers and teacher educators regarding the role of translanguaging in teaching English.

KEYWORDS:

• Translanguaging
• bilingualism
• foreign language learning
• learner perceptions
• EFL learners

1. Introduction

Foreign language teaching has moved away from an approach focusing upon analysing a language (e.g. Grammar Translation Method) to an approach focusing upon using language (e.g. Communicative Language Teaching) (Celce-Murcia 2001). Along with this paradigm shift from analysis to utility, monoglossic ideology (Bakhtin 1981) which refers to ‘proficiency in the two languages according to monolingual norms’ (Garcia 2009, 115) has been advocated, because the classrooms are the only places where the students are exposed to the target language (Littlewood and Yu 2011). This argument has enabled this monolingual principle to become the norm in the classes where foreign languages are taught (Wang 2019). Accordingly, language classrooms around the world continue to show monolingual bias, reinforcing traditional monolingual behaviour (Hu and McKay 2012) dominated by communicative language teaching (CLT henceforth). Here, the teacher encourages language use whereas in Grammar Translation Method (GTM hereafter), the target is language analysis. Thus the use of learners’ first languages has been systematically prevented in English as foreign language contexts, based upon the premise that classroom should reproduce the target language context, so students ultimately learn to communicate in that context (see Macaro 2009).

Nonetheless, language compartmentalization has recently been questioned by Garcia (2009), who argues that translanguaging, the simultaneous use and integration of languages in learning language, is an alternative model as a pedagogy of language teaching in line with bilingualism as an instructional and interactional model particularly at pre-school and K12 contexts (Moore 2018). According to Garcia (2009), the process of using both languages is mutually beneficial for instruction, as it engenders a more dynamic learning environment, allowing the freedom to use stronger (native-Turkish) and weaker (English in this study) language for separate but overlapping pedagogical purposes.

A number of studies report benefits for translanguaging (e.g. Moore and Nikula 2016). It can overcome the drawback mentioned by Inbar-Lourie (2010) that the current pedagogy of language teaching appears to ‘ignore or even suppress bilingual or multilingual options endorsing a predominantly monolingual policy’ (351). However, few studies have explored the impact of translanguaging practices on language learners (e.g. Nikula and Moore 2019), a research gap we address in our study by operationalizing translanguaging as a practical theory of language (Wei 2018), in which language is reconceptualised as multilingual, multi-semiotic and multi-modal resource developed by learners as a holistic multi-competence (Cook and Wei 2016) in line with heteroglossia (Bakhtin 1981). This study is also one of the few to focus on the potential impact of translanguaging in Turkey context (see, e.g., Karabulut 2019) by providing practice-based evidence.

2. Literature review

2.1. Monolingual-oriented teaching approaches

GTM, the dominant method till the mid-twentieth century, focuses upon grammatical rules, translation of texts, memorization of vocabulary and written exercises, aiming at grammatical competence (Richards and Rodgers 2014). Yet, GTM is criticised for not allowing the achievement of fluency and fostering communicative language use (Newson 1998).

CLT, proposed in reaction to GTM, has been the dominant theoretical model in English Language Teaching since the 1970s (Thompson 1996). CLT aims to develop communicative competence (Hymes 1972), developing language skills using functional communication activities (Littlewood 1981). However, communicative competence as an aim has received criticism around two core issues: (a) the linguistic content base of CLT, (b) the pedagogical treatment of linguistic forms in CLT (Celce-Murcia, Dörnyei, and Thurrell 1997).

2.2. Bilingualism and translanguaging

Garcia (2009) reconceptualises bilingualism as a dynamic process metaphorized as the way an all-terrain vehicle, moves by extending and contracting, flexing and stretching, in contrast to a motorcycle with only two balanced wheels. Her conceptualisation of translanguaging for pedagogical purposes was introduced by the Welsh educator, Williams (1996) and others (Baker 2011; Wei 2011). Translanguaging, or trawsieithu, referred to a pedagogical practice in bilingual education which deliberately allowed for the interchangeable use of the languages of input and output. Translanguaging has been described as digesting L1 language input as reading and discussing and processing L2 language output as writing, and thus, going beyond the two languages in a critical and creative process (Baker 2011; Wei 2011).

