“I Want to Be Born with That Pronunciation”: Metalinguistic Comments About K-Pop Idols’ Inner Circle Accents

Abstract

The Korean popular music (K-pop) industry, with its global popularity and increasing multilingual orientation, serves as a suitable context for exploring language perceptions. This research examines the metalinguistic commentary on K-pop idols’ English accents on YouTube. Specifically, we investigate how online users evaluate the idols’ English accents and how their metalinguistic comments communicate linguistic ideologies that favor a “native” way of speaking. Our dataset consists of 602 metalinguistic comments drawn from four popular YouTube videos featuring the evaluation of K-pop idols’ accents. We employ content analysis to first categorize comments into positive, negative, and neutral evaluations, then focus on aspects being evaluated in the users’ comments (e.g., social attractiveness and (non-)nativeness). The results indicate that a vast majority of comments (88.1%) convey positive evaluations, largely associating the idols’ accents with social appeal and native-like accents. Although a few neutral and negative evaluations exist, our result shows a dominant preference for inner circle accents and complex attitudes toward accented speech in digital spaces. We conclude by highlighting the influence of digital platforms in shaping language perceptions and the implications for linguistic stereotyping in the context of K-pop culture.

1. Introduction

Accents are powerful markers of identities; they influence how listeners perceive their interlocutors, thereby impacting interactions in profound ways. Research on accent perceptions highlights that listeners often make assumptions about a speaker’s intelligence, social status, and personality based solely on accent (e.g., Fuertes et al., 2012; Labov, 2006; Rubin, 1992). These judgments, with their extensive societal consequences, have frequently been explored in the context of linguistic stereotyping (i.e., listeners attribute stereotypical characteristics to speakers based on their speech patterns) and reverse linguistic stereotyping studies (i.e., listeners’ existing bias toward a person or group influence their perceptions of the speaker’s language or accent) (e.g., O. Kang & Rubin, 2009; Rubin, 1992, 1998, 2012).

With the rapid development of technology and the emergence of various forms of computer-mediated communication, a growing body of research has begun to focus on understanding how stereotypes are expressed, contested, and perpetuated in digital space. Importantly, studies have emphasized that accents remain an influential factor in shaping identities even in virtual environments (e.g., Chen & Herring, 2024; Cutler, 2020; Newman, 2014). Among the multitude of digital platforms available today, YouTube, a video-sharing platform, may play a crucial role in the context of accent stereotyping and identity construction. Cutler (2020) put an emphasis on the importance of identifying YouTube videos featuring regionally accented speech and examining online users’ reactions to those videos as they provide a valuable window into people’s “shifting language attitudes, enregisterment, and how people position themselves in relation to certain ways of talking” (p. 36). Specifically, viewers’ reactions to accent-focused content can reflect underlying personal, cultural, and societal perceptions regarding different accents.

Taking up Cutler’s (2020) call for further scholarly attention to the digital interactions on accents, the current study aims to explore online users’ metalinguistic commentary on English accents by considering YouTube videos that feature Korean popular music (hereafter referred to as K-pop) idols and, more often than not, the “nativeness” of their English accents. In expanding circle countries such as South Korea, speaking English with a native-like accent is considered a desirable learning outcome (Derwing, 2003). The K-pop industry has experienced an unprecedented global rise over the past decade, moving from the 28th largest global music market in 2004 to the 6th by 2020 (IFPI, 2021). With the remarkable increase in the number of fans outside South Korea (Korea Customs Service, 2023), K-pop evolved into a unique cultural intersection where international fandom appears in numerous online platforms that spark a wide range of discussions, including not only music genres but also political issues and language ideologies (e.g., Chun, 2017; J. Kang, 2023; Malik & Haider, 2023). Despite K-pop’s global presence, research on digital interactions regarding accents has not sufficiently explored the ways in which K-pop idols’ English accents are evaluated by online users nor how these evaluations reflect wider language ideologies.

Along with the global expansion of the K-pop industry, idols are often required to be equipped not only with musical talents but also with English proficiency to connect with diverse audiences; consequently, their linguistic performance is subject to scrutiny and evaluation online by fans. It is important to note that accents do not inherently reflect a person’s language proficiency. However, various English accents showcased by K-pop idols are frequently evaluated by their audiences as part of idols’ English proficiency. In this regard, K-pop, as a global trend, serves as a rich site to examine how linguistic stereotypes are portrayed and negotiated in digital spaces. In this study, we investigate how online users from various linguistic and cultural backgrounds perceive K-pop idols’ English accents. By analyzing online users’ metalinguistic comments on K-pop idols’ English accents, our goal is to uncover the online discourse surrounding linguistic stereotyping within K-pop culture.

2. Review of Literature

2.1. Language Attitudes in Digital Spaces

Since the seminal study by Lambert et al. (1960), which proposed the linguistic stereotype hypothesis emphasizing the role of speech patterns as influencing and shaping one’s social identity, an extensive body of research in linguistic stereotyping has emerged. Specifically, the extant studies on linguistic stereotyping focus on listeners’ perceptions of different accents and uncover generally negative views toward non-native and/or non-standard accents (e.g., Boyd, 2003; Brennan & Brennan, 1981; Grondelaers et al., 2019; Lambert, 1967; Pantos & Perkins, 2012; Ryan et al., 1977). For example, Brennan and Brennan (1981) explored the relationship between the level of accentedness in English spoken by Mexican American individuals, as assessed by native speakers of English, and the evaluative judgments made by those raters regarding the accentedness. In their research, participants listened to audio recordings made by Mexican American speakers and then completed their evaluation questionnaires featuring eight pairs of bipolar adjectives (e.g., educated—uneducated). The results indicated that as raters perceived the speech as more accented, the status rating (i.e., social prestige) given was correspondingly lower. In a similar vein, Pantos and Perkins (2012) investigated the attitudes toward the speakers with a U.S. accent compared to those with a “foreign” accent (e.g., Chinese-accented speech, Korean-accented speech, etc.), discovering an implicit bias (i.e., individuals’ subconscious bias) favoring speakers with a U.S. accent.

