Accent Anxiety: An Exploration of Non-Native Accent as a Source of Speaking Anxiety among English as a Foreign Language (EFL) Students

Speaking anxiety is a form of foreign language anxiety which may reduce students’ willingness to communicate orally. Despite accent being one of the most salient aspects of speech, there has been little research to date on the relationship between non-native accent and speaking anxiety. The purpose of this study is therefore to examine English learners’ perceptions and beliefs about accent, and also to explore the concept of accent anxiety, that is, speaking anxiety arising due to concern about one’s non-native accent. An anonymous online questionnaire was distributed to English students in a French university. The questionnaire sought to gauge the students’ attitudes both towards speaking and accent and gathered qualitative responses about the students’ experiences of accent anxiety and their coping strategies. A thematic analysis was then carried out on the 54 responses. It was found that the majority of the students did not believe attaining a native-sounding accent was essential to language learning and felt that comprehensibility should be the primary objective. However, many of these students nonetheless considered it a personal goal to sound more native-like. Furthermore, most of the students had at some point in their learning felt embarrassed or worried about their accents, with the two primary causes being fear of negative evaluation from their peers and fear of future communication issues. It was concluded that concern over how non-native accents sound is a potential source of speaking anxiety for learners of English. As these students highlighted the classroom as being the main location where this anxiety arises, the study concludes with some suggestions for educators as well as ideas for future research directions.


INTRODUCTION
The human factor of language learning has attracted much attention in current applied linguistics research, with the psychology behind language learning emerging as a key discipline in recent years. Research has focused on exploring the impact of individual differences on learning, such as motivation, personality, and emotion (Dörnyei & Ryan, 2015). The concept that the learner's emotional state, whether positive or negative, may influence their learning has seen particular interest in the last two decades (Dewaele et al., 2019), and is particularly relevant in the current context of the Covid-19 pandemic. Of all the emotions that learners may experience, foreign language anxiety is one of the most established in the current literature, having come to the forefront towards the end of the 1980s by works from, among others, Horwitz, Horwitz and Cope (1986) and MacIntyre and Gardener (1989). Their research demonstrated that students suffering from language anxiety are likely to exhibit similar symptoms to those who suffer from general anxiety, which include feelings of worry and dread, and difficulty concentrating. The current study aims to explore the relationship between this foreign language anxiety (FLA) and non-native accent.
Research on the role of accent in second language learning has faced the challenge of defining exactly what a native accent is, since all speech is accented to some extent. Derwing and Munro (2009) opt to define accentedness as the extent to which one's speech deviates from the local standard. Although first language accent is a relatively wellstudied area within the field of sociolinguistics, there is a distinct lack of research pertaining to accent in the foreign language, particularly in relation to the topic of speaking anxiety. While anxiety can affect many aspects of language performance, research has shown that speaking is the skill most affected (King & Smith, 2017). Horwitz et al. (1986) state that this may be rooted in the student's belief that if they cannot speak the language perfectly, they should not speak it at all. They also note that the perceived threat of speaking may be heightened due to the focus on communicative competences, which are even more emphasized in the current philosophy of communicative language teaching.
Speaking anxiety has been identified as the most common form of FLA as it allows for the learner to be publicly evaluated in comparison to traditional solitary tasks such as reading and writing (King & Smith, 2017). It has also been known to impact learner performance. In Kormos and Préfontaine's (2017) analysis of L2 speech tasks, participants reported feeling highly anxious when they had insufficient linguistic resources to allow them articulate task content. Zuniga and Simard (2022) further claim that learners have difficulty in finding their words preverbally due to the presence of negative emotions which restricts attentional resources.
Pronunciation has thus far been the focus within speaking anxiety research and accent has been considered a subfield of it. Thus, "accent anxiety" has not been hypothesized as an independent construct. While the concept of accent is closely related to pronunciation, it warrants further attention due to distinctions pointed out in this study.
"Pronunciation anxiety" was established as a specific construct by Baran-Łucarz (2014) who aligns findings with Dörnyei's (2005) L2 Motivational Self System. She says that learners may feel pronunciation anxiety because they fail to achieve their pronunciation learning goals, which in turn threatens the attainability of their ideal L2 self. Similarly, sources of pronunciation anxiety may come from the belief that sounding like a native is essential in language learning, which is connected to the ought-to L2 self. Based on this, Baran-Łucarz (2014, pp. 453-454) devised a fourcomponent construct of pronunciation anxiety comprised of: 1) pronunciation self-image 2) pronunciation self-efficacy and self-assessment 3) fear of negative evaluation 4) belief about the importance of pronunciation for successful communication.
Students' desire to sound native (1), combined with their low pronunciation self-assessment (2), their fears of judgement for their poor pronunciation, (3) and their belief that native-like pronunciation is necessary (4) result in feelings of anxiety about their pronunciation.
According to Moyer (2013), while often used interchangeably, there is a marked difference between accent and pronunciation. She states that while pronunciation relates to specific articulation, accent is concerned with suprasegmental features such as "intonation, rhythm, pitch, segmental length, tempo, and loudness" (p. 10). While the two terms are no doubt highly related to each other, this study will focus on the broader concept of accent referring to general speech fluency rather than individual sounds. With this in mind, one could hypothesize that "accent anxiety" may be considered an overarching construct encompassing the subset "pronunciation anxiety". However, further research is required on how they interrelate or if one is the subset of the other.

