MATSUURA・CHIBA : Students’ Attitudes Toward Unfamiliar Engl商ish学 A論cc集ent s a第nd8 T2h巻ei第r L14-号sp eak in2g0 B14ac年kgr3ou月nds

【 論 文 】

Students’ Attitudes Toward Unfamiliar English Accents and Their L1-speaking Backgrounds

MATSUURA Hiroko and CHIBA Reiko

Abstract

This study investigates 173 university-level Japanese English learners’ attitudes toward unfamiliar English accents, and the relationships between their attitudes and such sociolinguistic factors as gender, L1 dialectal background and region of residence as well as listening ability that might affect their attitudes. These students, all of whom were from Eastern , including the northern island of Hokkaido, were studying in their univer- sity either in (urban) Tokyo or in (rural) Fukushima. They were asked to listen to several Outer Circle English speakers and to indicate their evaluative impressions on the semantic differential scale. The results suggest that dialectal background is likely a significant factor that affects students’ evaluations of accents, whereas gen- der, region of residence, and listening ability are not. Interestingly, participants who had never spoken any regional dialects indicated more positive ratings toward unfamiliar English accents than those who were speakers of a regional dialect.

Keywords : language attitudes, English accents, Japanese students, L1 dialectal backgrounds

Introduction

Literature reviews Japan has traditionally been categorized as an EFL country, where English is not used on a daily basis while it is taught as a compulsory school subject, and as McKenzie (2008) summarizes, targeted teaching models in English language classrooms were initially RP (pre-World War II), and later (post 1945) standard US English. Jenkins (2009), however, differentiates EFL and ELF (English as a lingua franca) identifying that the norm in the ELF situation is no longer provided by native speakers. She illustrates that in an EFL context, “deviations from ENL are seen as deficiencies,” while in an ELF con- text, “deviations from ENL are seen as legitimate differences.” It seems more appropriate if Japan is placed in the realm of ELF rather than EFL. In an ELF context, where speakers and listeners in the majority of cases do not share their linguistic backgrounds, there are accordingly a wide variety of combinations of speakers and listeners used in research. The following are examples of research conducted with reference to Japanese learners of English.

Study Speakers Listeners

Matsuura et al. (1994) Malaysian, Chinese Malaysian, Bangladeshi, Japanese Sri Lankan, Micronesian, Hong Kong

Chiba et al. (1995) American, British, Hong Kong, Malaysian, Japanese Sri Lankan, Japanese

― 3 ― 商 学 論 集 第 82 巻第 4 号

Amano (2005) British, Japanese, Chinese, German Japanese

Fraser (2006) American, British, Scottish, Zimbabwean, Japanese Taiwanese, Japanese

Horie and Long (2007) Japanese (“native like”, “ English”) Japanese

McKenzie (2008) Glasgow Standard, Glasgow vernacular, Japanese Southern US, Mid-West US, Japanese (“heavily” and “moderately” accented)

Tokumoto and Shibata (2011) Korean, Malaysian, Japanese Korean, Malaysian, Japanese

Rivers (2011) British, Vietnamese, Taiwanese, Indonesian, Japanese Chinese, Indian, Thai, Japanese, Korean, American