The research into language attitudes has further expanded to a phenomenon referred to as reverse linguistic stereotyping (RLS), where listeners’ pre-existing biases about a speaker’s social group influence their interpretation of the speaker’s linguistic features (e.g., O. Kang & Rubin, 2009; O. Kang & Yaw, 2021; Kutlu, 2023; Rubin, 2012). O. Kang and Rubin (2009), for example, discovered that listeners often hold varying expectations based on the ethnicity of the speaker, which ultimately impacts their comprehension of the speech. By presenting photographs of speakers from two different ethnic backgrounds (Caucasian and Asian) in their experiment, the results indicated that the identical speech produced two different ratings depending on the photograph used; the speech accompanied by a Caucasian photo received a higher rating than with an Asian photo. Similarly, Kutlu (2023) conducted a study comparing American and Indian English using visual stimuli (Caucasian and South Asian), and the results indicated that accentedness judgments increased when the speech was shown alongside South Asian visual stimuli. In addition to examining listeners’ pre-existing biases, O. Kang and Yaw (2021) also incorporated listeners’ backgrounds as an additional factor in their analysis of accented speech. In their research, the authors identified listeners’ backgrounds, specifically related to their familiarity with different accents (e.g., studying abroad experience), and examined how this exposure influenced their perceptions of accented speech. The results indicated that listeners who had greater exposure to accented speech generally provided more positive evaluations of it. However, listeners with preconceived stereotypes about accented speech rated it as less prestigious and less socially appealing.

The body of research regarding language attitudes toward differently accented speech, overall, has primarily concentrated on English and its various accents (e.g., Brennan & Brennan, 1981; O. Kang & Yaw, 2021; Pantos & Perkins, 2012; Ryan et al., 1977). It is not surprising considering the role of English as a global lingua franca and its widespread use across the globe. According to Kachru (1992), English dialects and accents can be classified into three concentric circles: (1) the inner circle, representing native English-speaking countries (e.g., the United States, the United Kingdom, Australia, New Zealand, etc.); (2) the outer circle, including countries where English is used as an official language (e.g., India, Philippines, Nigeria, Singapore, etc.); (3) the expanding circle, encompassing countries where English is used as a foreign language (e.g., South Korea, China, Japan, France, etc.). In light of this, previous studies have shown that listeners, which include both native English speakers and those who speak English as a second or foreign language, often hold negative perceptions toward speakers whose accents differ from those of the inner circle, which can ultimately foster stereotypes toward non-native English accents (e.g., Ahn & Kang, 2017; Hendriks et al., 2018, 2023). Specifically, the research conducted by Ahn and Kang (2017) illustrates how these stereotypes create a fallacious association of non-native accents with various extralinguistic factors, such as social status, intelligence, academic achievements, gender, and ethnicity. Although several scholars have implemented listener training as an effort to reduce this stereotype (e.g., O. Kang, 2015; Lindemann et al., 2016; Manohar & Appiah, 2016), the preference for a native-like accent remains widespread.

While extensive empirical research has explored individuals’ perceptions and attitudes on various accents, the rise of digital communication has further broadened this field of study, encouraging researchers to examine accent stereotypes in digital spaces (e.g., Aslan & Vásquez, 2018; Chen & Herring, 2024; Cutler, 2016, 2020; Rymes & Leone-Pizzighella, 2018). Among the myriad online platforms, social media sites (e.g., YouTube, Facebook, and Bilibili) have widely been investigated as these platforms allow users to create and share content on a wide range of topics, including language. As Androutsopoulos (2015) argues, user-generated content regarding language results in context collapse, where diverse language attitudes toward multiple language varieties are expressed by individuals from various social backgrounds. In this context, social media platforms serve as dynamic spaces for enregisterment of language varieties—a process by which certain linguistic and/or non-linguistic features become social indexes (Agha, 2003, 2005; Cutler, 2016). Furthermore, the inherently interactive and diaglogic nature of platforms plays a crucial role in influencing public attitudes toward specific language varieties and accents (Heyd, 2016).

For instance, Aslan and Vásquez (2018) investigated metalinguistic comments made by online users on two YouTube videos that discussed a “non-standard” way of speaking English by a teenage girl in the U.S., uncovering how viewers evaluate the speech and connect it to various social categories (e.g., race, region, education, etc.). In their research, the “non-standard” speech was largely viewed negatively, often linked to lower-status social categories, “ghetto” or “hood”, highlighting the widespread prescriptive mindset within society. With a more concentrated focus on accent, Cutler (2020) examined four YouTube videos that featured “How to Talk Like a New Yorker”, focusing on the metalinguistic comments made by individuals regarding the New York accent. By analyzing users’ comments on these accent-focused videos, the results illustrated that the New York accent, which was historically tied to negative stereotypes, is now viewed in a somewhat favorable manner, indicating a shift in the perceptions of the New York accent. Along similar lines, Chen and Herring (2024) analyzed attitudes toward non-native Mandarin accents by analyzing online comments directed at two non-native Mandarin speakers on Bilibili (a popular Chinese online video platform). Their findings indicated a general negative perception of a non-native speaker who frequently made pronunciation mistakes.

In summary, scholars in this area have highlighted the important role of social media platforms in shaping and reflecting individuals’ attitudes toward various pronunciations and accents. According to Chen and Herring (2024), social media platforms not only reflect the prevailing language attitudes and ideologies but also have the potential to reinforce existing biases and perceptions, given their widespread use in contemporary society. Thus, the current study seeks to gain a deeper understanding of current attitudes toward English inner circle accents in our globalized world. Our focus is on analyzing comments made by online users in response to videos that featured K-pop idols speaking English with inner circle accents. To provide a context for our investigation, we outline the role of English in South Korea, especially in relation to the K-pop industry, in the following section.