Defining Foreign Language Anxiety
Foreign language anxiety (FLA) has evolved as a research domain over the past sixty years. It is generally accepted in the current literature that anxiety has mostly debilitating effects (Horwitz, 2017). A distinction was also made between trait anxiety and state anxiety, the former being an individual's tendency to be anxious and the latter relating to the experience of feeling anxious at a specific moment in time (Scovel, 1978). Horwitz et al. (1986) identified three principle dimensions of FLA: communication apprehension, test anxiety, and fear of negative evaluation. Communication apprehension relates to feelings of worry over general social exchanges, while test anxiety is rooted in fear of failure. Those who suffer from fear of negative evaluation worry about how they are perceived by others, and they may opt to avoid social situations rather than risk being embarrassed (Aida, 1994). Horwitz (2017) later clarified that these three dimensions are not the sole components of FLA, but rather are examples of specific anxiety-invoking situations. This type of anxiety was different from the trait and state distinctions outlined above, and was referred to as situationspecific anxiety (Horwitz et al., 1986). MacIntyre and Gardner (1989) suggested that this type of anxiety stems from repeated negative experiences and feelings of state anxiety in the language classroom which leads learners to associate that environment with nerves and apprehension.
More recent research has suggested that FLA, similar to motivation and attitude, is a dynamic system which can fluctuate significantly within a single learner (Gregersen, 2020). FLA is particularly associated with learners' concept of self. Horwitz (2017) states that students who seem unmotivated may in fact be experiencing anxiety about their learning, arising from the "inability to present oneself authentically and the resulting threat to self-concept" (p. 42). Self-concept is a common theme in both FLA and motivational research. Simsek and Dörnyei (2017) coined the term "The Anxious Self" after studying Turkish learners of English and finding that many of them suffered from anxiety in the classroom, but considered the anxious part of them to be distinct from their general personality. The idea that both FLA and motivation are tied to the individual's concept of self is particularly relevant to the current study, as accent has long been associated with identity, as further discussed in the next section.

Accent in EFL Learning
The perception of accent has changed in EFL research. In the past, the presence of an L2 accent was seen as something "undesirable" that had to be eliminated (Zárate-Sández, 2017, p. 237). The general consensus today is that accented speech is only an issue if it actively impedes communication or leads to unintelligibility (Moyer, 2013). Many researchers have also pointed out that the attainment of a native-like accent or pronunciation is not only extremely difficult, but may even be impossible (Derwing & Munro 2009;Long, 1990;Moyer 2013). Nevertheless, past studies have revealed that students learning a second language generally aim to speak with a native-sounding accent. In a study of EFL students in Canada, Derwing (2003) observed that 95% wanted to sound like a native speaker. These same students reported feelings of discrimination due to their foreign accents, and over half of them believed that they would be more respected by Canadians if their accent was more native-like. McCrocklin and Link (2016) found that EFL students wanted to speak with a native-like accent because they would feel more pride and acceptance in the target community. Similarly, a study of Russian learners of English living in America by Dolgova Jacobsen (2008) found that the individuals who wanted to integrate more into the local American community actively worked to minimize the foreign-ness of their accents and to adopt a more American-sounding accent. It is evident here that a nativesounding accent is likely to be considered desirable by foreign language learners. Further to this, there appears to be a strong link between both L1 and L2 accent and identity. Interestingly, Cutler (2014) suggests that L2 speakers are often aware that the sounds produced by them may be social marker indicators of native and non-native speech, and that identifying as a member of a racial and/or an ethnic minority community, can influence them when adopting the speech patterns of "ethnic others". In contrast, Moyer (2013) states that adults tend to consider their L1 accent as an integral part of their idea of self, and therefore may resist acquiring a new accent in the L2 as it ultimately means losing a part of their established concept of self. However, she claims that children are more likely to be less inhibited and speak with a more-native sounding accent. However, other studies have questioned the extent to which learners feel they must retain their L1 accent to preserve their identity. Derwing (2003) looked to examine this connection and explore whether students felt that they would lose aspects of their own identities if they aimed to speak with a native-like accent. The majority of the participants reported that they did not feel like they were losing their identity when speaking their L2, because they associated identity more strongly with their L1. Further research (see McCrocklin & Link, 2016;Piller, 2002) has also revealed that students may feel happiness and pride at being mistaken for a native speaker, demonstrating that they do not consider this a threat to their identity but rather a compliment on their language skills. Given these positive attitudes to native-sounding accents, it is reasonable to question whether the unattainability of such an accent may give rise to FLA, and more specifically, speaking anxiety.

