Pronunciation Teaching in the Second Language Classroom

The growth of research on pronunciation in L2 instruction has led to a better understanding of several pedagogical issues. First and foremost, such research has established that explicit pronunciation instruction enhances L2 learners’ pronunciation (^{Thomson & Derwing, 2015}). At the same time, a balanced approach, one that addresses both segmentals (vowels and consonants) and suprasegmentals (stress, rhythm, and intonation), is seen as facilitating pronunciation improvement (^{Sicola & Darcy, 2015}). Researchers are suggesting that the aim of pronunciation teaching in the L2 classroom needs to be intelligible (i.e., clear) speech rather than the unrealistic notion of attaining native-like pronunciation (^{Thomson, 2014}).

Research has also shown that pronunciation instruction-compared to other skill areas-is still given relatively infrequent attention and inadequate instructional time in the L2 classroom (^{Foote et al., 2016}; ^{Huensch, 2019}). In many cases, this reflects the lack of pronunciation-specific training available to L2 instructors, with teachers typically feeling reluctant and anxious about teaching pronunciation in the classroom (^{Baker, 2014}; ^{Couper, 2017}; ^{Foote et al., 2011}; ^{Henderson et al., 2012}). Instructors’ reluctance and anxiety about pronunciation instruction have been raised for two decades (see ^{Macdonald, 2002}, for some early work on this). Still, it continues to be a problem that needs to be addressed. Research has also demonstrated that pronunciation instruction tends to be mostly teacher-centred (^{Baker, 2014}; ^{Burri, 2021}; ^{Hismanoglu & Hismanoglu, 2010}), providing limited opportunities for learners to practice target features in a communicative environment. Besides the use of controlled (i.e., teacher-centred) teaching, in many cases, pronunciation instruction is unsystematic and ad hoc (i.e., unplanned) in response to learner errors (^{Couper, 2017}). Meanwhile, L2 instructors often perceive segmentals as easier to teach than suprasegmentals (^{Foote et al., 2016}; ^{Wahid & Sulong, 2013}) because of a common lack of confidence and training in teaching suprasegmental features (^{Foote et al., 2011}). Further limiting pronunciation instruction in the classroom is the substantial impact of contextual factors-such as curricular and time constraints, textbooks, and colleagues-on teachers’ practices, beliefs, and knowledge about pronunciation (^{Bai & Yuan, 2018}; ^{Burri, 2021}; ^{Burri & Baker, 2020}, ²⁰²¹).

One area of pronunciation instruction that has received considerable empirical attention is oral corrective feedback (OCF). In particular, recasts-defined as “the teacher’s reformulation of all or part of a student’s utterance, minus the error” (^{Lyster & Ranta, 1997}, p. 46)-have been frequently researched among the many feedback strategies available to L2 instructors (e.g., repetition, explicit correction, clarification requests, metalinguistic information, elicitation; ^{Ellis & Sheen, 2006}). As for classroom-based research, ^{Baker and Burri’s (2016)} study, for instance, showed that L2 instructors used recasts to address errors even though they viewed recasts as being ineffective in helping students overcome their pronunciation issues. Several other studies have demonstrated that pronunciation instruction is often limited to teachers’ use of recasts in the L2 classroom (^{Bao, 2019}; ^{Foote et al., 2016}; ^{Ha & Murray, 2020}; ^{Nguyen & Newton, 2020}), lending further support to the notion that pronunciation is perhaps the most challenging element of a language to teach (^{Setter & Jenkins, 2005}).

These studies have provided valuable understanding of pronunciation teaching in the L2 classroom, but the majority-if not all-of them have been done in one country and with one specific group of teachers. Research is, therefore, needed that examines pronunciation instruction across two (or possibly more) countries to see if there is a difference in the type of instruction provided in specific teaching contexts.

