English (pdf)
Article in xml format
Article references
Automatic translation
Send this article by e-mail
Cited by SciELO 
Cited by Google 
Similars in
SciELO 
Similars in Google 
Permalink
Introduction
One of the most relevant decisions that English as a foreign language (EFL) teachers must make is whether to restrict the use of the mother tongue (L1) in the foreign language (L2) classroom. This controversial issue has usually been approached from a prescriptive perspective, which focuses on what teachers should do rather than providing a description of what teachers actually do in their classrooms (Üstünel & Seedhouse, 2005). In line with this, many teachers avoid using the L1 in the L2 classroom, as they believe they should not prevent learners from accessing L2 input. However, many EFL teachers do use their L1 to teach L2 for a number of reasons such as providing the L1 equivalent to new vocabulary and giving key instructions in order to avoid misunderstandings. Even though the literature advocating for code-switching-generally defined as the use of L1 in the L2 classroom-over the last decades has increased, there are questions that have not been resolved regarding the amount and quality of L2 needed, and any decision on the matter is eventually left to the teacher’s judgment and intuition. Therefore, the present study aimed at analysing the functions that EFL nonnative speaker teachers accomplish with their talk and characterising the existing relationship between those functions and the language choices they make.
Background of the Study
The relevance of English as a lingua franca, that is, a language that is widely used by speakers of other languages to ensure successful communication, has prompted countries to introduce educational policies that include English as a second or foreign language in their national curricula. In the Chilean case, the national curriculum requires that English be the compulsory foreign language to be taught at school (114 hours per year) starting from 5th grade in primary education until 12th grade when students complete their secondary education. The compulsory nature of English is a reflection of the discourse that in the last 30 years has been introduced in governmental agendas, which have sought for their countries to become bilingual and thus be a competitive actor in a globalised market (Glas, 2008).
The first educational reform in the 1990s regarding the English curriculum focused on developing receptive skills; that is, 80% of the curriculum was devoted to listening and reading (Ministerio de Educación [Mineduc], 2009). In recent years, the Ministry of Education has emphasised the need for English to be taught without using L1 (Mineduc, 2019). However, specific guidelines in the national curriculum prompt teachers to focus on providing a great deal of information to be read and listened to by learners, which reduces their opportunities to produce language. In addition, national and international proficiency test results show that Chilean learners are lagging behind their Latin American counterparts (Gómez & Pérez, 2015). In 2018, Education First reported that Chile ranked 46 out of 88 countries/regions in the world in a language test, a ranking which belongs to the “low level” proficiency band. Although the reasons for this outcome may be related to some extent to curricular factors such as the insufficient number of hours allocated to learning English, particularly in municipal schools (Barahona, 2016), language-related factors such as the code-switching behaviours of EFL high school teachers may illuminate the decisions that those teachers make in the language classroom.
Code-Switching
Code-switching is a linguistic phenomenon that has been studied over decades and for which different definitions have been proposed. The broader definitions have used the term code-switching and code-mixing interchangeably, making them synonymous with “bilingual speech” (Schendl & Wright, 2011). More specific definitions have referred to code-switching as the use of two or more languages at the inter-clause/sentential level, and to code-mixing as the mixing of two languages at the intra-sentential/clause level (Lin, 2013; Üstünel, 2016). A more holistic definition is provided by Poplack (2010), who defines code-switching in the following terms: “[code-switching] refers to the mixing, by bilinguals (or multilinguals), of two or more languages in discourse, often with no change of interlocutor or topic. Such mixing may take place at any level of linguistic structure” (p. 15). This is in line with Schendl and Wright’s (2011) definition, which states: “If a person uses elements from both languages in a single discourse, be it insertion of single words or alternation of larger segments, this is referred to as code-switching” (p. 23). Regarding the EFL classroom, holistic perspectives on code-switching involve “the alternating use of more than one linguistic code by any of the classroom participants . . . and this can include both code-mixing (intra-clausal/sentential alternation) and code-switching (alternation at the inter-clausal/sentential level)” (Lin, 2013, p. 195).
