Sociolinguists have studied language attitudes and ideologies for decades. Attitudes are the socially shared beliefs of groups, and the basic beliefs underlying the more specific social attitudes of groups are ideologies (van Dijk 2009). A number of techniques have been adopted for the study of language attitudes, such as the matched-guise technique, direct elicitation from conceptual labels of language varieties, and the drawing of maps of linguistic variation and social meaning (see Ryan et al. 1988; Preston 1993). Ideologies, on the other hand, are examined through discursive and ethnographic analyses. As the concepts of “new media” and “big data” purport to identify a new era in which we are witnessing unprecedented change (Reyes 2014), studies in computer-mediated communication have also developed new ways to measure language attitudes by looking at user-generated, digital expressions of language attitudes in online media. Research in sociolinguistics, dialectology, and corpus linguistics has increasingly been using advanced quantitative methods to analyze larger and more complex online data sets to understand patterns of language use across regions, social groups, and communicative situations (e.g., Brezina et al. 2015). Meanwhile, research in computational linguistics has increasingly been concerned with integrating social information into natural language processing systems.
Following this trajectory, this study adopts text-mining techniques to analyze text data from a Chinese online forum regarding gangtaiqiang (Hong Kong-Taiwan accent), a socially recognizable form of mediatized Taiwanese Mandarin. The so-called gangtaiqiang has become a stereotype for many Chinese mainlanders who have little real-life interaction with Taiwanese people. Gangtaiqiang should, therefore, be conceptualized as a mediatized variety of Mandarin, rather than the actual speech of people in Hong Kong or Taiwan (Peng 2018). This article is the second half of a larger study on the social perceptions of gangtaiqiang. The first half employed survey data to conduct a qualitative analysis of how gangtaiqiang is represented in media and is perceived by Chinese mainlanders (Peng 2018). The present study compares the survey data with lexical network data to further unpack the social perceptions of mediatized Taiwanese Mandarin. To this end, this study aims to answer the following research questions:
- (1)What types of mediated masculinities are evaluated more positively and why?
- (2)How are tensions between traditional and newer mediated representations of masculinity received by Chinese audiences?
Central to these research questions is the complex relationship between language and masculinity. The majority of masculinity research focuses on how white, heterosexual, middle-class men use language to maintain their privilege and power (Connell 1995; Lawson 2020), with increasing interaction with queer (Baker and Balirano 2018) and nonhegemonic masculinities (Milani 2013; Brewer 2018). The ongoing discussion of masculinity has inspired linguistic inquiries from various perspectives. Sociolinguists explore the social practices and linguistic features associated with manliness or masculinity (e.g., Eckert and McConnell-Ginet 1992, 1999; Kiesling 2004, 2007, 2018; Lawson 2011, 2013). Discourse analysts examine how masculinity is constructed through discourse or media representation in various social contexts (e.g., Seidler 2006; Ehrlich and Levesque 2011; Milani 2013, 2015; Baker and Levon 2016). However, less work has been done in the context of East Asia. This study builds on Baker and Levon’s (2016) work on the racialized and classed “new men” image in the UK print media. The so-called “new man” is constructed as “a more sensitive, caring and anti-sexist type of man who worries about his own physical appearance and is happy to do his share of domestic labor” (Baker and Levon 2016, see also Chapman 1988; Rutherford 1988; Gough and Brendan 2001; Beynon 2002). The “new men” image is well in line with the concept of “metrosexuality,” a term introduced by British columnist Mark Simpson in 1994 to describe urban males who are meticulous about their grooming and appearance (Hall and Gough 2011; Hall 2015).
In the context of East Asia, similar concepts of metrosexuality have been constructed through televised media and then consumed transculturally (e.g., Iwabuchi 2002; Jung 2010; Yueh 2017). In Korea, for instance, popular culture is reshaping the norms of masculinity as mass media feature images of well-groomed and makeup-clad men (Jung 2010). In mainland China, new types of masculinity receive polarized opinions. Some call them girly and effeminate, while others view them as an expression of discontent against traditional ideas of manhood (see also SturtzSreetharan 2017). In Taiwan, new forms of masculinity go beyond a mere youthful countercultural fad and are increasingly recognized as nonconformist and cosmopolitan icons (see also Moskowitz 2009; Yueh 2017). Departing from the mediated representation of gangtaiqiang, we examine how the notions of “masculine” and “masculine language” are redefined and renegotiated in the Sinophone world. With these new types of masculinity rapidly emerging and widely consumed across East Asia, this study aims to engage with research on both Western and Chinese masculinities to further problematize the mismatch between media’s representation of gender and its perception by Chinese audiences.
