EMERGENCY REMOTE TEACHING IN INDONESIA: A MISSED OPPORTUNITY FOR GREATER LEARNER AUTONOMY

: The sudden switch to learning from home during the COVID-19 pandemic impacted teachers across the world. In Indonesia, schools were closed from early March 2020 onwards. This paper presents the findings of a qualitative research project that investigated how Indonesian teachers of English responded to the challenges of Emergency Remote Teaching (ERT) and whether ERT would lead to greater learner autonomy. Ten teachers responded to an invitation to participate in focus groups and individual interviews on Zoom and to contribute examples of their lesson plans from the lockdown period. All teachers found that WhatsApp was the most efficient and effective platform for remote teaching, allowing synchronous and asynchronous sharing of audio, video and text-based materials. Despite the challenges of poor connectivity and lack of face-to-face contact, the teachers were able to continue involving their students actively in integrated, communicative tasks that pushed them to extend their communicative competence. Unexpectedly, however, the move to online teaching did not herald a shift towards greater learner autonomy. The data from this research shows that English language teaching in Indonesia is still firmly teacher-controlled despite the affordances of online learning.

Learner autonomy means that learners take increasing responsibility for their own learning (Little, 2007). It is supported by the concept of lifelong learning: students who learn to become autonomous are empowered to go on learning long after they leave school -an attribute that is increasingly important in the rapidly globalizing and interconnected world of the 21st century. Studies have also shown that learner autonomy is closely correlated with good language learning outcomes (Jianfeng, 2019;Lengkanawati, 2017;Melvina & Suherdi, 2019;Myartawan et al., 2013). This study explores the role of learner autonomy in English language teaching in the first six months of ERT.
Early articulations of learner autonomy focused on language learning strategies, both cognitive and metacognitive. It was argued that through training in such strategies, students could gain the capacity to take more autonomous control of their learning. At an early stage, students would need guidance from their teacher; Littlewood (1999) called this 'reactive autonomy'. However, students would be able to take increasing control of their own learning. In a similar vein, Nunan (1997) elaborated five levels of learner autonomy: awareness, involvement, intervention, creation, and transcendence (see the description of each level in Appendix 1). Such views of learner autonomy suggest that students need to be self-motivated and have a sense of self-efficacy as well as a belief in an internal locus of control in order to seek out opportunities to learn and maintain intrinsic interest in learning (Ushioda, 2011). More recently, a study by Budianto and Mason (2021) involving three participants in an ESP (English for Specific Purposes) program at a non-English department at an Indonesian university identified four requirements for students in realizing autonomous learning: "1) a willingness to accept responsibility; 2) dedicated planning; 3) implementing effective strategies, and 4) monitoring progress." (Budianto & Mason, 2021, p.1).
Other studies focused on how to achieve autonomous learning from the teaching perspective. For example, Schweisfurth (2013), who discussed student-centred teaching in the context of developing countries, argued that student-centredness lies in a continuum: at one end of the continuum is the traditional, authoritarian 'chalk-and-talk' of transmission-based teaching. At this end of the spectrum, teachers maintain control of what and how students learn, even though they may take an interest in their students' welfare and try to make their classes relevant to the students' needs, their lives and social context and their interests. At the other end of the continuum, students are engaged in independent or group inquiry, negotiating meaning and making their own choices about what and how they learn. In other words, this end of the studentcentred spectrum merges with learner autonomy. Similarly, Li (2015), on the basis of her research into teachers' roles in fostering autonomy, suggested that teachers should provide English learning strategies instruction, monitor and evaluate students' learning, address the affective factors in language learning, and be prepared to vary their role in different contexts.