2.2. English in South Korea

Building on the global perspective of language attitudes influenced by digital environments, it is crucial to explore how these interactions manifest in specific cultural contexts, such as South Korea (hereafter Korea), where English serves not merely as a means of communication but also as a marker of social identity. In Korea, one of the countries in the expanding circle, English proficiency is pursued not only for communicative and practical purposes but also for its social indexicality as a class marker, allowing the speaker to showcase their connection to cosmopolitan (Lee, 2014). Furthermore, as Derwing (2003) points out, achieving a perfect native-like accent in English remains a desirable outcome for learners in numerous countries within the expanding circle (i.e., English as a foreign language (EFL) contexts). Consequently, it is not surprising that Korean L2 learners perceive English accents from the inner circle as more legitimate than English spoken with a Korean accent. For instance, Butler (2007) explored the impacts of accents performed by Korean elementary school teachers through experiments assessing students’ perceptions of American-accented English versus Korean-accented English. In his examination of the potential differences in students’ learning outcomes influenced by teachers’ varying English accents, the results indicated that there was no significant difference in students’ performance. Nonetheless, surprisingly, the research revealed a significant difference in student’s attitudes toward the two distinct English accents. Specifically, students believed that speakers of American-accented English excelled in their pronunciation and confidence when using English. Similarly, McDonald and McRae (2010) examined the attitudes of Korean English teachers toward a local variation of English, referred to as “Konglish”. The results indicated that a majority of teachers perceived Konglish as deviating from standard English, thus labeling it an incorrect way of practicing English.

Although earlier studies have shown that the perceptions of Korean L2 speakers toward local English varieties (e.g., Korean accented English) have progressed to appreciate the value of non-inner circle English accents (e.g., Ahn, 2014; Sung, 2019), the preference of Korean L2 speakers toward inner circle English accents remains pervasive (e.g., Ahn et al., 2020; Byun, 2016). For instance, Byun (2016), through her qualitative research with undergraduate students in Korea, discovered that Korean L2 learners hold complex perspectives regarding native-like accents. While students acknowledged that the English language is no longer limited to English-speaking countries, they also conveyed a strong preference for learning English from native speakers from the inner circle countries. Ahn et al. (2020) corroborated these findings in their analysis of perceptions of native speakers in Korean news and social media, revealing a prevailing belief that native speakers of English should serve as role models for learning English. In summary, while there is a growing recognition of non-native varieties as valid forms of English varieties, the deep-rooted preference for native speakers, particularly from the inner circle, highlights the complex attitudes of Korean L2 speakers toward English in today’s context.

2.3. K-Pop Industry in Korea and Its Global Impact

The unique role of English as both a lingua franca and a cultural capital has also been evident in the K-pop industries. K-pop, an abbreviation for Korean popular music, has gained a growing number of global fans thanks to its unique characteristics, including captivating melodies and beats, impressive choreography, and the visual appeal of its artists. Although K-pop encompasses a broad spectrum of sub-genres within the music scene of Korea, it is primarily associated with K-pop idol groups, such as BTS and New Jeans, which have emerged as the most popular sub-genres among the others. These idol groups are typically created and managed by production companies, which play an important role in various ways—from the selection of the members to the promotional activities.

In the early days of K-pop, idol groups were predominantly made up of Korean members (e.g., FIN.K.L, Wonder Girls). However, as the popularity of K-pop idols expanded beyond Korea, production companies began to recruit talent from various countries to enhance the group’s connectivity to the world, leading to over 30 foreign idols debuting in 2021 (Min, 2024). The increasing global popularity of K-pop has also led to a change in communication styles among idols. With an expanding number of international fans, many K-pop idols, whether they are Korean or foreign, have begun to speak English in many events outside of Korea, including interviews, concerts, and promotional events. This ability to communicate in English has become a charming point for international fans, giving a sense of connection and accessibility (Octaviani & Yamin, 2020). Furthermore, the idols’ frequent use of English has also encouraged fans in EFL settings to learn the language. In this regard, Faiza (2020) explored the K-pop community on Twitter (now called X), primarily focusing on fans who use English as a foreign language. The findings showed that fans were inspired to learn English while engaging with K-pop content and communicating with other fans around the world.

Faiza (2020) offers invaluable insights into the impact of K-pop culture on English language learning, highlighting the ways in which this musical genre promotes engagement with English in EFL contexts. However, it still remains unknown how the English-speaking practices of K-pop idol members are perceived and how this specifically influences their fans and learners of English. As K-pop continues to gain global popularity and maintains abundant digital activities across various platforms (e.g., Jin, 2021, 2024; Jung, 2014), it is important to investigate the effects of this phenomenon. Understanding how idol members‘ English proficiency, accents, and speaking styles are perceived by their audience could provide valuable perspectives.

2.4. Current Study

Language attitudes, particularly those pertaining to the speaker’s accent, have been extensively studied across various varieties and contexts. By broadening the research scope to digital spaces, the existing body of work has yielded new insights, showcasing the language ideologies that manifest through different online platforms. Regarding language ideologies related to English varieties, previous studies consistently highlighted a strong preference for either the “standard” accent or the inner circle accent (e.g., Brennan & Brennan, 1981; Hendriks et al., 2018; Pantos & Perkins, 2012). This trend has also been observed in Korea, where English serves as a symbol of social status (Lee, 2014). As one of the expanding circle countries where English is used as a foreign language, several studies have indicated that Korean L2 learners prefer inner circle English accents (e.g., Ahn et al., 2020; Butler, 2007; Byun, 2016). In this regard, this study aims to deepen our understanding of how the English accents of K-pop idols are perceived and evaluated in digital contexts. Considering the growing global influence of the K-pop industry and the various digital activities associated with it (e.g., Jin 2021, 2024; Jung, 2014), digital space is regarded as a valuable research area that can illustrate various language ideologies regarding English accents both within and outside Korea. Specifically, our research investigates how online users evaluate the English accents of K-pop idols, focusing on the following: (1) affect (i.e., positive, negative, neutral); (2) other social aspects related to the corresponding affect (e.g., social attractiveness, (non)-nativeness). Through the analysis of online users’ metalinguistic comments regarding K-pop idols’ English accents, our objective is to uncover the online discourse surrounding linguistic stereotyping within K-pop culture. This study, therefore, aims to contribute to the understanding of language attitudes in a contemporary digital space and highlight the dynamic interplay between K-pop idols’ use of English and how online users perceive it in the globalized context of K-pop.

