Strategies Used and Challenges Faced by Thai EFL Teachers When Eliciting Talk During Classroom Interactions in High School Contexts

Utilizing Conversation Analysis, this study examined classroom interactions to identify strategies secondary school teachers used to elicit EFL learners’ talk in four micro-pedagogical contexts. How strategies were used, and the challenges teachers faced were also explored. Four Thai teachers of English participated in the study, and approximately 7 hours of video and audio recordings were transcribed and analyzed alongside semi-structured teacher interviews. Overall, the most prevalent strategy used to elicit talk by the four teachers was display questioning, followed by some form of scaffolding strategy, and referential questioning. From classroom observations, the teachers’ consistent use of a variety of strategies apparently promoted learners’ active participation. Challenges included learners’ failure to respond, preference for passivity, lack of active verbal participation, selective answering, and shouted competing responses. It is suggested that L2 learning can be enhanced if teachers are aware of areas which may impede or promote L2 interactional opportunities in the classroom.


Introduction
With English being the official working language of the ASEAN Economic Community (AEC), in Thailand the Ministry of Education has legislated that all schools need to integrate English language learning into their curriculum. By doing so, it is hoped that English proficiency of Thai students will improve such that they are ready for local and global communication opportunities in the ever changing and increasingly competitive job market. However, despite many initiatives aimed at improving the English language competence of Thai students, Thailand is regionally shown to have very low proficiency, especially in oral English (see Atagi, 2011;Bruner et al., 2014;Educational Testing Service (ETS), 2017;EF Education First, 2019;Fredrickson, 2015;Khamkhien, 2010;Mala, 2016Mala, , 2019Nation, 2017;Noomura, 2013;Prapphal, 2001). Additionally, the National Institute of Educational Testing Service (NIETS) statistics from 2018 demonstrate particularly poor scores amongst high school students sitting the Ordinary National Education Test (O-NET). Grade 9 students managed to score, on average, just 30.45% for English, whilst Grade 12 were even lower at 28.31%.
There are myriads of reasons for such low English proficiency in Thailand, which all contribute to varying degrees. Noom-ura (2013) suggested that students' lack of responsibility for their own learning, low motivation, passivity during speaking activities, unchallenging English lessons, and L1 interference were all contributory factors affecting the learners' ability to master English language skills. One often noted reason for low scores is that students have very little exposure or opportunity to converse with English speakers, especially outside the classroom, thus being incapable of engaging in any extended English interactions (Boonkit, 2010). Other studies mentioned ongoing problems facing teachers such as large classroom sizes, lack of teacher training, insufficient time allocated for English language teaching, heavy workload, and inadequate teaching and learning resources (Baker & Jarunthawatchai, 2017;Bax, 2003;Hayes, 2010;Methitam & Chamcharatsri, 2011;Teng & Sinwongsuwat, 2015).
Therefore, amidst all these problems, it is crucial to examine what actually happens in the English language classroom, supposedly adopting Communicative Language Teaching (CLT), to determine how Thai teachers try to maneuver it to maximize learners' opportunities for interaction practice in the target language. With this in mind, researchers have also begun to consider other approaches to supplement current CLT-based instructional materials and strategies. Many are turning toward the integration of other sociocultural/interactional approaches to enhance CLT (Kirkpatrick & Ghaemi, 2011;Kramsch, 2014;Kustati, 2013;Sun, 2014;Wu, 2013;Young, 1999), and a number of studies in Thailand have suggested that CLT could be more effective if aspects of Conversation Analysis (CA) were incorporated into CLTinformed English lessons (Choopool & Sinwongsuwat, 2017;Kemtong et al., 2018;Sitthikoson & Sinwongsuwat, 2017;Teng & Sinwongsuwat, 2015;Waedaoh & Sinwongsuwat, 2018). By using CA to closely examine interactions occurring in the context of Thai classrooms, it is possible to identify which teaching strategies or practices may contribute to, or obstruct the development of Thai learners' interactional skills.
Observing the potential for a disciplinary fusion with keen interest, applied linguists have begun to adopt CA as a research tool for gaining insight into what actually happens in L2 classrooms (Bowles & Seedhouse, 2007;Koshik et al., 2002;Richards & Seedhouse, 2005;Seedhouse, 2005). Without a tool for close examination into actual teacherlearner classroom interaction, it would hardly be possible to determine real problems facing teachers in realizing their instructional goals. Therefore, by utilizing CA, it is possible to explore and breakdown interactions actually unfolding in Thai EFL contexts to identify the strategies teachers use to elicit learner talk and problem areas which may inhibit L2 interactional development. Moreover, utilizing CA allows a clearer understanding into how learner contributions can be shaped by teachers' interactional practices (Kemtong et al., 2018) and CA can, in fact, also be used as an analytical tool for probing and diagnosing problems learners encounter in using L2 in talk-in-interaction (Tantiwich & Sinwongsuwat, 2019).
In CA investigations of classroom interactional contexts, Seedhouse (2004) differentiated four areas of pedagogical focus which give rise to differences in classroom interaction; namely, procedural, task-oriented, form and accuracy, and meaning and fluency. In this study, the aforementioned taxonomy will be adopted and explained in more detail later in the paper. By employing CA as the analytical framework, this research will probe into what interactional strategies Thai teachers use to elicit learners' L2 talk throughout a lesson in English speaking classrooms in these four interactional contexts. Since there has been very little CA-informed research conducted in this area within Thailand, especially in the southern border region of the country, the outcomes of this research should uncover interesting and useful findings for the development of both students and teachers in the focal area and beyond.

Research Questions
The specific research questions to be addressed are as follows: 1. What strategies do (secondary school) teachers use to elicit Thai student's L2 talk in teacher-learner interactions in different contexts of an English-speaking class? 2. How do teachers use different elicitation strategies in each context? 3. What difficulties do teachers encounter when using these strategies to get the students to initiate/construct L2 turns, and how do they deal with these difficulties?

