Department of English
Bachelor Degree Project English Linguistics
Autumn 2020
Swedish compulsory school students’
attitudes toward English accents:
Exploring how
familiarity affects our language attitudes
Leonardo Hansson
Swedish compulsory school students’ attitudes toward
English accents: Exploring how familiarity affects our language attitudes
Leonardo Hansson
Abstract
This study will explore to what extent familiarity with English accents can influence compulsory school students’ attitudes towards them. Data from questionnaires completed by 98 students were analysed. The results show that the degree of familiarity with the English accent seems to affect the attitude attributed to it. More specifically, the results indicate that a higher degree of familiarity influences the ability to express an attitude. A lower degree of familiarity leads to similar attitudes being given to the accents, which shows a lack of differentiation between them. The results also indicate a bias towards RP. While it is not necessarily harmful, teachers should be aware of this and how their own teaching may influence how different accents are perceived. It is argued that teachers need to intervene in the process of stereotyping which will help develop an awareness of students’ language attitudes. To summarize, it is difficult to draw any wide conclusions from these results due to the study’s scope. Furthermore, the target group is not representative of Swedish compulsory school students as students from the chosen school generally finish with an above-average final grade. Further research is necessary to determine more specifically how familiarity affects attitudes of English accents and if these findings recur in other areas of Sweden where the final grade average is lower.
Keywords
English accents, language attitudes, Swedish compulsory school, Received Pronunciation, General American, Australian English, Irish English.
Contents
1. Introduction ... 1
2. Background ... 2
2.1 Varieties of language ... 2
2.2 Language attitudes ... 2
2.3 Familiarity ... 3
2.4 English in Sweden ... 4
3. Method ... 5
3.1 Data ... 5
3.2 Ethical considerations ... 8
3.3 Possible limitations ... 8
3.4 Analysis ... 8
4. Results ... 9
4.1 Familiarity, Exposure and Recognition ... 9
4.2 RP VS GA ... 11
4.3 AusE vs IrE ... 11
4.4 RP and GA vs AusE and IrE ... 11
5. Discussion ... 12
5.1 Implications for teaching ... 13
6. Conclusion ... 14
References ... 15
Appendix A ... 17
1. Introduction
According to Holmes, “...there is no universal consensus about which languages sound most beautiful and which most ugly, despite people’s beliefs that some languages are just inherently more beautiful than others” (Holmes, 1992, p. 410). These views on language are called language attitudes and are basically social attitudes to language (Eriksson, 2019).
Attitudes towards language can affect how motivated a person is to acquire a second language, which holds implications for language teachers (Holmes, 1992). There have been studies on upper secondary students’ attitudes towards English accents (Eriksson, 2019; Rindal, 2014) as well as on university students (Carrie, 2017). These studies found that British English (BrE) is usually seen as the formal variant of English, perceived by students as conveying status, competence, and intelligence. On the other hand, American English (AmE) is usually rated higher in its social attractiveness.
Rindal (2010) explains that these attitudes might derive from how accents are perceived to have formal or informal functions. BrE is associated with school, a more formal setting, while AmE is considered informal as its exposure predominantly happens outside of school.
But what is known about compulsory students and specifically Swedish compulsory school students in year nine? The Swedish school curriculum differs greatly between year nine and upper secondary school. The core content in compulsory school English in ninth grade includes “Spoken English with slight regional and social accents”
(Skolverket, 2008, p. 37) whilst upper secondary school English includes “Spoken language, as well as different social and dialect features” (Skolverket, 2018, p. 37). The difference between the core contents and the added time of exposure that students get could explain why upper secondary school students can recognize accents that compulsory school students cannot. The added exposure in upper secondary school English could also help the students differentiate between different accents. In addition, becoming older leads to more possible exposure in the students’ free time through media consumption.
Attitudes are not held by people in a vacuum they are developed towards languages displaying their views of those who speak it, including the “contexts and functions with which they are associated with” (Holmes, 1992, p. 410). Holmes (1992) argues that the assessments of people that listen to unfamiliar accents are totally random. Other research has shown that familiarity is a crucial determiner for individuals’ language attitudes as it displays social connotations towards a specific speech variety (Giles et al., 1974; Giles et al., 1979, as cited in Kircher & Fox 2019). However, Matsuura & Chiba (2014) suggest there may also be other factors that affect students’ reception of unfamiliar English accents. The results of their study suggest that one’s dialectal background is a more significant factor than gender, listening ability or region of residence in a student’s assessment of accents. They also found that participants without regional dialects gave more positive ratings to unfamiliar English accents than those participants with regional dialects. These findings suggest that there could be variables other than randomness attached to attitudes of unfamiliar accents, but as Matsuura and
Chiba (2014) acknowledge, the study findings remain speculatory due to limitations.
These findings validate the surveying of year nine compulsory school students’
language attitudes, at least theoretically, as attitudes towards unfamiliar accents can hold implications beyond randomness.