Matsuura (2012) American, Hong Kong Japanese

Matsuura, Chiba, and Yamamoto (1994) explores Japanese university students’ attitudes toward dif- ferent varieties of English (Malaysian, Chinese Malaysian, Bangladeshi, Sri Lankan, Micronesian, Hong Kong), and a strong preference for American English with less preference for Asian varieties of English among Japanese listeners is identified. With the basis of the result of the 1994 study, Chiba, Matsuura, and Yamamoto (1995) further examines Japanese students’ attitudes toward different accents, both native and nonnative varieties. In the study, the students’ attitudes toward six different varieties of English (American, British, Hong Kong, Malaysian, Sri Lankan, and Japanese) are investigated. The findings are that Japanese students’ attitudes towards speakers of English from the Inner circle (the U.S. and the U. K.) are more positive than to those of the Outer circle countries (Hong Kong, Malaysia, and Sri Lanka) as well as those of Japan, and the familiarity endorses positive attitudes toward native varieties, i.e. Ameri- can and British. On the other hand, their familiarity with their own variety, i.e. English spoken by Japa- nese students, does not lead to a positive judgment. Amano (2005) conducts research on how Japanese listeners differentiate their evaluation depending on the region from which they might assume the speakers come. The speakers in the study include British, Japanese, Chinese and German, and the listeners are Japanese junior high, senior high, under- graduate and graduate students in Aichi and Gifu prefectures. As the results show, Japanese listeners are not necessarily able to identify the country from which the speaker is from. In fact, the Chinese speaker is judged as a native speaker by more than half of the listeners. When the speaker is judged as “native” or “native-like,” the listeners evaluate the speaker highly as “intelligent” or of “socially high status.” Amano also reports that the Japanese speaker is not evaluated highly as “friendly” and “trust- worthy” in the study, contrary to the result of past research (Ryan, Hewstone, & Giles, 1984), Fraser (2006) investigates Japanese high school students’ perception of accents from six different regions (the U.S., England, Scotland, Zimbabwe, Taiwan, and Japan). The subjects in her study consider the American variety high in status, and the Japanese variety is perceived as being easy to understand ; however, the Japanese variety is not appreciated nor desired as a model to achieve for the subjects. Horie and Long (2007) employs the matched-guise technique to examine Japanese speakers’ atti- tudes toward their own English. In the study, the listeners are Japanese university students, and the speakers are two Japanese university students (1 male and 1 female) who read the passage in the two dif- ferent versions : one with their best “native-like” pronunciation and the other with “katakana-English” pronunciation. The results indicate that “native-like” English is perceived to be more favorable in pro- nunciation, fluency, general English level, intelligence, confidence and looks while “katakana-English” is favored in modesty, similarity and familiarity.

― 4 ― MATSUURA・CHIBA : Students’ Attitudes Toward Unfamiliar English Accents and Their L1-speaking Backgrounds

McKenzie (2008) also conducts a survey where speakers are native and nonnative speakers of 6 dif- ferent varieties, i.e. Glasgow Standard and vernacular, Southern US, Mid-West US, and heavily and mod- erately accented Japanese, and the listeners were Japanese students. He also finds rather negative attitudes of Japanese listeners toward Japanese accented English in terms of competence, yet he identi- fies their different reactions toward their own variety of English in terms of social attractiveness or solidarity. He points out that Japanese listeners indicated positive reactions to highly accented Japanese English indicating characteristics such as being gentle, pleasant, funny and modest. He summarizes those Japanese listeners’ conflicting attitudes towards their own variety, that is, rather negative in com- petence and positive in social attractiveness, in the following way. While “the more ‘Japanese’ the speaker is considered to sound, the less favorably she was evaluated” in competence, “the more ‘Japa- nese’ the speakers sounds, the more positively she will be judged” in social attractiveness. As the rea- son for these contrastive attitudes, influence of “media-transmitted stereotypes” and the little value given to Japanese accented English in the classroom are suggested for low evaluation in competence. On the other hand, a strongly accented Japanese English could make listeners have “in-group identity,” and hence they evaluate the accented Japanese English highly in solidarity. Tokumoto and Shibata (2011) explore the evaluative reactions of Japanese students to their own English, comparing them with those of other Asian groups (Korean and Malaysian). Their results indi- cate that Japanese negative attitudes toward their own variety were the strongest among the three groups, Malaysians being most positive toward their own variety. As Tokumoto and Shibata state, possi- ble reasons for Malaysians admiring their own variety of English include language situations in the country. Malaysia is a multilingual country where Malay, Chinese, Chinese dialects and Indian lan- guages are spoken, and also English is a second language that has been developed as its own variety. Rivers (2011) employs a semantic differential questionnaire to elicit Japanese students’ preference among different varieties of English (British, Vietnamese, Taiwanese, Indonesian, Chinese, Indian, Thai, Japanese, Korean, and American). The results show that rather strong Japanese students’ preference for native English speakers over those of other varieties of English, even when a speaker is not a native English speaker but the Japanese listener only perceives the speaker to be a native English speaker. Matsuura (2012) tries to identify possible factors that may influence the listeners’ preference for dif- ferent varieties of English. In the study, Japanese listeners’ reactions to Standard American English and Hong Kong English are examined. Among three factors, i.e. perceived comprehensibility, accentedness and fluency, the results indicate that perceived comprehensibility is a best predictor for preference. On the other hand, familiarity with a given English accent and speech rate do not appear to influence the lis- teners’ evaluative and affective ratings. In addition, she points out that the more time learners spent on English, the more negative views they may have toward non-standard English. While speakers were carefully selected taking their linguistic backgrounds into consideration, both native and nonnative, listeners sharing the same L1 are not usually clearly identified except for their pro- ficiency level, age and gender. When we closely examine the characteristics of listeners, a group of the “same L1” listeners are not necessarily homogenous with regard to their L1 dialects. Since dialectal backgrounds may affect their perception and attitudes toward varieties of languages, both their own language and English, in addition to proficiency, age, and gender, listeners’ dialects may need to be included as one of the listeners’ attributes. There have been some studies conducted on atti- tudes toward a variety of dialects in a given country. For example, Miyamoto (2010) reports Chinese university students’ attitudes toward ‘Shanghainese’ and standard Chinese, comparing reactions of infor- mants from Shanghai with those from distant regions. He identifies that the further the areas of infor-