3. Materials and Methods

3.1. Data Collection

Data were collected from YouTube using purposive sampling. In terms of our sampling procedures, we first employed the YouTube search engine with four key search terms in Korean in order to collect videos featuring idols’ performance of English accents and/or commentaries on their accents. The key terms employed in our search include “영어” (English), “발음” (literally, pronunciation; figuratively, accent), “아이돌” (idol), and “모음” (collection or compilation). We used the above search key terms as online users very likely offer metalinguistic comments on idols’ English accents in response to the video content, given both the focus of the videos (e.g., idols’ accents), as well as the participatory affordances of YouTube (Aslan & Vásquez, 2018). In addition, as YouTube allows the user to “filter” the search result generated based on criteria such as upload date (e.g., last hour, this week), relevance, view count, and rating, we selected relevance as a sorting criterion in order to retrieve videos that are most relevant to K-pop idols’ English accents. Among the videos in the returned results, we sampled the most popular videos as indicated by the number of likes, comments, and views the videos received. After screening the returned results based on the above criteria, we chose to sample the five most popular videos as the total numbers of comments under these videos exceeded 4584, which reached the sampling size in previous studies concerning online users’ metalinguistic comments on YouTube. For example, in Cutler (2020), a total of 2649 comments posted under four videos were sampled. For this reason, we consider the first five videos, with their considerable numbers of comments and views, to be sufficient for us to construct a rich dataset consisting of metalinguistic comments on idols’ English accents for the following data analysis.

We then sampled users’ comments in relation to accents under the five chosen videos. In this process, we eliminated one video from the data analysis as it did not have any accent-related comments. After data cleaning, users’ comments posted in response to four videos featuring K-pop idols’ English accents were analyzed in our study.

Table 1. Metadata of the focal videos.

No.	Account	The Number of Subscribers	The Title of the Video	Released Date	The Number of Views	The Number of Comments
1	Doyouram-Everyday K-Culture	21.6 M (million)	알고 들으면 더 신기한 아이돌 영어발음 특징 (RM, JOSHUA, ROSÉ, JENNIE, WENDY, MARK) [ENG] [Idol English pronunciation characteristics that is more surprising if you know it (RM, JOSHUA, ROSÉ, JENNIE, WENDY, MARK) [ENG]]	16 August 2021	203,435	431
2	JTBC Voyage	349 M	♨핫클립♨ 거침없는 영어실력 뽐낸 레드벨벳 웬디 (Red Velvet WENDY) (발음도 이뻐♥) #스테이지K #JTBC봐야지 [♨ Hot Clip♨ Red Velvet WENDY who boasted her fluent English skills (Pronunciation is pretty♥) #StageK #JTBC]	7 April 2019	4,143,521	1815
3	영크릿 | Young Cret	20.9 M	[BLACK PINK]블랙핑크 제니 영어 뉴질랜드 발음 VS 로제 호주발음 비교해보자 (영어공부|ENG SUB) [[BLACK PINK] Let’s compare BLACKPINK Jennie’s English New Zealand pronunciation vs Rosé’s Australian pronunciation. (English study|ENG SUB)]	6 July 2020	1,887,408	1232
4			[BLACKPINK, STRAY KIDS] 블랙핑크 로제 영어 VS 스트레이키즈 필릭스 영어 원어민 실화냐? (영어공부|ENG SUB) [[BLACKPINK, STRAY KIDS] BLACKPINK Rose English VS Stray Kids FELIX English native speaker, is it real? (English study|ENG SUB)]	18 July 2020	699,728	1106

shows the metadata of the four videos, including the name of the account, the number of subscribers of the account, the video title, the released data of the video, the number of views of the video and the number of comments of the video. In addition, we present the demographic information of the idols who appeared in the videos and their accents, as labeled by the content creators, in

Table 2. Idols and their accents described in the video.

No.	Idols’ Names	Gender	Demographic Background	Accents Described in the Video
1	Joshua	Male	Born and raised in the US.	Valley girl accent
2	Mark	Male	Born and raised in Canada	New York accent
3	Wendy	Female	Born in Korea, moved to Canada when she was 11 years old.	American accent
4	Rosé	Female	Born in New Zealand, moved to Australia when she was 7 years old.	Australian accent
5	Jennie	Female	Born in Korea, studied abroad in New Zealand when she was 9 years old and stayed there for five years	New Zealand accent
6	RM	Male	Born and raised in Korea	Undefined
7	Felix	Male	Born and raised in Australia	Australian accent

.

In the four videos, we first manually extracted all comments posted under those videos (n = 4584), and we reviewed the content of the comments to exclude comments that were not metalinguistic descriptions or discussion focusing on idols’ accents. To be clear, our consideration of metalinguistic comments aligns with Jaworski et al. (2004, p. 3), who define metalinguistic comments as those that express “‘common sense’ understandings of how language works, what it is usually like, what certain ways of speaking connote and imply, what they ought to be like”. Out of 4584 comments, 602 were identified as metalinguistic after screening the data. In them, we identified 197 (32.7%) as off-topic comments (Figure 1), as the evaluation is not directly focused on the idols’ accents but on other language-related aspects, such as grammar knowledge, lexical choices, capability of comprehending English, and so forth. Although our primary analytic goal is to examine explicitly accent-related metalinguistic comments, we also take off-topic instances into account in our analysis, as these “off-topic” metalinguistic comments are also motivated by videos featuring commentaries on the idols’ English accents and provide insights into online users’ perceptions toward idols’ linguistic performances. With this in mind, we address the off-topic comments as well to examine how online users’ evaluation of the idols’ speech may overlap with other aspects of language (e.g., grammar).

In terms of idols whose English accents are subject to evaluation in the videos sampled, it appears that the content creators overwhelmingly focus on idols who were either English native speakers (Joshua, Mark, Rosé, and Felix) or lived in an English-speaking country for a long time (Jennie and Wendy), while only one (RM) who was born and raised in Korea. However, the “skewness” toward English native speakers appears to be inherent in the dataset. Specifically, our initial consideration was to sample popular commentary videos on idols’ English accents regardless of the English varieties being evaluated in the videos. The returned search results showed that the majority of content creators who produced the focal type of video focused on idols who were somehow native speakers of English. For this reason, few idols who are non-native English speakers were subject to evaluation in our dataset.

Regarding the language used in the comments, out of the dataset of 602 comments, 446 were written in English, while 156 were in Korean, as shown in

Table 3. Metalinguistic comments written in different languages.

	Metalinguistic Comments (n = 602)
Languages of the Comments	English	Korean
Examples	Data ROSIE’s accent is so ELEGANT and MAGESTIC. its chef’s kiss	“남준 오빠 영어 발음 개섹시해 ㅠㅠㅠㅠㅠㅠㅠㅠ T: Namjoon’s English pronunciation is so sexy”
Video#1	82	53
Video#2	109	10
Video#3	143	31
Video#4	110	64
Tally	444 (74.1 %)	158 (25.9 %)

.