Speaking Anxiety and Pronunciation
Speaking anxiety has been linked to the learner's willingness to communicate (WTC), which has been defined as the intention to speak or not, if the choice is yours (MacIntyre, 2007). WTC is a concept which, similar to FLA itself, is now considered a dynamic system (MacIntyre & Wang, 2021). Naturally, the less willing the learner is to speak, the more likely they are to remain silent. In a study of silence in the Japanese classroom, King (2013, p. 337) claims that learners may be "unwilling to engage in the potentially embarrassing behavior of active oral participation for fear of being negatively judged by their peers." He says that in this case, classroom silence is a "defensive strategy" (p. 337). In a later study, King and Smith (2017) found that social anxiety was the primary cause of Japanese students' silence in the classroom.
Research to date has shown conflicted findings about which types of speaking situations invoke anxiety the most. In a study of EAP students entering Australian universities, Woodrow (2006) reported that their levels of anxiety were highest when conversing with native speakers. Similarly, Çagatay (2015) found that Turkish learners of English felt more anxiety when they were in contact with native speakers compared to with their peers and fellow learners, suggesting that learners may feel this way because they perceive natives as being more critical of their language skills, and also may feel intimidated as the native speaker represents their ideal learning goals. In contrast, Suleimenova (2013) reported that high school students learning foreign languages in Kazakhstan, felt anxious to speak due to negative speaking experiences in the classroom. The students were afraid of looking foolish in front of their peers, and they felt anxious when the teacher called on them in random selection.
While research pertaining to FLA deriving from nonnative accent is scarce, studies have been carried out examining the link between speaking anxiety and pronunciation. In interviews with Texan learners of French, Price (1991) found that many high FLA students were particularly worried about their poor pronunciation. They cited fears about embarrassing themselves in front of the class and negative past experiences as the reasons behind this pronunciation anxiety. Many of them reported feeling "ashamed" (p. 105) about their Texan accents and their frustration at their inability to pronounce French in a nativesounding way. Pronunciation anxiety is strongly influenced by issues about self-image when speaking the L2. Baran-Łucarz (2011) profiled two students, one with low FLA and the other with high FLA. She found that the low FLA student had a very positive image of her L2 self. She liked the way she sounded speaking English and was therefore highly motivated to try to attain a native-like accent. On the other hand, the high FLA student showed negative perceptions of his L2 self. He claimed that he was aware his accent sounded like a Polish speaker, and also reported feeling "ridiculous" saying words in a more "English-like" way (p. 508).
However, past research has demonstrated that anxiety may be more strongly associated with students' perceived pronunciation, rather than their actual pronunciation. Baran-Łucarz (2011) found that there was no significant correlation between FLA and actual pronunciation ability, as graded by native speakers. However, there was a relationship between FLA and perceived pronunciation. She surmised that while poor pronunciation ability may lead to anxiety about communication, students' perception that their pronunciation is poor is connected to a fear of negative evaluation. She states that this is in line with current theories about FLA which state it is more a symptom of learners' beliefs and self-perceptions rather than their actual linguistic ability.
While the concepts of accent and pronunciation have been outlined as distinct yet interrelated, there is an absence of a theoretical definition of accent anxiety distinguishable from pronunciation anxiety. While this study is primarily situated in the established theoretical field of speaking and pronunciation anxiety, there are two primary objectives identified here in relation to accent: firstly, to gain an understanding of students' perceptions about their own accents, and their opinions about native and non-native accents in general; and secondly, to explore whether feelings of embarrassment about their non-native accents create a form of accent anxiety within the learners. The research questions for the study are: RQ1: What is the nature of L2 learners' beliefs about accentedness?
RQ2: How do these beliefs interact with existing speaking and pronunciation anxiety constructs?

METHOD Participants
The target population were L2 learners who were studying English as a major or minor subject. A case study approach was taken, where a particular university in the north of France was chosen. The actual respondents were 54 students (40 female, 13 male, 1 non-binary). They ranged in age from 18-34, although the majority fell into the 18-24 bracket. Most (n = 48) had French as a mother tongue. In terms of university majors, 28 participants were studying English, 18 were studying a different language, and 8 were divided between humanities, music or commerce. Their overall English proficiency ranged from A2 to C2 on the CEFR scales, with most being B2 or C1.