It is often assumed that learner autonomy is a natural fit with online learning. Online learning opens the door to the vast resources of the internet from which learners can potentially choose materials and activities which will support their learning goals and preferences. For example, Nguyen and Stracke (2021) pointed out on the basis of a large, mixed-methods study in Vietnam that online learning can encourage learners to make choices, collaborate with others, and manage their own learning pace. Studies have also shown a correlation between learner autonomy and ICT skills. For example, a quantitative study with a non-experimental and cross-sectional design conducted by Nina-Cuchillo et al. (2021) revealed that students' ICT and learning autonomy correlate. Similarly, Evita et al. (2021) in a quantitative correlational study involving 97 participants from an English department in Indonesia found that students' degree of learning autonomy correlates with their ICT competence. Reinders and White (2016) point out that over the past two decades, developments in pedagogy in both ICT and autonomous learning have facilitated an increasing "fusion" or "convergence" of the two in ELT research and practice. They argue: "we are currently entering a phase in educational practice and thinking where the use of technology is enabling a shift of focus away from the classroom-and indeed in some cases formal educationtaking instead the learners' lives and their experiences as the central point for learning. Our understanding of how learners design their own learning experiences and environments and the role technology plays in this design are starting to merge, requiring a re-visioning of the role and shape of education" (Reinders & White, 2016, p. 2) Thus, with the move to remote and online learning during COVID, it could be envisaged that greater learner autonomy would be achieved in the Indonesian context. However, Indonesian teachers of English may not be ready to promote learner autonomy, given the traditionally teacher-centred, didactic nature of the Indonesian education system (OECD, 2015). Lengkanawati (2017), like researchers in many other contexts (e.g., Borg & Alshumaimeri, 2019 in Saudi Arabia; Kartal & Balcikanli, 2019 in Turkey;Nguyen, 2016 in Vietnam), found that Indonesian teachers valued learner autonomy but believed that it was not a feasible goal. They argued that students lacked sufficient language proficiency and relied on guidance from teachers and/or parents. They also argued that there was limited time in the curriculum, and the strong focus on the national exam militated against learner autonomy. Here it is important to note that the national examination in the Indonesian context has recently undergone a monumental shift from high-stakes testing to low-stakes testing (Defianty, 2021).
In fact, autonomous learning should have particular relevance in the Indonesian context, where classes are large and teachers have a heavy workload. As Smith et al. (2018) argue, working towards learner autonomy may be particularly important in contexts where classrooms are often over-full, under-resourced and typically teacher-fronted. Further, the COVID-19 pandemic has demonstrated that learning online can be an alternative to face-to-face learning, as documented by several studies about Indonesian students' readiness for online learning (Churiyah et al., 2020;Dwiyanti et al., 2020;Junus et al., 2021;Saintika et al., 2021;Wulanjani & Indriani, 2021). Findings from those studies confirmed Benson's (2013) argument that in the digitalized modern world, there are myriad opportunities for informal, autonomous learning and that students who engage in "translingual digital practices" (p. 841) as they make social contacts online, follow their hobbies, or enjoy popular culture may be learning a language without conscious intention. Indeed, Lamb (2013) found students in rural Indonesia were highly motivated to learn English and participated enthusiastically in "translingual digital practices" (Benson, 2013), accessing the internet, chatting online in English, watching movies and so on. However, they did not see this as language learning. This finding was similar to Nguyen and Stracke's (2021) study in Vietnam which found that first year university students were keen to learn autonomously outside class, but still regarded the classroom as the realm of the teacher.
Several potential factors can contribute to the development of autonomous learning in Indonesia. First is the current national curriculum, Kurikulum Merdeka, which explicitly calls for autonomous learning as one of the elements of Profil Pelajar Pancasila -the expected characteristics and competences of school graduates based on the Pancasila. Second, several studies in the Indonesian context document that learners have positive perceptions towards learner autonomy (Budianto, 2014;Jannah et al., 2022;Melvina et al., 2021). Third, several studies have also shown that autonomous learning is achievable in the Indonesian context (Padmadevi et al., 2020;Pondalos et al., 2022).
However, it is clear that if students are to benefit from the plethora of resources and tools available online, they need training and guidance. The literature on online learning and also on autonomous learning consistently stresses the importance of teachers' role in providing structured training and support for students' participation (Reinders & Hubbard, 2014). Teachers need to learn new skills in providing support for autonomous learning, in guiding students in setting goals and working towards them, and in curating online materials for their students' learning.
Given the potential of online learning, we wanted to learn how Indonesian high school teachers responded in the context of ERT. In particular, we asked: 1. How did teachers cope with instruction in the ERT context?
2. To what extent did teachers foster autonomous learning in their teaching of English during ERT?