Approaches to Code-Switching in the Language Classroom Context
The use of teachers’ L1 to teach L2 has been a matter of controversy in the literature as there are rather polarising perspectives in this respect. The position that supports the L2-only classroom environment contends that learners need to be exposed to a considerable amount of L2 input in order to learn the language and that the teacher represents the main source of target language (TL) input for those learners. Therefore, the main role of the teacher is to expose learners to the TL at all times, as this will contribute to increasing their TL proficiency (Turnbull, 2001). Although this prescriptive view represents an ideal scenario that is in line with the Chilean national curriculum (Mineduc, 2019), studies on code-switching in second language classrooms have reported that teachers differ in the amount of L2 use across language settings. For example, De la Campa and Nassaji (2009) assessed the amount of L1 used by teachers with different experience levels in two German-as-a-foreign-language classes. They found that the average amount of L1 use was 11.3% in both classes and concluded that L1 use “is dependent on the class circumstances and can be different from the average amount of L1 an instructor uses in all his or her classes” (p. 756). Similarly, Grant and Nguyen (2017) reported strong differences between the frequency of language switches produced by Vietnamese EFL teachers, ranging from 439 to 10 code-switching instances. Although no studies assessing the amount of code-switching have been conducted in Chilean EFL contexts, these findings suggest that the teachers’ selection of one language over the other may be determined by contextual and pedagogical factors, and that these differing pedagogical behaviours may take place in the same educational institutions and can be found in the same teacher. The contextual factors impinging upon the amount and quality of code-switching can be related to the students’ L2 proficiency level. Teachers in low-level classes tend to resort to the learners’ L1 significantly more than teachers who teach more advanced learners, and teachers’ use of the learners’ L1 has been found to be more frequent in state institutions/schools than in private schools, with teachers in the former group significantly encouraging more L1 activities in class (Grant & Nguyen, 2017). However, switching to the L1 has been found to be time consuming, discourages talented students, offers no TL environment, and prompts no effort towards using the TL (Promnath & Tayjasanant, 2016).
Another factor that has an impact on teacher’s language choice is the lack of competence on the part of non-native teachers (Üstünel, 2016). As Cook (2001) points out, “teachers resort to the L1 despite their best intentions and often feeling guilty for straying from the L2 path” (p. 405). This statement implies that educational institutions where a rhetoric adoption of communicative approaches such as communicative language teaching is carried out may still fail at achieving the L2-only approach (Wang & Mansouri, 2017). Furthermore, Copland and Neokleous (2010) argue that teachers are not always aware of either the quantity of L1 used in their classes or the purposes for which code-switching is used, which makes them underreport their code-switching practices. The ensuing contradiction between a stated belief and classroom behaviour regarding the use of code-switching may cause feelings of guilt as teachers struggle to reconcile pedagogic ideals with contextual realities. An approach that nurtures code-switching awareness is proposed by Cook (2001), who argues for a judicious use of L2, “maximizing” L2 in the classroom and emphasizing the “the usefulness of the L2 rather than the harm of the first” (p. 404).
Studies Addressing the Impact of Code-Switching in the EFL Classroom
Studies assessing the impact of code-switching in EFL learning have focused on linguistic skills in adults. In a Chinese context, Tian and Macaro (2012) assessed the acquisition of vocabulary in a group where the L1-equivalents of words was used, and another group where L2-only-explanations were provided. Results showed that all the L1-equivalent group scored better than the English-only explanation group, regardless of their proficiency levels. These findings were mirrored by Zhao and Macaro (2014), who argued that even though the L2-only instruction approach may be useful for the learning of some linguistic features and structures by adult learners, this may not be the case with vocabulary learning, as “the concepts that L2 represents have been established through the L1” (p. 77). In contrast, acquiring other abilities such as oral production skills may not benefit from L1-based instruction, as learners are not given the chance to decode and process oral L2 input (Haryanto et al., 2016). Other studies have reported prominent pedagogical functions behind the use of code-switching by teachers, such as explaining new vocabulary, illustrating grammatical rules, managing the classroom, eliciting learner talk, and providing task instruction. For example, Muñoz and Mora (2006) found positive outcomes in the use of L1 for communicative purposes in the EFL classroom. They video-recorded and qualitatively examined the interaction between one teacher and 31 learners as they code-switched in the classroom. The authors found that the use of code-switching by the teacher in the classroom accomplished a range of discourse functions such as affective, topic switch, interjection, and repetition. These functions sought to create a comfortable environment in the classroom where information could be clarified. Interestingly, the affective function was the most common function in the teacher’s discourse. This suggests that clarification may not be the main reason why teachers code-switch, and that the type of code-switching may be related to teacher personality. Similarly, Kim (2001) applied a descriptive approach to analyse the code-switching behaviour of eight beginner-level teachers who taught several languages: two Asian languages (Korean and Japanese) and two European languages (German and French). Kim designed the Functional Language Alternation Analysis of Teacher Talk (FLAATT) to analyse teacher talk in terms of a range of pedagogical functions. Results showed that L1 was used by teachers for substantial segments of the lesson and that the frequency of teaching acts for each teacher was different, though the function of marker (used to indicate the beginning or end of a topic) was the most frequent.