Design
The chosen methodology for this study was an anonymous questionnaire distributed and submitted online. The decision to use a questionnaire was based on both the subject matter of the study, and some practical considerations such as time restraints and geographical distance. Questionnaires have been the primary tool utilized in many studies regarding language anxiety since Horwitz, Horwitz and Cope (1986) developed the Foreign Language Classroom Anxiety Scale (FLCAS), consisting of 33 items on a 5-point Likert Scale. The scale was developed according to previous students' self-reported experiences and beliefs, and participants are requested to agree or disagree on the 5-point scale to statements such as: "I never feel quite sure of myself when speaking in my foreign language class." (p. 37). Many studies have adapted the scale for more specific purposes, such as for reading anxiety (Matsuda & Gobel, 2004) or for pronunciation anxiety (Baran-Łucarz, 2014; Kralova et al., 2016). In this study, Likert scale questions related to FLA were informed by the FLCAS.
The questionnaire opened with some demographic information and was then divided into five sections relating to students' general perceptions and beliefs about speaking and accents, feelings of accent anxiety, and strategies to overcome anxiety. In Sections B and C, several questions were asked about accent rather than solely concentrating on pronunciation. Most questions were designed as openended so that respondents have the space to fully explain their own opinions and personal experiences with the subject. The goal was to see if participants raised accent issues more broadly, rather than pronunciation alone. Section D of the questionnaire consisted of ten Likert scale statements, inspired in part from the FLCAS. The concept of "accent anxiety" as it relates to speaking and FLA was not measured using a pre-established theoretical underpinning, as one does not exist. The goal was to discuss findings in light of the previously proposed "pronunication anxiety" construct (Baran-Łucarz, 2014) and to identify preliminary accent issues, which can be followed up on with further research.
Google Forms was chosen as the platform for the questionnaire due to its accessibility. In order to reach as many students of all levels as possible, both an English and French version of the questionnaire were designed. The English one was created first and the questions were then translated to French in consultation with a native speaker.

Procedure and Analysis
Once the questionnaire was designed and translated, and the study received ethical approval, the next phase was the distribution. The questionnaire was first sent to three gatekeepers, all of whom are ex-colleagues of the lead researcher who are all employed in the English department of a university in France. These gatekeepers were asked to distribute the questionnaire to students across all disciplines of the university. Two weeks after the initial email, gatekeepers were requested to send a reminder to their students. The responses were recorded on Google Forms separately for the English and French questionnaires. Once responses stopped being accepted, the results from each version were downloaded.
Likert data allowed for trends to be identified and descriptive statistics to be generated. This data was also analyzed in conjunction with the demographic variables. A thematic analysis was then carried out on the qualitive responses. This allowed for in-depth analysis of emotions and thought-processes. Responses were analyzed by the first researcher and keywords were highlighted. Codes were then generated based on the identified keywords. The coded material was then grouped together under an organizing theme and finally each organizing theme was placed under a global theme. An interrater process was then underway, and the researchers discussed the preliminary codes and agreed on the identified themes. The analysis generated six organizing themes grouped under two global themes, as outlined in the Findings section.

FINDINGS
The majority of respondents (n = 49) chose to use the French version, meaning the quotations presented below are English translations.

Attitude Toward Speaking and Accent
In order to gain an understanding into the learners' general attitudes towards speaking, participants were asked to choose which language skill(s) they would most like to improve, and the situation(s) in which they felt at ease speaking English. Speaking was the skill most selected (52%), with speaking to natives being significantly more comfortable than speaking in class. The students were then asked whether they enjoy speaking in class, with 30 respondents saying they do. The classroom environment was a central factor for the 17 students who said they did not enjoy speaking in the classroom. Seven students said they enjoyed speaking in general, but in class it would depend on the number of students, and whether or not they felt comfortable with their classmates. Others felt that a negative environment impacted their speaking confidence: (1) "Speaking in front of everyone makes me anxious and I'm scared of making mistakes because of stress." (2) "It depends, because I'm often scared that people will mock me." (3) "No, out of fear of being judged by the other students or teachers." Of these 17 students who don't enjoy speaking, 6 respondents specifically highlighted their accent as a reason, with some statements being: Other reasons given included a lack of vocabulary, boring subjects, and general shyness.
Six of the Likert-scale statements related to attitudes and beliefs about accent, as demonstrated in Table 1. It was found that most of these students place more importance on grammar than pronunciation (statement 1), they consider accent to be part of their identity (statement 4), they care more about being understood than having a native-like accent (statement 5), but most of them want to sound like a native-speaker (statement 10). Opinions were more divided between the question of accent and proficiency (statement 2), and whether they consciously put on an accent (statement 8).
The participants were then asked two questions to gain an understanding of their beliefs about accent: • Is sounding like a native speaker one of your language learning goals?
• Do you think having a native-like accent is an essential part of learning a language?
Three primary themes emerged from these questions.

Accent is an added bonus
The vast majority (n = 48) felt that acquiring a native-like accent should not be considered an essential part of language learning, but many saw it as being beneficial. Two respondents mentioned that while accent is not essential, it does give you confidence which in turn makes you a better speaker. Similarly, another student felt that having a native-like accent can allow you to integrate into the language's culture more easily. Although the majority felt it was not essential, 30 respondents stated that sounding like a native is one of their learning goals, suggesting that these students do place a certain importance on acquiring such an accent. Some participants stated that while it is not their primary objective, they would like to acquire a native-like accent in the long-term. Other students referred to the prestige associated with having a native-like accent: (7) "Yes because people will take you more seriously if you sound like a native speaker." Others felt that accent is simply part of the language as anything else and so learners should make the effort to work on their accents. Four respondents said that while they would like to acquire this accent, they believed it to be impossible or they were unsure how to go about achieving this.