As the research of students’ language attitudes towards English has mostly been conducted for higher levels of education, this study aims to broaden the understanding of Swedish compulsory students’ attitudes towards English accents. This study also aims to explore to what extent familiarity with English accents can influence students’
attitudes towards them. The insights gained from the study can be useful for language education in Sweden, as the understanding of students’ attitudes allows us to notice and if needed prevent stereotypes.
2. Background
2.1 Varieties of language
All languages “exhibit internal variation” according to Wardhaugh, meaning that “each language exists in a number of varieties and is in one sense the sum of those varieties”
(2010, p. 25). The varieties themselves are “a set of linguistic items with similar distribution” (Hudson, 1996, p. 22, as cited in Wardhaugh, 2010). Wardhaugh states that this allows us to claim London English and Canadian English as varieties of a language (Wardhaugh, 2010).
The term variety is often used by linguists as it avoids the conflict of distinctions between the terms language, dialect, and accent which are hard to define (Melchers et al., 2019). It is explained that dialect encompasses differences in pronunciation, grammar, and vocabulary, while accents only differ in pronunciation (Melchers et al., 2019). Melchers et al. (2019) explain that variety loosely covers both dialect and accent which is why linguists prefer it.
In an attempt to classify English accents, Trudgill and Hannah (2008, as cited in Melchers et al., 2019) identified four main types of English: English-based, spoken in England, Wales, South Africa, Australia and New Zealand; American-based, spoken in the United States and Canada; Scottish-based, spoken in Scotland and Northern Ireland;
Irish-based, spoken in the Republic of Ireland. Notably, this categorization relates to standardized first-language varieties as mentioned in Melchers et al. (2019).
2.2 Language attitudes
Carrie (2017) defines language attitudes as an individual’s “evaluation of and disposition towards a speech variety and its speakers, consisting of thoughts, feelings and behavioural tendencies...” (2016, p. 430). Linguists argue that these dispositions are entirely subjective and state that there is no variety of a language better than any other (Wardhaugh, 2010). Even so, Holmes (1992) explains how certain communities regard one particular sound as good speech while another community can regard it as evidence of a lack of education. These attitudes lead to the widespread belief that a speaker can
be advantaged or disadvantaged both socially and aesthetically but also intellectually by the mere chance of speaking a variety of a language (Wardhaugh, 2010).
Most studies on language attitudes have surveyed Received Pronunciation (RP) and General American (GA) as these are considered to be reference varieties that broadly represent English spoken in England and the United States (Melchers et al. 2019).
Daniel Jones, a Professor of Phonetics who developed the concept of RP, did not believe it was better than any other accent simply argued that RP was easily understood (Melchers et al. 2019). In Melchers et al. (2019) General American is referred to as an American accent without great regional colouring.
Previous research on students’ language attitudes in Nordic countries has found that students seem to favour RP in competence ratings (Eriksson, 2019; Ladegaard &
Sachdev, 2006; Rindal, 2010). Eriksson (2019) and Carrie (2017) have argued that these findings derive from RP being the main variant of pronunciation being taught in school.
In social attractiveness, GA has been higher rated in Rindal (2010) but the opposite was found by Jarvella (2001).
In attitude research, the gender composition of a study’s participants may be useful in determining if the distribution is representative of the target group. Still, there are studies that do not display this information as seen in Chiba et al. (1995), Eriksson (2019), McKenzie, (2008) and Rindal (2010). Other studies such as Jarvella et al. 2001 and Ladegaard and Sachdev 2006 have included their sample’s gender composition and despite having a majority of female participants (85,7% and 75% respectively) it is not addressed how it may affect the studies’ findings. Hartikainen (2000, as cited in McKenzie, 2010) found that other background factors such as age and gender were not found to be significant in determining the attitudes of the informants. More recently, Matsuura and Chiba (2014) found that one’s dialectal background is a more significant factor than gender which may suggest that the sample's gender composition is less crucial in attitude research.
2.3 Familiarity
Previous research has shown that familiarity of specific varieties’ social connotations is an important determiner of an individual’s language attitudes (Giles et al., 1974; Giles et al., 1979, as cited in Kircher & Fox 2019). There has been no explicit definition of familiarity in the literature reviewed for this study. Accent familiarity (Winke et al., 2013) or familiarity with accents, as used in Chiba et al. (1995), is the measurement of how familiar an individual is with a particular accent. Familiarity as a determiner was used by Winke et al. (2013) and Saito et al. (2019) in their separate studies of L2 speakers of English rating oral performance of other L2 English speakers. While rating oral performance is different from language attitude research, the methods utilized by Wink also measured familiarity but with a scale in contrast to Chiba et al. (1995) who measured it through a binary recognition test. Other researchers like Eriksson (2019) found that upper secondary school students in Sweden seemed to prefer American English over British English as they were more exposed to it. Exposure is a prerequisite for familiarity, and according to Winke et al. (2013), the length and intensity of the exposure can lead to varied types and levels of familiarity.