METHOD
In order to investigate Indonesian English language teachers' responses to ERT and to answer the research questions, we conducted a qualitative, multisite case study involving two focus groups and 10 individual in-depth interviews at the end of the first two terms of ERT in July 2020. Because of the constraints of the COVID-19 pandemic, these interviews and focus groups were conducted and recorded via Zoom. We called for participants through a WhatsAppbased professional learning community. Ten English language teachers from different parts of Indonesia volunteered to participate in the study. All the participants were senior high school teachers, but their length of experience and qualifications differed (see Table 1). Before the research was carried out, the participants were informed about the focus of the research and assured that their personal information would remain unexposed. A consent form was signed by each participant prior to collecting data. In addition to the focus group discussions (FGD) and personal semi-structured interviews, the teachers also shared examples of lesson plans and samples of students' work. Pseudonyms are used in this article to protect the participants' identities and those of their students. Since all the participants were English teachers, with long experience of teaching English, the FGDs and personal interviews were conducted in English. Each FGD lasted for 120 minutes, and each personal interview ran for 30 to 60 minutes. This paper investigates whether the ERT context encouraged teachers to foster autonomous learning in their instruction. The FGD protocol was structured around broad questions such as: "How has the sudden switch to distance learning as a result of COVID-19 affected your teaching?" while the interview questions focused more specifically on learner autonomy, shown in Appendix 2.
Data analysis followed the procedure for grounded theory recommended by Creswell (2005, p. 230) and Braun and Clarke (2006) to draw out themes from the collected data. The FGDs and interviews were transcribed verbatim; the transcriptions were first annotated in Word and then coded using an Excel spreadsheet. Each response provided by the participants was ascribed multiple codes. These codes emerged from the data but were also guided by our reading of the literature on learner autonomy, as summarized in the literature review. Both researchers independently coded the data and then reviewed and aligned with each other's coding. Through this iterative process of coding, re-coding and clustering, a series of themes gradually emerged from the data.

Findings
This paper investigates teachers' teaching practice in the ERT context and whether or not the shift from onsite learning to online learning encouraged teachers to promote autonomous learning. Three themes emerged from the FGD and personal interviews: first, the challenges facing teachers and students in the new context of ERT; second, the teachers' ability to promote communicative language teaching in the online context; and third, the extent to which learner autonomy was realised.

Theme 1: The Challenges of ERT
As in many countries, the decision to send students home came with short notice. Hence, it is a little surprise that a major theme emerging from the data was the challenges which faced the teachers.
First of all, according to the teachers, many students found it difficult to access the internet. Many of them did not have devices at home, though some were able to borrow a phone from their parents or friends. Three of the participants stated that their respective schools are located in remote areas where the internet signal is unstable and consequently students often dropped out of online sessions with no warning. A significant factor for many students, both urban and rural, was the high cost of internet data. To address this problem, the teachers learned from their own and others' experience that the most appropriate platform to use was WhatsApp, which provided flexibility at minimal cost to the students. It enabled teachers to interact cheaply with students via messages, voice notes, photos and videos, both synchronously and asynchronously as the need arose. Through WhatsApp, teachers were also able to identify which students were online at the time they were giving the lesson, although it could be hard to distinguish who was who, when students were using their parents' or friends' mobile phones.
A second challenge that the teachers expressed was that ERT severely reduced the quality of teacher-student interaction. The teachers acutely felt the lack of direct face-to-face contact with their students. They explained that the lack of contact between teacher and students meant that it was hard to know what difficulties the students were facing, how they were responding, and how to motivate them. It was even difficult to know which students were actually participating in the classes. I feel that [my students] … have less motivation and some of them come to the class only [..] for some teachers that they like, …, sometimes they just come and then, after a moment, they turn off the video and do what they like. (00:24:52/Aldi) As Harry explained, he noticed that his students concentrated better in class because he could maintain eye contact with them. At home, they found it hard to balance their schedule, especially if their parents were not present. (00:26:09/Harry) The third challenge of ERT for these teachers was the workload they faced. Most of the participants were teaching 12 classes of 30 students each. Even though they maintained a separate WhatsApp group for each class, the volume of assignments, especially if they were returned through WhatsApp, was overwhelming. The challenges of ERT meant that achieving all the curriculum content was impossible and several teachers explained that that they had to lower expectations. For example, "For the teaching material, we follow the syllabus, although it could not be delivered 100%" (00:10:02/Harry) "Well, I don't really leave out materials … but lower down the standard" (00:03:08/Kira)