Accent is less important than other aspects
The implication of accent being a bonus is that other parts of language learning are deemed more important for these learners. When asked about their priorities in language learning, the majority of participants (n = 34) felt that vocabulary was the most important aspect, followed by grammar (n = 14), with pronunciation being considered the least important (n = 6).
A recurring response to several questions was that making oneself understood was far more important than sounding like a native speaker. Nineteen respondents stated that they believed fluency and being understood were more important to them than their accent, and that these aspects should take priority in language learning. Accent is only a concern for proficient speakers Another recurring viewpoint in responses the questions about whether a native-like accent is their learning goal and whether it is essential to learning was that only proficient learners of English should worry about their accents.
(11) "If one intends to teach the language in question, then it is essential to have an accent as close to the native speaker as possible." (12) "[accent is] not essential to learning, but a sign of good mastery [of the language]." (13) "I think if one is aiming for a C1 or better certification then yes it's important, but it's not critical to learning English." Overall, it appears that the respondents feel that accent is not a primary concern of language learning, but for their own personal goals they do aspire to sound close to a native speaker.

Accent Anxiety
In Section C, the participants were asked about their perceptions and feelings about their own accents. As can be seen in the chart below, the majority do think about how their accent sounds, be it often or sometimes.

Figure 1. When speaking English, how often do you think about how your accent sounds?
The students were also asked to rate their accent on a scale from 1 to 10, 1 being 'non-native-like' and 10 being 'native-like'. The majority placed their accent in the middle, between 5-7, with some rating it higher and others lower (Figure 2).
The participants were then asked to elaborate on their feelings about their accents in the following open-ended item: • How do you feel about your English accent?
Thematic analysis identified that 27 respondents reported positive feelings towards their accents, 22 respondents had more negative perceptions, and the remaining 5 respondents were neutral in their opinion.
The remaining four Likert scale statements further elaborated on the idea of accent and anxiety, as demonstrated in Table 2. This data shows that only 21 out of 54 participants felt less willingness to speak in class because of their accents. In comparison, 35 students reported feeling at times afraid to speak, but did not consider accent to be the cause. This suggests that while some students may suffer from accent anxiety, it is not the primary cause of FLA.

Figure 2. Rate your accent as sounding non-native-like (1) to native-like (10).
Having gained an understanding of the participants' general perceptions about their own accents, the following three open-ended questions aimed to uncover whether any of the students suffered anxiety about their accents: • Do you ever worry about what other people think of your accent when speaking English?
• Have you ever felt embarrassed about your accent when speaking English?
• Has your accent ever caused problems with communicating in English?
Thematic analysis of the responses identified that from the 54 participants, 40 of them worry about their accents either sometimes or often, suggesting that the majority of these students do suffer from accent anxiety to a certain extent. In order to uncover the sources of this, a thematic analysis of the responses to the questions above was carried out, with the resulting themes outlined below, along with their corresponding number of responses.

Comparison to others
In relation to whether participants compare their accents to others, the primary trend that emerged throughout the questionnaire (n = 21) was the participants' tendency to compare themselves to fellow learners. Interestingly, one participant felt this increased their confidence: (14) "I feel less shame than others with a stronger accent." However, the majority of comparisons resulted in negative effects. Many students reported that they worry about their own accent because their classmates' accents are better. They also stated that they feel judged in class and were afraid of being mocked. One respondent said that they purposely try to improve their accent for fear of being mocked and another felt less respected than others because of their accent. Some participants also reported fears about how their classmates view them: (15) "Yes [I worry] because other students have good accents." (16) "In a class with students who have really good accents, I feel embarrassed."

(17) "Yes [I worry] because certain students have better accents and surely they must judge us."
When reflecting on past experiences of feeling embarrassed about their accents, many wrote about times when they felt judged or ridiculed in class. One respondent mentioned that they felt embarrassed speaking in class after someone with a good accent and that this impacted their linguistic self-esteem. Of the 40 respondents who said they do at times worry about how their accent sounds, 21 of them cited reasons related to comparisons with peers. Interestingly, one student explained that they had at varying times been mocked for having a bad accent, but also for having a good accent (accusations of being pretentious or fake), and so they ultimately stopped worrying about how they sounded because people will judge regardless.