Researchers that have looked at language attitudes and familiarity directly (Chiba et al., 1995) found that familiarity with native accents leads to a more favourable view of the native speakers’ accent. Chiba et al. (1995) only operationalized familiarity with binary recognition tests. However, previous researchers have argued that the degree of familiarity is particularly important, as in Tauroza and Luk’s (1997) study on listening comprehension. While the degree of familiarity has not been operationalized with attitudes previously, Wink et al. (2013) suggest that accent familiarity can be defined and operationalized in many ways. Familiarity degree will therefore be measured in this study as the literature supports its viability in similar areas of research.
2.4 English in Sweden
The Swedish national test in English from 2019 utilized both British and American accents of English as seen in archives of the national test provided by Gothenburg University (2020). Therefore, at least in terms of comprehensibility, Swedish compulsory students must be able to understand these spoken varieties of English to meet the knowledge requirements. As previously mentioned, the English core content in years 7-9 should cover “Spoken English with slight regional and social accents”
(Skolverket, 2018, p. 37), leaving the teachers to decide what accents the students are exposed to. It is also possible that Swedish compulsory school students can recognize a wide array of accents as Sweden is placed fourth in EF’s (2020) English proficiency index.
Compulsory school students’ ability to recognize accents is not well known; there has been research on upper secondary school students in Denmark by Ladegaard (1998), displaying that 71% recognize GA, 66% RP, and 8% Australian English (AusE).
Another study in Denmark by Jarvella et al. (2001) on university students found that 91% recognized RP, 81% GA, and 54% Irish English (IrE). In contrast, a study in Turkey by Yaman (2015) found that university students lacked the ability to fully distinguish between the British and American varieties of English. While Yaman (2015) states that these findings cannot generalise the whole Turkish EFL context, it holds some implications. This result seems to be in line with EF’s English proficiency index which places Turkey in 69th place with a low proficiency rating. The proficiency level of English is vastly different in the world and even in Europe where countries range from moderate to very high proficiency according to EF (2020). Thus, the assumption that Swedish compulsory students are capable of both recognizing and differentiating between several English accents holds some ground with Sweden’s very high level of English proficiency. Even so, one can possess a high level of English proficiency and still not be able to recognize nor be familiar with a wide gamut of English accents as the individual received sufficient exposure. This could be the case for some accents, but it seems unlikely that Swedish students would not recognize RP and GA and they are likely similar to students in Norway, where Rindal (2014) found that students identified RP and GA with a 77-93% success rate.
3. Method
The purpose of this quantitative study is to explore to what extent accent familiarity affects compulsory students’ attitudes towards English accents. Previous research on language attitudes has utilized the matched-guise technique (MGT) and verbal-guise technique (VGT) which has shown to facilitate consistency to a high degree in data collection (McKenzie & Carrie, 2018, p. 832). In the MGT, the participants listen to recordings of an individual speaking the researched accents. On the other hand, the VGT utilizes recordings of speakers with accent and speaker. Garrett (2010) explains that after listening to each accent the respondents fill in an attitude rating scale of the accent. The results of this method aim to elicit the respondents’ attitudes to the different varieties of speech (McKenzie, 2010). The verbal-guise technique was chosen for this study instead of the matched-guise technique as the former allows the use of several speakers that represent each accent according to McKenzie (2010).
3.1 Data
The data of this study has been collected through an online questionnaire distributed at a Swedish compulsory school in Stockholm.
The questionnaire utilizes mostly close-ended questions as this technique provides a high standardization, which consequently strengthens the validity of the questionnaire results. There was one secondary question which asks if the participant has lived in an English-speaking country longer than 3 months. This question was included as the amount of exposure and familiarity for these participants are unlikely to be representative of the target group.
The study uses a bipolar semantic differential scale which has been used in previous language attitude research (Chiba et al., 1995; Bayard & Green, 2005). The scale also allows the degree of attitudes to be measured (McKenzie, 2010). This study uses a seven-point scale as according to Lemon (1973, as cited in McKenzie, 2010) it is believed to be the optimum number as a lower point scale displeases participants and larger ones create inadequate distributions; also, its unevenness allows for a neutral position on the scale. In practice, the participants listen to each accent and evaluate them on the bipolar semantic differential scale in regards to several personality traits (e.g.
honest/dishonest) (McKenzie, 2010). The result then reflects a representation of what the respondents’ attitudes are to the language or variety in question (McKenzie 2010).
Previous research has categorized attitudes into salient evolution categories to reduce the number of variables in the study (McKenzie, 2010). These have been narrowed into two salient evaluation categories which account for most attitude variance: competence and social attractiveness (McKenzie, 2010). This study used 7 of the original 17 adjectives used in the Evaluating English Accents WorldWide project (EEAW) a multinational collaborative project surveying the perceptions of English accents (Bayard
& Green, 2005). The chosen adjectives in the competence category were intelligent, reliable, and assertive. In social attractiveness, the adjectives friendly, humorous, and cheerful were chosen. The reduction of adjectives was made as some of them could
easily be interchangeable for the target group. In addition, Myrman (2004) also problematizes some of the adjectives as not all of them are clearly seen as personal traits.