Conversation Analysis (CA) and Classroom Interaction
The primary objective of CA is to discover how participants function successfully in socially constructed talk-in-interactions (Seedhouse, 2004) through the analysis of the organizational patterns of talk from their perspective, often referred to as an "emic" view (Pike, 1967). Taking this approach to studying how mutual understanding between the participants are reached (Hutchby & Woofit, 1998), analysts step into their shoes with the question "why this, in this way, now?" when scrutinizing observable episodes of participant talk (Mori & Zuengler, 2008;Seedhouse, 2004). Besides sociology, disciplines where CA has recently been applied include both interactional and applied linguistics (Couper-Kuhlen & Selting, 2001;Koshik et al., 2002). Its natural link with applied linguistics is suggested by Seedhouse (2008), as well as Sert and Seedhouse (2011), due to CA's fundamental premise of language as action and language learning through social participation. CA has played an important role in language teaching by having its resources applied to different domains of research and practices, for example, the reflexive relationship between pedagogy (taskas-work plan) and interaction (task-in-process), the organization of classroom interaction, learning assessment, and conversation teaching and learning (Koshik et al., 2002;Pourhaji & Alavi, 2015;Seedhouse, 2008;Sert & Seedhouse, 2011). It also holds great promise for research into various aspects of classroom-talk-in-interaction, including problems experienced in its dynamic micro-contexts (Sert & Seedhouse, 2011). CA has additionally been shown to serve as an effective method for identifying and explaining difficulties experienced by learners as they navigate through these contexts to become more competent L2 speakers (Bowles, 2006). Barraja-Rohan (2011) andFujii (2012) concur that CA is a powerful tool not only to analyze L2 interaction but also to identify causes of interactional problems.
With CA, learners can also reflect on their own interactional practices, including problematic interactional experiences (Clifton, 2011).
Moreover, by utilizing CA to minutely dissect relevant teacher-learner interactional episodes, it is possible not only to uncover how language use is shaped and understandings are negotiated between the participants, but also to witness strategies teachers use to elicit learner talk during different phases in each lesson. By analyzing the transcribed data, patterns of language can be observed which highlight different strategies used by teachers at different times, and of those strategies, which provide learners with the best opportunities to engage in conversation and promote learner talk.

Teacher Talk and the Investigation of its Roles in Promoting L2 Participation
The roles teachers play in the classroom can often help or hinder students' ability to develop L2 proficiency. In communicative EFL classrooms, teachers are supposed to play important roles, such as mediators, facilitators, and monitors (Richards, 2006), and are expected to be equipped with skills in talk, needed to assist students in classroom interaction, often referred to as interaction strategies. Central to providing opportunities encouraging students' active participation in both teacher-student and peer interactions (Li & Walsh, 2011;Richards, 2006;Seedhouse, 2011;Walsh, 2011), these strategies include questioning, repair or error treatment, collaborative or interactive learning, and scaffolding (Rido et al., 2014). Confirming that teachers' interactional strategies directly influence learners' willingness to participate in communicative activities, Lee and Ng (2010) suggest that as facilitators, teachers can actually utilize strategies such as referential questions, scaffolding, and back-channel feedback to encourage active learner participation in class activities. Although previous studies have concurred that teachers' beliefs determine how communicative approaches are implemented, Chen (2020) suggested that the teachers' understanding of the approaches, and especially the strategies used, was a more determining factor.
During linguistic exchanges, a variety of interactional strategies are used by teachers to elicit answers from learners (Lee, 2006;Waring, 2012), and this is typically achieved by asking questions. A common feature of L2 classrooms are question and answer sequences, whereby the teacher asks in order to assess learning, check comprehension, introduce a new topic or material, and develop L2 ability (Jafari, 2013). Two common types of questions, display and referential, were discussed by Walsh (2011) as a way of teachers controlling the discourse. Display questions, are questions the teacher already knows the answer to, and often used to check learners' knowledge. Referential questions are genuine questions, whereof the teacher does not know the answer, and in language classroom settings are seen as being able to elicit more "natural" responses from learners, due to their ability to elicit extended learner turns and create opportunities for additional learner talk (Brock, 1986;Lee, 2006;Tsui, 1995;Walsh, 2006Walsh, , 2013. These strategies also serve as a basis for the analysis in the current study. The framework used for this research aligns with Seedhouse's (2004) perspective on the following classroom interactional contexts.
Procedural context. Every lesson features an element of procedural context, acting as a mandatory precursor to a following context, or as a summary at the end of lessons. The teacher's aim in this context is to inform learners of the organization of activities encompassing the framework of the lesson, usually accomplished via a multi-unit turn (MUT). Where turn-taking is concerned, a "traditional" L2 classroom ensues, where the teacher is the center of attention who controls the learner's turn sequences and participation. Although there are exceptions when turn-taking sequences vary due to emerging questions regarding clarification of the proceedings, flexibility of turn-taking procedures are able to be modified by teachers asking display questions, enabling learners to take turns, both at the beginning and end of lessons.
Task-oriented context. During this context, the teacher is required to direct the pedagogical aim toward the implementation and completion of various tasks requiring the learner to interact (hopefully in the target language) as independently as possible. Turn-taking sequences during these instances of interaction take on a reflexive nature. Any occurrences of repair are ultimately achieved by the learners themselves through negotiation of meaning and understanding, established during completion of the task. The fundamental purpose is task accomplishment through interaction, and any linguistic failures during these interactions are generally ignored. Task-based activities can efficiently promote communicative language use, as long as they are well planned and monitored by the teacher to facilitate co-operative peer work in the target language.
Form and accuracy context. Central to this context will be particular linguistic forms presented by the teacher, and expected to be accurately reproduced by learners. The main focal points are aimed solely on the accuracy of form production, and often involve repetition or rote learning, associated with methods such as audio lingual. When instruction is delivered, instances of turn-taking are formalized to the extent that the teacher acts as a conductor steering the sequencing patterns of interactions. These rigorously controlled turn-taking events are usually organized around adjacency pairs, and teacher prompt/student production patterns are optionally followed-up with some form of evaluation, or feedback, found in IRF/IRE (Initiation, Response, Feedback/ Evaluation) sequences.
Meaning and fluency context. Here, the maximization of student interaction is the driving pedagogical factor. Opportunities for language use and exchange are accentuated in order for students to engage in speaking activities related to the pedagogical environment they are studying in. To help facilitate learner interaction, pair or group work is often utilized as an organizational tool, whereby learners have opportunities for freedom of expression without fear of having to adhere to precise conventions of language. Therefore, fluency has the freedom to develop holistically as learners will be free to construct turns. They will also have more flexibility into managing their interactions due to the teacher taking a less "hands-on" role, acting more as a facilitator and monitor during the lesson. Role-plays are a popular and often effective speaking activity allowing for spontaneous language production. During role-plays, peer correction (which can help students develop language proficiency), may be witnessed.