― 5 ― 商 学 論 集 第 82 巻第 4 号 mants from Shanghai get, the lower the evaluation of Shanghainese becomes while standard Chinese receives positive reactions regardless of the regions. As possible factors affecting the attitudes, he points out the influence of the economic status of the area where a particular dialect is spoken along with cultural and historical backgrounds. The language policy promoting standard Chinese by the govern- ment, on the other hand, may contribute to a positive evaluation regardless of regions. Japan is generally classified as a monolingual country where Japanese is the only language, and other languages including English are used among limited groups of people and places. Thus, in the research done with Japanese people, the subjects are considered as homogeneous linguistically. In reality, how- ever, there are a variety of dialects spoken in Japan, and their pronunciation and vocabulary are distinctive and widely different. People who are bidialectal and switch between standard Japanese and dialects depending on the context are not uncommon. Long (1997) points out that dialects in Japan are basically regional and different from those in the U.K. and U.S. where societal dialects have been developed with industrialization and urbanization. Gottlieb (2005) summarizes that a dialect spoken in the Kanto area, specifically in Tokyo, is com- monly considered to be standard Japanese, and dialects in Japan vary not only in their accents but in lexi- cal items as well. As she points out, the use of dialects among people in Japan has changed in that more children have been educated in standard Japanese and when those children become parents who speak more standardized Japanese, dialects are spoken only on limited occasions. Okamoto (2001) investigates how the use of the dialect may influence listeners’ evaluation of the speaker, employing the matched-guise technique where a male and a female Japanese speaker read the monologue in two different versions ; one with standard Japanese and the other with Nagoya dialect. While the male speaker is evaluated as higher in intelligent, active and appearance categories when using standard Japanese but more sociable using the dialect, for the female speaker, using standard Japanese solicits better evaluation in most respects. Carroll (2001) illustrates different Japanese reactions toward their own dialects between respon- dents from urban areas and those from rural areas citing the results of the NHK (Nippon Hoso Kyokai, Japan Broadcasting Corporation) survey. According to the survey, respondants from urban areas indicate more favorable and positive attitudes toward than those from rural areas. A question of whether those Japanese attitudes toward their own dialects has an influence on their judgment of unfa- miliar non-native varieties of English may elicit insightful aspects of language awareness.