Lewis, Jones, and Baker (2012) distinguished between the novel concept of ‘translanguaging’, especially in the classroom context, and codeswitching, and translation, which could occur while analysing languages and/or communicating in foreign languages. They pointed out that translanguaging offers novel functions through its promotion of learner engagement in active meaning-making. Garcia (2009) indicates that translanguaging ‘goes beyond what has been termed code-switching … although it includes it, as well as other kinds of bilingual language use and bilingual contact’ (45). Mazak (2017) argues that translanguaging in formal classrooms challenges traditionally held theories of second-language acquisition, and the importance of a rigid separation of languages. In this sense, translanguaging runs counter to the traditional understandings of second-language acquisition, which positions teaching in L2 as the only acceptable method, viewing other approaches as deficient or inadequate (Mitchell 1988).

Translanguaging can have positive effects, allowing for purposeful pedagogical alternation of languages in spoken and written, receptive and productive modes (Baker 2011). There are few investigations of the effect of translanguaging on language skills in TESOL and EFL contexts – speaking (Galante 2020), writing (Sun and Lan 2020), listening (Galante 2020), reading (Vaish and Subhan 2015) – and on students’ perceptions (Rivera and Mazak 2017). We aim to support these arguments with practice-based empirical evidence on the effectiveness of translanguaging pedagogies.

3. Methodology

This mixed methods research investigates the impact of translanguaging upon the development of four language skills– i.e. listening, speaking, reading and writing of learners whose instruction is dominated by grammar-translation instruction and by Communicative Approach, and investigates the following research questions:

1. Does the translanguaging experience have any statistically significant effect upon the students exposed to grammar translation, in terms of their receptive and productive foreign language skills?

2. Does the translanguaging experience have any statistically significant effect upon the students exposed to communicative language teaching in terms of their receptive and productive foreign language skills?

3. What are the impacts of translanguaging pedagogy on learners’ perceptions of language learning?

3.1. Design

The study adopted an embedded quasi-experimental design (Creswell and Plano Clark 2007) to identify the impact of the intervention based on translanguaging, using convenience sampling. At the beginning of the academic year, students were placed in pre-intermediate (14–15 years old) level or upper-intermediate (16–17 years old) level, according to the requirements of the school curriculum. In order to verify that students were at these levels, a sample Preliminary English Test (KET; Carne, Hashemi, and Thomas 1996) was given to the pre-intermediate classes, and a sample Cambridge English (FCE; Ireland 2000) to the upper-intermediate ones. Four intact classes participated, two groups each at the pre-intermediate and upper-intermediate levels. At each level, classes were randomly selected for the experimental or the control condition. The two control groups received traditional EFL instruction (i.e. one, grammar translation, and one, communicative approach), and the two experimental groups, translanguaging instruction-based on Garcia’s (2009) model. Our students were aware that L1 was purposefully integrated into teaching, but we did not negotiate the concept of translanguaging. All four classes were taught by the first author (see Appendix A for the percentages of the use of languages).

3.2. Context and participants

The study was conducted with 60 pre-intermediate and 60 upper-intermediate students in the 2018–2019 fall term in a high-school in the north of Turkey. The students were in at least the 97th percentile in a high-stake test, and placed into prestigious high schools. 10 students from each class of 30 volunteered to be interviewed. The pre-intermediate students (14–15 years old) had been exposed to grammar-translation method for 2 years, while those (16–17 years old) at the upper-intermediate level had received 2 years of instruction-based on communicative language approach. Being native speaker of Turkish, the teacher who delivers these classes has a good command of English, French, German, and Spanish, with 17 years of English teaching experience with training on bilingualism and translanguaging during his doctoral studies. The teacher used Turkish and English during the translanguaging process because the students do not speak any other languages.

3.3. Instruments

3.3.1. Skills test (reading–writing–listening–speaking)

We prepared the skills tests according to the curriculum’s A2 level objectives for pre-intermediate students and B2 level objectives for upper-intermediate students. The test was piloted with another group of students. Mean of item difficulty index of the test was 0.61; mean of item distinguishing index was 0.80 and Cronbach Alpha internal consistency coefficient of test items was 0.98. The writing scores were evaluated with a rubric including overall organization and coherence, vocabulary range, grammatical accuracy, content, format & length, and the speaking skill was assessed based on the following criteria: pronunciation, fluency and coherence, lexical resource, grammatical range and accuracy. The same skills test was used as a pre- and post-test for all groups and evaluated by two English teachers independently and the interrater reliability analysis (Pearson correlation) between the pre- and post-tests was found to be (r = 0.87; r = 0.90) for speaking and (r = 0.75; r = 0.79) for writing respectively, which proved a relatively high agreement between the two teachers’ scores.

