Articles Survey of the Teaching of Pronunciation in Adult ESL Programs in Canada, 2010 Jennifer A. Foote, Amy K. Holtby, and Tracey M. Derwing This follow-up study reexamines the state of the teaching of pronunciation in ESL classes across Canada. The purpose of the survey was twofold: to gain a snapshot of current practices and to compare this with the picture of 10 years ago. We based the current work on Breitkreutz, Derwing, and Rossiter’s (2001) survey asking teachers about resources, approaches, and beliefs about teaching pronunciation. We also asked for background information about the instructors’ formal education and teaching experience. For the most part, instruction in pronunciation in Canada has not changed substantially in the last decade. More training oppor- tunities are available, although these are still not enough according to many of our respondents. The number of pronunciation courses offered in English-as-a- second-language (ESL) programs has also increased. Teachers’ beliefs about pro- nunciation instruction remained largely the same, with a similar focus on suprasegmentals and segmentals. However, we did find a slight difference in how teachers approached these two aspects of pronunciation. Ten years ago, teachers reported emphasizing both aspects in class, whereas today there seemed to be a slightly greater focus on segmentals. Finally, we offer several recommendations for TESL programs, ESL programs, and ESL instructors. Ce suivi porte un nouveau regard sur l’état de l’enseignement de la prononciation dans les cours d’ALS de par le Canada. L’enquête avait deux objectifs : obtenir un aperçu des pratiques courantes et le comparer avec les pratiques d’il y a dix ans. Nous avons fondé le travail actuel sur l’enquête de Breitkreutz, Derwing, et Rossiter (2001) adressée aux enseignants et portant sur les ressources, les ap- proches et les croyances relatives à l’enseignement de la prononciation. Dans l’ensemble, l’enseignement de la prononciation n’a pas changé de façon significa- tive au cours des dix dernières années. Même s’il existe beaucoup plus d’occasions de formation, plusieurs des répondants estiment que ce n’est pas suffisant. Le nombre de cours de prononciation offerts dans le cadre des programmes d’ALS a également augmenté. Les croyances des enseignants par rapport à l’enseignement de la prononciation sont demeurées en grande partie les mêmes, mettant l’accent sur la production segmentaire et suprasegmentaire. Nous avons, toutefois, repéré une légère différence dans la façon dont les enseignants abordaient ces deux as- TESL CANADA JOURNAL/REVUE TESL DU CANADA 1 VOL. 29, NO 1, WINTER 2011 pects de la prononciation. Il y a 10 ans, les enseignants ont indiqué qu’ils por- taient leur attention sur les deux aspects alors que de nos jours, on semble insister un peu plus sur la production segmentaire. Nous terminons en offrant plusieurs recommandations relatives aux programmes d’enseignement de l’ALS, aux pro- grammes d’ALS et aux enseignants en ALS. Ten years ago, Breitkreutz, Derwing, and Rossiter (2001) conducted a survey to determine the nature and extent of pronunciation instruction in English-as-a- second-language (ESL) classrooms in Canada. The era of communicative lan- guage teaching (CLT) had marked a “fall from grace” for pronunciation instruction (Isaacs, 2009, p. 2). Not only had CLT encouraged the belief that learners could improve their pronunciation through input alone (Breitkreutz et al.; Isaacs), but studies such as that of Purcell and Suter (1980) suggested that teaching pronunciation was ineffective. However, as the importance of pronun- ciation became clear to practitioner/researchers (Morley, 1991) and the noticing principle for second-language learning was increasingly accepted (Schmidt, 1990), some researchers began to focus on the possibility that teaching pronun- ciation could have an effect on overall intelligibility (understandability) of ac- cented speech, as well as on comprehensibility (the effort required of a listener to understand accented speech, Derwing, Munro, & Wiebe, 1997, 1998). Al- though this argument had been made in earlier eras (Abercrombie, 1949), there was little empirical evidence to sustain support for pronunciation instruction. Furthermore, during the 1990s, many new resources became available for both ESL teachers and learners (Celce-Murcia, Brinton, & Goodwin, 1996; Dauer, 1992), which made it easier for ESL teachers to find pronunciation activities to incorporate into their classes. Breitkreutz et al. sought to determine “to what extent the recent renewal in interest in pronunciation that is reflected in research and teacher resource books is also evident in classroom practice” (p. 51). They found that although some instructors were indeed teaching pronunciation, many also expressed a desire for more training and better materials. Although today pronunciation instruction continues to be underrepresented in second- language acquisition (SLA) research (Deng et al., 2009), several studies have been published since Breitkreutz et al.’s research that have furthered our un- derstanding of adult second-language (L2) learners’ pronunciation issues (Couper, 2003, 2006; Field, 2005; Hahn, 2004; Munro & Derwing, 2006; Zielinski, 2006). Resources for stand-alone courses in pronunciation have flourished, with new editions and new entries (Dale & Poms, 2005; Gilbert, 2005; Grant, 2010; Hewings, 2004) Furthermore, a survey of current teaching materials indicates that publishers now incorporate more pronunciation activities into their gen- eral-skills textbooks than previously (Derwing, Diepenbroek, & Foote, in press). In the light of these developments, we determined that it was time to survey Canada’s ESL instructors to find out if significant changes have occurred in ESL pronunciation teaching practices in the last decade. 