Research questions The purpose of this study is to examine what are the sociolinguistic variants of English learners that might affect acceptance of English accents that are totally unfamiliar to them in their English language learning and using environments. These variants are such sociolinguistic aspects as L1 dialectal back- grounds, the regions where they live, and gender. In addition to these, English proficiency, which is assessed by a listening comprehension test is included as an independent variable. Participants in this study are solely English learners whose proficiency is still limited, and the investigators o.f this study thus considered that their English level might affect their reactions when they encounter unfamiliar English. We hope to identify what types of learners are more likely to accept unfamiliar accents, and, in turn, what types are less likely. In this study, we recruited English speakers from India, Sri Lanka, Kenya, and Ghana, all of which are Outer Circle countries where English is used as their second language as well as their official language. These English varieties are also unfamiliar to the vast majority of Japanese EFL learners.

― 6 ― MATSUURA・CHIBA : Students’ Attitudes Toward Unfamiliar English Accents and Their L1-speaking Backgrounds

Unfamiliar English can be seen as an English variety in which listeners perceive some unfamiliarity in such linguistic aspects as phonology, morphology, semantics, and pragmatics. Among these diverse lin- guistic aspects, this study focuses on phonological unfamiliarity only. The research questions addressed in this study are as follows : 1) Which group of students is more likely to accept unfamiliar English accents : those who speak accented Japanese themselves or those who speak standard Japanese only ? 2) Which group of students is more likely to accept unfamiliar English accents : those who live in an urban area or those who live in a rural area ? 3) Which gender of students is more likely to accept unfamiliar English accents : male or female ? 4) Which proficiency level of students is more likely to accept unfamiliar English accents : higher or lower ?

Method

Questionnaire The questionnaire consisted of two parts. The first part was to elicit participants’ attitudes toward unfamiliar English accents. This study employed a verbal guise technique, which has been widely used in language attitude studies in Japan and other parts of the world (e.g., Matsuura et al., 1994 ; Chiba et al., 1995 ; McKenzie, 2008 ; Rindal, 2010 ; Rivers, 2011). This first part featured a list of six adjectives arrayed on a seven point bi-polar rating scale. A list of adjectives used in McKenzie’s study (2008), which investigated Japanese students’ reactions toward English accents, were found to be grouped into two semantic categories : competence and social attractiveness. Rindal (2010), following the line of Ladegaard (1998), states that the semantic category of status and competence involves intelligence, edu- cation, ambition, leadership skills, self-confidence, social status, formality, and politeness, while the social attractiveness category refers to reliability, likeability, generosity, sense of humor, popularity, and attractiveness. Following the line of McKenzie (2008) and Rindal (2010), the investigators of this study selected six adjectives : intelligent, confident, and polite for the category of competence and reliable, lik- able, and kind for social attractiveness. After the survey, we ran a factor analysis on participants’ ratings regarding these six adjectives. And as a result, two main factors were extracted (Table 1). Factor 1 was named as social attractiveness and Factor 2 as competence. These two factors explained the 69.84% of the total variance. The results suggested our six adjectives were applicable for the social attractiveness and competence categories.

Table 1 : Results of factor analysis

Factor 1 Factor 2

likability 0.814 reliability 0.812 kindness 0.783 politeness 0.731 confidence 0.790 intelligence 0.675

The second part of the questionnaire asked such general backgrounds of the participants as age, gender, and major as well as their dialectal backgrounds, which specifically asked where they were from

― 7 ― 商 学 論 集 第 82 巻第 4 号 originally, where they were living at the time of the survey, and whether or not they were speaking any regional dialects of Japanese. The entire questionnaire was written in Japanese so that no participants would misunderstand it. As seen from participants’ TOEIC scores (range=350-800, average=457.68), the L2 proficiency of many was still limited, and the investigators considered that lower level students might be more comfortable with the Japanese questionnaire.