Thus, although the “optimal” position (Macaro, 2014)-a stance where the teacher is aware of the advantages and disadvantages present in using the L1 that may promote or hinder learning-would seem adequate for EFL learners, the literature has not reached consensus in terms of the frequency and the type of code-switching that should be provided in classes with homogeneous L1s (Liu et al., 2004). These findings highlight the importance of exploring the pedagogical functions that are accomplished by EFL teachers as they code-switch in their classrooms.
Method
This section will describe the process through which the data for the study were gathered and analysed. The study attempted to answer the following research questions:
Research Question 1: What language is the most frequently used by EFL teachers in a Chilean high school context in two classroom modes?
Research Question 2: What types of functions are performed and how often by these EFL teachers?
The present study is descriptive in nature and is informed by Kim and Elder’s (2005) focus on functions performed by teachers with their code-switching. The present study employed nonparticipant observation procedures, as audio-recordings were made by the participant teachers. The main aim was to describe and explore the amount of L1 (Spanish) used in relation to the pedagogic functions performed by two EFL teachers. Another goal was to establish whether there is any systematicity in the preference of one language over the other when accomplishing pedagogical functions.
Context and Participants
The data for this study were taken from two female EFL teachers-whose names have been changed-working at a high school in Santiago, Chile. Clara is in her twenties and is less experienced than Ester, who is in her forties. Ester has been teaching in public and semi-private high schools for 14 years, whilst Clara has been working in high schools and language institutes for six years. Both teachers are native speakers of Spanish. The students in their classrooms are 14-15-year-old learners who are taught a 35-minute EFL lesson four times a week, and who belong to the “real beginner” Common European Framework of Reference level, according to a placement test taken at the institution. In the Chilean public high school context, teachers tend to deliver their L2 lessons by using L1, particularly with those learners who seem to lack commitment. Indeed, research in the Chilean high school context has suggested that whilst students appreciate the value of English as a global language, they do not show commitment towards learning the language at school (Gómez & Pérez, 2015).
Data Collection Procedures
Each teacher audio-recorded two of their own groups in three different lessons (35 minutes each). The two teachers started recording their lessons over a period of three weeks in order to ensure that they taught the same content and had similar class objectives. In total, there were 12 sessions recorded and transcribed (420 minutes). Analysis of transcriptions focused on two specific classroom modes put forward by Walsh (2006), namely, the managerial mode and the materials mode as they contained specific pedagogical goals that were more relevant for the context in which the data were collected. The managerial mode is in place when the teacher seeks to organise the physical environment and introduce or conclude an activity, and it is characterised by extended teacher turns and the presence of transitional markers (e.g., “okay,” “now”). The materials mode seeks to provide language practice around a piece of material, and it features display questions, error correction, and modelling. Other modes in Walsh’s taxonomy (skills and systems mode and classroom context mode) were not included in the analysis as the nature of the lessons did not warrant their inclusion (i.e., these modes were not represented in these teachers’ talk). Ten minutes of each mode, for each teacher, were extracted from the data. Pauses longer than three seconds were not considered in the total time to be included for analysis. Only whole-class teacher talk was included in the analysis, as the study aimed at characterising teachers’ code-switching behaviour in relation to their pedagogical functions. When the ten minutes of data for a particular mode were completed with a teacher, the researchers stopped including data for that mode. Thus, 40 minutes of data (ten minutes for each mode, for each teacher) were extracted from the lessons.
Method of Data Analysis
The FLAATT instrument. The FLAATT scheme was adapted from Kim (2001) to analyse the data gathered. The categories in the instrument were mainly drawn from Duff and Polio (1990) and Sinclair and Coulthard (1975). The “language used” category contains five types of “language code,” which provide information on the nature of the utterance in terms of the amount of L1 and L2 used by the teacher (see Table 1).
Table 1 “Language Used” Category in the FLAATT Scheme (Kim, 2001)
| Category | Language code | Definition |
|---|---|---|
| Language used | L1 (Spanish) | The unit entirely consists of Spanish |
| L1c | The unit mainly consists of Spanish with one word of morpheme in English | |
| L2 (English) | The unit entirely consists of English | |
| L2c | The unit mainly consists of English with one word of morpheme in Spanish | |
| Mix | The unit is a mixture of Spanish and English, to which the above categories of L1c and L2c cannot be applied. |
In order to identify language codes, teacher utterances were separated by means of the basic unit of analysis in the FLAATT, which is the “analysis of speech unit” (AS-unit). These units are defined as
A single speaker’s utterance consisting of an independent clause, or sub-clausal unit (a segment that cannot stand alone, although can be expanded into a full clause by reconstructing the omitted elements) together with any subordinate clause(s) associated with either. (Foster et al., 2000, p. 365)
The AS-unit is more suitable for this type of analysis than other types of units such as the T-unit and the C-unit, because it was created especially for analysing oral discourse and, therefore, it can account not only for sentences and phrases but also for utterances even when they seem to be fragmentary on the surface. There are a number of rules for the segmentation of the data into AS-units. For example, the independent clause in the AS-unit must minimally include a finite verb. There are also rules for considering minor utterances (e.g., “yes” and “now”) as AS-units, as long as they perform a full function. Space constraints prevent the researchers from including the set of rules, which may be found in Foster et al. (2000). Once the AS-units were identified, a language code for each was determined. This was done by following a set of guidelines adapted from Kim (2001), and included procedures for treating proper nouns, interjections, and word fragments, as well as rules for classifying the AS-unit as one of the five language codes existing in the FLAATT instrument.