2 JENNIFER A. FOOTE, AMY K. HOLTBY, AND TRACEY M. DERWING Breitkreutz et al. (2001) surveyed 67 instructors and/or program coordi- nators, most of whom were from Alberta, British Columbia, or Ontario. The goal of the research was to uncover to what extent pronunciation instruction was being incorporated into curricula and which approaches and materials were being used. Respondents’ attitudes toward, and beliefs about, teaching and learning L2 pronunciation were also probed. They found that many in- structors expressed a desire for training in how to teach pronunciation. Most respondents favored a mixture of segmental instruction (i.e., individual sounds such as /p/ and /b/) and suprasegmental instruction (i.e., broader aspects of pronunciation such as stress and intonation). The respondents also indicated a need for more materials and pronunciation-related curricu- lum development. Although the study described above is the only Canadian survey of pro- nunciation-teaching practices of which we are aware, similar studies have been conducted in other English-speaking countries. Burgess and Spencer (2000) surveyed instructors in the United Kingdom to discover their teaching practices and attitudes toward teaching pronunciation. They found that al- though instructors recognized the importance of suprasegmentals, they found them difficult to teach. Some instructors also indicated that their learners had difficulty perceiving sounds that were not in their first language (L1). Finally, the results indicated that most instructors reported integrating pronunciation instruction into their classes rather than providing stand-alone pronunciation lessons. MacDonald (2002) conducted interviews with eight instructors in the Adult Migrant English Program (AMEP), a large, federally funded ESL pro- gram similar to LINC that focuses on pronunciation teaching practices and views. The results indicated a need for better materials and a stronger cur- riculum to support pronunciation teaching. Further, the study found that as- sessing and monitoring pronunciation were both problematic for instructors. Finally, Burns (2006) conducted a study in Australia with instructors in the AMEP. The survey inquired about instructors’ teaching experience and their confidence levels in teaching pronunciation, as well as the materials and ap- proaches that they favored. Burns found that instructors preferred the teach- ing of segmentals over suprasegmentals. In addition, her findings indicated that although instructors were fairly confident in their abilities, many desired more professional development in teaching pronunciation. Although pronunciation does not receive as much academic attention as other aspects of SLA (Deng et al., 2009), it is of great concern to many L2 learners in Canada. Derwing and Rossiter (2002) surveyed 100 adult ESL- learners to learn about their pronunciation difficulties and strategies. Over 50% of the learners reported that pronunciation contributed to breakdowns in communication. Although it has been argued that L2-speakers may wish to maintain their accents as part of their identities (Golombek & Rehn Jor- dan, 2005), Derwing (2003) found that 95% of L2 immigrant respondents in- TESL CANADA JOURNAL/REVUE TESL DU CANADA 3 VOL. 29, NO 1, WINTER 2011 dicated that they would choose to speak like native speakers (NSs) if they could. Approximately a third of these same respondents stated that they had experienced some discrimination due to their accents, and 53% felt that “Canadians would respect them more” if their pronunciation were better (p. 555). Other studies have also shown that foreign-accent discrimination is a problem faced by many L2-speakers in North America (Lippi-Green, 1997; Munro, 2003). In fact, Davila, Bohara, and Saenz (1993) and Reitz and Sklar (1997) demonstrated that an L2 accent is associated with an economic penalty for some individuals. Although many ESL learners may wish to speak like NSs, the reality is that most people who learn a second language as adults will retain some de- gree of a foreign accent (Abrahamsson & Hyltenstam, 2009; Flege, Munro, & MacKay, 1995). Some researchers believe that it is possible for adults to achieve native-like speech (Bongaerts, Summeren, Planken, & Schils, 1997). However, studies in support of the claim that adult learners of English can pass as native speakers are limited to research about people whose L1 is closely related to their L2 (Bongaerts et al.) or people who have had extensive interaction in their L2 through marriage (Piller, 2002). Although pronunciation instruction is unlikely to lead to native-like speech, it can help L2 speakers improve their intelligibility. Research on the effectiveness of pronunciation instruction is limited, but