Second Language Learners’ Perceptions of Pronunciation Teaching and Learning

Research on L2 students’ perceptions of pronunciation is relatively scarce. The few available studies have shown that L2 students value and desire pronunciation teaching (e.g., ^{Baker, 2011}, ^{Dao, 2018}; ^{Derwing & Rossiter, 2002}), and they appreciate receiving OCF in the classroom (^{Ha & Nugyen, 2021}; ^{Ha et al., 2021}). Learners often find pronunciation challenging to acquire (^{Cenoz & Garcia-Lecumberri, 1999}) but are generally motivated to improve with the aim of attaining native-like pronunciation (^{Nowacka, 2012}; ^{Smit & Dalton, 2000}). Also, students do not necessarily want native-speaking teachers but instead value knowledgeable and proficient instructors (^{Levis et al., 2017}), while advanced learners tend to express stronger concerns for pronunciation than their lower-level counterparts (^{Huensch & Thompson, 2017}). Other research has established the reluctance of L2 learners to strive for native-like pronunciation in fear of facing social pressure due to ethnic group affiliation (^{Gatbonton et al., 2005}). It has also been suggested that when students feel anxious about their pronunciation, they are less likely to engage in communication (^{Baran-Łucarz, 2014}). In an often-cited study by ^{Derwing and Rossiter (2002)}, adult immigrant students in Canada reported not receiving enough beneficial pronunciation instruction in the classroom and using strategies such as paraphrasing, self-repetition, spelling/writing, and adjustment of volume to overcome communication breakdowns.

This line of inquiry has provided important insights into L2 learners’ perceptions of various pronunciation issues, but according to ^{Levis (2021)}, more research is urgently needed to understand better L2 students’ thoughts and feelings (i.e., perceptions) about pronunciation.

Levis’s call for more research on students’ perspectives of pronunciation can also be extended to inquiry into the comparison of L2 learners’ perceptions with their teachers’ practices. Several studies have explored the relationship between L2 students’ perceptions and their teachers’ beliefs, goals, attitudes, and practices (e.g., ^{Graus & Coppen, 2017}; ^{Mackey et al., 2007}; ^{Ruesch et al., 2012}; ^{Yoshida, 2010}; ^{Zhou et al., 2014}); yet, thus far, research on the connection between students’ perceptions and their instructors’ pronunciation practices has only been examined in the Vietnamese context. Findings suggest that Vietnamese learners of English and their teachers both value explicit, systematic, and communicative pronunciation teaching (^{Nguyen & Hung, 2021}; ^{Nguyen et al., 2021}), and, generally, they hold favourable views towards explicit feedback provision (^{Ha & Nguyen, 2021}; ^{Ha et al., 2021}). These studies have provided important insights into the Vietnamese context. However, research examining the relationship between L2 learners’ perceptions and their teachers’ pronunciation teaching practices must also be conducted in other countries to obtain a more in-depth understanding of a vastly unexplored area.

Teacher-Participants

As mentioned previously, three L2 instructors teaching in two countries (Australia and Japan) were included in the current study. The selection of the three participants was a sample of convenience, for the teachers were part of a larger research project examining the longitudinal process of learning to teach English pronunciation (^{Burri & Baker, 2021}). The teacher-participants were highly experienced instructors and possessed an MEd with a specialisation in TESOL from an Australian university.

Linda (52 years of age) taught in an Intensive English Center (IEC) in Australia for five years and held two administrative positions (assistant head teacher and integration network coordinator) for 1.5 years. Before working at the IEC and studying for her MEd, she was a primary and high school teacher for 20 years. The theme of the two lessons observed was “ancient Egypt,” requiring students to use all four skills in learning about the history of Egypt.

Adil (32), from Iran, worked as an L2 instructor and administrator in an English language program at a vocational institution in Australia. His full-time job included 1-2 days of teaching speaking/listening or reading/writing courses per week, curriculum design, and coordination of the entire program. Preceding his move to Australia to enrol in an MEd program in 2012, Adil had taught English at the tertiary level in Iran for eight years and ran his own language school. Practicing simple past tense (through the use of all four skills) was the focus of Adil’s lesson that was observed.

Aya (35), from Japan, taught general English courses at a private junior high school in Japan for four years. Before commencing her MEd studies in Australia in 2012, Aya had taught English at a Japanese senior high school for five years. The goal of the two observed lessons was to have students practice “can” vs “can’t” through speaking, listening, and reading.

It must be noted that the instructors were aware of the focus of the research, and the classes they taught were part of general skills courses without a specific focus on pronunciation teaching.

Student-Participants

As with the three instructors, selecting the student participants was a sample of convenience. Linda had 16 students in class at the time of data collection. The learners were predominantly migrants and refugees with significant schooling gaps and trauma-related issues. Their residency in Australia ranged from one to 14 months. They were between 12 and 18 years old and came from Burma, China, Congo, Eritrea, Indonesia, Papua New Guinea, Syria, and Thailand. The average English proficiency of the learners was at an intermediate level. Adil taught 16 Middle Eastern migrant and refugee students. Some of Adil’s learners were long-term residents of Australia (6+ years), while others had been in the country for just a few months. Adil’s learners, who spoke Assyrian or Arabic as their first language, were in their 40s or above, with mostly beginning-level English proficiency. Lastly, Aya taught 17 junior school students who were 13 years old. The length of their English studies in Japan varied from six months to over six years, but they all possessed pre-intermediate English proficiency.