The original FLAATT scheme included three goals of classroom interaction proposed by Ellis (2012). These are core goals (teaching the language itself), framework goals (related to classroom management), and social goals (maintaining social relationship between interlocutors). They were replaced by Walsh’s two classroom modes in order to characterise the pedagogical moments of the lessons before the functional analysis was carried out and to provide information regarding the interrelatedness of the pedagogic goals of the teacher and the language used.
Teaching acts. The FLAATT instrument provides a number of columns where the teaching acts (i.e., functions) created by Sinclair and Coulthard (1975) may be used to identify the AS-units. New pedagogical functions can be created depending on the range of functions performed by teachers. Two examples of the most common functions found in the data are provided below (Kim, 2001).
“well,” “okay,” “now,” “good,” “right,” “all right,” and a close class of items represent this category that marks the beginning or end of a topic or move boundary.
3 (Ya [Well]
4 students
5 so, here are your worksheets about physical description
Realized by a statement, a tag question, or a phrase. Its function is to expand, justify, exemplify, or provide additional information to the response given by a student.
2183ah ya
2184 (pero eso is not “look like” porque estamos hablando de appearances
[But that’s not “look like” because we are talking about appearances]
Results
This section will present results regarding the type and the nature of the functions performed by the two teachers as part of the managerial and materials mode, which allowed for establishing similarities and comparisons between the teachers regarding their language choices.
L1 and L2 AS-Units: Managerial Mode
As shown in Table 2, most of the AS-units produced by Ester and Clara belonged to the opposite categories in the classification (L1-only and L2-only). The categories of L1c, Mix, and L2c accounted for only 2% of the total amount of AS-units (21 instances out of 1,084). Both Ester’s groups (G1 and G2) and Clara’s groups (G3 and G4) produced more AS-units in L1 than in L2 in this particular mode. Ester produced L1-only units in 319 instances, compared to 217 L2-only units, whilst Clara produced 396 L1-only units, compared to 131 L2-only units. Ester’s percentage of L2-only units reached 40%, and Clara’s L2-only output only reached 24%. Ester tended to produce more L2-only units than Clara; however, the percentage of occurrence of L2-only AS-units for both teachers is less than 40%, in all groups, with one of Clara’s groups (G4) reaching the lowest percentage (15%). Overall, the two teachers produced 32% of L2-only AS-units in this mode. Finally, L1-only units were produced 715 times in all groups, which represents 66% of the units in this mode. These findings suggest that these teachers do not use L2 in their classrooms 66% of the time in this mode. There is a clear difference in the number of AS-units produced in L2 in each of Ester’s and Clara’s groups, although this difference is more marked in the latter.
Table 2 Language Codes by Groups: Managerial Mode
| L1 | L1c | Mix | L2c | L2 | Total AS-units | ||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| AS-units | AS-units | AS-units | AS-units | AS-units | |||||||
| Teacher | n | % | n | % | n | % | n | % | n | % | N |
| Ester G1 | 178 | 54 | 3 | 0.9 | 1 | 0.3 | 3 | 0.9 | 140 | 43 | 325 |
| Ester G2 | 141 | 64 | 0 | 0 | 3 | 1 | 1 | 0.5 | 77 | 35 | 222 |
| Total | 319 | 58 | 3 | 0.5 | 4 | 0.7 | 4 | 0.7 | 217 | 40 | 547 |
| Clara G3 | 166 | 64 | 1 | 0.3 | 3 | 1 | 0 | 0 | 90 | 35 | 260 |
| Clara G4 | 230 | 83 | 2 | 0.7 | 0 | 0 | 4 | 1 | 41 | 15 | 277 |
| Total | 396 | 74 | 3 | 0.6 | 3 | 0.6 | 4 | 0.7 | 131 | 24 | 537 |
L1 and L2 AS-Units: Materials Mode
Table 3 shows that, unlike the findings in the managerial mode, Clara was the teacher who produced more L2-only units in the materials mode (305 AS-units amounting to 54%) in her two groups, compared to Ester who produced 254 (36%) in her two groups. The teacher who produced more L1c, Mix, and L2c units was Ester (89 units) in comparison to Clara (32 units). More language mixing is seen in this mode (9%) than in the managerial mode (2%), which can be explained in terms of the language that is needed to follow textbook activities (i.e., translation of specific words). This might also explain to some extent the lower number of L1-only units in this mode (47%, with 606 units out of 1,285). The groups displayed similar numbers within each teacher’s group, which did not occur in the managerial mode. The different classroom management approaches taken by teachers may have influenced the amount of L2 use in that mode. The number of L2-only units that both teachers produced in the materials mode was higher than in the managerial mode (559 instances in the former, and 348 in the latter).