3.3.2. Attributional interview questions

The interview includes five questions, which elicited the students’ perceptions toward the use of translanguaging. 20 volunteer students were interviewed after the post-tests.10 were chosen from each group to ensure the widest possible range of post-test scores, in order to explore the potential multiple factors.

3.4. Procedure

Initially, the pre-test means of both experimental and control groups at each level were found to be equal in paired samples t-test. After being informed about the intervention, all were eager to experiment with the new instruction method. We later began to implement the translanguaging instruction lasting for 10 weeks - 4 h of teaching per week.

In control groups, the students received instruction based on GTM, offering a continuously translated instruction into Turkish, whereas those in the communicative group received English-medium instruction only. In the experimental groups, on the other hand, the input and output were deliberately switched in accordance with translanguaging principles (Garcia 2009), including systematic, equal distribution of both languages for the following activities: collaborative dialogues, reading texts, listening/visual resources, project learning, research, preview–view–review, comparing texts, and word walls (Celic and Seltzer 2001).

The instructional bilingual practices for each skill included: reading (using bilingual reading material, accessing and building background knowledge, allowing mutual assistance as language partners, building four-box graphic organizer & Frayer model), listening (using bilingual characters for dialogues and stories, listening stations in L1, language structures checklist, making small group instruction, allowing mutual assistance as language partners), writing (note-taking, pre-writing and completing graphic organizers using all language resources, allowing mutual assistance as language partners, making connections), speaking (accessing and building background knowledge, audio recording ideas, allowing mutual assistance as language partners, small group work, individual conference). Also utilized were bilingual resources as bilingual dictionaries, picture dictionaries and translation apps and websites, and word walls including frequently used terms in Turkish and English, with visuals and examples (see Appendix B for a sample of translanguaging practices). Instructional practices in GTM group were translation of a literary passage, reading comprehension questions, antonyms/synonyms, cognates, deductive application of rule, filling in the blanks, memorization, using words in sentences, and composition. The activities in CLT classroom were role plays, interviews, discussions, information gap activities, language games, language learning simulations, problem solving tasks, quizzes, and surveys. The students were unaware of the research questions throughout the intervention.

3.5. Data analysis

3.5.1. Quantitative

Our data indicated a normal distribution, linearity and homogeneity of variances, typical assumptions of parametric tests (Thompson 2006). Therefore, we used paired samples t-test. Moreover, Cronbach Alpha internal consistency coefficient of test items in the achievement test was 0.98, i.e. scores were reliable and valid. Having conducted the pre-test and intervention, we applied the post-test to all groups, and the difference between groups at each level was compared with a one-way analysis of covariance (ANCOVA) on the dependent variable (post-test) while statistically controlling the covariate (pre-test). Partial eta square was used to measure effect size. Accordingly, 0.01, 0.06, and 0.14 are considered to have a small, moderate, and large effects, respectively (Green, Salkind, and Akey 1997).

3.5.2. Qualitative

We employed grounded theory to analyse our interview data (Charmaz 2006) to explore the potential impact of the translanguaging instruction. We prepared a table containing all 166 responses and coded the data iteratively and simultaneously to induce the preliminary open codes, which were categorised to induce categories, a process of axial coding used to create conceptual families from the summaries, which yielded first 23, then reduced to 15 categories. Finally, we re-assembled these categories into four selective codes as the major themes, namely, constructive, cognitive, interactive, and affective dimensions, which allowed us to induce a theoretical understanding (Saldaña and Omasta 2017) of the potential impact of translanguaging. We also counted the frequency of emerging themes in order to determine the representativeness of each of the 15 categories. Overall inter-coder reliability was 90%. We resolved coding disagreements through negotiation and reached consensus.

We addressed four criteria of trustworthiness (Lincoln and Guba 1985). The first author’s emic stance in the classroom was supported by the etic stance of the second author, which ensured credibility. We also provided thick descriptions of the context, participants, and the instruments to maximize transferability. We addressed dependability and confirmability by detailing the research procedures such as data collection and analysis, which involved several rounds of debriefing and member-checking of the data. This aimed to minimize possible bias of the first author caused by his presence in the classroom context.

4. Findings

The first statistical analysis was conducted to check the extent to which the groups are equal. Table 1 and 2 show these pre-test scores.

Table 1. Pre-intermediate experimental and control group's pre-test scores.

Groups	N	Mean	SD	df	t	p
Experimental	30	74.77	6.94	58	.87	.38
Control	30	76.20	5.69

*p < .05.