Fear of communication breakdown
The second most prominent source of accent anxiety (n = 12) was the fear of communication breakdown. Some respondents worried that their accents would lead to a lack of understanding on their interlocutor's part. 12 students said they had at some point felt embarrassed about their accents because of communication issues, be it with their teacher, native speakers, or fellow students. One person recalled having to repeat themselves multiple times because of their poor pronunciation, and another felt that their flawed intonation impeded the nuances of their intended message. Another stated that they feel embarrassed about their accent when a native speaker can immediately identify their nationality. Others also worried that people would equate their nonnative accent to poor English skills. Interestingly, although fear of communication breakdown was listed as a primary reasons for accent anxiety, when asked Has your accent ever caused problems with communicating in English?, the majority of respondents (n = 42) answered no, suggesting that accent causing communication breakdowns is more of a hypothetical fear than an actual language issue.

Negative thoughts
The final common sources of accent anxiety have been grouped under the term negative thoughts, relating to cases where participants blamed their general feelings of worry, lack of self-confidence or negative emotional state on the anxiety they were experiencing when speaking in the classroom (n = 7). One felt that the pressure of their English course aroused feelings of anxiety about how they sounded, and two others said that heightened stress cause them to speak with what they considered a poor accent, which in turn increased their levels of anxiety. One student reported feeling too shy to attempt a nativesounding accent, and another reported feeling constant uncertainty about their pronunciation. Another participant said that they felt embarrassed about the number of mistakes in stress and intonation they make when speaking, which resulted in them not daring to speak anymore. One student also said that they felt most anxious about their accent when delivering presentations, but felt completely at ease with their accent when speaking in a relaxed environment. This section has demonstrated that 40 out of 54 of these students at times feel worried or embarrassed about their accents, with the three primary reasons being comparison to their peers (n = 21), fear of communication issues (n = 12), and emotional unease (n = 7). The following section will now explore whether demographic variables affect the extent to which accent anxiety occurs.

Demographic Variables
There are four demographic variables of interest in this study: age, gender, English level, and university major. However, given that 50 out of 54 respondents were in the 18-24 age bracket, the decision was taken to not analyze age as a variable. There were no significant differences in response patterns found between the males and females with reference to the themes discussed above. Again, it should be noted that the gender dispersion was not wide in the sample, with just 13 males compared to 40 females (and one non-binary).
The participants' self-reported English level was analyzed as a variable. The aim here was to identify if proficiency, as estimated by the participant, correlated with levels of speaking anxiety. Overall, it was found that the level had little effect on the extent to which the students suffered from worry or embarrassment about their accents. While the C1 and C2 respondents rated their own accents higher than the lower levels, they exhibited similar levels of anxiety about their accents (Figure 3 and  4). Slightly more B2, C1 and C2 students felt that a nativelike accent was one of their learning goals compared to the B1 and A2 students, but the consensus across all levels was that such an accent is not essential to the learning process.

Figure 3. Accent Rating per Level
The final variable was university discipline. As mentioned in the opening section, there was an almost even split between English majors and alternative majors, with most of the alternative majors being another language. Similar to the level variable, the major seemed to play little role on accent anxiety. Although the English majors rated their accent at the upper end of the scale compared to more mid and lower end ratings for the alternative majors, they reported similar beliefs about accents and harbored comparable worries and experiences about their own accents (Figures 5 and 6).

DISCUSSION
The findings of the study generated six primary observations.

Observation
1: Despite believing that comprehensibility is more important, many learners nevertheless aim to acquire a native-like accent.
It is clear from the findings that these students do not consider it a priority to acquire a native-like accent. Many of them repeatedly highlighted an extensive vocabulary as being the most important aspect of the language, followed closely by correct grammar, with accurate pronunciation being deemed the least important. The participants felt strongly that a native-like accent should be seen as a bonus objective, rather than as a principal languagelearning goal. The majority believed that if the speaker can make themselves understood in social interactions, their accent is of little importance.

(24) "I don't think it's essential, if your grammar is good, if you have a lot of vocabulary and your accent is good enough to be understandable then it's not that important. Nevertheless it's great to improve it."
This is in line with the contemporary consensus on accent in foreign language learning as outlined by Moyer (2013), namely that a non-native accent is only an issue if it actively impedes comprehensibility.
However, although the vast majority of students believe it to be unnecessary, over half of them would still like to speak with a native-sounding accent. Some believed that advanced learners should aim for a nativesounding accent, with many fearing that their non-native accent would lead people to believe that they had low proficiency in the target language.