Speakers Of ten Outer Circle English speakers who provided us with their voice recordings, five speakers were selected as readers for this study. In the speaker selection procedure, four English instructors, two native Japanese speakers and two native English speakers, evaluated these ten speakers in terms of their accentedness and fluency. In order to assess these two aspects, we prepared two statements “This speaker’s accent is different from North American English accent” and “This speaker read the passages fluently,” and a seven-point rating scale which ranged from 1 “strongly disagree” to 7 “strongly agree.” By tallying accentedness and fluency ratings of four English instructors, the following five speakers were selected as readers : a male from Kenya, a male from Ghana, a female from Sri Lanka, a female from India, and a male from India (Table 2). All of them were graduate students studying in U.S.

Table 2 : Backgrounds of the speakers

age gender country L1 length in US

Speaker 1 26 M Kenya Swahili 8 years Speaker 2 44 M Ghana Akan 9 years Speaker 5 26 F India Bengali 2 years Speaker 6 26 F Sri Lanka Sinhalese 1 year Speaker 10 31 M India Hindi 3 years

Listeners A total of 173 undergraduate students, 82 (47.4%) males, 88 (50.9%) females, and two (0.1%) unre- ported, participated as listeners. All of these students are native Japanese speakers, taking at least one English language course at their university. Their major varied and their general English proficiency level varied (87 students reported their TOEIC scores, which ranged from 350 to 800) as well. Initially, 186 Japanese students participated in this study, but 11 of them indicated in the questionnaire that they experienced living abroad for more than one month (one month to six years), and another two students answered that they were exposed to Kenyan English, a variety we employed in this study, almost everyday. We excluded these 13 students from having been participants in this study. Among the 173 participants, 99 (57.2%) were enrolled in university in Northeast Japan and 73 (42.2%) in the Tokyo metropolitan area. They were living in either of these two areas, rural and urban, at the time of the survey. Regardless of the city where these participants originated, students studying in the urban area were regularly encountering standard Japanese, whereas students in the rural area were fairly frequently experiencing local Japanese accents. The prefectures where the participants were originally from varied but all of them were from either Northern or Eastern Japan. For example, 29 were from Fukushima, 29 from Tokyo, 16 from Miyagi, 13 from Saitama, 9 from Tochigi, 6 from Kanagawa, 4 from Aomori, and 3 from Hokkaido. Accordingly, L1 dialects these people were speaking varied as well. However, 94 (54.3%) reported that they had never spoken non-standard Japanese (they

― 8 ― MATSUURA・CHIBA : Students’ Attitudes Toward Unfamiliar English Accents and Their L1-speaking Backgrounds were speaking standard Japanese only), whereas 78 (45.1%) were speaking or had spoken a regional dia- lect previously. The latter group of students could speak standard Japanese as well, depending on the situation and interlocutors. Therefore, these 78 students were categorized as a group speaking standard Japanese plus non-standard Japanese. Fifteen listening comprehension questions from the Official TOEIC Bridge Guidebook and Official TOEIC Bridge Workbook were given to participants to assess their English proficiency. The reading passages for these questions were read by a native speaker of North American English, a variety most familiar to Japanese learners in general. After listening to each passage, participants read the question and selected the best option out of four within a time limit of 10 seconds. The mean score of this test was 10.62 (SD=2.394) with the range of 2-15, and the medium was 11.00. The investigators then divided them into three groups based on their test score : higher (12-15), medium (10 and 11) and lower (2-9). The higher group consisted of 64, medium 64, and lower 45 students.