The study uses a familiarity scale, as this allows the study to measure a degree of familiarity. It is also a seven-point scale as it is believed to be the optimum by researchers. While a familiarity scale has not been operationalized in attitude research previously, it has been used in studies on rating oral performance of L2 speakers of English (Winke et al., 2013; Saito et al., 2019). Previous attitude research has utilized familiarity but in binary format (Chiba et al., 1995), but as this study seeks to explore degrees of familiarity a scale is needed. Despite not being directly connected to attitude research, researchers argue that studying the degree of familiarity is useful in areas such as listening comprehension (Tauroza & Luk, 1997). To measure familiarity more accurately the participants were asked to identify the speaker's accents as the misidentification of the accents could, according to McKenzie (2008, p. 141), “render the data more difficult to interpret”. Furthermore, recognition questions are highly valuable in studies that involve non-native speakers’ evaluations of native speakers, as they have likely gotten lesser exposure to varieties of L2 speech than natives, and therefore could be less familiar and have difficulties in identifying different varieties (McKenzie, 2008).
The questionnaire itself contains an introductory description and five sections, four of which contain an audio clip with an English accent, and the final section contains a background question. Before answering the questionnaire, the participant was prompted to read the description, explaining its content. They were also informed that the questionnaire is anonymous and that they are free to opt-out anytime. The first four sections contain an audio clip of an English accent and request the participants’ attitude on the accent by rating seven adjectives on a scale of 1-7 as seen in figure 1 below:
Figure 1. Attitude scale from the questionnaire.
In addition, the participants were asked to answer to what extent they hear it in their free time and in their English class. There was also an open question where the participants were asked to guess where the accent originated from. And while open questions are not particularly suited to quantitative research as Dörnyei (2016) states, some have merit, like the specific open question which asks about a concrete piece of information. In the last section, the participants answered a few questions about their language background.
Lastly, at the end of the questionnaire, the participants were asked to give their consent to be a part of the study before submitting it. The audio clips that were used in the study were from the International Dialects of English Archive (2020). The first speaker representing GA (International Dialects of English Archive, 2021b) is a middle-aged man, but an exact age was not found. The second speaker representing RP (International Dialects of English Archive, 2021d) is 73 years of age currently however, the date of the recording is unknown. The third speaker representing AusE (International Dialects of English Archive, 2021a) was 25 years of age during the recording. The fourth speaker representing IrE (International Dialects of English Archive, 2021c) was 18 years of age during the recording. All the speakers in the audio clips read the same text by Honorof et al. (2020). As there was not a standardized version for the AusE and IrE accents, the speakers that represent those accents might possess some regional variation of their respective variants. This difference is not as important for this type of study as the Swedish curriculum does not demand students to be hyper-aware of slight regional variation of these accents. The criteria for the clips were that they should generally reflect an individual from that area.
The target population of the study consisted of Swedish compulsory school students from Stockholm. The gender make-up in Stockholm is male 50,57% and 49,43% female according to Länsstyrelsen (2018). The gender composition of the target group is likely similar to Stockholm as a whole as students who live in relative proximity of the school are prioritized over other applicants (Stockholms stad, 2019). The participants were at least 15 years of age to ensure that they were able to consent to be a part of the study.
This specific target group was chosen as the current knowledge of Swedish compulsory school students’ attitudes was less researched compared to the higher levels of education.
The questionnaire was produced in English but also used instructions in Swedish when deemed necessary as English proficiency varies among students. Before the data collection, the questionnaire was tested by a pilot group consisting of five former upper secondary school students who reported no concerns. Dörnyei (2016) supports piloting the research instruments prior to the investigation as it ensures a better outcome in the particular context.
The data was collected in the following procedure. The students at the compulsory school in Stockholm were asked to participate in the study, as long as they were at least 15 years of age and agreed to it. They were gathered in the school’s assembly room and got the survey link on their school email. As this school is formatted for online education, all students possess their own personal school computer. The students were first introduced to the project and if agreeing to participate, the participant would do so in the questionnaire’s consent form. After that, all participants listened to the first audio clip collectively before filling out the questionnaire. This process was repeated with all
the accents until the end of the questionnaire where the students answered the secondary questions and lastly submitted it.
The study had 98 participants which exceeded the minimum requirement of 30 participants that Dörnyei (2016) stated is necessary for quantitative studies to attain a normal distribution. Generally, it is argued that the greater the sample size, the better (Dörnyei, 2016). There were two participants who could not participate in the study as they were not 15 years of age. Similar to other research (Chiba et al. 1995; Eriksson, 2019; McKenzie, 2008; Rindal (2010) this study did not request the participants’
gender. In a Finish study gender among other background factors were found not to be significant in determining the attitudes of the informants (Hartikainen (2000, as cited in McKenzie, 2010). Matsuura and Chiba (2014) study suggest that gender is less of a crucial factor in terms of students’ evaluations of accents in contrast to the dialectal background.
3.2 Ethical considerations
The study was designed in accordance with the four ethical requirements for research:
information, consent, confidentiality, and utilization (Patel & Davidsson, 2011). The study fulfilled the information requirement as the participants were briefed on the study’s purposes before participation. Secondly, the questionnaire was entirely optional meaning that the participants decided to take part in the study voluntarily. That fulfilled the consent requirement which states the right of deciding participation. Thirdly, as the questionnaire is anonymous in nature it also fulfilled the confidentiality requirement which states that all the surveyed information should be confidential and safely stored.