Theme 2: Communicative Language Teaching
An interesting finding from the study was that teachers continued to promote communicative language learning despite the difficulties of interacting with their students. Data from the study showed that despite the challenges, the participants were striving to improve students' communicative competence. Some participants explained that they commonly integrated the four macro skills in language learning, moving from input (reading or listening) to output (speaking and writing). For example, Delfina provided this description of an oral presentation task which began with reading and discussion, and then moved into preparing the presentation: "I told them to make one minute speech on sustainable development. There are 17 issues from UNDP, for example, zero poverty. So I told them to make a one minute speech…. So, they make it like writing at the beginning, so writing skill, and after that I give them several days to memorize it, and when we are ready for delivering the speech, they should say it in only a minute, that's for speaking. So, speaking and writing are integrated in delivering the speech… I told them to deliver the speech, and make a video about that." (00:09:34/Delfina) Another example offered by Delfina was an integrated task which involved presenting a videoed book review. Her students first read the book (with some scaffolding) and then discussed it together via WhatsApp. Next, Delfina explained the genre of the book review, and the students brainstormed points for and against the book, based on their earlier discussion. They also brainstormed examples of prefabricated chunks of language which they could use in presenting their own reviews. Then, individually, they prepared their reviews and recorded a draft review. This was shared with a peer who provided feedback. The students were then able to practice recording and re-recording their reviews, and finally posted the completed review on Delfina's YouTube channel. For the final element of the integrated task, the students viewed each other's reviews and wrote a reflective paragraph on their achievements.
Nevertheless, the teachers faced numerous difficulties in teaching communicative skills. Listening was particularly problematic. If the teachers posted videos, or links to YouTube videos, students often complained that they couldn't understand. So, some teachers fell back on teaching vocabulary.
"if I, --sent some videos but still many of them can't [understand] …, "Miss? I don't understand". So yeah. "Just try to listen and then grab some words". So yes, mostly I'm focusing on vocabulary." (00:05:44/Iris) A productive strategy for teaching listening was to use songs: an example given by Firman was "Hero" by Mariah Carey. In the interview he explained the purpose of using songs as learning materials: "I asked them to find out what is the song talking about. And after they understand the song, they have to sing it. Just for listening and … good pronunciation." (00:02:02/Firman) All the teachers strongly encouraged the use of English, even though the class was conducted online, recognizing the importance of their own speaking as listening input: "While I'm having the online class, I always have my class in English. When they don't understand they can raise their hand, so I can try to translate or try to find other description with my explanation by using English. And if they still don't understand it, I try to use bilingual in Bahasa Indonesia." (Aldi: FGDS1 00:46:33) Teaching speaking was also difficult during ERT. Most teachers began their online class with a friendly warm-up, and found that students were happy to respond. Although some teachers managed to organize class discussions via WhatsApp, it was often hard to encourage all students to participate: "In a normal classroom in a face-to-face classroom, you would elicit answer from everybody to make sure that everybody gets a turn. How does that work online? Some students volunteer and some students remain very quiet." (Aldi: FGDS1 00:51:14) Fauzi overcame this problem by numbering the students and getting student 1 to ask a question to student 2 and so on. Staged turns were found to be more effective than allowing students to volunteer. However, as Iris pointed out, if students were learning at home, they were often shy about practicing speaking in front of their families.
"Yes, it's quite hard for speaking. They don't have any self-confidence, especially if we have a video and then they are in their family rooms or with other friends. They tend to be ashamed. So yes, try and yes, try to talk" (00:05:44/Iris). Nevertheless, some teachers had found excellent ways to encourage students to practice speaking in ERT. For example, Rachmah asked her students to create a dialogue with a partner about a recent trip, record it, and send it to her as a voice-note. Similarly, Firman asked his students to prepare comprehension questions from a reading passage, record them on VoiceNote, and send them via WhatsApp.
The participants also expected their students to participate in reading. An activity that reflected standard classroom-based practice was reading comprehension exercises. The students were assigned to read, for example, some passages from the textbook. In some cases, these passages were photocopied at school and the students or their parents could pick them up at school to take home. Otherwise, teachers sent a link to certain webpages, or, in Aldi and Delfina's case, to their personal blogs. Then the students would do a reading comprehension quiz about the text they had been asked to read. In class, this might have been a quiz on Quizziz or Kahoot, but these tools proved difficult to apply via WhatsApp. Moreover, the teachers were not satisfied that the students had completed the quiz without cheating, perhaps by getting their friends to do the quiz for them or by sharing the answers with each other.
Another participant, Iris, asked her students to Google different countries' responses to Covid-19, and write a summary. She also mentioned that she required her students to read simplified classics, such as Oliver Twist, which they could source online as a PDF. To demonstrate their understanding, the students had to write a summary. Again, she doubted whether the students wrote the summaries themselves, as it was possible to find summaries on the internet. Nevertheless, as Aldi commented, this did at least mean that the students were interacting with Google! It appeared that the several teachers found it easier to teach writing than other skills during ERT, perhaps because they could assign a task via WhatsApp and have students submit it either via Google Classroom, or via WhatsApp itself. Fauzi shared a lesson plan demonstrating how he scaffolded students' writing skill in the online context: first he showed the students a prompt picture via WhatsApp. Then, in the WhatsApp chat, students brainstormed vocabulary stimulated by the picture. Next, they moved on to posting phrases describing the picture. The next stage was to write short paragraphs of connected sentences, again posted on WhatsApp chat. These could be adjusted by the teacher as a form of immediate feedback. (Lesson plan: Fauzi) However, most teachers confessed that it was impossible for them to give feedback during ERT. With 12 classes of 30 students all submitting assignments, the most they could do was to record that the assignment had been completed. Even though they had a separate group for each class, it was an almost impossible task to keep track of all the assignments, especially if students were using their friend's or their parents' device. In the context of Covid-19 the teachers considered that it was sufficient to record whether the task had been completed or not.