Related Studies on L2 Talk in Classroom Interaction
In Thailand, a group-oriented culture of learning is often the norm, where individual learning or individual work is secondary to cooperative and group work, wherein learners excel due to their collectivist cultural orientation (Bruner et al., 2014). The use of a culturally-appropriate cooperative learning approach is promoted by Phuong-Mai et al. (2005), suggesting that Asian learners not only prefer this approach, but also are proven to perform better in groups, therefore allowing more opportunities for L2 interaction. If questioned individually, they are often reticent to respond to teacher questions in the target language for fear of giving an incorrect response, and in turn, losing face (Cortazzi & Jin, 1996). According to Burden (2004), this is one reason Thai students prefer not to participate in individual communicative tasks, ask and answer questions, or express opinions. Therefore, it is essential that teachers build rapport with learners to help create an environment with low anxiety levels, since this can affect communicative language development.
To analyze classroom interactions, an apparatus is needed. Without a tool for close examination into actual teacherlearner interaction, such as CA, it would be difficult to determine real problems facing L2 teachers in realizing their instructional goals. By utilizing CA to breakdown interactions occurring in Thai EFL contexts, it is possible to identify strategies teachers use to elicit learner talk, and identify problem areas which may inhibit L2 interactional development. Moreover, CA allows a clearer understanding into how learner contributions can be shaped by teachers' interactional practices (Kemtong et al., 2018) and can, in fact, be used as an analytical tool for probing and diagnosing problem areas learners encounter in using L2 in talk-in-interaction (Tantiwich & Sinwongsuwat, 2019).
Despite an expanding number of CA-inspired studies related to L2 learning and teaching conducted in the West, (Barraja-Rohan, 2011;Gardner, 2012;Kasper, 2006;Mori, 2007;Seedhouse, 2004;Sert, 2010;Sert & Seedhouse, 2011) in the East, the emergence of "applied-CA" or "CA-like" studies has been more gradual, especially in the Thai EFL context (see, e.g., Choopool & Sinwongsuwat, 2017;Kemtong et al., 2018;Makeh & Sinwongsuwat, 2014;Nookam & Sinwongsuwat, 2010;Pitaksuksan & Sinwongsuwat, 2019;Rodpradit & Sinwongsuwat, 2012;Sitthikoson & Sinwongsuwat, 2017;Teng & Sinwongsuwat, 2015;Ussama & Sinwongsuwat, 2014;Waedaoh & Sinwongsuwat, 2018). Most of these previously mentioned studies concluded that the implementation of CA principles to the teaching of English in conjunction with CLT cannot only help raise learners' awareness of naturally-occurring English conversations, but positively impact their conversation abilities. However, there remain needs in this context for the investigation into teachers' interaction with the learners to see how it shapes their interactional skill development. Incorporating CA principles into the study of the four contextual areas of language classrooms allows us to see which elicitation strategies should be recommended in each context, and which should be discouraged, thereby giving learners the potential for better interactional opportunities and the ability to further their communicative development.
Therefore, this research aims to add to the existing findings in order to promote Thai learners' L2 interactional development opportunities by highlighting elicitation strategies used by the Thai teachers in the four contextual areas of focus.

Methodology
This is a primarily qualitative, exploratory study of interactions between teachers and learners in an EFL classroom context. CA was applied as institutional discourse methodology to examine interactions within a Thai high school environment. The focus was not on random or spontaneous language exchanges, but on activities designed for the specific purpose of promoting and encouraging teacher-learner and learner-learner interactions in this particular context.

Research Focus
The pedagogical area of focus concentrates on turn and sequence patterns in the four different language exchange contexts previously discussed: procedural, task-oriented, form and accuracy, and meaning and fluency.
Thai EFL teacher-learner interactions in these contexts were examined in order to identify the strategies teachers used to elicit students' L2 talk in an English-speaking class, and to determine how these teachers used different elicitation strategies in each of these contexts in order to encourage learners' L2 talk. Additionally, the study also explored difficulties faced by the teachers when using different strategies to encourage learners to initiate relevant L2 turns and maintain L2 interactional sequences.

Research Setting and Participants
Classroom interaction data were collected from eight different English-speaking classes in two high schools in the South of Thailand, one in the province of Pattani, and the other in Songkhla. In the Songkhla school, two teachers were observed teaching two different classes for approximately 1 hour and 30 minutes each. The two Grade 10 classes ranged in size from 16 to 48 students. In Pattani, two teachers were observed teaching six different 45-minute classes for 4.5 hours. The Grade 10, 11, and 12 classes, had 16, 20, 23, and 24 students.
The public (government) schools were in relatively low populated areas; even though Songkhla is the provincial city, it is still considered a low-density area. The school in Pattani province is considered a semi-rural location, in a sub district town. Very little English is spoken outside the language classroom in either area. Although Songkhla has more opportunities, and some schools and universities have English programs with NES, or foreign teachers, but still very few compared to places which attract English speaking tourists, or expat communities.
In Pattani, the area is shaped by Muslim culture, with no apparent "western" influence. Outside school, Malay is the dialect often heard spoken by the locals. The region is not a destination much visited by foreign travelers, apart from Malaysians, and NES teachers are exceptionally rare. Therefore, the general English communicative proficiency of students in urban areas, such as Bangkok, sometimes with higher school spending budgets for equipment and teachers, more international schools, and better opportunities to practice English outside the language classroom with NES and other foreigners, could be considered as being better placed to encourage EFL learning than the rural, or semi-rural schools featured in this study. As suggested by Alhamami (2020), the communities language learners live in may influence their beliefs and attitudes, and, in turn, affect their L2 learning motivation.
The schools were selected using convenient sampling, with four Thai NNES teachers agreeing to partake in the research. They were approached based on their interest in improving their communicative English teaching, and were known as previous university research students willing to participate. All had similar qualifications and at least a few years of teaching experience.
The first two teachers are teachers A and B, for ethical purposes of confidentiality. Both are female, 24 and 25 years old respectively, and have taught for approximately 2 years at a high school in Songkhla, with around 1,614 male and female students. They taught in English for communication, and grammar. Classes were Grade 10 to 12. Students in each class varied, from 16 to 48.
The second school is located in Pattani. It has approximately 526 male and female students. Teacher C, a 26-yearold male, has taught at the school for 4 years, and teaches Grades 10 to 12. Teacher D is a 24-year-old female who also teaches Grade 10 to 12. The smallest class had 16 students, the largest, 21. The study focused on years 10 to 12. The reasoning was, having had between 9 and 12 years of EFL instruction, classes should have good opportunities for rich and informative data collection.