Lastly, the collected data was only used for its intended research purposes and therefore fulfilled the utilization requirement.
3.3 Possible limitations
The choice of only male speakers in the audio clips was intentional as having both male and female speakers would create more variables to analyse. Another limitation is the differences between the speakers' age and voices which could also contribute to how they are perceived. The speakers are also in a formal context when they read the text and not speaking like in everyday conversation, which also narrows the scope of the study.
There is also the possibility that a majority of the participants will not be familiar with all of the accents which would mean that they have not developed attitudes towards them. The results could still hold implications for teaching as the curriculum requests exposure to a slight variation of English.
3.4 Analysis
The primary data were the participants’ attitude to each accent and the familiarity rating of each accent were both calculated into means. The mean score reflects what the participants’ attitudes are to the accents and how familiar they are with them. The two
data sets are then compared with each other to see if either higher or lower familiarity of accents reflects a difference in attitude. These data sets were put into T-tests to determine if there were significant differences in attitudes towards the accents and also if the accents were significantly different in familiarity. As there were over one hundred T-tests, most of the results will be summarized and only a selected few T-tests will be mentioned.
The question “How often do you hear accent 1 in your free time / English class?” will be used in conjunction with the other background questions to interpret the result. The recognition question will be using response validation criteria meaning that answers for RP would be considered correct with answers like “British”, “Great Britain”, “UK”,
“England”. However, while “UK” would be seen as correct, answers like “Scotland” or
“Wales” would not. In terms of GA answers such as “USA”. “American”, “United States” were seen as correct. In IrE answers with “Irish” or “Ireland” were only seen as correct and in AusE only “Australian” was seen as correct.
4. Results
The result section will be divided into two major sections. These sections are familiarity and exposure, and attitudes. The first section will discuss familiarity, exposure, and recognition of the accents. The latter section will divide the accents into groups with similar ratings of familiarity and compare them in terms of their attitude rating.
4.1 Familiarity, Exposure and Recognition
As seen in Table 1 RP has the highest mean familiarity. GA has only a slightly lower mean and a t-test showed no significant difference in familiarity between the two accents. RP and GA are both significantly more familiar than AusE and IrE. As seen in Table 1 AusE has higher mean familiarity than IrE and the t-test result showed a significant difference between the accents.
Table 1. Displays the Mean Familiarity and Standard Deviation of Each Accent.
Accent Familiarity:
Mean
Familiarity:
Standard Deviation
GA 5.23 1.35
RP 5.28 1.39
AusE 4.01 1.42
IrE 3.52 1.51
Note. The highest rating is marked in bold and the lowest in italic.
Table 2 presents the participants’ mean ratings in how often they were exposed to each accent in the English classroom and in their free time. The result is similar to familiarity as there is no significant difference between exposure of RP and GA in free time or school contexts. The exposure of RP and GA is significantly higher than AusE and IrE both inside and outside the classroom. However, in contrast to familiarity, there is not a significant difference in exposure between AusE and IrE inside or outside the classroom.
Table 2. Shows the Mean exposure and Standard Deviation of Each accent. The highest rating is marked in bold and the lowest in italic.
Accent School:
Mean
School:
Standard Deviation
Free time:
Mean
Free time:
Standard Deviation
GA 3.90 1.70 4.30 1.86
RP 4.09 1.83 4.21 1.62
AusE 2.16 1.12 3.36 1.57
IrE 2.26 1.46 3.17 1.67
Note. The highest rating is marked in bold and the lowest in italic.
As seen in Figure 2, RP was recognized by 82% of the participants, GA by 78%, AusE by 46%, and IrE by 27%. This is similar to how the average familiar rating is as RP and GA are recognized the most and are close to each other. AusE is less recognized than RP and GA, but even so, AusE is still 19 percentage points higher than IrE. While it is hard to compare the metrics of familiarity and recognition, the result falls in a similar trajectory of the familiarity scale where AusE is also more familiar than IrE. The recognition test appears to confirm the result of the familiarity scale.
Figure 2. Displays what percentage points of the participants recognized the individual accents
4.2 RP VS GA
As seen in Table 3, RP has a higher mean than GA in all attitudes. The result of the t- test showed RP being significantly higher than GA in all attitudes. In terms of familiarity, RP also has a higher average mean than GA (Table 1). However, this difference in familiarity is not significant with a p-value of 0.84 (p<0.05). In the recognition test, RP was recognized by 82% of the students and GA by 78%. What is seen between RP and GA is that despite there being no significant difference in their familiarity, there is a significant difference in attitude between them. This result could indicate that a high degree of familiarity does not exclusively dictate how this target group rates these particular accents in terms of attitudes.
Table 3. Shows the Mean Attitudes of Each Accent.