Theme 3: Promoting Learning Autonomy
The third major theme emerging from the data concerned learner autonomy. We had assumed that the difficulty of maintaining contact in the online context and the increase in online participation might lead teachers towards delegating greater responsibility for learning to the students themselves. However, it seemed that this was not the case.
Data from the research revealed that generally in their online teaching, the participants mainly tried to replicate their classroom teaching using the online environment. They continued to position themselves as controllers of the students' learning. They took a roll-call of attendance to ensure that students were present in online classes; they determined the content and sequencing of classes; they set assignments and marked them, even if that only meant recording that the assignment had been done. The teachers continued to plan the students' learning activities and determine the learning materials. Students had little or no choice in the learning materials: for example, it was Firman who chose the song "Hero" as listening practice for his students and prescribed reading materials were distributed either in hard copy or online with set reading comprehension questions.
The teachers claimed that they were indispensable to the students' learning. As Delfina explained, during the school lockdown, "it's hard for them to get good score because they don't have anybody beside them to talk, to share with. It's just like … they just like to have me as a problem solver" (00:26:08/Delfina).
The teachers explained that they did not believe that their students had the competence and motivation to learn autonomously. Aldi, for example, commented that students' learning needed to be monitored either by teachers or parents, as they were "unable to manage their schedule by themselves". Similarly, Endang said, "I don't think [students can learn autonomously] because you know, … only 25% who have high motivation in learning by themselves." (00:20:04/Endang) Some participants argued that their students were not able to learn autonomously in the current situation because they were overwhelmed by assignments given by other teachers. As Firman said, most teachers in the school responded to home schooling by setting multiple assignments, so students were too busy: "They also have another task from the other subjects.
[They] learn 16 subjects. So probably they got too much. They got too much homework". (00:08:01/Firman) Nevertheless, there were some indications in the interviews of a move towards greater autonomy. For example, the use of WhatsApp enabled teachers to send asynchronous messages which students could access in their own time wherever they chose. This meant that some students were able to access classes more flexibly, allowing them, for example, to help their parents in the rice fields during the day and to study at night (PI/10:59/Firman).
Some teachers also set more flexible assignments which involved students in making some choices. Iris, for example, allowed her students to choose one of the simplified classics for their book review assignment (although she considered that many of the students copied and pasted their reviews from online sources). In Delfina's class, students could choose which of the 17 Millenium Development Goals they would discuss in their video presentation. This type of assignment also gave students flexibility in preparing the assignment: they could record it over and over again if they wished, rather than make the presentation once-off in real time. Another aspect of Delfina's classes was a move towards self-assessment and peer assessment, giving students more responsibility for monitoring their own learning.
Six of the participants suggested links to sources that their students could access for extra work. So, although these participants were hesitant about students' willingness and capacity to direct their own learning, they still attempted to foster learning autonomy among their more motivated students.
There was also evidence that at least a few of the students were taking the initiative to direct their own learning. For example, Aldi found that one of his students had started to read Shakespeare during the lockdown (00:27:39/Aldi), and Kira recounted that some of her students actually performed better during the period of learning from home. As she said: "What is really interesting is that the students who are not really active during face to face classroom are more active during distant learning. They do the assignment faster than others because they don't really need to think what other people would think about him, they just really need to think about 'I have to submit the assignment on time.'" (FGD2/00:49:49/Kira).