Data Collection

To obtain insights into the instructors’ pronunciation teaching practices and, ultimately, to compare this data set with their students’ perceptions, video footage from classroom observations and a 45-minute semi-structured interview were used. Linda was observed twice (each observation lasted 105 minutes). Adil was observed once for 120 minutes,¹ and Aya’s two observed lessons were 45 minutes each. The video camera was positioned so that only students that provided consent were included in the recording.

^{Richards’s (2011)} teaching competence framework, comprising a holistic and sociocultural perspective on L2 pedagogy, informed the creation of the interview questions. Linda was interviewed in her office, Aya at a local coffee shop in Tokyo, and Adil on Skype. The interviews were audio recorded and then transcribed verbatim upon the completion of the data collection. Transcribing the video footage was deemed unnecessary, given that the analysis of the classroom observation data included counting the number of pronunciation teaching techniques and instances of OCF provision (see data analysis below).

The questionnaire ^{Baker (2011)} developed in her research was utilized and modified to elicit L2 learners’ perceptions of pronunciation teaching and learning.² Due to logistical difficulties, the researcher could not visit the classes before observations, so the instructors were asked to give the questionnaire to the students who had consented to participate. While this is not ideal, it shows the challenges of conducting research in intact classrooms. The anonymous questionnaire contained three multiple-choice items on general pronunciation instruction, four multiple-choice items on error correction/OCF, and four open-ended questions that asked the students to provide short written answers about their instructors’ pronunciation practices (see Appendix).

It must also be noted that the timing of the data collection varied due to the teacher participants’ availability and the researcher’s ability to travel and visit the research sites. In Linda’s and Aya’s cases, the two classroom observations, the semi-structured interviews, and the student questionnaire were conducted within a week (Linda in June 2019 and Aya in October 2018). The questionnaire was administered before the interview and the classroom observations to ensure that students reported on practices before the researcher observed any teaching. However, due to family-related circumstances, Adil’s situation was more complicated, so he was interviewed in December 2018, six months before his students completed the questionnaire, and the researcher carried out the observation in May 2019.

Data Analysis

While coding the observation data, the researcher watched each video footage several times and identified and counted instances (i.e., instructors’ use) of language awareness, controlled, guided, and free techniques. As ^{Baker (2021)} explains, language awareness serves for a teacher to facilitate her learners’ phonological awareness of target features (e.g., listening discrimination activities). Controlled techniques allow learners to produce target features, with the instructor continuing to exert a dominant role in students’ production (e.g., drills, reading aloud). Guided techniques enable students to practice features in a less teacher-centred environment (e.g., pair work and information-gap activities), and free techniques provide students with opportunities to produce language more creatively with no or little influence by the instructor or materials (e.g., games, discussions).

Besides the coding of techniques, all instances of OCF were counted in the video footage. Once done, the students’ responses to the multiple-choice and open-ended questionnaire items were collated in an Excel spreadsheet to obtain an overview of the perceptions of all 49 student participants. Themes were noted, and each questionnaire was explored separately to examine differences between the three classes. Lastly, each interview script was read several times and coded thematically according to the techniques and OCF identified in the observation data and themes in the students’ questionnaire responses. This process enabled the researcher to attain an in-depth understanding of the relationship between the instructors’ practices and their learners’ perceptions of pronunciation teaching and learning.

Linda

As shown in Table 4, Linda used 28 techniques, with 97% (27/28) of them being controlled in nature. These controlled techniques were repetitions/drills, explanations, modelling, and reading aloud, most targeting segmental sounds in newly learned vocabulary.

In the interview, Linda explained that “in the actual teaching practice in the classroom, it’s the segmentals that are important,” especially “final consonant sounds” and “medial vowel sounds.” The only instance of guided practice occurred when she asked the class about the vowel sound in the stressed syllable of the word “archaeology.” As for her use of OCF, 85% of 40 instances (34/40) were recasts of segmental sounds in target words taught during the two observed lessons (plus 5% explicit corrections and 10% prompts). Almost all of these recasts were provided during two oral reading tasks in which individual students were asked to read out in front of the class several sentences from the textbook. As such, the majority of these explicit recasts can be classified as didactic recasts because they refer “to a reformulation of a student’s utterance in the absence of any indication of a communication breakdown” (^{Ha & Murray, 2020}, p. 3).