Accent Intelligent Reliable Friendly Humorous Assertiv e
Competen t
Cheerfu l
GA 4.59 4.42 4.45 2.10 4.01 4.61 3.96
RP 5.42 5.04 4.94 3.27 5.22 5.07 4.03
AusE 3.91 3.74 4.59 3.51 3.72 3.73 3.51
IrE 3.82 3.81 4.35 3.31 3.74 3.71 3.66
Note. The highest rating is marked in bold and the lowest in italic.
4.3 AusE vs IrE
AusE and IrE are similarly rated in attitudes as seen in Table 3 and there is no significant difference found between any of the attitudes. As seen in Table 1, AusE has a higher average familiarity than IrE and there was also a significant difference found in a t-test with p.0.02 (p<0.05). While familiarity is significantly higher in AusE than IrE the accents were not significantly differently rated in attitudes, which indicates that familiarity of this degree does not solely steer how the accents are rated.
4.4 RP and GA vs AusE and IrE
RP and GA are both significantly more familiar to the students than AusE and IrE. In terms of attitudes, RP is significantly higher rated in all attitudes than AusE and IrE, except in “Humorous” where there is not a significant difference between RP and AusE or RP and IrE. GA is significantly higher than AusE and IrE in the attitudes
“intelligent”, “reliable”, “competent”, and “cheerful”. In “Assertive” and “Friendly”
there was not a significant difference found between GA, AusE, and IrE. In
“Humorous” both AusE and IrE are rated significantly higher than GA.
This result seems to display that a high degree of familiarity does not have to equal a significantly higher attitude rating. However, what can be observed is that accents with higher degrees of familiarity like RP and GA differ significantly in chosen attitudes.
Furthermore, accents with lower degrees of familiarity, AusE and IrE, lack a significant difference in attitudes. Even so, AusE was significantly more familiar than IrE which might indicate that the degree of familiarity IrE was rated at, leads to similar attitudes, but at a higher degree of familiarity i.e., RP and GA, the attitudes clearly diverge.
5. Discussion
The results of this study reflect Carrie’s (2017) statement on L2 speakers having a better ability to recognize and sporting a greater familiarity with RP and GA compared to other varieties of English speech. Similar results are also seen in Denmark where RP and GA are recognized easier than Irish and Scottish accents (Jarvella et al., 2001).
Another study in Denmark reports GA and RP being recognized higher than AusE and Scottish Standard English (Ladegaard 1998). There were however differences in the ability to recognize different accents between students across the world.
In contrast to previous research, this study shows 46% of the compulsory school students recognizing AusE compared to Ladegaard’s (1998) study where about 8% of the upper secondary school students recognized the Australian speaker. This study’s result showed that these compulsory school students could distinguish between RP and GA which differs from Yaman’s (2015) study where university students did not fully distinguish between American and British varieties of English. This study finds that Swedish compulsory students at least in the examined target group are able to discern between a wider gamut of accents. Admittedly further research on compulsory school students’ ability to recognize English accents is required as the target group consists of students from one school with a high final grade average. Still, the findings could be echoed outside of Stockholm as Winke et al. (2013) states that different degrees of exposure can lead to varied types and levels of familiarity. Nevertheless, further examination in other areas in Sweden, preferably ones with a lower grade average could give insights into whether these findings recur despite the difference in grade average.
In terms of attitude rating, the findings of this study were similar to previous research (Carrie, 2017; Eriksson, 2019; Ladegaard & Sachdev, 2006; Rindal, 2010) as RP was highest in competence ratings. These attitudes may derive from formal/informal functions as discussed in Rindal (2010), with RP being predominantly taught in school which is a formal setting.
Another interesting result is how GA was lower rated in social attractiveness than RP which is in line with the findings of Jarvella et al. (2001) but in Rindal (2010) the opposite was found as students rated GA higher in social attractiveness. Carrie (2018) argues that despite these contradictory attitudes on RP and GA found in these studies, there is a pattern found in L1 and L2 speakers indicating the improbability of a speech
variety scoring highly in both competence and social attractiveness. This is not seen in this study’s result which on the contrary shows RP as being significantly higher rated in both competence and social attractiveness.
Considering both familiarity and attitudes, the results of this study indicate that at the higher degrees of familiarity tested in RP and GA, attitudes can differ significantly. This is not seen in the lower degrees of familiarity where AusE and IrE do not differ significantly in attitudes even though AusE is significantly more familiar than IrE. A partial explanation for this result could be that the degree of familiarity with accents is not enough to create a social attitude. This could be partially explained by Holmes’
(1992) statement that attitudes of unheard accents are random. A higher degree of familiarity would be required with AusE and IrE for them to possess unique attitudes.
This would also explain why RP and GA are rated significantly different in attitudes despite their shared high level of familiarity. There seems to be a point of threshold needed for an individual to develop an attitude to an accent. This threshold of familiarity might be higher than the degree of familiarity seen in either AusE or IrE which explains why the participants’ attitudes are not significantly different.
In terms of limitation, this study could have benefitted from using speakers that were more similar in terms of voice quality and age which could have yielded a different result. But due to the study’s limited scope, these possibilities were narrowed down.
Despite this, the study had similar findings to previous research such as RP being rated highest in competence.