Table 1 shows no statistically significant difference between the pre-test mean and standard deviation of the experimental group (Mean = 74.77; SD = 6.94) and the pre-test mean and standard deviation of the control group (Mean = 76.20; SD = 5.69). Accordingly, pre-intermediate experimental group and control group are equal based on the pre-test scores before the intervention.

Table 2 reveals no statistically significant difference between the pre-test mean and standard deviation of the experimental group (Mean = 75.40; SD = 4.21) and the pre-test mean and standard deviation of the control group (Mean = 75.87; SD = 4.98). Thus, after ensuring that both groups are equal based on the pre-test scores, we continue to present our findings for each research question.

Table 2. Upper-intermediate experimental and control group's pre-test scores.

Groups	N	Mean	SD	df	t	p
Experimental	30	75.40	4.21	58	.39	.70
Control	30	75.87	4.98

*p < .05.

4.1. The impact of translanguaging on the students’ receptive and productive foreign language: the grammar translation group

The difference between pre-test and post-test scores between the learners in the translanguaging group and the grammar-translation group was calculated and analysed. The descriptive statistics of pre- and post-test scores are illustrated in Table 3.

Table 3. Pre-intermediate experimental and control group's pre- and post-test achievement.

		Experimental Group					Control Group
Tests		N	M	SD	t	p	N	M	SD	t	p
L2	Pretest	30	74.77	6.94	21.55*	.00	30	76.20	5.69	3.91*	.00
Skills	Posttest	30	90.07	5.34			30	84.63	5.41

*p < .05.

Table 3 displays significant differences between the pre-intermediate experimental group students’ pre-test scores (M = 74.77; SD = 6.94) and post-test scores (M = 90.07; SD = 5.34) in favour of post-test scores (t₍₂₉₎ = 21.55; p < .05). There were significant differences between the control group students’ scores (M = 76.20; SD = 5.69) and post-test scores (M = 84.63; SD = 5.41) in favour of post-test scores (t₍₂₉₎ = 3.91; p < .05). The existing significant value (.000) is smaller than the significance level (.05), i.e. the learners’ performance in the experimental group significantly improved after the translanguaging instruction. Thalheimer and Cook’s (2002) effect size calculator was used to calculate the Cohen’s d value for effect size of this difference between the control and the experimental group. Cohen’s d value was calculated as 1.03, indicating that the translanguaging instruction had a large effect on the scores for four skills.

4.2. The impact of translanguaging on the students’ receptive and productive foreign language: the communicative language teaching group

The descriptive statistics of pre- and post-test scores of the translanguaging and communicative groups are illustrated in Table 4.

Table 4. Upper-intermediate experimental and control group's pre-test and post-test achievement.

		Experimental group					Control group
Tests		N	M	SD	t	P	N	M	SD	t	p
L2	Pretest	30	75.40	4.21	19.10*	.00	30	75.87	4.98	14.30*	.00
Skills	Posttest	30	87.60	5.31			30	84.87	5.22

*p < .05.

Table 4 reveals significant differences between the upper-intermediate experimental group pre-test scores (M = 75.40; SD = 4.21) and their post-test scores (M = 87.60; SD = 5.31) in favour of post-test scores (t₍₂₉₎ = 19.10; p < .05). There were also significant differences between the control group scores (M = 75.87; SD = 4.98) and post-test scores (M = 84.87; SD = 5.22) in favour of post-test scores (t₍₂₉₎ = 14.30; p < .05). The existing significant value (.000) is smaller than the significance level (.05).

Cohen’s d value was calculated for effect size of this difference between the control group and the experimental group using Thalheimer and Cook’s (2002) of effect size calculator. Cohen’s d value was calculated as .53, implying translanguaging pedagogy has a medium effect on students’ achievement scores. Additionally, as the assumptions had not been violated, one-way analysis of covariance (ANCOVA) was conducted. According to Dornyei (2007), in quasi-experimental studies, the use of ANCOVA contributes to the reduction of the initial group differences. Table 5 demonstrates a summary of the results of the one-way ANCOVA and between-subjects’ effects at each level.

Table 5. Analysis of Covariance (ANCOVA), Dependent Variable: Post-test.