Table 3 Language Codes by Groups: Materials Mode
| L1 | L1c | Mix | L2c | L2 | Total AS-units | ||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| AS-units | AS-units | AS-units | AS-units | AS-units | |||||||
| Teacher | n | % | n | % | n | % | n | % | n | % | N |
| Ester G1 | 184 | 49 | 25 | 7 | 21 | 6 | 5 | 1 | 141 | 38 | 376 |
| Ester G2 | 188 | 55 | 10 | 3 | 20 | 6 | 8 | 2 | 113 | 33 | 339 |
| Total | 372 | 52 | 35 | 5 | 41 | 6 | 13 | 2 | 254 | 36 | 715 |
| Clara G3 | 121 | 41 | 4 | 1 | 2 | 1 | 7 | 2 | 164 | 55 | 298 |
| Clara G4 | 113 | 41 | 6 | 2 | 4 | 2 | 8 | 3 | 141 | 52 | 272 |
| Total | 234 | 41 | 10 | 2 | 6 | 1 | 15 | 3 | 305 | 54 | 570 |
To sum up, the overall number of L1-only units in the two classroom modes (56%) represents a substantial amount of L1 produced by these two teachers. L1 was the most used language by Ester (691 out of 1,262 units, with 55%) and Clara (630 out of 1,107 units, with 57%). These numbers show that even when there were differences between the two teachers across modes, both produced a very similar number of L1-only units. Across teachers and modes, the most frequently used language by these teachers was L1-only (56%), followed by L2-only (39%), Mix (2.3%), L1c (2.2%), and L2c (0.5%).
Functional Analysis: Managerial Mode
Due to space constraints, analysis of the functions performed by these two teachers are presented in terms of the five most frequent teaching acts performed in their classrooms, as well as their dominant type (L1 or L2) in the managerial and materials mode. Table 4 presents this information for the managerial mode. Regarding Ester’s data, the most frequent function in her two classes was marker, and her most dominant language was L1 in one group and L2 in the other. However, the number of instances in which the function is performed in both groups (37 out of 65 for the L1-dominant group and 29 out of 56 for the L2-dominant group) suggests a similar number of code-switching instances. In contrast, Clara’s data show that the most common function was not the same in her two classes: the most common functions for Groups 3 and 4 were directive and comment, respectively, and were mostly performed in the L1. Ester produced three of her most common functions with L2 as a dominant language (marker, directive, nominate), with nominate reaching the highest percentage of L2 dominance (65%). Thus, she used the L2 profusely when disciplining students and giving instructions (some of which were reinforced by an L1 translation). Clara did not have L2 as a dominant language in the most frequent functions performed by her.
Table 4 Most Frequent Functions Performed and Their Dominant Language Type: Managerial Mode
| Teacher | Overall L2 percentage | 1 | 2 | 3 | 4 | 5 | |||||
| N | N | N | N | N | |||||||
| Ester G1 | 43 | Mar (L1) | 37/65 | Dir (L2) | 34/52 | Met (L1) | 19/37 | Che (L1) | 30/36 | Nom (L2) | 20/34 |
| Ester G2 | 35 | Mar (L2) | 29/56 | Met (L1) | 29/39 | Dir (L2) | 15/27 | Che (L1) | 19/21 | Com (L1) | 16/18 |
| Clara G3 | 35 | Dir (L1) | 28/53 | Mar (L1) | 18/31 | Dis (L1) | 26/28 | Com (L1) | 23/26 | Gqu (L1) | 14/19 |
| Clara G4 | 15 | Com (L1) | 47/48 | Mar (L1) | 28/42 | Dis (L1) | 27/28 | Gqu (L1) | 22/22 | Che (L1/L2) | 12/24 |
Note: Mar = marker, Dir = directive, Com = comment, Met = meta-statement, Dis = discipline, Che = check, Gqu = genuine question, Nom = nominate, N = number of instances in which the function is performed in the dominant language in relation to the total number of AS-units. Categories where the L2 is the most dominant language are displayed in bold typeface.