Data Collection
Seven hours of recordings were collected, which is regarded an acceptable number where findings can be reasonably generalized and conclusions can be formulated for any L2 research purpose (Seedhouse, 2004).
In the classrooms, two video cameras were positioned at the front and rear, and two audio recorders centrally, to capture as many students as possible. Although the researchers were aware of the observer's paradox, it was considered an unavoidable part of data collection. To minimize this, the equipment was introduced into the classroom as far in advance as possible, so the students had some familiarity with it being there.
The teachers were interviewed to ascertain why certain strategies were used, and what were believed to be the most successful in achieving learning objectives. Difficulties encountered when trying to initiate learner talk, and proposed or attempted solutions were also discussed.
The interviews were cross-referenced with findings from the CA to discover if any consistencies or irregularities were observed, and if they supported the interpretations of the findings. The interviews lasted approximately 1 hour, using audio recording and note-taking. The segments of data were transcribed and selected to help build a better picture of the findings.
The teachers and students participated voluntarily, and were informed that they could terminate their research participation at any moment without prejudice. Moreover, the participants were assured that collected data would remain confidential, and would only be used within the study context.

Data Analysis
The collected data were transcribed and analyzed through the lens of CA, as described in the literature review. The analysis involved selecting segments of interactions relevant to answering each research question, and transcribing each line of relevant material turn-by-turn, whilst utilizing Jefferson's (2004) glossary of transcript symbols as a model (see Appendix), and demarcating interactional sequences with respect to the taxonomy of classroom interactional contexts outlined previously, alongside determining the elicitation strategies used by the teachers in each context. The findings will be abbreviated using the following method to describe who is speaking line-by-line in the transcribed excerpts: Teacher (T) A, B, C, D; Learner (L) 1, 2, 3, etc.; and Learners (Ls).
Definitions of elicitation strategies. The following section will describe some of the strategies most utilized by teachers, and give examples of their use. Both display and referential questions are sub-categories of epistemic questions (questions seeking information).

Strategies Teachers Used to Elicit L2 Learner Talk
The findings related to the first research question will be examined in this section from the perspective of the previously discussed four contexts. The first contextual area of focus, procedural context, was identifiable at both the beginning and end of lessons. The findings will be presented as such.