some studies have demonstrated that instruction can make a positive difference. Derwing et al. (1997) conducted an experiment in which 12 weeks of pronunciation instruc- tion were provided to 13 L2 speakers who had been living in Canada for be- tween two and 20 years (with a mean length of residence of 10 years). The instruction focused much more heavily on prosodic features than on seg- mentals. The learners recorded speech samples at the start and end of the course; these samples were then rated by listeners for accentedness and com- prehensibility and transcribed to measure intelligibility. The researchers found that of the 13 speakers, eight showed significant improvement on at least one of the measures of pronunciation. Couper (2003) gave pre- and post- tests to 15 learners who received pronunciation instruction (focusing on both segmental and suprasegmental features) for approximately two hours per week for 16 weeks as part of an ESL course. He too found that the learners’ pronunciation had improved. Three years later, Couper (2006) conducted a study focusing on epenthesis (inserting an extra sound) and absence (drop- ping a sound) in L2-learners’ speech. He found that after receiving two weeks of special pronunciation instruction interspersed with regular English classes, learners made significant improvements. Finally, in a comparison of form- focused pronunciation instruction focusing on English /ɹ/ with and without corrective feedback, Saito and Lyster (2011) determined that the group that received feedback improved, whereas the group that had no explicit correc- tion did not. 4 JENNIFER A. FOOTE, AMY K. HOLTBY, AND TRACEY M. DERWING In the last two decades, there has been a shift from focusing primarily on segmentals to a greater emphasis on suprasegmentals in many pedagogical materials. Derwing et al. (1998) tested the effectiveness of three types on in- struction on 48 ESL learners. The learners attended ESL classes for 20 hours per week for 12 weeks. One group received no pronunciation instruction, an- other received regular intervals of segmental training, and the third received regular intervals of suprasegmental training. Both groups receiving pronun- ciation instruction showed improvement when reading individual sentences aloud, but only the group that had suprasegmental instruction showed im- provement when speaking extemporaneously. Hahn (2004) investigated the role of primary stress (sentence stress) on intelligibility. NS participants lis- tened to one of three lectures given by a NNS. In one lecture, the speaker used appropriate primary stress, in another incorrect primary stress, and in the third no primary stress. Hahn found that listeners both retained more in- formation and evaluated the speaker more highly when listening to the ut- terances with appropriate sentence stress. She concluded that sentence-level stress is important to ESL learners’ intelligibility. Field (2005) investigated the role of lexical (word-level) stress on pronunciation and found that incor- rect placement of lexical stress had a negative effect on intelligibility. The increasing evidence that suprasegmental instruction is effective does not suggest that segmental instruction has no place in a pronunciation cur- riculum. However, given the limited class time usually allocated to pronun- ciation instruction, it is important to know which sounds should receive the most attention (Munro & Derwing, 2006). One of the most common activities used to practice segmental distinctions is minimal pairs, words that differ by only one sound such as bug/rug or hat/hate. However, in an examination of minimal pairs commonly used in textbooks, Levis and Cortes (2008) found that in half of the minimal pairs that they examined, either one or both of the words was rarely used in speech. This does not mean that minimal pairs should not be used, but rather that they should be selected carefully. A con- cept that is useful in determining which sounds to teach is that of functional load (FL, Catford, 1987). A phoneme with a high FL is more likely to be im- portant in distinguishing between two words than one with a low FL. Munro and Derwing examined the effects of high and low FL errors on the compre- hensibility and accentedness of Cantonese speakers of English. They found that sentences with high FL errors were rated as more accented and less com- prehensible by NSs of English. Given the research findings over the past 10 years, we now address the following questions. 1. What pedagogical training in pronunciation do ESL instructors in Canada have or have access to? 2. How much, how, and with which materials is pronunciation being taught? TESL CANADA JOURNAL/REVUE TESL DU CANADA 5 VOL. 29, NO 1, WINTER 2011 3. What are instructors’ beliefs and attitudes toward pronunciation instruction? In answering these questions, we determine whether pronunciation instruc- tion in Canada has changed markedly in the last decade. Method The Survey The survey was based on Breitkreutz et al.’s (2001) study with some changes and additions. It contained 45 questions that elicited information on several topics including: background information on participants and ESL programs (including classes and students), training opportunities available, resources and activities used, the nature