Table 5 displays the most frequent functions in the managerial mode, by teacher. The functions of marker and comment were the most frequent for Ester and Clara, respectively, and were performed mainly in the L1.
Table 5 Most Frequent Functions Performed and Their Dominant Language Type per Teacher: Managerial Mode
| Teacher | Overall L2 percentage | 1 | 2 | 3 | 4 | 5 | |||||
| N | N | N | N | N | |||||||
| Ester | 39 | Mar (L1) | 66/121 | Dir (L2) | 49/79 | Met (L1) | 48/76 | Che (L1) | 49/57 | Nom (L2) | 27/47 |
| Clara | 25 | Com (L1) | 70/74 | Mar (L1) | 46/73 | Dir (L1) | 45/72 | Dis (L1) | 53/56 | Gqu (L1) | 36/41 |
The number of marker functions in Ester’s discourse (121 in total) is higher than Clara’s, and most of them are realized in the L1 (55%). Directives were very frequent in Ester and Clara’s data; however, the predominant language for this function was L2 and L1, respectively. Directive is the only function with an overall higher percentage of L2 units (50.33%) in the managerial mode. Clara’s most frequent was comment, which was delivered in L1 95% of the time, mainly to organise the classroom and the attendance.
Functional Analysis: Materials Mode
As can be seen in Table 6, the most frequent function in one of each teacher’s groups was model correct scaffolding (MCS), which had L2 as the dominant language. In line with the findings presented for the managerial mode, Ester’s use of the marker function permeated her data in this mode. However, this function ranked fourth in one of Clara’s groups (Group 3) and was not present in the other (Group 4). In Clara’s data, the number of most frequent functions with L2 as the dominant language increased from one in the managerial mode to five in the materials mode. In contrast, Ester’s output in this respect decreased from four in the managerial mode to one in the materials mode. Regarding other salient functions, in this case, as a total of the two classes, Ester’s overall number of check functions in this mode were 44, 42 of which were produced in the L1. Interestingly, out of the 42 AS-units, 40 were produced by means of “ya” (with a rising intonation).
Table 6 Most Frequent Functions Performed and Their Dominant Language Type: Materials Mode
| Teacher | Overall L2 percentage | 1 | 2 | 3 | 4 | 5 | |||||
| N | N | N | N | N | |||||||
| Ester G1 | 38 | Mar (L1) | 30/42 | Gtr (L1) | 18/42 | Eva (L1/L2) | 17/34 | Com (L1) | 14/23 | Dir (L1) | 11/21 |
| Ester G2 | 33 | MCS (L2) | 39/49 | Mar (L1) | 31/43 | Dqu (L1) | 18/32 | Com (L1) | 22/30 | Che (L1) | 22/24 |
| Clara G3 | 55 | Dir (L2) | 18/36 | Com (L1) | 25/32 | MCS (L2) | 28/29 | Mar (L1) | 14/27 | Eva (L1) | 20/26 |
| Clara G4 | 52 | MCS (L2) | 34/41 | Dqu (L2) | 24/33 | Dir (L1) | 19/27 | Com (L1) | 17/22 | Che (L2) | 15/18 |
Note: Gtr = Genuine translation, Dqu = display question, Eva = evaluate.
Table 7 presents the most frequent functions and their dominant language by teacher. Although the MSC function is more frequent in the materials mode (Walsh, 2006), Ester’s most frequent function is marker, which confirms the pervasiveness of marker functions in her talk in both modes. Both teachers performed their most frequent function in the L2.
Table 7 Most Frequent Functions Performed and Their Dominant Language Type per Teacher: Materials Mode
| Teacher | Overall L2 percentage | 1 | 2 | 3 | 4 | 5 | |||||
| N | N | N | N | N | |||||||
| Ester | 35 | Mar (L2) | 61/85 | MCS (L2) | 48/67 | Gtr (L1) | 29/67 | Com (L1) | 36/53 | Dqu (L2) | 26/44 |
| Clara | 57 | MCS (L2) | 62/70 | Dir (L1) | 43/63 | Com (L1) | 45/72 | Mar (L2) | 26/47 | Dqu (L1) | 27/47 |
Discussion
Overall, there were particular differences between these two teachers in terms of the type of language used and the functions that were performed in the two modes analysed. Discussion will be presented in relation to the research questions posited in the study.
RQ1: What language is the most frequently used by EFL teachers in a Chilean high school context in two classroom modes?
Results show that L1-only was used in the majority of the AS units present for both teachers, in the managerial mode and the materials mode. In line with Hosoda (2000), the use of L1 not only accomplishes a number of social functions, but also fulfils an important interactional role. In EFL classes where levels of proficiency and motivation are low, teachers tend to code-switch to L1 in order to support understanding. For example, although Ester produced the highest number of L2-only directives in the managerial mode, analysis of particular extracts suggests that in order to reinforce such directives and make them understandable for all the students, she included immediate translations of their utterances, both from L2 to L1, and vice versa, as can be seen in Extract 1.