Level	Source	Type III Sum of Squares	df	Mean square	F	p	η2
Pre-	Pre-test	1202.938	1	1202.938	144.689	.000	.717
Intermediate	GROUP	618.258	1	618.258	74.364*	.000	.566*
	Error	473.895	57	8.314
	Total	459,921.000	60
Upper-	Pre-test	43.314	1	43.314	1.885	.175	.032
Intermediate	GROUP	387.210	1	387.210	16.853*	.000	.228*
	Error	1309.619	57	22.976
	Total	459,704.000	60

*p < .05.

As Table 5 depicts, after we adjusted for pre-test scores, the results of the one-way ANCOVA on the skills test yielded significant differences between the achievement of the students in the two conditions (i.e. experimental and control) at pre-intermediate level on the post-test, with the experimental group outperforming the control group on the post-test in terms of language skills (F(1, 57) = 74.364; p < 0.05). Partial eta square (η²) of this difference was calculated as .56. This value both indicates a large effect and explains variance in post-test scores of the experimental group students. That is, the translanguaging instruction has a large effect on the experimental group language skills and explains 56% of variance in post-test scores when pre-tests were controlled. Similarly, after controlling the pre-test scores in the upper-intermediate level, the experimental group post-test scores were significantly higher than those of the control group (F(1,57) = 16.853; p < 0.05). Partial eta square (η²) of this difference was calculated as .22. This value both displays a large effect and explains variance in post-test scores of the experimental group. In other words, the translanguaging instruction has a large effect on the experimental group and explains 22% of variance in the post-test scores when the pre-tests were controlled.

4.3. Students’ perceptions towards their language learning through translanguaging

The third question focused on students’ perceptions of translanguaging pedagogy. Below are the themes and subthemes (see Appendix C for the distribution of codes) that emerged on the basis of the interviews with the students in both of the experimental groups who participated in the translanguaging experience.

4.3.1. Constructive dimension

4.3.1.1. Making meaning

Some students reported that translanguaging enabled them to make meaning and understand the topics, as well as convey messages verbally and by writing. Student 11 delineated: ‘The materials were the most important ones. I could understand the topic and discuss with friends’, by increasing understanding, translanguaging seems to ease reading comprehension and communication flow. Similarly, student 19 expressed: ‘I enjoyed Turkish-English texts because I could understand better and increase my vocabulary knowledge too because I always had problems with words. Especially scientific terms are clearer in my mind. Also, building four box graphic organizers helped me a lot to remember and use them’. The quote highlights the positive impact of translanguaging on vocabulary knowledge besides reading comprehension. Student 2 stated: Brainstorming should be this way. I mean in any language. Also writing in any language initially helped me a lot to organize my thoughts and write easily, which was a big trouble for me. This quote shows that translanguaging enabled improvement in writing skills.

4.3.1.2. Promoting autonomous learning

Autonomous learning was a significant benefit. Student 10 expressed ‘I had the autonomy to determine the language to use and I learnt more by using both of the languages’. Likewise, student 6 highlighted this priority: ‘focusing on learning and using Turkish and English contributed to my success’. Student 14 also stressed improvement by concentrating on authentic conversation: ‘real communication occurred because we could speak both Turkish and English. We knew both and we practised it. So we developed both’.

4.3.2. Cognitive dimension

4.3.2.1. Accessing full linguistic repertoire

When focusing on learning a language, some students highlighted that they came to realize their potential by accessing their full linguistic repertoire, thus promoting communication. Student 2 is now capable of distinguishing real learning facilitated by translanguaging practices: ‘I thought that we were not realizing real learning since we did not do any translations. However, once I realized that I could do more than translation, I became happier’. In the same fashion, student 10 said: ‘I came to understand that I could speak and understand English. It happened for the first time, and I realized that I had not attempted to talk in English beforehand’. This quote shows that the student was able to realize oral production for the first time.

4.3.2.2. Discovering the language system

The cognitive dimension means inquiring, and creating sense is often related to cognition. In this regard, the student 8 noted: ‘I regarded English lessons as not just a regular lesson. Indeed, it was like something to discover and implement’. Translanguaging was implied as having a crucial role in triggering inquiry into the language system.

4.3.2.3. Bilingual awareness raising

Another aspect of the cognitive dimension seems to be awareness-raising. Student 7 mentioned that being made aware of her bilingualism engendered positive feelings: ‘I felt more powerful when I could practise two of the languages. I came to understand that I am bilingual. It is a good feeling’.