Procedural context
Beginning of a lesson. The aim of the teachers at the start of all the lessons was to transmit information and organize the physical learning environment, introduce new topics, or reintroduce previous material. In the excerpts shown, teacher D's (T'D) class had 24 Grade 10 students (excerpts 1, 2, and 5), focusing on re-introducing a previous topic about telling the time.
Teacher C (T'C) had 23 Grade 11 students (excerpts 3 and 4) and introduced a new topic about the importance of the English language, and 16 Grade 12 students (excerpt 6) asked to describe the contents of their upcoming role-play.
Both teachers used display questions to check, or confirm the learner's knowledge at the present time. An example of such questioning is shown in lines 1 to 2 of the excerpt below, taken from T'D class.  By using this strategy at the beginning of the lesson, the teachers were also able to activate the learner's schema by reviewing material previously taught via a MUT, a turn with multiple turn construction units (TCUs), as can be seen in lines 13 to 28 of excerpt 2 below, also taken from T'D's class: The exception in this case, was T'B, who used this stage of the lesson to introduce new material to the learners via MUT, after a short ice-breaking activity. It was also noted that both T'A and B delivered their MUT in L1. This was not the only time L1 was used by our teachers during the procedural stage; T'A, B, and C all utilized L1 to translate during interactional exchanges with the learners to explain or reinforce understanding.
Scaffolding was also widely used by three of the teachers to encourage learner talk, usually in the form of Word Search Strategy (WSS), although, T'C extended on this by choosing to reformulate questions: Teacher echo was a strategy employed by two of the teachers (A and D), whilst acknowledgement tokens such as okay, right, and aha, or repetition were utilized by T' A, C, and D in order to give indirect feedback to correct learner utterances, or as an attempt to encourage further learner talk, as excerpt 6 from teacher C demonstrates below: For all four teachers, display questions in the form of confirmation checks were the most frequently and successfully used strategy for eliciting talk in this context, and when asked, some learners were happy to respond although many seemed happy enough to be passive observers and offered no individual oral input. Scaffolding also proved to be a popular strategy for eliciting talk for all but one of the teachers, with WSS being a favored and successful method of delivery. All but one teacher used L1 in one form or another to ensure or reinforce learner understanding, and the one exception (T'D), when interviewed, said that as a language teacher, exclusive use of L2 during class was an important strategy.
End of a lesson. At the conclusion of the lessons, and after a MUT, T'A, with 16 Grade 10 students (excerpt 7), aimed to reinforce the fruit and vegetable related vocabulary featured in the class by eliciting from the whole class answers to display questions in order to confirm comprehension. In another class, T'B aimed to recap the oral form focused lesson with 48 Grade 10 students (excerpt 8) by delivering a MUT in L1, before finally asking the boisterous class if they had any questions about the content material.
A MUT was the most used strategy during this context, with all four teachers utilizing this strategy to explain, deliver, or transmit information or instructions to the learners. However, the language used to deliver this information varied between all the four teachers. Teachers A and C both chose to use L1 and L2 to transmit information, with teacher A the only teacher taking the opportunity to review the material taught during the lesson. Teacher B chose to deliver the MUT exclusively in L1, whilst conversely, T' D used only L2. It is evident that three of the four teachers used confirmation checks with their students, as they did at the start of the lesson, however, only one teacher used display questions as a supplement, as can be seen below in lines 1 to 4 of extract 7, unlike at the start of the lesson wherein every teacher used this strategy: It is also worth noting that T'B used neither display questions nor confirmation checking as a strategy in this context, and asked only one formulaic expression in the form of a referential-related question as a conclusion to the lesson, as shown in lines 8 to 11 below: Interestingly, as was previously seen in context at the start of the lesson, L1 for translation and explanation purposes was used once again by the same three teachers, with only T'D choosing to use only L2 when interacting with students. Additionally, scaffolding, often utilized in the previously explored interactions, was utilized much less on these occasions. Only T'A and D decided to utilize scaffolding this time, with T'A, again relying on WSS, whilst T'D chose to change scaffolding strategies on this occasion, by modeling language and reformulating questions instead of using WSS, as they did previously. It is worth noting that positive feedback, which was a frequently used strategy by three of the four teachers at the beginning of the lesson, was only utilized by T'A in the closing stages of the lesson, with good and very good being used to motivate and encourage learner participation.
The observations of all four teachers during the procedural context in eight different classes, reinforced the information given during teacher interviews, which indicated that the short time available for classes (an average of 45 minutes), meant lessons had to be moved along quickly, leaving few opportunities for lengthy introductions of topics or materials, or teacher-learner interaction at the beginning of classes, or summaries at the end of lessons.
Task-oriented context. In these excerpts, T'A (excerpts 9 and 11) organized the 16 grade 10 students into four groups, with the aim of eliciting speaking practice around a piece of material. Each group's task was to think of different fruits or vegetables for other groups to guess by asking descriptive questions such as: Does it have seeds? Is it green? Is the skin rough, or smooth? Only one question could be asked by each group at a time, until a group attempted to answer correctly. There was a lot of interaction within the groups and between students, with the teacher taking the role of facilitator. Due to the competitive nature of the students, interaction and participation in both L1 and target language was witnessed, which indicates that the pedagogic goals appeared to have been met.
Teacher C (excerpt 10), focused on the topic of describing feelings with the aim of eliciting L2 interaction between the 16 grade 11 students, through the use of pair-work in preparation for role-play activities. Display questions in an IRF pattern were predominately used in order to check understanding. The interactional episodes between peers and the teacher were spontaneous, and the majority of students looked comfortable adding input before, during, and after the proceedings. A lot of laughter was heard during the interactional exchanges, and although there was some L1 used, the students seemed to use the target language in a natural way, and looked able to manipulate and adjust their L2 output, when needed in order to complete their task.
When introducing the material, interactions between teacher and learners were predominantly Initiate Response Feedback (IRF) in nature, with T'A using display questions in the form of confirmation checks, whilst T'C preferred to utilize referential questions to elicit information from the learners working in different groups. Scaffolding was an elicitation strategy extensively used by both teachers, with Word Search Strategy (WSS) being the preferred method for T'C, whilst T'A extended on this strategy by also choosing to reformulate questions, which can be seen in lines 1 to 5, and also used corrective repair in line 12 to further learner understanding in an attempt to encourage learners to elaborate and offer more talk, as can be seen in the following excerpt: Another strategy which both teachers employed was teacher echo. Teacher A favored this approach to confirm learners' correct utterances and as a sign for the learners to initiate another turn in an expanded sequence, which is shown here in lines 8, 12, and 16 of Excerpt 9, above. Whereas, T'C used this strategy as a type of confirmation for the students of a correct answer, and follows on to give positive feedback in lines 53 to 54. Both teachers also used positive feedback throughout, with acknowledgement tokens such as; good; very good; that's perfect; and excellent (as illustrated below in excerpt 10). T'A also chose to use L1 as a translation strategy which encouraged learners to supply correct information using L2, once the correct answer was provided, the teacher would encourage the learners to translate back into L1, to confirm and reinforce learner understanding. This strategy worked very well with this class and the majority of students were keen to participate in providing answers in both L1 and L2, as shown in excerpt 11 below. Teacher A utilized L1 again to perform a MUT prior to the start of the group work stage, to express satisfaction that the students understood the content material and were ready to move onto the next stage of the lesson.
Both teachers tried to build a rapport with the students to make them feel comfortable when sharing their work with the rest of the class, which appeared to enhance the communicative language exchanges between teachers/students, and students and their peers. According to McFarland (2012), learning is mastered easily when students learn in a positive environment where they feel they are supported in whatever they are doing. Participation for students in the observed classes was generally not a problem, each group seemed to be well versed on what was expected of them, and set off to complete the assigned task without hesitation, once group understanding was achieved.
Form and accuracy. In this context, teacher B's lesson had a large class of 48 grade 10 students. The main goal was practicing correct forms, especially focusing on elicitation of oral patterns, intonation, and pronunciation. The students seemed to be quite boisterous, with many shouted competing responses whenever a question was asked. Moreover, the same students seemed to interact, whilst the rest were passive observers, or simply not focused on the lesson. On the other hand, T'D's class consisted of only 20 grade 11 students, and was much more orderly. The focus was on verb forms associated with ability. The goal was to produce previously taught forms accurately, by giving examples. Most responses were elicited from individual students rather than the whole class, although, when the whole class were asked display questions, the same three or four learners answered most of the time. Albeit, different in size and focus, both classes favored memorization through repetition and rote learning. Teacher talk was observed more during this context than any other, with little opportunity for any spontaneous learner L2 outside the area of focus.
Both teacher B and D used display questions to check learners' present understanding of the lesson content, with T'B also choosing to use referential questioning on one occasion in order to elicit opinion from the students:

Excerpt 12
[T'B, L Cls_48 Ss, G10, Tm_32:56-33:11] 12 T: "who do you think is the best?" ((lots of class participation, calling out)) 13 L3: number three 14 T: "okay we will agree with the popular one" Teacher echo was used by both teachers, although, more favored by T'D on these occasions, and used as a type of confirmation to signify correct learner utterances. Teacher B used this strategy to provide examples of correct language use, asking the whole class to repeat and echo to produce correct forms of intonation, as seen in the following excerpt: and echo to produce correct forms of intonation, as seen in the following excerpt: A MUT was only used by T'D, to directly repair a previously incorrect student utterance and end the turn with a display question, before moving on to shift the focus in another direction. As mentioned previously, display questions were utilized by both teachers throughout the excerpts, although, T'B's students generally delivered whole class responses, whereas, the students from T'D's class added to this, and were also more inclined to offer responses and participate individually.
Both teachers utilized a form of direct repair when dealing with incorrect utterances. Teacher B, used this to instruct a learner to alter their intonation when pronouncing a word at the end of a sentence, whilst T'D chose to highlight a grammar point to reinforce learner understanding, by reminding students that be able to can be used in any tense, seen in lines 7 to 9 of excerpt 14 below: Direct repair can be used to ask the learner to repeat the correct form presented to them by either the teacher or a peer. On these occasions, both teachers chose not to do so, and simply chose to highlight a point, or problem, as a type of direct feedback without learner contribution.
Positive feedback, employed as a motivational tool to encourage learner participation, was not used by T'B on this occasion, with T'D only using sparingly. The few instances of use came from T'D in the form of acknowledgement tokens such as, okay, seen in line 7 of excerpt 14.