Procedure Data were collected during the last 20 minutes of regular class time, although student participation was voluntary. The investigators explained that they were going to be given two tasks : evaluation of others’ accents and a listening comprehension test. Students were also told that they could withdraw at any time. Since most of the participants had never indicated their impression using a semantic differential scale, participants practiced how to answer on the scale in Part 1. They were asked to listen to one of the speakers (Speaker 10) reading a short passage, and to choose a number which they felt the closest to their impression. And any questions were cleared at this stage. Then they listened to the second speaker. The CD containing the audio file was stopped while students were answering on the evaluat- ing scale. The CD was played only once. The data from Speaker 10 were not used for analyses. Participants were then asked to answer the second part of the questionnaire, which was on such general background as their age, gender, and major as well as such linguistic backgrounds as where they were originally from, where they were living, and whether or not they were speaking any regional dia- lects and/or accented Japanese at the time of the survey or prior to it. In addition to these, participants indicated their TOEIC scores, experiences of living abroad, and experiences with English varieties, if they had any. After the task of evaluating reading passages, participants were asked to take the listening compre- hension test recorded by a speaker of North American English, a variety they were most familiar with. They listened to each of fifteen short passages, and answered a question. It took approximately 10 min- utes to complete this listening test.

Treatment of the data The investigators of this study did not consider that the Outer Circle speakers who contributed their voice recording were representatives of their own English variety. Nor did we intend to compare acceptability levels of English accents among the speakers. They were employed for this study mainly due to their availability. Our interest was solely in individual differences among listeners in evaluating English speakers whose accents were unfamiliar to them. Thus, evaluation ratings for each speaker were tallied and averaged for the use of analyses.

― 9 ― 商 学 論 集 第 82 巻第 4 号

Results

Our first research question asked which group would be more likely to accept unfamiliar English accents : those who were speaking accented Japanese themselves or those who were speaking standard Japanese only. In order to examine this, a t-test was run on their ratings of the four speakers combined. The results found that the group who had never spoken accented Japanese was more favorable toward unfamiliar English accents compared to the group who were speaking regional dialects themselves (Table 3).

Table 3 : Standard only vs. standard plus non-standard

N Mean SD t p

standard only 94 103.11 10.82 2.019 0.045 standard & non-standard 78 99.53 12.43

The second research question was concerned with area of residence at the time of the survey. Which group of students would be more likely to accept unfamiliar English, the group living in an urban area or the group living in a rural area ? These two regional distinctions were determined by the loca- tion in which their university was : Tokyo or Fukushima. The participants in this study were all resid- ing in either of these two cities or in its suburbs. The results indicated that although students in Tokyo tended to indicate a higher average than those in Fukushima, no significant difference was obtained between these two student groups.

Table 4 : Urban vs. rural

N Mean SD t p

urban 73 103.04 10.748 1.508 0.133 rural 99 100.33 12.251

The third research question asked which gender would be more favorable toward unfamiliar English accents : male or female. As indicated in Table 5, although the female group tended to indicate higher ratings than the male group did, no significant average difference was found between the groups.

Table 5 : Male vs. female

N Mean SD t p

male 82 100.54 11.762 -1.125 0.262 female 88 102.56 11.646

The last research question was with regard to participants’ English proficiency assessed by a listen- ing comprehension test recorded in North American English, an English variety most familiar to them. The Pearson Correlation Coefficients found the test scores and evaluation ratings were not significantly correlated (r=0.057, p=0.230). The investigators then divided the participants into three groups depending on their test scores : higher, middle, and lower proficiency groups. A t-test run between the higher and lower proficiency groups did not find significant average difference (Table 6). It appeared that English proficiency was less likely to affect English students’ evaluations of unfamiliar accents.

― 10 ― MATSUURA・CHIBA : Students’ Attitudes Toward Unfamiliar English Accents and Their L1-speaking Backgrounds