4.3.2.4. Facilitating learning

Some students emphasized that translanguaging played a pivotal role in facilitating the learning process. Student 9 describes their positive experiences emphasizing the practical learning gains from the lesson: ‘Lessons made learning easier and all the lessons were practical. I could put the knowledge into practice’, while student 15 highlights how the use of materials in two languages contributed to comprehension: ‘I think I can understand things in graphs and shapes easily. All these materials enabled me to understand better’.

4.3.3. Interactive dimension

4.3.3.1. Developing interactional language use

Some students highlighted the interactional language use of translanguaging. Student 1 ‘had opportunities to practise English with my teacher and my friends. Turkish and English were used simultaneously’. The novel feeling of enjoyment was expressed by student 2: ‘talking to each other was a good feeling as I had not experienced it before’. Additionally, student 17 drew attention to the scaffolding, as well as complementary aspect of translanguaging:

Dialogues and discussions were of great help. We put into practice what we learnt both in Turkish and in English. Every activity completed each other.

4.3.3.2. Authentic language use

Another aspect underscored was authentic language use, exemplified by students 16 and 19, respectively:

‘Translanguaging was realistic and made me feel comfortable while asking for clarification. I was relieved when I could understand whole thing. My peers and teacher could move faster and could express our ideas without being forced to talk in English’.

‘I did not have to pretend as if I could not understand Turkish. It made me feel realistic and led me to talk in English’.

Some students pointed out that authentic language use made the language a part of daily life; it became a more tangible and concrete concept. In this regard, student 8 emphasized the concrete aspects ‘I realized that I had regarded English as an abstract thing beforehand. Now I can use it in my daily life at school and at home. It is in my life’.

4.3.3.3. Promoting communicative abilities

Some students indicated that the attention was on communication because they were allowed to clarify their opinions, using their full linguistic repertoire. Students 13 and 14 underlined their newly-developed communicative abilities gained:

‘We kept communicating and got the idea or the feeling across. Lessons flowed like water. We were in constant communication with our classmates anyhow using the languages concurrently. It felt great’.

‘I enjoyed the lessons because my concern was on communicating and understanding the message and conveying my message. I could do it somehow’.

Moreover, some students, such as student 19, began to gradually increase their participation:

‘Turkish and English were the parts of classes and I noticed that this situation improved my understanding and it encouraged me. I started not to shy away from the discussions and dialogues alike’.

4.3.4. Affective dimension

4.3.4.1. Feeling secure during in-class communication

Some students expressed how translanguaging practices made them feel more secure about expressing ideas in English only, through the freedom to use either language. Student 4 said: I did not have to worry about conveying my ideas because I could use what I got both in English and in Turkish, and student 14 was similarly relaxed: I was relieved when the teacher did not force me to speak in English. Student 16 noted how the flexibility led to self-confidence, explaining that ‘I enjoyed the lessons because I could join the lessons and feel confident’. Student 15 expresses how he was freed from inhibitive effect of forced activities: ‘I felt relaxed because I was not forced to do anything’.

4.3.4.2. Developing sense of comfort while learning English

Another aspect the feelings of comfort and relief felt during translanguaging practices. Students 19 and 20 underlined the increased comfort in the lessons. They said respectively

‘Since nobody felt nervous before, while and after the English classes, we could get into contact comfortably using both of the languages’ and ‘Most importantly, I felt myself comfortable because I used to feel nervous in English lessons beforehand’.

4.3.4.3. Sustaining motivation to speak

Maintaining motivation is of significance for foreign language learners. Student 11 stressed how translanguaging increased his motivation and created genuine interest: ‘at the end of week one, I got accustomed to it, and translanguaging enabled me to be motivated to take part in activities and stop feigning. I used Turkish and/or English when necessary and learnt’. Feigning or pretending may be common when students are prevented from being actively involved due to their low level of English, blocking self-expression.

4.3.4.4. Volunteering instead of being forced to engage

Voluntary participating, rather than being pressured is also important in developing intrinsic motivation to learn. To student 15, translanguaging helped her develop a voluntary participation and engagement: ‘I liked the lessons because I was not forced to talk in English. When I felt ready, I realized that I automatically did it, and I expressed my thoughts in English’.

4.3.4.5. Experiencing enjoyment of learning

The students emphasized fun as one of the key aspects of language learning. Student 6 agreed: ‘Being able to understand the teacher and the friends and being able to talk in English were so helpful that I started to enjoy English classes’. Student 4 framed the enjoyment in terms of the difference between previous and present learning methods: ‘It was great. It was fun really. I understood that English is not just memorizing and translating’.