Meaning and Fluency
These classes aimed to promote oral fluency by allowing students to express themselves via contextual role-plays. Teacher D's lesson (excerpt 15) was based around the topic of relationships, whilst teacher C (excerpts 16 and 17) focused on sports activities. Both classes consisted of 16 grade 12 students. All 3 excerpts feature teacher-learner interactions at either the beginning or end of learner role plays. Teacher C preferred students to work in pairs, whilst T'D chose small groups, even though only two students produced any meaningful language during the interactions; therefore, in hindsight, using pair-work would have been a better choice in this class. Overall, even though some of the role-play dialogues were pre-planned by the learners, they looked quite natural as the students performed them without having to refer to any scripts for guidance.
In these instances, at the beginning of the interactions between the teacher and learners just prior to the commencement of the role-play, both T'C and T'D chose to utilize referential questioning, seen in line 1 of excerpts 15 and 17, for the purpose of eliciting information from the learners in relation to the content of their upcoming role-play presentation. The teachers provided opportunities for extended learner turns, and both chose to elicit talk by giving learners space to interact, utilizing strategies such as the acknowledgement token "uh↓ huh↑" which can be seen in line 6 of the following excerpt. The occasions provided to the students allowed them to interact with both their teacher and peers. A good example of this can be seen in line 8 of excerpt 15, where L2 self-selects and joins in the interaction without feeling the need to be invited. The students all looked and sounded comfortable whilst interacting in L2, with no noticeable communication breakdowns witnessed by the researchers during either of the classes.
As the teacher-learner interactions progressed, both teachers, again, used acknowledgement tokens as a feedback strategy, by utilizing expressions such as; okay, right and aha. Moreover, positive reinforcement was another feedback technique both teachers employed to motivate and encourage learners to extend turn-taking episodes, with language such as; okay, good, wow!, and very good, which can be seen at the end of the following excerpt in line 61: Scaffolding, a strategy often featured in other contextual classroom extracts in this research, was only witnessed on one occasion during these two meaning and fluency excerpts. Teacher C chose the single use of this strategy through reformulation of a question, which succeeded in its intention via a correct learner utterance, as seen in line 3. Once student role-play presentations concluded, both T'C and D utilized content feedback as a strategy to summarize the learner's performance, process, and content of material delivered. Both chose to dissect the interaction along with the whole class in order to clarify and reinforce different sections of conversation by interacting and re-enacting passages contained in the role-play. When asked what students said, or how they responded at certain times during the presentation, most learners offered answers which correctly corresponded to the role-play, with plenty of active participation from the students in both classes.
To summarize the role-play interactions, both teachers asked questions such as; "how did they begin the role-play?", "how did they close the role-play?" and "how did they answer?" The use of confirmation checking questions to ask what was said, and by whom, at a particular time during the role play was responded to actively by most of the learners in both classes either individually, or as a whole, with a level of competence that would suggest a good general understanding of the material presented during these particular extracts, alongside evidence of fluency in the use of L2 by a number of students during both lessons.

How Strategies are Used to Elicit L2 Learner Talk
The findings related to the second research question of this study will be examined in this section, focusing on the how teachers used different elicitation strategies in each of the previously discussed four contexts.

Procedural context
Beginning of a lesson. Here, all four teachers used a MUT to either review material or introduce new material at the start of the lesson. There was much less student participation compared to other contexts, particularly at the start of a lesson, as the pedagogical goal at this stage was to transmit information, organize the physical learning environment, introduce or summarize the contents or materials, or conclude an activity (Walsh, 2012). As MUT's were used by all four teachers, few opportunities existed for student language production, as they were mostly required to listen.
Both T'A and T'B delivered their MUT in L1, whilst T'C and T'D preferred to use L2. Benson (2004) suggests that when L1 is used, learners feel more confident in expressing themselves, whilst the teacher is able to discern what has been learnt, what remains to be taught, and if any students require further assistance, explanation, or clarification of anything related to the lesson. When interviewed, most teachers confirmed their agreement, and stated the use of L1 helped, especially if something was needed to be explained.
Display questions were the most used strategy by all four teachers, to check or confirm comprehension. Lightbown and Spada (2006) state that they are often used, in fact not because the teacher is interested in the answer, but simply to check the learners' knowledge of the language being discussed at that time.
Scaffolding was utilized by all but one (T'B), with WSS most used. Applebee (2002) described scaffolding as accomplishment of new or difficult tasks with the assistance of a more knowledgeable "other." Teacher C chose to extend this strategy by modeling language or reformulating questions. During these interactions, there was evidence of L1 usage by three teachers (A, B, and C) to help translate and/or explain information.
Teachers A, C, and D, used positive feedback to encourage or motivate the learners, whilst, T'A and T'D used teacher echo to help reinforce correct learner utterances. However, Walsh (2012) cautioned that even though echo is a commonly used phenomenon, teachers should be aware of how often they use this strategy since it may actually disrupt the flow of communication in the classroom, thereby depriving learners' opportunities for interaction.
End of a lesson. At the close of lessons, all teachers, again, delivered a MUT, with T'A choosing to review material in both L1 and L2. The three remaining teachers chose to give instructions, or explain what to expect in the next lesson, with T'B employing L1, T'C using both L1 and L2, and T'D choosing to use L2 exclusively (which they also did at the beginning of the lesson). Previous studies have shown that the occasional use of L1 by both teachers and students can increase L2 comprehension and learning (Cook, 2004;Tang, 2002).
Display questions and confirmation checks were again, the most used strategy by all but T'B, who chose not to use them. Referential questions were also used by both T'B and T'D to seek learner's opinion. In CLT, referential questions are generally preferred over display questions to elicit language due to the higher order thinking required (Richards & Schmidt, 2009;Wright, 2016). However, their use should be dependent on the level and ability of the students being taught.
Scaffolding was only used by T'A and T'D in this case, with WSS favored by T'A, whilst T'D preferred to model language and reformulate questions. Lastly, T'A was the only teacher to extended strategy use by choosing to employ both positive feedback and teacher echo.
Task-oriented context. The strategies utilized during these episodes were often found to be aligned with IRF interactions. According to Walsh (2011), IRF is the most commonly featured discourse structure found in classrooms worldwide. During these sequences, the teacher usually controls the topic, allocates turns, and decides how the focus of attention is divided between students (Erickson, 2004). The observed exchanges proved to be a useful way to involve students, both individually and collectively, by practicing language related to the material.
The students were attentive when listening to instruction, and interacted with the teachers before, and their peers during, the group task stage. Teacher A used display questioning in the form of confirmation checks, preferring to seek both whole class, and individual responses. The same strategies were used by T'C, although more often with individual students. Although employed, referential questioning was used much less. Scaffolding, via WSS was used by both T'A and T'C; moreover, T'A chose to extend on this by reformulating questions and using corrective repair.
Teacher echo and positive feedback were favored by both T'A and T'C during these episodes, although, as Walsh (2012) previously suggested, teacher echo may disrupt communication. Conversely, it may also be given for the benefit of the class to help ensure the whole class understands in union, and moves forward at the same time, so that everyone is kept "in the loop." Teacher A, again, extended strategy usage by employing L1 to translate and explain with whole class participation. After analyzing all episodes of teacherlearner interactions in this study, T'A's lesson was proven to have the highest number of different strategies used in any one class. This is an interesting finding, as the observations also confirmed that the participation level of learners definitely appeared to increase more than in other classes.