2 Media representations of gender ideologies in East Asia
The mass media plays a major role in communicating and transmitting social values (Lippi-Green 1997: 133; Queen 2012). Similarly, the social perceptions of a language and its speakers are often subject to their portrayal in the media. Therefore, getting a sense of how Taiwanese people are represented in the media is pivotal to understanding the formation of Chinese mainlanders’ ideologies regarding gangtaiqiang. The majority of literature on mass media looks at nonfictional, less scripted genres as the venue of investigation, because language use is often considered more spontaneous and “authentic” in those genres than in narrative media (Queen 2015: 160). It might be true that characters only exist in a narrative or fictional context, but the embodiment of the character is real (Queen 2015: 160). The characters’ language patterns, costumes, gestures, and facial expressions are linked to the broader language ecology of human communities, reflecting the broader demographics of the community the character inhabits. Such linkage is known as characterization.
The goal of characterization is to achieve individual distinctiveness. It generates stereotypical but unique ideas about the character and the group of people that the character is representing. Language becomes part of characterization through its connection to the social attributes that we might broadly consider as the character’s “identity” (Queen 2015: 155). Similar to the concept of characterization is the notion of iconicity. Irvine and Gal (2000) describe iconization as “the socio-symbolic appropriation of a linguistic form not merely as an index or marker of a group, but as an iconic representation of the intrinsic qualities of that group” (Irvine and Gal 2000: 35–84). For instance, women are associated with certain language forms or behavior patterns, and these forms and behavior patterns represent not only the identity of women but also the social qualities thereof. Whether the character in narrative media truthfully represents the “real identity” is central to the issue of authenticity (Bucholtz and Hall 2003; Coupland 2003, 2007; Queen 2015). In other words, authenticity is concerned with the realness of the language used in narrative media. In what follows, we discuss different types of narrative media across Asia-Pacific and then narrow my focus to how images of Taiwanese people are represented and perceived in narrative media.
Many recent studies of language and gender in the Asian-Pacific context (e.g., Lin and Tong 2008; Hiramoto and Teo 2014; Chen and Kang 2015) revolve around Confucian traditions, according to which women are expected to conform and be selflessly supportive of their men. Many of these traditional and stereotypical features are realized in the female characters of Kung-Fu movies (Hiramoto and Teo 2014). Many female Kung-Fu practitioners – also referred to as nüxia in Chinese – are portrayed in a way that conforms to Confucian values. They usually play a supporting role to the male characters even if they are no less capable of fighting or protecting themselves. On the opposite end of the spectrum from traditional Confucian values lies the category of “Kong girl,” a type of materialistic girl who abandons traditional values and embraces capitalism and materialism with open arms (Chen and Kang 2015). The “Kong girl” identity has been widely discussed online, because it disrupts Confucianism’s social norms for East Asian women. The Kong girl stereotype appears to have been sparked by an angry Hong Kong girl’s online post in 2005 about her boyfriend not paying for snacks on a date. The discussion later went viral, and the girl’s sense of entitlement together with her ostentatious and demanding persona became the demeanor associated with the so-called “Kong girl” stereotype.
The Korean Wave (hallyu) has taken East Asia by storm. Its dramas present an image of modern Asian women striving to balance between traditional virtues and cosmopolitan living despite the hardship and adversity they encounter in life (e.g., Lin and Tong 2008, 2014). Although often portrayed as tender and humble, the female protagonists in these dramas reject the blind acceptance of Confucianism’s traditional gender roles. Instead, they are empowered to adapt qualities of modern Western femininity such as “strong,” “independent,” and “tough” (Lin and Tong 2014). Korean dramas represent women on the traditional side of the spectrum who also aspire to acquire a form of Western modernity. In a similar vein, Taiwanese idol dramas provide their mainland audience with a window into the image of Taiwanese people and further generate ideologies about what Taiwanese people should be like. These dramas generally feature a modern urban romance of college students or young professionals. They often depict women’s aspirations to have it all – to achieve both career and family success – and also feature a meek and caring “male friend” character that the female protagonist can fall back on after heartbreak. Known as nuannan “warm man” in Mandarin, this type of “good man” is often feminized and seldom expresses his masculinity (Fiske 2011, cited in Lin and Tong 2008). This “good man” stereotype typifies media representations of Taiwanese guys for many mainland audiences. As traditional East Asian women embrace Western modernity with open arms, men are also welcoming different types of masculinity (e.g., Jung 2010; Asher 2018; Gao 2019; Rapp 2020).