The predominance of English comments may be due to the global popularity of K-pop and fans’ use of English as lingua franca (Faiza, 2020) and the bilingual representation of the video content (e.g., all four videos provided English subtitles, and three videos indicated the availability of English subtitles in their title).

3.2. Data Analysis

The overarching analytic approach we adopted in this study is content analysis, as our research is an exploratory and data-driven study. Content analysis is a method for “the subjective interpretation of the content of text data through the systematic classification process of coding and identifying themes or patterns” (Hsieh & Shannon, 2005, p. 1278). Following Aslan and Vásquez’s (2018) coding procedure of metalinguistic commentary on YouTube, our analysis consisted of two stages: we first categorized the data into a tripartite division, including positive, negative, and neutral comments. While standards for neutrality can differ among individuals, our emphasis was on explicit evaluations. Thus, both positive and negative evaluations were frequently accompanied by evaluative adjectives, and any mere description of the accent through labeling the English varieties (e.g., Australian accents) was regarded as neutral. We present examples of the above categories later in the next section in

Table 4. The overall tendency in terms of general evaluation.

General Evaluation	Percentage (Number)	Example
Positive	88.1% (357)	Overall positive evaluation (e.g., I want to be born with that pronunciation), Adjectives of positive evaluation (e.g., Rose and her aussie accent is so elegant)
Neutral	10.6% (43)	Mere description of accent (e.g., They both speak aussie accent)
Negative	1.3% (5)	Adjectives of negative evaluation (e.g., I find Jennie’s kiwi accent to be very pretentious)

. We then mapped out emerging and reoccurring subcategories concerning what accent-related aspects (e.g., (non-)nativeness) are subject to the evaluation in users’ metalinguistic comments and how those accent-related aspects are associated with the idols as an individual (e.g., social attractiveness).

On a practical note, we each read through all accent-related comments to familiarize ourselves with the data. This, in turn, allowed us to identify recurring themes emerging from the data. Meanwhile, we annotated the data with memos and notes to help ourselves identify repeated patterns relating to aspects being evaluated in the comments. Prior to formally undertaking the data analysis, we conducted a pilot analysis focusing on 100 comments from the dataset. Based on our pilot analysis, we decided on initial codes and the tripartite division together. Then, we went through several rounds of coding independently, drawing on the agreed-upon categories. In this way, we constructed the codes inductively and recursively by re-reading and analyzing all the metalinguistic comments under examination. Specifically, both authors first coded the data independently based on the initial categories. Following the independent coding, the two authors compared their codebooks and, after discussing any discrepancies, reached 100% agreement on the results. Codes used in the data analysis included social attractiveness, (non-)nativeness, phonetic features, speech, motivation, proficiency, comprehensibility, and metalinguistics. Additionally, we used “simultaneous coding” (Saldaña & Mallette, 2017) for some comments when there are overlapping codes. In other words, a single comment is coded with multiple labels when its evaluation has references to multiple accent-related aspects (e.g., social attractiveness and (non-)nativeness).

In the following sections, we present findings based on our research foci, which are as follows: (1) affect (i.e., positive, negative, neutral) of the metalinguistic comments; (2) dominant aspects related to the corresponding affect (e.g., social attractiveness, (non)-nativeness). Following the above section, we offer a discussion regarding off-topic comments identified in our dataset.

4. Findings

4.1. Affect Toward Idols’ English Accents

In a total of 602 metalinguistic comments, as illustrated in Figure 1 above, 67.3% (n = 405) of them indicated evaluations regarding English pronunciations and/or accents of idols. We examine the overall sentiment expressed toward the idols’ accents and pronunciations, as detailed in Table 4. In addition, Figure 2 illustrates the overall tendency in terms of affect indicated in online users’ comments toward the K-pop idols’ English accents. As seen in Figure 2, we find that a majority—88.1% (n = 357)—expressed a positive evaluation with varying degrees of aspects out of a total of 405 accent-related comments. It is crucial to consider the linguistic backgrounds of the idols featured in our study; out of seven idols, six either grew up in or studied in inner circle countries at a young age (see Table 2). Thus, the overall positive evaluation of these idols’ English pronunciations and accents reflects predominantly favorable attitudes toward inner circle accents among the audience. In contrast, a smaller portion of the comments, accounting for 10.6% (n = 43), reflected a neutral stance regarding the idols’ English accents and pronunciations. Furthermore, only 1.3% of the comments (n = 5) expressed a negative evaluation of their English-speaking practices. This distribution of sentiment highlights a general trend: the audience predominantly favors the idols’ native-like speaking styles and their inner circle accents. In our further examination in terms of various aspects of evaluation, as illustrated in

Table 5. The overall tendency in terms of aspects of evaluation.

Aspect	Examples	%(n) *
Social attractiveness	젠틀섹시 조슈아... 영어할 때 더 잘생겨짐 (T: Gentle sexy Joshua...Becoming more handsome when speaking English)	33.6% (136)
(Non)-nativeness	Even native Australia got shock hearing Felix’s think Aussie accent	30.3% (123)
Phonetic features	Rose’s voice is very soothing and smooth (coming from an Australian, I don’t really understand why people love our accent because to us it sounds not that unique haha). It’s also interesting to see Rose’s accent sometimes switches to a more American pronunciation eg) rolling r sound in some words over time. Maybe from being with Jannie for so long with her half kiwi half American accent	4.7% (19)
Speech	I am very impressed with Rose, and not because of her accent or how fluent she is or how she pronounces the words. I am impressed because of what comes out of her mouth. They are not only instantaneous responses, but very intelligent as well. She doesn’t just open her mouth to speak, they are all carefully selected and well thought, and she does it in split second.	3.7% (15)
Motivation	I watched the video because of their accent but finished it with so much inspiration that I too can achieve my dreams if I work hard for it.	3.2% (13)
Proficiency	I love Wendy’s accent, like she’s so fluent in English. not like me though I’m not good at English and I’m not that so fluent and accent at it...	2.2% (9)
Comprehensibility	Jennie’s English is so soothing to hear. The words are so clear, like we don’t need subtitles to understand her.	2.2% (9)
Metalinguistic	Many people can relate with rose, so do I. we will show our different character based on the language we speak. I love	0.7% (3)

* Since comments can be coded under multiple evaluation aspects, the overall percentage exceeds 100%.