5.1 Implications for teaching
This study’s findings echo previous research on RP being the highest in competence ratings (Eriksson, 2019; Ladegaard & Sachdev, 2006; Rindal, 2010). In the previous research it is argued that this can originate from RP being the primary variant taught in school. However, according to the students of this study, this did not seem to be the case as they have reported that GA and RP had equal amounts of exposure in the classroom.
Eriksson (2019) similarly states that the British variant of English no longer dominates education in Sweden. However, RP is still rated higher in intelligence and competence and rated higher in all attitudes in this study. The attitudes towards RP are not necessarily harmful but teachers should be aware of this and how their own teaching may influence how different accents are perceived. It is argued by Bradac and Giles (1991) that teachers need to intervene in the process of stereotyping which will help develop an awareness of their language attitudes as linguists state that there is no variety of a language better than any other (Wardhaugh, 2010).
Teachers need to be more aware of their own language attitudes, and as previous research argues, if educational programmes place too strong of a focus on a particular variety of English, it may implicitly educate the students that this is the correct variety of English (Eriksson, 2019).
6. Conclusion
The main aim of this study was to examine if the degree of familiarity affects compulsory school students’ attitudes towards English accents. The study’s results indicate that the degree of familiarity seems to affect attitude in English accents. More specifically, the results indicate that a higher degree of familiarity influences the ability to express an attitude, while a lower degree leads to similar attitudes across accents.
This interpretation can be supported by Holmes (1992), as attitudes are developed towards languages, meaning that exposure is required for the individual to familiarize themselves with it and create social attitudes towards them. Furthermore, if an accent is unfamiliar its assessment is totally random (Holmes, 1992), which would also coincide with a lower degree of familiarity meaning an attitude has not been developed.
However, there is the possibility that individuals may associate aspects of unfamiliar accents to those they are familiar with. And as seen in Matsuura and Chiba (2014) individuals without regional dialect gave more positive ratings to unfamiliar English accents than those participants with regional dialects. Another interesting finding was the extent of recognition of AusE English by the students, as previous research in the neighbouring country Denmark showed a lower rate of recognition in upper secondary school students.
In terms of teaching implications, it seems that RP is rated higher in terms of competence but also overall, even though GA gets a similar amount of exposure in a classroom context, and in the students’ free time. This indicates social connotations towards RP which might not be directly harmful but well worth addressing. Admittedly, it is hard to generalize from this sample size alone but as other researchers have generally found RP being higher rated in competence, it validates these findings. Based on this study’s result in conjunction with previous research, it would seem that creating awareness of these language attitudes are essential for teachers and students alike, as a means to raise awareness of social connotations as well as prejudice linked to language.
In conclusion, it is difficult to draw any wide conclusions from this result due to the study’s scope and the target group not being fully representative of all compulsory school students as the school students have an above-average final grade. However, a high final grade average does not equal a high level of English proficiency and the proficiency does not exclusively determine an individual's ability to recognize English accents. Further research is necessary to determine more specifically how familiarity affects attitudes of English accents and if these findings recur in other areas of Sweden where the final grade average is lower.
References
Bayard, D., & Green, J. A. (2005). Evaluating English accents worldwide. Te Reo, 48, 21–28.
Bradac, J. J., & Giles, H. (1991). Social and educational consequences of language attitudes.
Moderna Språk, 85(1), 1–11.
Carrie, E. (2017). ‘British is professional, American is urban’: Attitudes towards English reference accents in Spain. International Journal of Applied Linguistics, 27(2), 427–447.
doi:10.1111/ijal.12139
Carrie, E., & McKenzie, R. M. (2018). Implicit–explicit attitudinal discrepancy and the investigation of language attitude change in progress. Journal of Multilingual and Multicultural Development, 39(9), 830–844, doi:10.1080/01434632.2018.1445744
Chiba, R., Matsuura, H., & Yamamoto, A. (1995) Japanese attitudes toward English accents.
World Englishes 14, 77–86. doi:10.1111/j.1467-971X.1995.tb00341.x
Curriculum for the compulsory school, preschool class and school–age educare (revised 2018).
(2018). Skolverket.
Dörnyei, Z. (2016). Research Methods in Applied Linguistics: Quantitative, Qualitative, and Mixed Methodologies. Oxford: Oxford University Press.
EF (2020). EF English proficiency index. EF Education First Ltd. Retrieved January 24, 2021, from: https://www.ef.com/wwen/epi/
Eriksson, L. E. (2019). Teachers’ and students’ attitudes and perceptions toward varieties of English in Swedish upper secondary school. In B. L. Egeland, T. Roberts, E. Sandlund, & P.
Sundqvist (Eds.), Klassrumsforskning och Språk(ande): Rapport från ASLA–symposiet i Karlstad, 12–13 April, 2018. (pp. 207–233) ASLA:s skriftserie.
Garett, P. (2010). Attitudes to Language. Cambridge: Cambridge University Press.
Göteborgs universitet. (2020, October 29). Projektet nationella prov i främmande språk – Nafs.