3 New models of masculinity: a transcultural wave
On September 5, 2018, BBC news posted an article titled “Flowerboys and the appeal of ‘soft masculinity’ in South Korea” (Asher 2018). The article engages with a new image of soft masculinity – also known as “flowerboy” – represented in Korean Media. In Korea, images of men wearing makeup are permeating mass media and are widely consumed via the influence of Korean pop. The pretty-boy image is glorified in television commercials, dramas, and billboards and has replaced the traditional images of tough and macho southern Korean men. However, as Jung (2010) argues, it is not feminine; the phenomenon should rather be conceptualized as hybrid or versatile masculinity – soft yet manly at the same time – which is different from effeminized (see Rapp 2020). Similar notions include inclusive masculinities (Anderson 2009; Dashper 2012) and caring masculinities (Hanlon 2012; Elliott 2016; Hunter et al. 2017) in which normative forms of masculinity are replaced with nontraditional characteristics, such as interdependence, rationality, and empathy (Lawson 2020). The emergence of versatile masculinity broadened the traditional spectrum of masculinity by reshaping what is considered “masculine” in the context of East Asian cultural norms.
Due to mass media and the proximity of shared Confucian values, the Korean trend soon turned into an inter-Asian transcultural flow (Chua 2004; Jung 2010). Middle-aged Japanese women started to consume this newly constructed form of Korean masculinity with the premiere of Korean romance drama Winter Sonata in early 2002. The male lead of Winter Sonata played by Bae Yong-Joon soon became a symbol of metrosexuality characterized by his tender charisma, purity, and politeness (Jung 2010). Situated in Korea’s postcolonial context, the character embodied the long-lost traditional Japanese values that many middle-aged Japanese women longed for. These images of Bae Yong-Joon upended negative Japanese perceptions of South Korean men (Jung 2010).
Versatile masculinity also hit Taiwan, pushing the already androgynous Taiwanese pop culture even further from traditional gender norms. The Taiwanese ideal of manhood was predicated primarily on familial responsibilities and relationships (Wong and Yau 2016). The meek male identity – perhaps androgynous by Western standards – was often portrayed in Taiwanese Mandopop (Mandarin popular music) and posed a stark contrast to the hypermasculine ethos of Beijing Rock (Moskowitz 2009). With the new trend of hybrid masculinity, “flowerboy” images started to permeate idol dramas in order to seize the attention of the younger generation. These idol dramas present a wide spectrum of new masculinities to audiences in both Taiwan and mainland China, from the “new men” to “flowerboys.” These new forms of masculinity became increasingly recognized as youthful icons representing a nonconformist and cosmopolitan identity (see also Moskowitz 2009; Yueh 2017; Peng 2018). On the mainland, the government’s condemnation of this “pathological culture” backfired, instigating fierce debates. While some claimed that it was “eroding social order,” others expressed their frustration with the more rigid traditional conception of manhood (Gao 2019).
4 Gendered representation of gangtaiqiang
Agha (2011) suggests that mediatized experiences are preceded by nonmediatized ones. Mediatized representations of cultural practice presuppose shared understandings of sociological types or memberships. Media plays the role of disseminating and reinforcing the reflexive links between communication and commoditization. Such experiences, or semiotic encounters, are subsequently invoked in real-life. In other words, mediatization reinforces the semiotic links between social and communicative signs by drawing upon prior off-line social experiences, and these links are retrieved in subsequent real-life encounters. The media experience will subsequently be invoked or extrapolated to become stereotypes of real-life Taiwanese Mandarin speakers.
Taiwanese Mandarin has been overrepresented in Chinese-speaking television media since China’s economic reform in the 1970s. While the content of mass media often addressed ideological themes such as patriotism, socialism, and modernization (Gold 1993), songs and television programs from Hong Kong and Taiwan seized the attention of the younger generation by expressing innocuous feelings about youthful dreams, personal identity, romance, unrequited or forbidden love, and other seemingly “trivial” things in life. For many young Chinese mainlanders, Hong Kong and Taiwan represented a prosperous and modern cosmopolitan lifestyle and a new urban identity to which they aspired (Zhang 2005, 2017). Gangtaiqiang became a semiotic symbol of cosmopolitanism. However, the past 30 years have witnessed the rise of China’s political and economic power as well as changing power dynamics between China and Taiwan. This has been accompanied by linguistic reversals. As discussed earlier, gangtaiqiang has come to be perceived as effeminate and effete as the younger generation on the mainland gravitates toward their homegrown variety of Mandarin as the new model of masculinity. For many Chinese mainlanders, features of Taiwanese Mandarin, such as sentence-final particles and the replacement of neutral tones by full tones, have become associated with femininity, whereas features of Putonghua (standard mainland Mandarin) remain unmarked (Peng 2018, 2020).