Discussion
This paper examines the role of learner autonomy in the context of ERT. The data shows that there was little move in this direction: the teachers' strategy in coping with ERT was to replicate their classroom teaching as closely as possible in the online environment. In terms of the five levels of learner autonomy described by Nunan (1997): awareness, involvement, intervention, creation and transcendence, the teachers did not even begin to embark on these levels. As Lengkanawati (2017) also found, teachers firmly believed that their students were not ready for learner autonomy.
The online environment demanded a raft of new skills from teachers, as they worked out how to identify appropriate tools to maintain communication with their students, and appropriate strategies to maintain students' progress in terms of the curriculum. It was understandable in these conditions that teachers were not quick to take advantage of the students' increased interaction with ICT.
In fact, the study highlighted the dedication of these teachers and their resilience in the face of the unprecedented demands made upon them. In particular, their ability to continue to promote communicative skills was impressive. They also attempted to maintain a student-centred approach, getting their students to interact with each other online, and trying to select materials which they thought would engage their students' interests. However, in terms of Schweisfurth's (2013) continuum of student-centredness, these teachers remained at the traditional, 'chalk-andtalk' end of the spectrum, believing that they were indispensable to students' learning. It was not surprising, given the long tradition of Indonesian teacher-fronted education, that they positioned themselves at the transmission end of this spectrum when the move to online teaching was imposed. They continued to decide what curriculum goals should be addressed, what materials should be used, when the students should be in class, and how they would be assessed -the antithesis of allowing learner autonomy.
Nevertheless, two of the teachers, Delfina and Rachmah, positioned their teaching a little further along the Schweisfurth spectrum, allowing their students some degree of choice. Littlewood (1999) referred to this approach as 'reactive autonomy': students are still closely guided by their teachers while being allowed a certain degree of choice.
In fact, allowing the students greater autonomy would have eased many of the difficulties the teachers were facing in the move to ERT. The first challenge we identified, as described above, was the students' difficulty in accessing the internet because of the often-unreliable network and their lack of access to devices. This would have been much less of a problem if the teachers had relied less on synchronous teaching and allowed their students more flexibility in scheduling their study at times when data was cheaper or more reliable. It would also have been easier for students to access online devices by arranging their study at times when their parents and siblings were not using the family's devices. The teachers worried that students were incapable of managing their own time, but if students are not allowed some choice in when, where and what to study, they cannot learn to manage their own priorities. If the teachers had offered some tools and strategies, such as a list of tasks to be achieved over a period of time, they could have supported their students in gradually becoming more autonomous (Reinders & White, 2016).
Another challenge that the teachers faced was the pressure of workload, caused partly by the large class sizes. Again, allowing greater learner autonomy could have helped to ease this problem. For example, encouraging students to self-assess or to do peer-assessment could have helped to overcome the enormous burden of marking and also the impossibility of giving formative feedback. Indeed, Delfina had begun to use this practice. In her book review task, the students uploaded their oral reviews to her YouTube channel, viewed each-others' work and then wrote a reflection on their own achievement in relation to others. Providing a rubric for selfassessment -or better still, developing a rubric for self-assessment in collaboration with the students themselves -is a helpful strategy for supporting greater learner autonomy.
In terms of CLT, teaching listening was the most difficult challenge. Partly this difficulty was exacerbated by the teachers' positioning themselves as "knowers", typically picking out and explaining vocabulary for their students, as Fauzi and Iris described, rather than focusing on building their students' capacity to make meaning for themselves. Autonomous learners would have chosen the songs or videos that they themselves would like to listen to, and known how to use tools such as sub-titles, online dictionaries and Google Translate to assist in their meaning making. The teacher's role, then, shifts from "knower" to "adviser" or "consultant" (Reinders & White, 2016).
Similarly, rather than prescribe reading texts for the students, students could have chosen their own reading challenge. This does not mean that teachers would simply send the students away to trawl the internet without guidance and support; rather, students could choose from a range of suggested texts or prepared sites at an appropriate level, or choose their own topic to research, perhaps writing a book review or report to crystallize their learning. As reported above, Iris had tried a teaching strategy along these lines by asking her students to read and summarize one of the simplified classics. She was concerned, however, that students had simply copiedand-pasted summaries, suggesting that this task was overwhelming for students with little English competence. There are implications here: first, as Little (2007), Reinders and Hubbard (2014) and others point out, moving to learner autonomy is a process which requires careful scaffolding. Students need to acquire appropriate learning strategies, and systems need to be put in place to support their learning. Delfina's book review task was an example of how students can be encouraged to move towards learner autonomy. After completing the first closely scaffolded book review, they could have moved on to choose from a range of suggested texts and complete further reviews following the same genre. In this way, with support, students can gradually become more independent.
A key feature of learner autonomy is that learners can choose goals, materials and learning activities to suit their learning styles and interests. Choice has been shown to be an important aspect of motivation, and for learning autonomy to work well motivation is essential. Allowing students greater choice (and supporting it well) contributes to motivation in a virtuous cycle. As Ushioda (2011, p. 223) eloquently argued, "if students are involved in the management of their own learning and in shaping it according to their own personal interests, they are also exploiting and nourishing their intrinsic motivation." In terms of speaking, another challenge for the teachers was to get their students to participate in speaking in synchronous classes. Fauzi's response -at the controlling end of Schweisfurth's (2013) spectrum -was to have his students take numbered turns. Rachmah, on other hand, allowed her students more agency by having them work in pairs to record a dialogue about a recent trip and send it to her as a voice note. This was not only a more communicative, authentic, and personalized task, but it also gave the students an opportunity to practice multiple times and monitor their own production: a good step towards learner autonomy.
Altogether, the move to ERT meant that teachers had to relinquish their hold on students to a certain extent. As Aldi pointed out, they could no longer monitor which students were engaged and attentive as they could in the classroom. They had to trust that their students were following the synchronous classes and not cheating on the assigned asynchronous tasks. The plethora of tasks assigned by many teachers was an indication that they found it difficult to trust their students: by setting multiple assignments they attempted to make sure the students were "ontask".