(25) "I am under the impression that if I don't have a good accent I don't speak 'good' English."
(26) "But sometimes people give less credibility to someone with a strong accent, regardless of the content of their remarks." Certain students in the current study felt that a native-like accent connects to authenticity, in the sense that speaking with such an accent means one speaks a version of the language which is more real than a typical learner. This may suggest that acquiring a native-like accent is more of a goal for students with integrative motivation who wish to assimilate into the target community, rather than those learning due to instrumental reasons. One student in the current study did not feel accent was important for their English learning, but felt otherwise about their other language: (27) "I would not feel embarrassed to have a bit of a French accent when talking to an English native speaker as sounding French is still prideful, but when it comes to Japanese I have a strong desire to sound the most possible like a native speaker as I want to feel acknowledged and respected. How much you love the language also matters I think." This relationship between accent, identity and motivation will be further discussed in the fifth observation below.
Observation 2: Accent is a potential source of speaking anxiety which can affect even higher level students.
The finding that 40 out of 54 participants sometimes worry about how their accents sounds suggests that accent is indeed a potential source of language anxiety. Even the students who said that they liked speaking reported feeling worried or embarrassed at times because of their accent: (28) "I like to speak English but as I'm self-conscious about my accent I don't do it as often as I would like." (29) "I like interacting but I'm not very confident because of my accent." Matsuda and Gobel (2004) theorized that higher level students may feel more language anxiety due to increased pressure to perform well. This study did not support this finding with relation to accent anxiety, as the students' estimated level of English competence and their university major had little effect on whether they worried about their accents. This is in line with Horwitz et al. (1986) who claim that level is not an adequate predictor of FLA. The fact that such accent anxiety seems to affect all levels is significant. As a relatively under-researched area, it is possible that accent has been overlooked in past research on speaking anxiety.
It is also worth considering whether the data discussed here on accent anxiety is comparable to Baran-Łucarz's (2014) model of pronunciation anxiety, a phenomenon she says derives from four interrelated aspects: the desire to sound like a native speaker; the belief that one's pronunciation is poor; the fear of being negatively evaluated because of one's pronunciation; and the belief that accurate pronunciation is important for successful communication. The findings of the current study have demonstrated that although these students do not believe it necessary to attain a nativelike accent, many of them wish to do so, which equates to the first aspect of the pronunciation anxiety model. Their fears about communication breakdown suggest that they believe accent is important for communication, linking with the fourth aspect. It is clear that fear of negative evaluation (discussed below) is another primary source of accent anxiety. The only aspect of the pronunciation anxiety model that is not consistent with the findings of the current study on accent is the idea that anxiety stems from poor self-assessment. Although some students who rated and described their accents as poor did exhibit signs of accent anxiety, many students who rated their accents highly and felt like they spoke quite well nevertheless reported feeling worried or embarrassed about their accents at times. Overall, it appears that Baran-Łucarz's (2014) pronunciation anxiety model can also be applied to accent anxiety, with the caveat that accent anxiety can affect even those students who feel confident and rate their own accents quite highly.
It is possible that this difference is due to the fact that pronunciation can be considered a language skill as any other that can be practiced and refined, whereas accent is perhaps not as easy to alter and also has connections with identity and self-image. Indeed, many students expressed the desire to improve their accents, but felt this was difficult, if not impossible, a viewpoint which is also held by many researchers in the field (Derwing & Monro, 2009;Long, 1990;Moyer, 2013). It should also be noted that many participants may have assumed that accent and pronunciation are the same thing and thus used the term 'accent' when describing concerns about their pronunciation. With accent and pronunciation being so interlinked, it is inevitable that the terms may be used interchangeably in literature and studies. It is important that future research on accent makes the distinction between the two concepts clear to participants. The next three observations will now examine the dimensions of accent anxiety, with particular emphasis on its sources.
Observation 3: Accent anxiety stems from internal fears rather than from external feedback.
An interesting finding from the study is that most feelings of anxiety about accent derive from internal worry or fear rather than from negative feedback from speaking interactions. This is in contract to Baran-Łucarz's (2013) finding that pronunciation anxiety is frequently triggered by external feedback. The aforementioned study was based on control-value theory (see Pekrun, 2006) which claims that anticipated negative feedback is a key factor in provoking negative emotions. In the current study, there was no question related to anticipated negative feedback, however participants appear to be placing significant self-imposed pressure on themselves. One of the primary reasons cited for worrying about accent was the fear that it would lead to problems communicating, in which the interlocutor would not be able to understand the speaker.
(30) "I'm afraid they won't understand certain words and that communication will therefore be difficult." This connects to Horwitz et al.'s (1986) communication apprehension, as well as to MacIntyre et al.'s (1998) model of Willingness to Communicate, in that doubts about communication efficacy can increase FLA and unwillingness to communicate. However, a noteworthy finding was that although a significant number of students worried about such imagined communication breakdown, almost none of them reported any significant issues in this regard, suggesting that this fear is hypothetical and not a commonly occurring issue for this sample. Baran-Łucarz (2011) states that pronunciation anxiety stems from the speakers' perceived poor pronunciation rather than their actual pronunciation ability. Although it was outside the scope of the current study to investigate such links with accent, the fact that most accent worries were related to fears of future communication issues suggests that accent anxiety also arises from perceived weaknesses rather than from actual problems. The students in Price's (1991) study who felt anxious about their pronunciation highlighted negative teacher feedback as one of the key reasons for such anxiety. This did not apply to the current study, as no student mentioned their teacher as the cause of their anxiety. However, the classroom environment was labelled as a key contributor, as shall be discussed below.
Observation 4: Learners fear negative evaluation of their accents from their fellow classmates, but not from native speakers.
Perhaps the most surprising finding of the study was the significant evidence that accent anxiety almost exclusively affects students in the classroom rather than in their everyday language use. The results of the study show that Horwitz et al.'s (1986) third dimension of language anxiety, fear of negative evaluation, applies strongly to accent anxiety. Most students felt worried about how their accents sounded because they feared mockery from their fellow classmates. They repeatedly pointed out that they did not fear this same negative evaluation from native speakers. Similar to this fear of negative evaluation, the students tended to compare their accents to that of their classmates. While some felt increased confidence due to believing others had worse accents, most were of the opinion that their accents were weaker than their fellow students.
(34) "When I find myself, for example, in English class with people who have a really good English accent, I feel a bit lousy compared to the others." (37) "I feel inferior to those with a better accent." It is important to highlight, however, that most learners' fears were related to imagined mockery rather than to specific experiences of being ridiculed in class. This perceived judgement from their peers appears to be unwarranted, which again seems to support the third observation: accent anxiety arises from fear of hypothetical situations, not from actual negative experiences.
Observation 5: Accent anxiety may derive from threats to the learner's portrayal of L2 self-image.
The idea that accent is associated with self-image arose several times in the study. While the previous observation was concerned about negative evaluation of skill, this finding relates to how they will be perceived more generally when using their L2. Most of the students' fears involved damage to their selfimage: (38) "Yes [I worry about my accent] simply because of the image that the other person has of me." (39) "I'm afraid that people will think I'm a moron." (40) "I'm afraid that other people will judge my accent and say that I don't speak English well." As stated in the preceding section, the students also tended to compare themselves to others and worry about losing face because of their accents, particularly in front of their peers. It seems that having a poor accent, whether perceived or actual, weakens learners' confidence in their L2 self-image. According to MacIntyre et al. (1998), linguistic self-confidence plays a significant role in the learner's willingness to communicate, and this confidence ultimately implies a lack of anxiety. This connection between accent and L2 self-image is also reminiscent of Dörnyei (2005)'s L2 Motivational Self System in which learners are actively motivated or demotivated by concepts of their ideal L2 self and ought-to L2 self. Simsek and Dörnyei (2017) suggested that FLA operates in a similar manner to this system with learners associating their anxiety with levels of self-identity. Interestingly, certain students in the current study felt that anxiety about their accents motivated them to actively improve: Despite these comments, the majority of the students in the current study felt that their accent anxiety negatively impacted their learning, supporting Horwitz's (2017) view that FLA is essentially debilitating.