Exempel på uppgiftstyper för årskurs 9. https://www.gu.se/nationella-prov-frammande- sprak/prov-och-bedomningsstod-i-engelska/engelska-arskurs-7-9/exempel-pa-uppgiftstyper- for-arskurs-9
Honorof, D. N., McCullough, J., & Somerville, B. (2021, January 24). Comma gets a cure.
International Dialects of English Archive. https://www.dialectsarchive.com/comma-gets-a- cure
Holmes, Janet. (1992). An Introduction to Sociolinguistics. New York: Longman Group UK Limited
International Dialects of English Archive. (n.d.–a). Retrieved October 4, 2020, from https://www.dialectsarchive.com/
International Dialects of English Archive. (n.d.–a). Australia 27. Dialects Archive. Retrieved January 15, 2021a, from https://www.dialectsarchive.com/australia-27
International Dialects of English Archive. (n.d.–a). General American 1. Dialects Archive.
Retrieved January 15, 2021b, from https://www.dialectsarchive.com/general-american-1 International Dialects of English Archive. (n.d.–a). Ireland 13. Dialects Archive. Retrieved
January 15, 2021c, from https://www.dialectsarchive.com/ireland-13
International Dialects of English Archive. (n.d.–a). Received Pronunciation 3. Dialects Archive.
Retrieved January 15, 2021d, from https://www.dialectsarchive.com/received- pronunciation-3
Jarvella, R. J., E. Bang, A. L. Jakobsen, & I. M. Mees (2001). Of mouths and men: non-native listeners identification and evaluation of varieties of English. International Journal of Applied Linguistics 11.1: 37–56.
Kircher, R., & Fox, S. (2019) Attitudes towards Multicultural London English: implications for attitude theory and language planning, Journal of Multilingual and Multicultural Development, 40(10), 847–864, doi: 10.1080/01434632.2019.1577869
Ladegaard, H. J. (1998). National stereotypes and language attitudes: the perception of British, American and Australian language and culture in Denmark. Language and Communication 18.4: 251–274.
Ladegaard, H. J., & Sachdev, I. (2006). ‘I like the Americans...but I certainly don’t aim for an American accent’: Language attitudes, vitality and foreign language learning in Denmark.
Journal of Multilingual and Multicultural Development, 27, 91–108.
Länsstyrelsen. (2018). På tal om kvinnor och män - Stockholms län 2018. SCB.
https://www.lansstyrelsen.se/stockholm/tjanster/publikationer/2018/pa-tal-om-kvinnor-och- man---stockholms-lan-2018.html
Matsura, H., & Chiba, R. (2014). Students’ attitudes toward unfamiliar English accents and their L1-speaking backgrounds. The Commercial Review, 82(4), 3–13.
McKenzie, R. M. (2010). The Social Psychology of English as a Global Language: Attitudes, Awareness and Identity in the Japanese Context. London: Springer
Melchers, G., Shaw, P., & Sundkvist, P. (2019). World Englishes: Vol. Third edition. London:
Routledge.
Myrman, F. (2004). Swedish People’s Attitudes to Different Accents of English.
Patel, R., & Davidson, B. (2019). Forskningsmetodikens grunder: att planera, genomföra och rapportera en undersökning (5th ed.). Lund: Studentlitteratur.
Rindal, U. (2010). Constructing identity with L2: pronunciation and attitudes among Norwegian learners of English. Journal of Sociolinguistics, 14(2), 240–261.
Rindal, U. (2014). Questioning English standards: learner attitudes and L2 choices in Norway.
Multilingua, 33(3–4), 313-334. doi:10.1515/multi-2014-0015
Saito, K., Tran, M., Suzukida, Y., Sun, H., Magne, V., & Ilkan, M. (2019). How do second language listeners perceive the comprehensibility of foreign-accented speech?: Roles of first language profiles, second language proficiency, age, experience, familiarity, and metacognition. Studies in Second Language Acquisition, 41(5), 1133–1149.
doi:10.1017/S0272263119000226
Skolverket. (2011). English. Retrieved December 29, 2020, from https://www.skolverket.se/undervisning/gymnasieskolan/laroplan-program-och-amnen-i- gymnasieskolan/amnesplaner-i-gymnasieskolan-pa-engelska
Stockholms stad. (2019, December 20). Regler för antagning. Grundskola Stockholm.
https://grundskola.stockholm/sok-skola/regler-for-antagning/
Tauroza, S., & Luk, J. (1997). Accent and second language listening comprehension. RELC Journal, 28(1), 54–71.
Winke, P., Gass, S., & Myford, C. (2013). Raters’ L2 background as a potential source of bias in rating oral performance. Language Testing, 30(2), 231–252.
doi:10.1177/0265532212456968
Yaman, İ. 2015. “Exploring ELT students’ Awareness of the differences between the British and American varieties of English.” Ondokuz Mayıs Üniversitesi Eğitim Fakültesi Dergisi, 34(1), 153–164. doi:10.7822/omuefd.34.1.9.
Appendix A
Stockholms universitet 106 91 Stockholm Telefon: 08–16 20 00 www.su.se