, two aspects, “social attractiveness” and “(non-)nativeness”, stood out as predominant aspects throughout our dataset. These elements reflect how the audience perceives idols’ English-speaking practices in relation to other aspects.

4.2. Social Aspects Related to the Corresponding Affect

4.2.1. Social Attractiveness

Out of a total of 405 comments, a notable 33.6% (n = 136) focused specifically on evaluating idols’ English accents and pronunciation by linking these linguistic attributes to their overall social appeal. For social attractiveness, we categorized comments that link idols’ English-speaking practices to the other physical and personal attributes that are socially perceived as attractive. Importantly, all 136 comments that addressed this particular aspect expressed a positive sentiment toward the idols’ accents. In these comments, commenters praised idols’ physical attributes, such as voice and general appearance, in addition to their native-like accents. In these evaluations, commenters frequently employed evaluative adjectival phrases (e.g., elegant, classy, sexy) or metaphors (e.g., “she sounds like a Disney princess”) to praise the idols’ accents and their social attractiveness. Below, we discuss several examples that highlight social attractiveness in addition to positive evaluation of the idols’ English accents and pronunciation.

(1)	Korean original:	젠틀섹시 조슈아…영어할 때 더 잘생겨짐
	English translation:	Gentle sexy Joshua...Becoming more handsome when speaking English

In Example (1), the commenter provided a general evaluation of Joshua’s English-speaking practices, who is part of the Korea-based idol group Seventeen. It is a well-known fact on the Internet that Joshua, being a Korean-American, spent his early years in Los Angeles, United States, until he turned 19. Given his background, it is normal for him to communicate in English since it is his first language. Nevertheless, his English-speaking practices, described as “Valley girl accent” in the video, frequently come under evaluation. In particular, Example (1) indicated that his English-speaking practices contribute to enhancing his attractiveness by rendering him “more handsome”.

While Example (1) provided a broad positive evaluation, Example (2) concentrated on pronunciation, differentiating between what was presumably the idol’s English proficiency and pronunciation; although the commenter did not praise the former, the latter received an extremely favorable evaluation using superlative. The idol subject to evaluation in this comment is Wendy, who was born and raised in Korea until she was 11 years old and moved to Canada. Thus, the user of this comment acknowledged the fact that she is not a native speaker of English. However, it is important to note that her pronunciation is still being assessed from the viewpoint of a native speaker. Overall, the commenter in Example (2) demonstrated a positive evaluation regarding the idol’s pronunciation and social appeal by performing the following: (1) explicitly distinguishing pronunciation from general English-speaking skills/proficiency; (2) evaluating from the perspective of a native speaker of English; (3) connecting her personal attributes to the evaluation.

(2)	She doesn’t get every word exactly right, but her tone and pronunciation are the best I’ve ever heard from a non-native English speaker. Couple that with her kindness and intelligence, and Wendy really is an incredible idol and person.

Consequently, examples (1) and (2) illustrate not only the commenters’ positive evaluations of the idols’ inner circle accents but also the connection between inner circle accents and social appeal, emphasizing that how idols’ English accents and pronunciation are perceived can greatly affect their overall image. This relationship between social appeal and favorable evaluation toward the inner circle accent is not coincidental; the connection between accent and social appeal has been emphasized in this field of research. Zahn and Hopper (1985) developed a speech evaluation instrument that includes three key dimensions: attractiveness, superiority, and dynamism. These dimensions serve as a foundational tool for assessing listeners’ perceptions of speakers and have been widely applied in research exploring linguistic stereotyping and reverse linguistic stereotyping (e.g., O. Kang & Rubin, 2009; Rubin et al., 2016). For example, in their research, O. Kang and Rubin (2009) found a link between the dimension of superiority and the social attractiveness of speakers, particularly in terms of their evaluations of L2 performance. Consequently, the insights from previous studies, in conjunction with our findings, highlight the idea that accents and pronunciation connotate much more than mere linguistic features. They are intricately tied to the social identities that the listener ascribes to the speakers. This relationship suggests that inner circle accents are frequently associated with social attractiveness. In the context of the K-pop industry, where the social appeal of idols is directly tied to their overall fame and popularity, and consequently their success, this relationship may suggest that the pronunciations and accents of idols, which are unrelated to their language proficiency nor easily altered by effort, can shape how fans perceive them, potentially impacting their career trajectory in a highly competitive entertainment industry.

4.2.2. (Non-)Nativeness

Following social attractiveness, the second dominant aspect associated with idols’ English pronunciation in our dataset is (non-)nativeness, which accounts for 30.3% (n = 123) of the examples. For (non-)nativeness, we categorized comments that link idols’ English speech to a “native” way of speaking under this aspect. Such metalinguistic commentary is often expressed by, for example, making explicit claims about idols’ native speaker identity and comparing their pronunciation to that of inner circle speakers. The extent to which the idols sound like native speakers is frequently referenced when commenters offer positive evaluations of their speech. In a few cases, an idol’s native speaker identity is highlighted by the commenter to rationalize the idols’ native-like pronunciation; in those cases, the commenter does not subject idols’ accents to compliment but instead considers idols’ pronunciation as insignificant to comment on unless the idol is perceived to have made a lot of intentional efforts to achieve the accent. Below, we discuss several examples that highlight the nativeness, or the native speaker identity, of idols in a positive, neutral, and negative light. Given the statistical predominance of positive evaluations, we present two examples featuring positive metalinguistic comments and one example from each of the remaining subcategories in our dataset.