Nonetheless, the questionnaire data showed that all of Linda’s students wanted her to correct their pronunciation, while 12/16 students preferred correction in front of the class and 14/15 as a group. The findings, therefore, provided evidence that her students’ preferences for receiving corrective feedback aligned with Linda’s practices. Yet, the incongruity between her predominantly teacher-centred approach to pronunciation instruction and her students’ desire to improve their pronunciation and their goal of using spoken English in their daily life (16/16) and in their future university studies (11/16)-warrants further examination.

The interview data suggested that the curriculum and Linda’s beliefs about L2 learning were reasons for the incongruence between her practices and her students’ perceptions. The objective of the curriculum was for students to produce academic texts, and Linda believed that too much content to cover and subsequent time constraints hindered her ability to teach pronunciation in the classroom. As the following statement demonstrates, she also believed that pronunciation learning was a long process and was, therefore, of limited priority to her students:

If it’s an issue where I think they’re not going to be getting it in a hurry, say for example, a Thai student, I will let it go, because their language learning journey is a long one. It may be five, 10, 15 years before they actually are going to be able to replicate that sound so why harp on about it now, because what we’re wanting them to do a whole lot of things; we’re wanting them to speak and write correctly, in sentences, and really, the aim of the IEC is for them to be able to produce academic texts. That’s what we’re aiming for…So, if they can’t quite nail that little sound, or that particular word, and if they’re constantly saying that particular word wrong, maybe that’s just not a priority right now.

Adil

Adil, who also taught in Australia, used fewer techniques⁴ but more instances of OCF than Linda (see Table 4). A total of six techniques included four controlled (three reading-aloud tasks and one explanation) and two guided techniques (elicitation of the word ending “-ed” and a discussion with the class about the sounds /p/ and /b/ occurring in English and Arabic). As for his provision of OCF, a total of 55 instances were counted. Except for six instances, all of the OCF occurred during the reading-aloud tasks, but unlike Linda, Adil mainly used prompts to address his students’ errors. Thirty-four of the 55 instances were prompts, with 26 questions to have students self-correct their pronunciation (e.g., “Do you mean ‘spend’ or ‘spent’?”). Besides prompting, and somewhat similar to Linda, he also mainly used didactic recasts (17/55) and explicit correction (4/55), focusing on “accuracy problems rather than communication breakdowns” (^{Ha & Murray, 2020}, p. 20). Overall, his OCF targeted mostly segmental sounds, which was done individually between the teacher and the student. The questionnaire data suggested that this matched his students’ preference: 14/15 liked error correction in front of the class. Nonetheless, given that all of his students expressed a strong desire to be taught and improve their pronunciation (11/14 wanted their teacher to correct their pronunciation, and 15/16 indicated that they used English in their daily lives), the potential discrepancy between the students’ perceptions and their teacher’s somewhat limited approach to pronunciation-at least from a communicative point of view-requires further inquiry.

As in Linda’s program, the curriculum appeared to play a substantial role in contributing to this incongruity. Besides Adil’s OCF provision being aligned with the program outcome of having learners self-correct their pronunciation issues, he explained in the interview that the lower levels focused on grammar and pronunciation was therefore thought to be of less importance:

At this stage, it’s enough for [students] to be able to make a correct sentence grammatically rather than being able to pronounce it correctly; as long as it does not make any other meanings or bad meaning, it should be fine. As long as everyone understands [them], that’s the correct pronunciation.

Aya

Notably different from Linda and Adil’s ad hoc (i.e., unplanned) and reactive approach (^{Couper, 2017}; ^{Foote et al., 2016}; ^{Nguyen & Newton, 2020}), Aya’s pronunciation instruction (provided in Japan) was predominantly pre-planned and phonics-based to help her students “understand the connection between pronunciation and also spelling.” As seen in Table 4, she used 50 techniques in her two lessons, with 74% (37/50) being classified as controlled. Of these 37 controlled techniques, 28 were repetitions/drills, with almost all directed at the whole class rather than individual students, which supports previous research suggesting that Japanese junior high school teachers frequently use “listen and repeat” in their English classrooms (^{Uchida & Sugimoto, 2018}). However, the data analysis also revealed that Aya employed a more expansive repertoire of techniques than the other two participating teachers. Besides the controlled techniques, she used five language awareness techniques (e.g., songs, teacher-read sentences, and students circled correct word on the handout), five guided techniques (e.g., elicitation of sounds/sentences, pair work), and three free techniques (e.g., students asking and answering questions about a picture in the textbook). These techniques more or less gradually progressed from language awareness to less teacher-controlled techniques, suggesting that Aya was perhaps aware of her learners requiring extra scaffolding in learning pronunciation due to their relatively low English proficiency level. Few techniques allowed the students to practice the target features (stress and rhythm) in a communicative context, but all her students indicated in the questionnaire that they would be using English in Japan and not in English-speaking contexts; therefore, fluency development was probably less of a priority at this point.