This attitudinal shift can perhaps be best explained by how gangtaiqiang is represented in the media and in particular by how it is styled and stylized in response to the recent surge of previously mentioned “versatile masculinity” (Jung 2010). Many male actors use this fashion and speech style to appeal to mass female audiences, leading Taiwanese Mandarin to its iconic girly stereotype in the eyes of many Chinese mainlanders (Peng 2018). Similar forms of Korean versatile masculinity were also constructed and consumed through Taiwanese media (e.g., Iwabuchi 2002; Chua 2004; Jung 2010; Yueh 2017). Taiwanese pop culture was inevitably hit by the transcultural flow of hybrid masculinity. “Flowerboy” and meek men images permeate idol dramas, seizing the attention of the younger generation. These idol dramas present a more fluid form of masculinity to audiences in both Taiwan and mainland China, assigning new meanings to Taiwanese Mandarin and its speakers. Behind these idol dramas are the rise of the middle class and a strong consumerist culture where men’s increasingly image-conscious practices have begun to challenge traditional notions of masculinity and how people construct their sexual identity. Scholarly interests have also been drawn to atypical gendered spaces and activities in relation to a normative hegemonic masculinity (e.g., Hall 2015), including gendered linguistic practices.
5 Chinese masculinity
The concept of Chinese masculinity is constructed in relation to Western hegemonic masculinity, which is often embodied in the appearance of physical strength, independence, nonconformity, and a strict heterosexual identity (Carrigan et al. 1985; Kiesling 2004). Hegemonic masculinity is also found to be correlated with aloofness and messiness, while friendliness and neatness are often associated with effeminacy (Campbell-Kibler 2011). Social attributes such as soft, gentle, and caring can come across as weak, dependent, and effeminate when displayed by men in American culture. Expressions of love and affection are usually associated with femininity. Linguistically, features such as a creaky voice (Yuasa 2010), high rising terminals, and the discourse marker “like” are semiotically linked with femininity in American culture, as these features are commonly used by young women. Features such as /s/-fronting are also perceived as effeminate or gay sounding (Campbell-Kibler 2011; cf. Levon 2006).
Meanwhile in China, even though men traditionally had dominance over women, the domination was not predicated on physical strength and power but rather on social prestige. Media representations of Chinese masculinities are mostly characterized by “filiality, brotherhood, and loyalty” (Kam 2002, 2014) but are still relatively removed from the transcultural flow of versatile masculinity. While there has always been an official push for traditional masculinity, cosmopolitan desires in mass media have emerged as a competing force against more traditional sentiments. The spectrum of Chinese masculinity is expanding from the image of wartime heroes to the emergence of metrosexual urbanites to negotiate new meanings of gender. Xiaoxianrou “little fresh meat,” an apparently sexualized concept, became an Internet buzz word in China to describe a young male rising star known for more effeminate or androgynous looks (see Figure 1). Images of androgynously clothed and makeup-clad men sparked fierce debate on the Internet. Many slammed such portrayals of men as “sissy pants” or “pathological,” rejecting this new type of cosmopolitan masculinity, while some advocated for a more fluid form of masculinity. As a journalist posted online, “the ridiculous condemnation of ‘sissy pants’ men shows the gender ideology of a patriarchal society that equates toughness with men and fragility with women” (Gao 2019).
Similar to the three waves in sociolinguistics studies, the study of masculinities also progresses in three waves: first from the sex-role paradigm (foil for women/women as a foil) to a focus on hegemonic masculinity (interrogating male power) and then to a more identity-driven exploration of normative manhood (Song and Hird 2013). While the same themes permeate throughout, the focus has shifted from social roles to power structures to personal identity. And thanks to globalized mass media, the third wave has brought a surge of scholarly interest in analyzing non-Western societies from a cross-cultural perspective. Meanwhile, Chinese masculinities have also become more fluid and hybrid, evolving from historically hegemonic models to the selfless and asexual Maoists and then to the post-Mao heteronormative practices permeating current popular media (Song and Hird 2013). The shifting paradigms of masculinity are not only represented in media but also reflected in attitudes toward gendered speech acts such as the burikko style1 (Miller 2004) and the metrosexual males.