and amount of pronunciation instruction taught, and beliefs about pronunciation instruction. The sections that dealt with materials and approaches were separated based on whether pronunci- ation was taught in stand-alone classes or integrated into general ESL classes; respondents answered applicable section(s). The section that dealt with be- liefs about teaching pronunciation was required of all participants. Multiple- choice questions, yes/no questions, checklists, Likert scales, and open-ended questions were included in the survey, which was piloted by two ESL in- structors. The survey took approximately 20-25 minutes to complete and was delivered using an online Web tool SurveyMonkey (2011, for a copy of the entire survey, please contact one of the authors). Participants In total, 201 individuals responded to the survey. However, because some people abandoned the questionnaire after completing only the first section (demographic information), each response form was manually checked and those including few answers were removed from the data set, leaving a total of 159. At a minimum, instructors needed to answer the short school demo- graphic information section and at least one of the pedagogical sections: either the stand-alone, integrated or “all teachers” section. Generally, the instructors whose responses we analyzed had answered all the questions, leaving only the stand-alone section blank if it did not apply to them. We included re- sponses from program directors although several questions were left unan- swered (i.e., many of the teaching questions did not apply to them). Most of the respondents were instructors (n=129, 85%), whereas the rest were either program coordinators (n=13) or both (n=9). A few fell into other categories such as assessoror team lead. We received responses from eight provinces, but over 80% were from Alberta, British Columbia, and Ontario, the three provinces that receive the largest numbers of ESL newcomers (see Table 1). The breakdown of participants between provinces differed from the earlier survey and was also not perfectly representative of immigration patterns in Canada. For example, fewer respondents were from Ontario and British Colum- bia than from Alberta although the former provinces receive more immigrants. 6 JENNIFER A. FOOTE, AMY K. HOLTBY, AND TRACEY M. DERWING Table 1 Percentage of Respondents from Provinces Province Current Study (%) Previous Study (%) Alberta 33 21 British Columbia 24 31 Ontario 27 42 Nova Scotia 7 – Manitoba 4 – Saskatchewan 3 – New Brunswick 0.7 – Quebec 0.7 – Other – 5 Most of the participants in the current study were in their 40s or 50s (63%), with 16% over the age of 60 and 21% under the age of 40; most were female (89%). Twenty-eight respondents reported a first language other than Eng- lish; the others were native English-speakers. Almost half of the respondents (48%) had been teaching for 10 years or less. The others were divided fairly equally among the following categories: 10-15 years (20%), 15-20 years (16%), and more than 20 years (16%). In terms of TESL preparation, almost half (49%) held a TESL diploma from a college or university, whereas fewer than a quarter (21%) held a master’s in TESL. One person held a PhD in TESL, and 4% had a BEd in TESL (see Table 2). Table 2 TESL Training Levels in Alberta, British Columbia, and Ontario All AB (%) BC (%) ON (%) Provinces (%) No formal TESL preparation 19 0 6 10 A non-credit TESL certificate 19 10 17 16 TESL Diploma from a college 40 71 51 49 or university BEd in TESL 5 0 3 4 MEd/MA in TESL or applied linguistics 17 19 20 21 PhD in TESL or applied linguistics 0 0 3 1 TESL CANADA JOURNAL/REVUE TESL DU CANADA 7 VOL. 29, NO 1, WINTER 2011 Because our sample population was not completely representative of im- migration patterns in Canada, we also compared the education levels of re- spondents from Ontario, Alberta, and British Columbia to see if there were major differences. We found that whereas respondents from Ontario gener- ally had higher overall levels of TESL education, respondents from British Columbia and Alberta were more likely to have received high levels of pro- nunciation-specific training such as a credit course focused specifically on pronunciation instruction (see Table 3). Procedure First, we approached provincial ESL organizations across Canada and asked them to distribute the survey to their members. In some provinces, potential participants were then sent an e-mail request explaining the pur- pose of the study, the participants’ rights, and the procedures. How, when, and whether the TESL organizations chose to share the information with their members is reflected in the response rates. Some organizations made it easier for their members to respond than did others. For example, some provided a direct link, and others required their members to go to another Web site. Some distributed the information right away in spring 2010, whereas others waited until after approval was obtained at a board meeting and sent the information in the summer when fewer respondents were at work. In some provinces, we received no response from the professional organizations; in these instances, we approached programs directly and re- Table 3 Pronunciation-Specific Training in Alberta, British Columbia, and Ontario* All AB (%) BC (%) ON (%) Provinces (%) None 4 3 14 5 Sporadic workshops at conferences 67 71 68 66 As part of a general TESL or 51 74 54 59 linguistics course Linguistics courses 53 68 38 52 A credit course that focuses specifically 22 26 14 20 on pronunciation teaching A combination of linguistics courses 20 23 14 18 and a pedagogical pronunciation course *Respondents were able to check more than one answer. 