Form and accuracy.
Here, display questions were the most used strategy, employed by both T'B and T'D. T'B chose to extend on this via referential questioning. With a large class of 48 students, T'B usually asked the whole class for their responses to questions, which created a lot competition via shouted responses from the noisy class. The class took longer to settle down and appeared less disciplined when compared to the smaller sized classes. Finn et al. (2003) suggested that smaller sized classes had fewer episodes of student misbehavior or disruptiveness due to learners having more opportunities for social interaction with the teacher, and less for messing around, or unruly behavior. Teacher echo was also evident with both T'B and T'D choosing to utilize this strategy, usually to confirm correct utterances. When repair was deemed necessary, both teachers decided upon the use of direct repair, which Fan (2012) suggests, is used to foster teacher-student and student-student interaction by asking students to repeat the correct form presented to them, either via a teacher or peer. During these classes, the only positive feedback was given by T'D.
Lastly, T'D initiated a MUT in L2 to correct a learner response and clarify a grammar point. When interviewed, she stated her belief in the importance of limiting L1 use to maximize the potential for target language acquisition. Conversely, T'B used L1 on many occasions and believed it was beneficial when explaining new topics or unknown concepts quickly, especially with limited lesson time. According to Ellis (2008) and Turnbull (2001), using L1 can motivate learners to learn L2, as it not only saves time, but also can help shape their conceptualization of learning.
Meaning and fluency. The aim of the teachers was to allow students to express themselves clearly, which, Bowles and Seedhouse (2007) agree, is an important pedagogical goal. The most frequently used strategies by both T'C and T'D were quite similar in scope. Firstly, referential questions were often used, followed by display questioning in the form of confirmation checks. The question-and-answer sequences were kept relatively short compared to other contexts, and the teachers indicated that this was for two reasons. Firstly, they wanted to allow the maximum amount of time for learner talk as possible, and secondly, the short time for each class did not give the teacher any choice other than to move the class along quickly to cover the class material. By allowing students more freedom and opportunity to interact, the observations indicated spontaneous language use, and fluency between most students. Acknowledgement tokens were used occasionally by both T'C and T' D as they attempted to extend learner talk and to provide feedback. The use of these tokens assures the current speaker a turn space (Wong, 2000), thus helping encourage the learners to continue talking.
Feedback was also frequently used by T'C and T'D, with both using positive and content feedback. Hyland (2006) claims that providing students with whole-class feedback is one of the most important teacher tasks, as the type of individual attention given is often rarely possible during class proceedings. Lastly, scaffolding, often used in other contexts, was only used by T'C at this time.