(3)	Even native Australia got shock hearing Felix’s thick Aussie accent

In Example (3), the commenter directs their evaluation specifically to Felix, an Australia-born and -raised, Korea-based rapper, while simultaneously implying the unexpected “thick” Aussie accent the idol displayed. Instead of directly evaluating Felix’s native-like accent, the commenter highlights the role of the listener (i.e., “native Australian”) and uses the adverb “even” to emphasize the unexpectedness of Felix’s accent. The unexpectedness is further reinforced in the comment: hearing his “thick” Aussie accent, the “native Australian was shocked”. In doing so, the commenter conveys the idea that Felix’s accent is not only noticeable due to its prosodic features (i.e., a thick Australian accent), but also unexpectedly and extraordinarily “authentic”, as even native speakers were shocked by it. In addition, the comment implies the linguistic ideology that “native Australians” represent the authoritative performance of Australian-accented speech and, by extension, Australian English as an English variation. Native speakers, therefore, are seen as the best judges of Australian-accented speech, and Felix’s accent is considered “thick”—and “authentic”—enough to surprise and impress those who represent the authentic norm. However, the commenter likely identifies Felix as a non-native speaker, as Felix is actually a native English speaker who was born, raised, and educated (at the college level) in Australia. Lacking this background knowledge and perhaps instead speculating based on Felix’s Asian ethnicity, the commenter evaluates his Aussie accent, which might otherwise be considered an “unmarked” aspect of his English pronunciation (i.e., speaking Australian-accented English).

(4)	Rosé is a native English speaker but I’m still fascinated whenever she speaks English. I just melt whenever her voice become low

Unlike in Example (3), in Example (4), the commenter makes an explicit native identity claim for Rosé (i.e., “Rosé is a native English speaker”), who was born in New Zealand and raised in Australia. However, the commenter stresses their appreciation of her English pronunciation despite acknowledging Rosé’s native speaker identity (i.e., “but I’m still fascinated whenever she speaks English”). At the same time, the commenter’s remark implies that it is somewhat contradictory to compliment Rosé’s English pronunciation while knowing that she is a native speaker, as her pronunciation should be perceived as “unmarked”, “natural”, and “authentic”, and therefore rarely subject to judgment. However, Rosé’s “unmarked” accent is still subject to evaluation by the viewer, even by those who are aware of her native identity, possibly due to other social categories Rosé may be associated with, such as her ethnicity (i.e., Asian) and bilingual (i.e., being able to speak both Korean and English). In addition, Rosé’s English speech is marked by the low pitch of her voice when speaking the language (i.e., “I just melt whenever her voice becomes low”). Here, the comment on Rosé’s voice in English as “becoming low” implies a contrast between the voice she uses when speaking other languages (e.g., Korean) and the one she uses when speaking English.

While examples (3) and (4) are both positive metalinguistic comments, Example (5) represents instances where the idols’ native accents receive a neutral evaluation.

(5)	Jennie has the most neutral accent (unless she does her NZ accent) among all four... neither too ‘black’ not too ‘white’ nor too ‘Asian’ be it singing or speaking

In Example (5), the commenter evaluates the English accents of the four-member Asian girl group Blackpink. The commenter specifically points out the “neutral” accent of Jennie, who was born and raised in Korea and studied in New Zealand for five years when she was 10 years old. Jennie’s English is described as “accent-free”, as her pronunciation does not lean toward any particular accent but is considered “the most neutral” among all four members of the group. Additionally, in parentheses, the commenter notes an exception: Jennie’s English only has an accent when “she does her NZ accent” (New Zealand accent). In this way, the commenter demonstrates metalinguistic awareness that a speaker can choose which English accents to perform (e.g., “NZ accent”, New Zealand accent), and that an English accent is not inherently or naturally tied to the speaker’s identity (e.g., ethnicity, race), characteristics, or personal traits. Instead, the same speaker (in this case, Jennie) can and does perform multiple accents, which do not necessarily confirm the social categories the speaker identifies with.

Despite showing the metalinguistic awareness of accent as a linguistic performance, the evaluation that follows by the commenter reinforces the artificial association between accents and social categories. Specifically, Jennie’s accent is regarded by the commenter as “neither too ‘black’, nor too ‘white’, nor too ‘Asian’, be it singing or speaking”, which further reinforces the commenter’s prior evaluation that Jennie has the “most neutral” English accent in her group. Meanwhile, the comment shows that accents are closely associated with, and are salient indicators of, ethnicity. Jennie’s accent—being perceived as distancing itself from those ethnic English varieties (i.e., Black, White, Asian)—becomes the most “neutral”, or unidentifiable, in terms of its indexicality to social categories. Example (5), thus, represents a paradoxical linguistic ideology in which accents are seen both as a performance that is not naturally or exclusively tied to ethnicity, race, or nationality (e.g., Jennie using a New Zealand accent) and as a strong indicator of a particular social category, such as ethnicity or race.

A few examples in our dataset negatively evaluate the idols’ performance of native accents. Example (6) below is one of those instances.

(6)

I find Jennie’s kiwi accent to be very pretentious because she was only in New Zealand for 5 years and lived in Seoul for 19 to 18 years. She obviously knows how to speak English without the accent. Also I believe her accent isn’t truly kiwi anymore since she has live outside of the country for so long she would lose the authentic dialect.

As presented above, the commenter first categorizes Jennie’s English accent as a “kiwi accent”, referring to the New Zealand accent, and regards her native-like accent as “very pretentious” due to Jennie’s limited, non-permanent stay in New Zealand and her more stable, long-term presence in a non-English-speaking country, Korea (i.e., “she was only in New Zealand for 5 years and lived in Seoul for 18 to 19 years”). The commenter alludes to the idea that sojourners who perform a native-like accent are not authentic representations of the English variant and that it is not “natural” but pretentious for sojourners to use the accent, as they “obviously know how to speak English without the accent”—while presuming that the idol knows a completely neural way of speaking English. The remark that follows further reinforces the close association between inhabitant status and linguistic authenticity. Specifically, the commenter indicates that Jennie lost her once-possessed Kiwi accent after ceasing her sojourn in New Zealand (i.e., “I believe her accent isn’t truly Kiwi anymore since she has lived outside of the country for so long she would lose the authentic dialect”). Thus, the linguistic ideology communicated here is that speakers need to constantly inhabit an English-speaking country to maintain “authenticity” in their accent and that they are otherwise being “inauthentic” or “pretentious” in their accented English speech.

4.2.3. Off-Topic

In this section, we address a brief discussion of “off-topic” comments that do not directly point to idols’ English accents but are closely associated with their linguistic performance and/or English language, such as English proficiency, because those comments, to some extent, illustrate the linguistic ideologies we mentioned in the previous analysis as well.