CONCLUSIONS
The precipitous move to ERT was an immense challenge for teachers across the world. As Pu (2020) commented, this sudden switch to online learning was in effect a global experiment in education. The Indonesian teachers of English who participated in this study were able to transfer their classes online and, despite the considerable challenges, maintain a strong focus on communicative language learning, engaging their students in interactive tasks and achieving outcomes which may not have seemed possible at the outset of ERT. They have our considerable admiration for their dedication and perseverance.
However, it was surprising that these teachers did not do more to foster greater learner autonomy given the opportunities -and challenges -of the online environment. Perhaps the suddenness of the shift did not allow teachers to imagine the potential of allocating greater agency to students, and perhaps, if this research had been conducted six or twelve months later, we might have found different results.
This study contributes to the growing literature on learner autonomy in Indonesia, which has already shown that autonomous learning is possible, even in rural areas of the country (Lamb, 2013). Like previous studies (Budianto, 2014;Jannah et al., 2022;Melvina et al., 2021), it confirms that teachers are interested in learner autonomy, even if they still believe that their students are not yet ready to take responsibility for own learning. The study suggests that before learner autonomy can become more of a reality in Indonesia, teachers will have to: learn to trust their students more, freeing them up to study more independently; help students develop productive language learning strategies which will allow them to benefit from the plethora of materials available online; develop tools which can support their students' development of autonomous learning habits and strategies; and re-imagine their role as "language advisers" rather than as directors of learning. Future research may show that the experience of teaching remotely during the COVID pandemic has sown the seeds of change, and that eventually Indonesian students of English will benefit from a move towards greater learner autonomy.

Level
Learner action Content Process 1 awareness Learners are made aware of the pedagogical goals and content of the materials they are using.
Learners identify strategy implications of pedagogical tasks and identify their own preferred learning styles/strategies. 2 involvement Learners are involved in selecting their own goals from a range of alternatives on offer.
Learners make choices among a range of options.

intervention
Learners are involved in modifying and adapting the goals and content of the learning program.

creation
Learners create their own goals and objectives.
Learners create their own tasks.

5
transcendence Learners go beyond the classroom and make links between the content of classroom learning and the world beyond.
Learners become teachers and researchers.