Elicitation Difficulties and Strategies to Overcome Them
The findings related to the third research question will be examined, focusing on difficulty's teachers encountered when encouraging learner talk, and how difficulties were dealt with. During the teacher/learner interactions, some elicitation challenges were observed, such as failure to respond to questions; lack of active participation; only specific students/groups answering/participating; students shouting out simultaneously (especially in large classes); and a preference for learner passivity.
The following two excerpts from T'C and T'D's classes show examples of elicitation challenges they faced, and the strategies chosen to motivate students to produce the target language.
In excerpt 18, T'C finds the learners lacking in active participation when asking students' opinion regarding presenting. In line 6, when the teacher does not feel the learners participating, "come on" is used as encouragement for a learner turn. An incomplete learner turn occurs in line 10, followed by L3's utterance "er, er, concentrate," which the teacher acknowledges and extends with "you must have" in line 15, again as a form of scaffolding (WSS), which encourages firstly L3, then Ls and teacher echo responses of "concentrate" in lines 16 to 18. This is then rewarded with positive feedback of "good" in line 18 to motivate further learner interaction.  Weimer (2009) suggests that the lack of active participation amongst some students may be due to overly motivated or overly participating students in a classroom, inhibiting other's chances to speak or act. This was consistent with the classes observed, with particular learners or groups of learners interacting with the teacher the majority of times. During the teacher interviews, T'A concurred that during one class, despite knowing the correct answer, one student admitted that she was not confident enough to speak due to fear of making a mistake in front of her peers. This was the general consensus of all four teachers, when interviewed.
Other factors to consider are learners not talking freely due to culture or background, where students are expected to listen rather than interact with the teacher. Bruner et al. (2014) suggests that Thai learning styles are affected by group-oriented Thai culture, where the preference is for cooperative and group work, as opposed to individual learning and individual work. Due to Thais' collectivist cultural orientation, students progressed better when participating in group work, and individual communication, such as asking or answering questions, expressing opinions, or volunteering, are often shied away from for fear of "losing face" (Bruner et al., 2014). In order to minimize learner anxiety, Burden (2004) proposes creating a learning environment where teachers play an active role in building rapport between learners.
In excerpt 19, T'D begins with a referential question (lines 1 and 2) to elicit learner's opinion regarding the role play performance of their peers. When no response is forthcoming, she reformulates the question in line 4 with "have some suggestions?" responded to in line 6 (although the voice was only loud enough to be heard by the learners very close by). After a brief pause, she acknowledges the utterance with "okay" in line 10, then asks which pair they liked the most. She then gives a shortened version of the question with the teacher echo "which pair?" (line 12). When the learners fail to respond again, she resorts to using L1 to translate the question "which pair do you like the most?" (line 14), prompting some muttering, laughing, and pointing/ gesturing amongst the students, (although no learner talk is forthcoming). In line 18, she again asks a referential question regarding why the learners chose a particular pair, proving successful when L2 answers in line 20 with "it's natural". The teacher echo's this response in line 22, along with the acknowledgement token "okay", before moving on. During interview, teachers A and B explained that when the students are being taught by a NES teacher, they are forced to use L2, and are more comfortable doing so, whereas when the teacher is a Thai/NNES, they focus more on grammar form and accuracy when they speak, and are less inclined to use L2, which may be one reason why L1 was frequently observed during the lessons. Teacher D believed the exclusive use of L2 during class to be beneficial for the students, but found it difficult, as some students may think the teacher is not friendly. Therefore, she felt using L1 at certain times reinforced understanding and encouraged participation. Some scholars have actually stressed the benefits of using L1 in EFL classrooms, such as Cook (2004), who suggested that using L1 achieved more comprehensible input and target language production. Furthermore, Li (2000) maintains that code switching between L1 and L2 can be viewed as bilingual interaction, and a sign of strength rather than deficiency in either of the languages.
One strategy not often witnessed was extended wait time. Mak (2011) proposed that students have a better chance of developing confidence when teachers pause for 3 seconds or more, in turn, affecting their willingness to participate. The wait time during classroom interactions between participants was seldom that long, with teachers often choosing 1 second or less before initiating another turn. This may be due to time constraints, as each teacher, mentioned a busy schedule each day which needed to comply with the school curriculum, thus limiting the time available. Additionally, with average lessons lasting 45 minutes, a sense of urgency to keep the class moving along quickly was observed.

Conclusion
The purpose of this qualitative study was to provide empirical evidence of EFL teacher-learner interactions in four micro-pedagogical contexts and fill a gap in CA-informed research concerning classroom interactional practices in Thailand. The findings suggest that, in general, the teacher orchestrated strategies aligned with the pedagogical goal in the four different contexts during most of the classroom interactions, and by using CA as the framework in this study, it was possible to see patterns of elicitation strategy use at the micro level.
The observations discovered, when teachers used a wider variety of strategies, there appeared to be more active learner participation. The findings also provided insights into how particular strategies were utilized, and showed that display questions were the most used strategy between the four contexts. Some elicitation challenges were faced, and strategies used to overcome them were witnessed, such as scaffolding or L1 usage, alongside encouragement and motivation strategies. It should be noted that some factors which may influence learners' language development, and are outside the teacher's control, such as large class sizes, curriculum constraints, short classes, lack of resources and students' motivation to learn English, were all evident during classes.
Moreover, teachers L1 use swung from prohibition in some classes, to extended periods of integration. This often depended on the teacher's opinion regarding the benefits of L1 use, the material taught at a particular time, or the comprehensibility level of learners. This area of classroom interaction is under-researched, especially in the southern border region of Thailand, where more than one L1 is often used. It would be interesting to study how these mother tongues are positively integrated into lessons as a tool used by both teachers and students to help maximize L2 development, and help to bridge the gap between the unfamiliar and familiar, thus lessoning L1 dependence. Previous studies have shown the benefits of L1 helping to scaffold learning, although, the researchers are unaware of any such studies in this diverse region, and future research could potentially provide interesting and useful findings.
It is hoped the insights gained from this research will be helpful to EFL teachers and educators in general, to assist in improving the implementation of interactional classroom practice, and add to EFL research already undertaken in this expanding field. It is supported that if teachers are aware of strategy use and development, they will be able to not only improve their strategies, but use language convergent to the pedagogical focus during different contextual episodes, thereby giving learners better opportunities and space to participate in, and develop interactional competence. Nevertheless, there were some limitations of utilizing CA particular to this research, such as the selective nature of data analysis in the four contexts, alongside the inability to comprehensively analyze everything collected due to the restricted time available; therefore, the generalizability of the findings are subject to certain limitations due in part to happenings which appeared in this research.
[ point of simultaneous speaking (of two or more people) ] end point of simultaneous speaking = talk by two people which is contiguous (i.e., not overlapping, but with no hearable pause in between) OR continuation of the same turn by the same speaker even though the turn is separated in the transcript (0.2) the time (in 10ths of a second) between utterances (.) a micro-pause (one tenth of a second or less) wo:rd sound extension of a word (more colons = longer stretches) word.
fall in tone (not necessarily the end of a sentence) word, continuing intonation (not necessarily between clauses) wor-an abrupt stop in articulation word?
rising inflection (not necessarily a question) word (underline) Emphasized word, part of word, or sound ↑word rising intonation ↓word falling intonation°w ord° talk that is quieter than surrounding talk hh audible out breaths .hh audible in breaths w(hh)ord laughter within a word >word< talk that is spoken faster than surrounding talk <word> talk that is spoken slower than surrounding talk (word) approximations of what is heard ((comment)) analyst's notes' "word" idiomatic translation of Thai utterances

Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This study was supported by the Graduate Studies Scholarship Contact under the Thailand's Education Hub for the Southern Region of ASEAN Countries (TEH-AC) Scholarship Program Graduate School, Prince of Songkla University (0997/2651), alongside research grant (6111121008). The funding source had no role in the study design, analysis, data interpretation, or decision to submit for publication.

Ethical Approval
This study was approved as part of the longitudinal research project titled "Reforming English conversation teaching with a Conversation Analysis-(CA-) informed Teaching Model (CA-T Model)," with the approval number PSU IRB 2019 PSU L 006, issued by SBSIRB, Prince of Songkla University.