Extract 1. Ester
126 SX, di una descripción de S1 [Give a description of S1] (…)
127 ( describe S1
128pero [but] look the vocabulary
129 (mira tu vocabulario [look at your vocabulary]
As pointed out by Kim (2001), these “translation” episodes occur before or after the L2 unit was produced, and were frequent in the data for both modes. They may prompt learners to resort to the teacher’s translation rather than to their ability to decode the L2 message. In the present study, these instances were labelled restate instruction (ReIn), in order to address the frequent restatement of instructions, whether it be from L1 to L2, or vice versa. This represents a strong pedagogical argument against using L1 when teaching L2 (Ford, 2009). Many teachers hold the belief that transmitting messages in L1 is faster in order to cover all the contents needed, and that since students are beginners, their talk must be “sheltered” in order to avoid frustration and demotivation in learners (Promnath & Tayjasanant, 2016). Although Butzkamm (2003) states that clarifications in the mother tongue can help leaners increase their confidence in using foreign language expressions, when too much time is spent prompting students to focus on single L2 words or phrases and then providing the translation, opportunities for more meaningful output may be lost. As can be seen in Extract 2, Ester code-switches to L1 by translating the phrase “next class,” and in doing so, she code-switches to L1 on five occasions, all of which focus on the translated phrase. This linguistic behaviour may prevent learners from formulating and testing hypotheses about the TL, as they are given the translation and do not benefit from the code-switching provided by the teacher (Kim & Elder, 2008).
Extract 2. Ester
556 next class we are going to continue to work in the worksheet
557 (¿qué dije? [what did I say?]
SX: que escribiéramos los… [that we should write the…]
558 (next class) next class we are going to continue working in the worksheet
560 (ya [okay]
561 (pero [but] “next class”
562 (¿qué es [what is] “next class”?
563 ( “next class” es la próxima clase [means next class]
564 we are going to continue working in this worksheet
565 (¿ya? [right?]
Results also highlight the high number of AS-units produced in L1 in the managerial mode. In this study, this represent 66% of all units in that mode, confirming that regarding managerial issues such as controlling students’ behaviour, the L1 is a “heavyweight ally” (Lee, 2007). The high number of L2-only directives produced by Ester in the managerial mode is explained by the frequent repetition of the same directive in the L2 to tackle disruptive classroom behaviour.
RQ2: What types of functions are performed and how often by these EFL teachers?
Teaching acts in the managerial mode. An L2 approach to instructions involves training students in various listening strategies and becoming acquainted with instructional and formulaic expressions, along with the “modelling, reiteration, conscious recycling of key phrases and vocabulary, effective use of handouts and board, as well as the need for patience as opposed to using the L1 when students do not understand something the first time” (Ford, 2009, p. 71). In contrast, an approach that includes code-switching advocates for using L1 when providing complex instructions to beginners (Forman, 2012). In the present study, the data seem to represent the latter approach, as a significant presence of L1 marker and directive functions in the managerial mode was found. The functions were produced mostly in L1, particularly in Clara’s data. Clara used L1 to convey rather simple directive functions, which suggests that an inclusion of L2 in those instructions should be considered in order to provide learners with more instances of L2 input. As has been shown, the ReIn function was present in these teachers’ data, which suggests that these EFL teachers could promote more active learner behaviour if they did not provide immediate L1 translations to their utterances, as learners would then focus on decoding the L2 message and retrieving meaning.
This comment function was also frequent when the teacher introduced the class and gave more detail on the steps that had to be followed with a certain assessment or task, as a complement of the meta-statement function. The reason why Clara had the lowest number of L2-only units in Group 4 in this mode may be explained by the abundant presence of comment functions produced in L1, as Extract 3 shows.
Extract 3. Clara.
900revisé todas las pruebas y no, (en este curso no) en este curso no hay muchas malas notas [I’ve checked the tests and no, (not in this class) there aren’t a lot of bad marks in this class]
901¿Recuerdan el trabajo de contestar las 10 preguntas que ya hicimos? [Do you remember the task we did where you had to answer 10 questions?]