(7)	Korean original:	한국은 이상한게 영어 잘 하면 무슨 엄청난 인텔리인 걸로 착각하는데 이런 풍조 좀 없어졌으면 좋겠다. 미국에서 오래 살았으니 영어 잘 하는거야 당연한거고, 미국 길바닥 거렁뱅이도 영어 잘 함.
	English translation:	It’s strange that Korea mistakes for great intelligence when it comes to good English, but I hope this trend will disappear. Since he/she lived in the U.S. for a long time, he/she is good at English, of course. It’s natural, and homeless people in the US are also good at English.

In Example (7), the commenter begins by problematizing the common association between intelligence and English proficiency in Korea (i.e., “It’s strange that Korea mistakes for great intelligence when it comes to good English”). In addition, the commenter regards the acquisition of good English as an expected, natural process if the speaker resides in an English-speaking country for a considerable amount of time (i.e., “since he/she lived in the U.S. for a long time, he/she is good at English, of course”). In other words, acquiring good English is seen as a natural consequence of living in a country where English is predominantly spoken, regardless of one’s intelligence, class, or living circumstances, as indicated in the commenter’s following remark (i.e., “it’s natural, and homeless people in the U.S. are also good at English”). This, in turn, points to a misconception that English proficiency is solely linked to one’s immediate linguistic environment, where the speaker can acquire the language as a “natural” consequence.

(8)	Korean original:	둘 다 영어 잘해서 개멋있다. 영어 잘하는 사람 ㅈㄴ 멋있어
	English translation:	Both of them are so cool because their English is so good, People who are good at English is fucking cool.

On the contrary, Example (8) illustrates a stark contrast to Example (7) by linking one’s English proficiency with favorable traits, such as being perceived as “cool”. While it is evident that most of the idols who appeared in our selected videos acquired their English language skills naturally rather than through formal education, the commenter nonetheless perceives their high English proficiency in a positive light. This stance is not limited to individual admiration but extends to a broader, generalized perspective, suggesting that individuals with high levels of English proficiency are often viewed as possessing positive attributes, as expressed in Example (8) “fucking cool”.

5. Discussion and Conclusions

Our analysis showed that K-pop idols who performed inner circle English accents are predominantly associated with positive evaluations in our dataset (88.1%, n = 357), which aligns with findings in previous studies (e.g., Y. Kang, 2012; O. Kang & Yaw, 2021). Specifically, the idols’ accents are positively evaluated in connection to two dominant aspects—social attractiveness and nativeness. For social attractiveness, idols’ accents are largely associated with physical appearance (e.g., handsome), personal traits (e.g., sexy, elegant, and kind), and overall linguistic performance. As shown in Example (2), the commenter highlighted English pronunciation as an evaluative aspect of the idols’ linguistic performance despite being aware that accents are irrelevant to the idol’s English proficiency. However, accents are subject to positive evaluation and contribute to the idols’ overall social attractiveness. Thus, our investigation indicates that there continues to be a widespread favorable view of English inner circle accents in digital space. In particular, in the context of the K-pop industry, the important influence of accents, although unrelated to language proficiency, in evaluating one’s English proficiency, and by extension, the person as a whole, suggests that this linguistic feature can greatly impact the idols’ appeal and consequently their overall career success.

The second dominant aspect (i.e., (non-)nativeness) (30.3%, n = 123) shows that idols’ native-like accents are predominantly evaluated in a positive way, despite a few cases in which the idols’ nativeness is subject to neutral and negative comments. Interestingly, many commenters showed their metalinguistic awareness toward the performative nature of accents, which is irrelevant to the overall proficiency, while at the same time showing a favorable stance toward idols’ native-like accents. The notion of nativeness is still noticeable when it comes to the evaluation of one’s language proficiency. Considering the impact of K-pop culture on English language acquisition, as highlighted by Faiza (2020), it is important to recognize that the widespread preference for nativeness can profoundly shape the language attitudes of K-pop fans worldwide as these fans may develop an unrealistic aspiration to master inner circle accents. Our analysis also touches upon “off-topic” comments that do not directly pinpoint accents and/or pronunciation but are closely associated with other linguistic features. In those instances, commenters showed contradictory attitudes toward the capability of speaking fluent English. While some viewers perceive speaking fluent English as a “natural” consequence for those who live in an English-speaking country, others consider it as symbolizing desirable characteristics.

With all that being said, our study is limited in a few aspects. First, the focal video genre appears to emphasize K-pop idols who perform inner circle English accents as only one K-pop idol (i.e., RM) was neither a native English speaker nor had lived abroad. The overwhelming attention given to—and the positive evaluation of—K-pop idols with inner circle English accents by the content creators of the focal video genre, to some extent, leads commenters to focus on inner circle accents due to the underrepresentation of non-inner circle English accents in the videos sampled. With this in mind, we acknowledge the limitation of our study in that non-inner circle accents performed by K-pop idols’ are rarely subject to online users’ metalinguistic commentary in our dataset. Thus, future studies with a focus on online users’ evaluation of non-inner circle English accents would shed new insights into the focal phenomenon with less “skewness” toward nativeness in data analysis. In addition, there are likely other factors that are associated with online users’ metalinguistic comments on the K-pop idols’ accents, such as online users’ linguistic backgrounds (e.g., expanding circles vs. inner circles), as well as the idols’ gender as there may be different social expectations toward female and male idols in terms of linguistic performance (e.g., sounds cute vs. sounds handsome). However, in this study, we focus on linguistic ideologies reflected in the metalinguistic comments written by online users, regardless of their demographic background or the aforementioned extralinguistic aspects. Future studies with access to demographic information of online users will contribute to deeper understandings of linguistic ideologies associated with accented speech.

In conclusion, our analysis suggests that inner circle accents receive favorable evaluations and are often linked to other positive characteristics of the speakers by online users. This phenomenon illustrates a wider social preference that favors certain accents over others, often associating them with higher intelligence, credibility, and social status (e.g., Fuertes et al., 2012; Labov, 2006; Rubin, 1992). However, it is important to note that a few users demonstrated metalinguistic awareness, challenging the prevailing stereotypes, as seen in Example (7). By focusing on online users’ metalinguistic commentary toward idols’ English accents on YouTube, our study broadens the scope of studies concerning L2 speech perception. For future studies, we call for research looking into different digital media platforms, video genres, and/or online communities to see whether online users’ perception toward inner circle accents and/or accented speech may vary across different social groups and media contexts.