Table 6 : Higher vs. lower proficiency

N Mean SD t p

higher 64 101.31 12.038 0.680 0.498 lower 45 99.64 13.391

Discussion

A notable finding in this study was that participants who had never spoken accented Japanese were more likely to accept unfamiliar L2 accents, compared to bidialectal speakers of Standard Japanese and their own regional dialect. It was likely that participants’ positive or negative attitudes toward L1 regional dialects influenced their attitudes toward L2 accents to some extent. Although we did not ask which city they were originally from, non-dialect speakers in this study were assumed to be city dwellers possibly from their birth or, at least, from their early childhood. Such participants seemed to originate in the Tokyo Metropolitan area or a relatively large city in Northern and Eastern Japan. Carroll (2001) cited an NHK survey, which found that respondents from the more urban areas tended to place greater emphasis on dialects when asked whether education should place value on dialects or put more energy into the standard language. As this NHK survey suggests, people in urban areas are more likely to accept Japanese dialects than those who actually live in rural areas. Carroll (2001) explains that this is partly because urban people tend to feel nostalgia for a furusato, or an image of a rural home village. She continues that this image is associated specifically with rural dialects, which are perceived to be warm, expressive, gentle, for being among friends, etc. As this NHK survey suggests, city dwellers appear to be rather tolerant to L1 dialects and accents, and their general tolerance to accents seems to positively influence their reactions to L2 accents when they actually encounter any. Rural people, on the other hand, apparently do not feel that way as actual speakers of their respective regional dialects. In the history of Japanese language education, there were days when dialects were treated negatively. The post war school curricula issued by the Ministry of Education in fact state that correct forms should be used and that slang and dialects should be avoided as much as possible. Those days when regional dialects were treated with rather negative connotations continued from right after World War II through the early 70s. It is natural to assume that rural people are still humiliated by their own dialect, and hence try to avoid using it, particularly in front of non-dialect speakers. It is likely such negative views of dialect speakers in regard to L1 accents affect their perceptions of L2 accents rather negatively. The finding of this study that attitudes toward English accents significantly differed between dialect speakers and non-dialect speakers did not appear to support the results of McKenzie’s 2008 study, which detected no significant difference between urban and rural people’s attitudes in listening to various Eng- lish accents. This is probably because McKenzie, in his questionnaire, rather focused on the regional provenance of his subjects and did not ask about their dialectal backgrounds, i.e., whether or not they spoke any regional dialects. Regional provenance does not necessarily tell to what extent subjects are involved with dialect-speaking situations, as a speaker and/or listener, in that region. Nowadays, younger generations of people in many rural areas tend to speak standard Japanese, particularly in public, while speakers of heavily-accented Japanese are mostly from older generations. Perhaps age is another important influential factor for researchers to consider in their attitude studies. Also which part of Japan respondents are from might make results different because people’s perceptions with regard to their own dialect vary depending on region. For example, people in Tohoku, which refers to the prefectures of

― 11 ― 商 学 論 集 第 82 巻第 4 号

Aomori, Iwate, Akita, Yamagata, Miyagi, and Fukushima, tend to hesitate to use their dialects when speaking to interlocutors from a different region whereas Kansai (the prefectures of Osaka, Kyoto, Shiga, Nara, and Wakayama) people generally do not. Because of these variations, more needs to be taken into account by researchers if they hypothesize the use of L1 regional dialect affects listener attitudes toward L2 accents. This study is limited to the investigation of English learners’ evaluative reactions to several English Outer Circle English accents, which are unfamiliar to the vast majority of Japanese students. The Outer Circle English speakers who contributed voice recordings for this study were fluent English speakers but they were not as fluent as Inner Circle English speakers in general. The investigators of this study did not investigate students’ attitudes toward native English accents, both familiar and unfamiliar ones. A future study should examine how attitudes toward unfamiliar Inner Circle English are different from those toward familiar Inner Circle English. And if there were any difference in attitudes, researchers should further examine what kind of influential factors are involved there. This should be done possibly from different angles and with different methods.