While Aya used a wider variety of techniques than Linda and Adil, the observation data also revealed that she provided fewer instances of OCF than the other two instructors (see Table 4). A total of 18 instances (10 recasts and eight explicit corrections) were identified, and, except for one recast, all of the OCF was directed at the entire group of students, not at individual learners, as in Linda’s and Adil’s cases. This whole class-based approach to feedback provision mirrored her use of repetitions/drills, but it appeared to align with her students’ preferences: Eight of the 17 students indicated in the questionnaire that they did not like to be corrected individually or in front of the class, and six were somewhat uncertain about whether they wanted the teacher to correct their pronunciation. Nevertheless, even though Aya’s OCF provision appeared to accommodate her students’ preferences, the following interview excerpt demonstrates that the rare occurrence of OCF in Aya’s classroom was most likely the result of her doubts about her ability to correct her students’ pronunciation. When asked how she corrected her students’ errors, she replied: “I’m not [a] skilful pronunciation teacher, so that’s my challenging point. I don’t even know how to do it correctly, so I’m still struggling.” The lack of OCF provision is problematic given ^{Saito and Lyster’s (2012)} proposition that corrective feedback is needed in conjunction with pre-planned form-focused instruction for pronunciation work to be effective with Japanese learners of English. Nonetheless, the findings suggested that Aya’s practices aligned with her learners’ perceptions more closely than with Linda and Adil. It is, of course, speculative, but Aya’s general English course in which speaking featured more prominently than in Linda’s and Adil’s courses, and the fact that she shared the same L1 with her students and, therefore, most likely possessed a more in-depth understanding of the students’ needs and goals (^{Cook, 2016}), may have contributed to this closer alignment.

Teachers’ Practices Versus Students’ Perceptions

As this study illustrates, the relationship between the teachers’ practices and their learners’ perceptions of English pronunciation is complex. On the one hand, the learners in all three classes strongly desired to be taught and improve their pronunciation. However, the observation and interview data showed that all three teachers used predominantly, albeit to a varying degree, controlled (i.e., teacher-centred) techniques that provided learners with few opportunities for communicative pronunciation practice. On the other hand, while the limited range of pronunciation techniques highlighted an incongruity between the students’ perceptions and their teachers’ practices, the teachers’ OCF provision appeared to meet the students’ preferences.

Adding to the complexity was the notable variation among the teachers’ use of techniques and OCF, with the students indicating in the questionnaire that their instructors successfully spent ample time teaching pronunciation. However, as shown in Table 5, the perceived time of pronunciation instruction varied widely among students, ranging from 3-50 minutes (Linda’s and Aya’s students) to 3-4 hours (Adil’s students) per lesson, but some of these numbers did not match the amount of pronunciation instruction that was observed (and coded) in the study. The students also considered a broad range of activities to help improve their pronunciation (e.g., reading aloud, listening to music, videos, homework, pair work, body language, tests, correction, speaking, stories, handouts, textbook, repetition). At the same time, the observation data showed that all the activities the students mentioned in the questionnaire were used in their classrooms. Furthermore, as Table 5 illustrates, the large majority (44/49 or 90%) felt optimistic about their teachers’ approach to pronunciation instruction (e.g., “I feel good/happy when the teacher teaches pronunciation,” “I feel like I am learning something,” “I like it,” and “very easy to understand and it motivates me to study pronunciation more”), with only three students expressing mixed feelings or difficulties with pronunciation. Therefore, the findings derived from the questionnaire suggested that L2 students desire pronunciation instruction irrespective of the type of delivery provided in the classroom. Yet, as Linda’s following remark implies, the relationship between instructors’ pronunciation teaching practices and their students’ perceptions of classroom occurrences is a complex issue and, therefore, in need of more research:

I spend plenty of time on pronunciation, in their view, but I think I don’t spend that much time. But they saw it differently; they saw it as when I read to them, or when they read to me, or when they listen to a reading text, they get the opportunity to practice pronunciation. They look at it differently, so what you understand as pronunciation instruction and what they understand as pronunciation instruction is different.