All these gendered speech styles are performative and create personae by locating meanings in social practice (Eckert 2019). How meanings are made and disseminated, therefore, becomes central to the discussion of style. The attitudinal shift toward gangtaiqiang pertains to how stylistic innovation is mediatized online by the mass communicator (e.g., television celebrities) and to how meanings are made off-line by the audience. When the new trend of metrosexuality makes its presence via stylistic innovation in the Chinese-speaking media, China is being pushed to renegotiate the traditional notions of masculinity. The mismatch between different models of masculinity thus results in the realignment of attitudes toward gangtaiqiang. The realignment of attitudes toward gangtaiqiang among Chinese mainlanders can therefore be conceptualized as a reconciliation between the changing models of cosmopolitanism and masculinity.
The methodology of this study was inspired by Baker and Levon’s (2015) triangulatory approach to the representations of different types of men in the British press. Their study compared the results of the corpus-based approach and of Critical Discourse Analysis, a type of discourse analytical research that primarily studies social power abuse, dominance, and inequality (van Dijk 2001). The triangulation approach to the research question cross-checks for higher reliability and validity. This approach is particularly useful for discourse analysis to avoid the problem of intentionally selecting the data to prove a preconceived point. This study is part of a larger project that explores the social perceptions of gangtaiqiang with qualitative surveys and text-mining techniques as a way of triangulation. The qualitative study was previously reported in Peng (2018). This study focuses on the text-mining approach with larger data sets to ensure more reliable findings to indicate a commonly realized discourse. The data are analyzed through network analysis, which is useful mainly for exploratory analysis, data visualization, and grounded theorizing (Glaser and Strauss 1967).
Network analysis has gained scholarly attention among social scientists (e.g., Borgatti et al. 2009; Palla et al. 2007; Evans and Aceves 2016; Kawa et al. 2018; Brezina et al. 2020). It is a family of methods that can be used to visualize and measure the intertwined relationships between textual or nontextual variables. For example, methods have been used to map the evolution of political rhetoric in the State of the Union discourse (Rule et al. 2015) as well as the connections between faculty placements and where they received their PhD (Kawa et al. 2018). Brezina et al. (2020) developed #LancBox, a software package that serves to analyze and visualize language data and corpora. Modeled on Rule et al.’s (2015) study, we adopt network analysis to visualize and unpack the statistical connections between qualifiers associated with gangtaiqiang, which we select from Peng’s (2018) study. This study takes a co-occurrence approach to identify relevant and interpretable units of meaning in the online corpora (i.e., the scraped comments). In other words, the relationship of two words is constructed based on how frequent the two words appear in the same comments.
The text-mining approach draws upon online user-generated data to explore Chinese mainlander’s perceptions of gangtaiqiang. A total of 13,062 entries and comments with the keyword gangtaiqiang were scraped using Python from the Tianya Forum, one of the most popular discussion forums in China. The comments were used as a corpus with each comment stored as a “bag of words.” Bag-of-words representations ignore word order and only indicate whether a given word is present in the document. “Stop words” such as prepositions, conjunctions, sentence-final particles, etc. were removed as they are not relevant to our research questions.
The comments were stored as rows in a table. Each column represents a single word in the vocabulary of the corpus. For each comment, the “value” of each word is represented by the number of times it appears in the comment divided by the number of documents that include the word (a.k.a. term-frequency, inverse document frequency, or TF-IDF; e.g., Rule et al. 2015; Evans and Aceves 2016; Kawa et al. 2018). This assigned higher weights to the rarer and potentially more contextually relevant words that appear in fewer comments. To uncover more specific attitudes rather than filler words, we included only the top 1,000 words that appeared in less than 1% of comments. The cutoff point of 1% was selected for ease of interpretability after experimenting with a range of different thresholds. Cutoff points between 3% and ½% produce qualitatively similar findings but with less specificity or more noise, respectively. To visualize the context in which these qualifiers are being used, a “lexical network” is constructed with links between words and phrases that co-occur in the data set. Clustering this network of words serves to unravel distinct themes or conceptual categories that are used to discuss gangtaiqiang.