902yo ése no tenía intención de ponerle décimas ni nada por el estilo pero lo voy a hacer considerando que fue un trabajo de revisión que lo hicimos post prueba y que en realidad está increíblemente relacionado con lo que ya hicimos [I didn’t have the intention of giving you extra points or anything like that but I will do it considering that it was a review task we completed after the test and it’s highly related to what we already did]
Teaching acts in the materials mode. A higher number of AS-units were produced in this mode due to the nature of functions such as the MCS function. These functions had the L2 as a dominant language, which is explained by the error correction and modelling that takes place when scaffolding learner utterances. Ester produced a high number of marker functions in the L2, contrasting this use with the marker functions in the managerial mode, where she performed them mainly in L1. Although she kept performing marker functions in the materials mode (unlike Clara), she may not have needed to repeat them in this mode in L1, as learners were focused on a piece of material and knew what to do with it as she nominated students. Regarding the evaluation function, it is typically composed of formulaic chunks (Lee, 2007). They represent “low cost” language, in the sense that they are easier to understand and it is common to find them in the teacher talk, so the majority of the evaluation AS-units should be conveyed in L2. The data from Ester and Clara show that evaluation was performed mostly in L2 by means of a discrete group of short expressions that are easily understood (e.g., “yes,” “no”). However, the accept function, which is very similar to the evaluation function (e.g. “okay,” “yes,” “no”), was usually performed in L1. This suggests that occurrence of these functions in either language may not be a conscious choice, and the decision is not made in a systematic fashion by these teachers. Noteworthy is the fact that the comment function was important in this mode and its dominant language was L1, similar to results in the managerial mode. Finally, the data shed light on the idea that the type and frequency of certain AS-units and functions are influenced by idiosyncratic speech. For example, out of the 44 check functions performed by Ester, 42 of them correspond to the lexical token “ya,” behaviour that was not present in Clara’s data. This linguistic behaviour may be part of Ester’s repertoire of L1 “pet words” that permeate her discourse, which is influenced by affective and idiosyncratic factors (Muñoz & Mora, 2006).
Conclusion
The present study aimed at identifying and portraying the code-switching behaviours of two teachers and the types of functions performed across two classroom modes. Results indicate that there is no consistency in the choice of one language over the other while conveying certain pedagogical functions. Both teachers used functions that had either L1 or L2 as dominant languages when conveying exactly the same function, in the same classroom mode. This inconsistency suggests that the decision on whether to code-switch and the manner in which it is done is made rather subconsciously and without previously thinking about the impact those decisions might have. In addition, the finding that L2 use was more prominent in one teacher in comparison to the other does not imply that the former teacher was better prepared, or that her output was understood. Teachers can produce a great number of L2 units, but their discourse may mainly consist of repetitions and markers, as the data have shown.
In many EFL contexts such as the Chilean one, the impact of delivering an L2-only class is still a matter of controversy, since learners in these settings are not adequately exposed to the target language (Barahona, 2016) and teachers face linguistically heterogeneous classrooms (Muñoz & Mora, 2006). In order for EFL teachers to make informed decisions regarding the use of L1 in the classroom, guidelines should be provided for teachers regarding ways in which to incorporate L2 when performing a number of pedagogical functions, and also in relation to how teachers can increase their language awareness as they code-switch in the classroom (Grant & Nguyen, 2017). The overall 56% of L1-only units produced by these two teachers is worrying, as EFL teachers must take advantage of the limited time they have in their classrooms by providing more instances for students to be exposed to the TL. Teachers may benefit from being aware of the way in which certain linguistic structures such as formulaic expressions can be used to increase L2 exposure, and that certain instructions can be delivered by means of modelling and exemplifying a task (Ford, 2009). Indeed, there are functions that could be performed most of the time in the L2, such as marker, directive, modelling, and checking, and a principled delivery of such structures can be beneficial for learners (Promnath & Tayjasanant, 2016). However, a teacher’s L2 output should not be made up solely of repetitions of functions such as markers, as this prevents learners from making an effort to decode more complex L2 messages. This finding has implications regarding the work that is being done by the English Open Doors programme, and initiatives such as “English in English.” This initiative was recently put forward by the Ministry of Education and seeks to improve the linguistic abilities of 5th and 6th grade teachers and learners by means of workshops and activities implemented by bilingual teachers (Mineduc, 2019). Whilst this is a step in the right direction towards making teachers more cognisant regarding the use of the L2, municipal or subsidised school teachers may have more difficulties differentiating an “L2-only” approach from an approach that takes into account the quality of the output produced by these teachers. Indeed, prescriptive perspectives to code-switching-such as the English in English initiative, and strong versions towards native-speakerism (Lowe & Kiczkowiak, 2016)-may lead teachers to avoid L1 use without a principled approach that encourages reflection on the language used to achieve learning objectives. Thus, it becomes crucial to consider linguistic, contextual, and idiosyncratic factors in the discussion on L2 use. Recent guidelines that seek to improve general English language proficiency and stress the relevance of taking an L2-only approach should encourage reflective processes regarding the functions that are accomplished by teachers as they code-switch in the EFL classroom.