4.3.4.6. Developing sense of real language learning

The superficial, nonauthentic mode of course materials gives students a sense of artificiality in learning, whereas translanguaging made the students feel honest to themselves. Student 19 emphasised this: ‘Stopping behaving as if we were speaking English made me feel honest. When we were put into groups and told to discuss the topic in the text, we used to talk in Turkish but when the teacher was near us, we used to say just ‘really’, ‘I think so’ etc. without understanding anything. It was not real’.

5. Discussion

This study sought to address three research questions. The findings revealed that the students exposed to translanguaging instruction outperformed those exposed to grammar translation, regarding both receptive and productive skills. The translanguaging instruction had a large effect (η²: 0.56) on language skills, based on the ANCOVA statistics.

Our translanguaging instruction helped mediate learner understandings (Lewis, Jones, and Baker 2012), as well as co-construct meaning and demonstrate knowledge (Garcia 2009), revealing a more realistic and complete picture of our pupils’ linguistic behaviour, thus engendering language learning. Additionally, translanguaging could be regarded as an inclusive approach (Garcia and Wei 2014) rather than the additive-subtractive dyadic (May 2017) for all pupils, irrespective of their linguistic background. Our study showed increased communication between teachers and students, and among students themselves (Nussbaum 2014). Furthermore, the research revealed that four-fifths of the interviewees reacted positively to translanguaging. The systematic facilitation of translanguaging in the classroom made them feel more secure, motivated and open to learning, and increased comprehension (Dikilitas and Mumford 2020) by overcoming obstacles to the pedagogic task (Creese and Blackledge 2010), and promoting a deeper understanding (Baker 2011), as revealed in our quantitative findings. The only criticism, expressed by one-fifth of the interviewees, concerned their reluctance to use Turkish because they stressed that learning English should not involve other languages, which reflects their deeply-rooted monoglossic attitude.

In addressing the second research question, we found that the translanguaging group outperformed the CLT group in both receptive and productive skills measurements with a large effect (η²: 0.22) on foreign language skills, based on the one-way analysis of ANCOVA. CLT favours L2-only instruction, while translanguaging allows for the flexible and systematic use of L1 and L2 for instruction and interaction, rather than monoglossic (Littlewood 2014) and monological thinking (Weinreich 1974). This approach ‘upends traditional language ideologies’ (Poza 2017, 113), which dominated the field of applied linguistics and the professional discourses on the practice of language teaching. Additionally, Macaro (2005) argues that exclusive use of the target language is based on a dogmatic monolingual approach, resting on empirically untested assumptions, for example, that L2 learning is a natural process tantamount to first language acquisition. In line with our positive findings, the use of L1 can be helpful in comprehension (Butzkamm and Caldwell 2009) and raising awareness (Scott and Fuente 2008) as regards linguistic differences between languages. Translanguaging can serve different purposes – making meaning, facilitating learning, promoting communicative abilities – as shown in this study. Translanguaging has been found pedagogically effective since ‘both languages are used in a dynamic and functionally integrated manner to organise and mediate mental processes in understanding, speaking, literacy, and, not least, learning’ (Lewis, Jones, and Baker 2012, 1). It also allows the pupils to be more successful in language skills through integrating languages and supporting new language forms (Gort and Sembiante 2015). The findings about the first and second research questions imply that translanguaging is likely to have more effective outcomes for emergent bilinguals (Garcia and Wei 2014), corroborated by the empirical statistical results of this study, i.e. a large effect (η²: 0.56) at GTM group, a large effect (η²: 0.22) at CLT group.

The third research question explored the qualitative impact of the translanguaging experience on learners’ perceptions. The affective dimension is the most cited, which shows favourable reactions to the implementation of translanguaging, and the development of positive perceptions of simultaneous and interchangeable language use (Martin–Beltrán 2010). Our analysis revealed that translanguaging can be linked with positive perceptions and relatively better learning outcomes. In this regard, emotions (i.e. experiencing enjoyment of learning) should not be ignored (Dewaele 2015). Translanguaging played a pivotal role in creating a positive, comfortable, and friendly teaching environment, providing the resources to encourage learning (Peercy 2016, 165), and promoting cognitive engagement in content-matter learning (Duarte 2019).