Conclusion

This study investigated tertiary-level Japanese English learners’ attitudes toward unfamiliar English accents, specifically focusing on such sociolinguistic factors of gender, L1 dialectal background, and region of residence that might affect their attitudes. The effect of listening comprehension ability of reading passages recorded in North American English, i.e., a variety most familiar to Japanese English learners in general, was investigated as well. The results suggest that dialectal background is likely a significant factor that affects students’ evaluations of accents, whereas gender, region of residence, and listening ability are not. Interestingly, participants in this study who had never spoken any regional dialects indi- cated more positive ratings toward unfamiliar English accents than those who were speakers of a regional dialect (although they were speakers of Standard Japanese as well). A possible explanation of this is that dialect speakers were less tolerant to L1 accents, and they might have held similar negative attitudes when they encountered nonnative English accents which were totally unfamiliar to them. Non-dialect speakers, on the other hand, seem to be more tolerant to accented Japanese possibly with their nostalgic images with regard to regional dialects in general. It is likely they tended to hold similar tolerant atti- tudes toward L2 accents as well. However, due to the limitations of this study, these explanations still remain as speculation. Future studies are needed to confirm the findings of this study and for deeper understanding of people’s attitudes, which appear to be rather complex in reality.

References

Amano, S. (2005). Nihonjin eigo gakushusha no hibogoeigo ni taisurutaido [Japanese learners’ attitudes toward nonna- tive English : Influence of the nonnative accents on the evaluation]. Tagenbunka 5, 1-14. Carroll, T. (2001). Language planning and language change in Japan. Richmond, Surrey, UK : Curzon Press. Chiba, R., Matsuura, H. & Yamamoto, A. (1995). Japanese attitudes toward English accents. World Englishes 14, 77-86. Fraser, S. (2006). Perceptions of varieties of spoken English : implication for EIL. In R. Kiely, P. Rea-Dickins, H. Woodfield & G. Clibbon (Eds.), Language, Culture and Identity in Applied Linguistics (pp. 79-97). London : British Association for Applied Linguistics in association with Equinox. Gottlieb, N. (2005). Language and Society in Japan. Cambridge : Cambridge University Press.

― 12 ― MATSUURA・CHIBA : Students’ Attitudes Toward Unfamiliar English Accents and Their L1-speaking Backgrounds

Horie, C. & Long, C. (2007). Japanese attitudes toward their own English : A matched-guise study. Paper presented at Shakaigengo kagaku 19th conference. Jenkins, J. (2009). World Englishes. London : Routledge. Ladegaard, H.J. (1998). National stereotypes and language attitudes : the perception of British, American and Austra- lian language and culture in Denmark. Language and Communication 18, 251-274. Long, D. (1997). Hogenkaramita nihongorashisa [Japaneseness of dialects]. Nihongogaku, July Issue, 6-13. Matsuura, H., Chiba, R. & Yamamoto, A. (1994). Japanese college students’ attitudes towards non-native varieties of English. In D. Graddol & J. Swann (Eds.), Evaluating Language (pp. 52-61). Clevedon : Multilingual Matters. Matsuura, H. (2012). Japanese learners’ evaluative reactions and affective responses toward English accents. The Shogaku Ronshu 80, 4, 21-32. McKenzie, R.M. (2008). Social factors and non-native attitudes toward varieties of spoken English : a Japanese case study. International Journal of Applied Linguistics 18, 63-88. Miyamoto, D. (2009). Chugokujin no gengo hyoka [Language Evaluations of Chinese University Students from Beijing, Tianjin, Shanghai, and Hangzhou]. The Japanese Journal of Language in Society 2, 55-68. Okamoto, S. (2001). Effects of the Nagoya Dialect on the Impression of Speakers : Using the Matched-guise Technique. The Japanese Journal of Language in Society 3, 2, 4-16. Rindal, U. (2010). Constructing identity with L2 : pronunciation and attitudes among Norwegian learners of English. Journal of Sociolinguistics 14, 240-261. Rivers, D.J. (2011). Intercultural processes in accented English. World Englishes 30, 375-391. Ryan, E.B., Hewstone, M. & Giles, H. (1984). Language and intergroup attitudes. In J.R. Eisher (Ed.), Attitudnal Judg- ment (pp. 135-158). New York : Springer-Verlag. Tokumoto, M. & Shibata, M. (2011). Asian varieties of English : Attitudes towards pronunciation. World Englishes 30, 392-408.

― 13 ―