CONCLUSION
Although not considered a learning priority, attaining a native-like accent was seen as desirable by the majority of students in the current study. It was also found that many have at some time or other experienced embarrassment about how their accents sound. It is reasonable to conclude therefore that accent is a potential source of speaking anxiety, in that feeling anxious about how they sound may prompt the learner to avoid speaking. This is in line with Baran-Łucarz's (2014) finding on pronunciation and speaking anxiety. The two primary explanations for this accent anxiety here were fear of mockery from classmates and fear of communication breakdown, both of which are consistent with research on general foreign language anxiety.
The finding that this concept for the most part only affects students in the classroom suggests that teachers should be made aware of accent anxiety and work to foster a learning environment in which students are not afraid of being mocked, which may be achieved by ensuring classmates know each other well and employing group activities aimed at supporting each other rather than competing with each other. It was decided that the current study would focus on speaking anxiety within the traditional classroom setting in order to gain a general understanding of the nature of accent anxiety. However, given the current educational situation, it would also be fascinating to see whether accent anxiety is positively or negatively impacted by online-learning. One student in the current study mentioned that they felt less anxiety online, but also did not believe they were making as much progress compared to face-to-face learning. Assuming that online-based learning will become more and more prevalent in the future, more research of this kind is needed.
As pointed out in the introduction to this study, research on speaking anxiety has focused on pronunciation, and accent has generally been considered a subfield. This study points out distinctions between accent and pronunciation, however it must be noted that participants and researchers often use these terms interchangeably. Up to now, "Accent anxiety" has not been hypothesized as an independent construct within L2 speaking anxiety literature. Further research is required to determine if pronunciation anxiety is actually a subfield of the hypothesized construct "accent anxiety," or if they are independent constructs.
There are inevitably certain limitations to this study which may affect the findings, such as the small sample size of 54 participants based in one university, and the time and location constraints caused by the Covid-19 pandemic. Nevertheless, the findings presented here have demonstrated that accent anxiety exists and may contribute to students' unwillingness to speak. It is hoped that this study, along with further research in the area of accent anxiety, will ultimately help to develop understanding into the nature of speaking anxiety, and eventually inform students and teachers alike about how best to overcome and prevent this obstacle to successful language acquisition.