8 JENNIFER A. FOOTE, AMY K. HOLTBY, AND TRACEY M. DERWING quested that the introductory letter be distributed to ESL teachers on staff. SurveyMonkey provided counts for responses on closed questions; for open-ended questions, each response was coded, entered into a spread- sheet, counted, and double-checked. Results Background Information Most of the participants taught in large programs: 37% were from programs with 11-20 classes and 27% from programs with more than 20 classes. When asked to describe their students using a checklist, the respondents indicated that most were immigrants (81%) and federally funded (67%). Other students (with a higher than 30% response rate) included: exam-prep (IELTS, TOEFL), non-federally funded students (e.g., provincially funded), fee-paying, aca- demic bridging, occupation-specific (e.g., internationally trained nurses), refugees, and international students. Most of the respondents taught at levels between CLB 3 and CLB 6, although all the levels from literacy to CLB 10 were reported. The highest percentage reported teaching CLB 4 and CLB 5. A large number of respondents (85%) indicated that they had students who had requested pronunciation instruction or classes. Of the approximately 40 languages that the respondents suggested were challenging for English pro- nunciation, eight were mentioned 10 or more times (see Table 4). Six of the eight languages listed are from Asian countries; this could reflect the large number of Asian students in Canadian ESL classes. Table 4 Most Troublesome Languages, in Terms of Pronunciation Language Mentions Chinese* 98 Vietnamese 61 Spanish 28 Korean 27 “Asian” 23 Japanese 12 Arabic 10 Thai 10 *The terms Chinese, Mandarin, and Cantonese were used by the respondents. We collapsed them into Chinese. TESL CANADA JOURNAL/REVUE TESL DU CANADA 9 VOL. 29, NO 1, WINTER 2011 Pedagogical Training Fifty percent of respondents indicated that some of the instructors in their institutions had special training in pronunciation, 18% indicated that none did, and 32% were unsure. Respondents were also asked to indicate what types of pronunciation training they themselves had received; most had re- ceived sporadic pronunciation professional development at conference pre- sentations or workshops (66%). Fifty-nine percent had received instruction as part of a general TESL or linguistics course, 52% had taken linguistics courses such as phonetics or phonology, and 20% had taken a credit course at university that specifically focused on L2 pronunciation instruction. In terms of accessibility, most (81%) indicated that instructors could access con- ference presentations, over a third (38%) said that they were able to attend in-house seminars or workshops, and 18% indicated that they could take commercial courses offered by private businesses. Just over half (51%) indi- cated that instructors could access college or university courses that focused on the teaching of pronunciation. Approaches and Materials Integrating pronunciation into general ESL classes. The respondents were rela- tively positive about whether most instructors at their institutions were able to integrate pronunciation instruction into their classes, with 70% reporting that they could. Only 12% said that they could not, and 18% were unsure. However, can and doare not the same, and fewer than half (46%) of the re- spondents agreed that the instructors at their institution incorporated pro- nunciation into their regular classes. A third (36%) were not sure what their colleagues did, and 18% said that their colleagues did not incorporate pro- nunciation. When reporting on their own practices, teachers indicated that they regularly integrated pronunciation into their general ESL classes (86%) and that they regularly corrected mispronunciations (73%). As for the amount of time spent on pronunciation instruction, a percentage was calcu- lated for each teacher when possible. Respondents were asked to provide the total number of teaching hours per week and then select the amount of time spent on pronunciation from a list (less than 15 minutes, 15 minutes, 30 min- utes, 60 minutes, 90 minutes, 120 minutes, and more than 120 minutes). We were unable to calculate exact percentages when respondents selected less than 15 minutes or more than 120 minutes, but we determined rough percent- ages to gain an idea of time spent on pronunciation. First, we examined only those responses that had exact numbers (n=99). We then determined the ranges and means for the total number of instructional hours per week, the total number of hours spent on pronunciation, and the percentage of class time per week devoted to pronunciation instruction. The 99 teachers in this subset taught an average of 17.13 hours per week (Range=2-35 hrs/week). Less than one hour on average was spent on pronunciation (M=0.86, 10 JENNIFER A. FOOTE, AMY K. HOLTBY, AND TRACEY M. DERWING