In conclusion, these results imply that translanguaging might be a more effective pedagogical approach than monoglossic teaching approaches. Our findings also resonate with the tendency to adopt ‘multilingual turn’ in classroom practices (Cenoz and Gorter 2011). We also concluded that translanguaging impacts students’ learning processes and overcomes the constraints of monolingual instruction and language separation (Bunch 2013). This led to the opportunity for the enhancement of pupil–teacher relationship, and for meaningful participation (Palmer 2008), thus empowering learning (Garcia and Wei 2014, 79–80). Translanguaging is likely to foster bilinguals’ pride in their emergent skills (Fallas Escobar 2019) as seen in our qualitative themes. Garcia (2009) emphasizes the necessity for constant relationships with others and other contexts, as well as with other times, as expressed in the saying ‘I am because you are; you are because we are’ (Makalela 2019).

6. Implications

This translanguaging instruction provided novel insights into high school foreign language teaching. Based on Garcia’s (2009) model, it demonstrates the improvement in foreign language skills, which promotes the concept of translanguaging as a practical theory, the language instinct metaphorizing the inherent capacity to acquire languages by providing a translanguaging space (Wei 2018) to experience constructive, interactive, cognitive, and affective insights. The study also offers implications for those teacher educators and researchers desiring to reconsider ‘the trans turn in Applied Linguistics’ as suggested by Carroll and Mazak (2017, 5). The context-specific implication for all teachers and teacher educators is that the systematic shuttling between Turkish and English as instructional and interactional language should be encouraged as beneficial rather than detrimental (Wach and Monroy 2020). As shown, the systematic flexible use of Turkish and English led to an improvement in scores, and qualitative novel insights into English language learning. Additionally, we suggest that policy makers and educators should reconsider one-language polices, and experiment with integrating translanguaging practices into curriculum, materials, and even evaluation criteria. It might also be significant to inform and train teachers about potential uses and benefits of translanguaging to be able to promote more comprehensive use of target language. Nonetheless, we acknowledge that the study involved highly performing high school students. The first author was inevitably cognizant of the research questions, possibly influencing the process of instruction and data collection. However, the quantitative scores would minimise any researcher subjectivity.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Notes on contributors

Muhammet Yasar Yuzlu

Muhammet Yasar Yuzlu is a Ph.D. candidate at Bahcesehir University and works as an English language teacher at Turkish Ministry of Education. His research interests include in-service teacher education and development of bilingual pedagogy through translanguaging.

Kenan Dikilitas

Kenan Dikilitas is a Professor of University Pedagogy at the University of Stavanger in Norway. Kenan's research interests include teacher education, mentoring and investigating action research, as well as in-service teacher development in various contexts.

Appendices

Appendix A. The percentages of the use of Turkish and/or English according to the groups.

Table

Groups	Levels	Turkish	English	Intervention
Experimental group 1 (n30)	Pre-intermediate GTM	50%	50%	Translanguaging as a pedagogy
Experimental group 2 (n30)	Upper-intermediate CLT	50%	50%
Control group 1 (n30)	Pre-intermediate GTM	80%	20%	None
Control group 2 (n30)	Upper-intermediate CLT	0%	100%

GTM, Grammar translation method; CLT, Communicative Language Teaching.

Note: The equal percentages in translanguaging groups represent the non-hierarchical and equal status of the languages in the teacher’s stance, and the space for the languages in designing instructional practices (Garcia, Johnson, and Seltzer 2017). Taking into account the flexibility, the teacher made shifts in accordance with the translanguaging corriente (Garcia, Johnson, and Seltzer 2017).

Appendix B. A sample of translanguaging practices.

Table

Translanguaging practices	Examples
Accessing and building background knowledge Mutual assistance as language partners and pre-writing	Students were asked to brainstorm holiday types orally in the language of their choice, and created a semantic map in target language in pairs. Students were asked to write a good memory of theirs in the language of their choice and they read each other’s prewriting in pairs and made observations on how to sequence the events. Then, they started to write the story in the target language

Appendix C. The density of the emerging main and sub-themes.

Table

Main themes	Density	Sub-themes	Density
Constructive	17	Making meaning	8
Dimension		Promoting autonomous learning	9
Cognitive dimension	34	Accessing full linguistic repertoire	9
		Discovering language system	7
		Bilingual awareness raising	10
		Facilitating learning	8
Interactive dimension	35	Developing interactional language use	15
		Authentic language use	6
		Promoting communicative abilities	14
Affective dimension	53	Feeling secure during in-class communication	9
		Developing sense of comfort while learning English	10
		Sustaining motivation to speak	5
		Volunteering instead of being forced to engage.	5
		Experiencing enjoyment of learning	18
		Developing sense of real language learning	6