The network was first constructed without focusing on any particular constructs of interest (i.e., unseeded). However, these initial results demonstrated only the most “central” words used by commenters without illustrating distinct themes or clusters as can be seen later in Figure 3. Most of the selected phrases were not relevant to Taiwanese Mandarin. Next, the visualization was refined by seeding the model with 11 key qualifiers from Peng (2018): polite (有礼貌), soft (软), cutesy (萌), nonstandard (不标准), uncomfortable (彆扭), unpleasant (不好听), gentle (温柔), pretentious (做作), alright (还行), babyish (嗲), and pleasant (好听). The network was constructed using comments containing at least one of these keywords to investigate the contexts, associations, and emergent themes among the keywords.
Finally, additional 15 contextually relevant qualifiers were added to include a wider spectrum of qualifiers that might have been overlooked in Peng (2018). These qualifiers are cosmopolitan (都会), polite (礼貌), young (年轻), well-mannered (客气), cultured (文化), amiable (亲切), cute (可爱), refined (气质), well-behaved (有教养), stylish (时髦), fashionable (时尚), Kangxi (康熙),2 idol drama (偶像剧), and Taiwanese drama (台剧). These qualifiers are what the author – as a native speaker of Taiwanese Mandarin – perceived to be the most common ideologies Chinese mainlanders hold about mainland and Taiwanese Mandarin. The links between words were calculated based on their frequency of co-occurrence. The more comments they co-occur in, the stronger the link is. After trial and error, the top 150 links were selected for interpretability of the network visualization and to prevent the graph from turning into a “hairball” (Nocaj et al. 2014), which is not helpful for interpretation.
For the network visualization, different clusters in the network were color coded to indicate the distinct “themes” of the online discourse. In the visualization, distance between clusters bear no significance, especially for unconnected components. The visualization is created using a standard physics-based engine, where each link acts as a “spring” pulling connected nodes together, and each node naturally repels other nodes (to space them out). The strength of each spring does correspond to the strength of correlation between words (so the thicker links tend to be shorter, i.e., pulling nodes closer together).
A few caveats must be included concerning the limitations of this approach. It is difficult if not impossible to collect commenters’ demographic information. As Tianya is a China-based discussion forum and all the discussions are in simplified characters, it is assumed that its general user population consists of Chinese speakers from the mainland. However, more detailed information as to their region of origin, age, gender, etc. remains unknown. Another drawback is concerned with the bag-of-words model. Since discourse analysis unpacks the texts’ cultural meaning and power relationships, text analysis should be a human undertaking (van Dijk 1980). The bag-of-words technique treats the words as independent of the original word order and dislocates content from its context. It is helpful for document classification, frequency counts, or word clustering as it reduces complexity. However, if the context in which the words are used is an important component of the text, removing words from this context eliminates important components of the data (Sinclair 1991; Schuelke-Leech and Barry 2017). This study adopts this approach as a triangulation to complement and validate the previous qualitative study (i.e., Peng 2018) with a much larger data size. This method also overcomes the observer’s paradox as these data were generated spontaneously without supervision.
7 Results and discussion
The results of the lexical network analysis are presented in the following order: (1) raw data, (2) seeded data, and (3) more contextually relevant data. Figure 2 shows a network without any preselected key words. The model was fed with unseeded (raw) data. While indicative of the conversations in the source data, the initial result was uninterpretable as most of the words appear to be irrelevant to gangtaiqiang (see Figure 2).
Words that co-occur often are linked together in the graph. The words that show up in this network are those with the strongest links (i.e., are highly correlated) in the raw data. The colors were applied using “spectral clustering,” a network clustering method that identifies densely connected components. Each color corresponds to a densely connected community in the graph, which can be representative of discursive topics or themes. Some of them are more clearly disconnected from the rest of the graph (e.g., the yellow section on the bottom in Figures 2 and 3), while some seem to be overlapping (blue and red). Each coloring corresponds to a distinct theme.
In order to refine the results, the model was fed 11 key qualifiers identified in Peng’s (2018) study. Figure 3 shows the network after filtering out comments that do not contain one of the 11 key qualifiers. The qualifier “pretentious” appears to be in the center of one of the main thematic clusters in the network visualization. Comments using the word “pretentious” were also very likely to include words such as “celebrity,” “intentional,” “babyish,” and names of celebrities. Moving further, there are “accurate” “feigned,” “wrong,” “special,” and “unnatural.” “Celebrity” implies that the perceived pretentiousness is associated with television celebrities. This cluster reveals how television celebrities are perceived when affecting a Taiwanese accent. The speech styles of television show hosts and television celebrities contribute greatly to their gangtaiqiang stereotypes as Chinese mainlanders are exposed to Taiwanese Mandarin mostly through televised media. Other perceptual themes of gangtaiqiang include “soft” (the yellow cluster close to the bottom) and “polite” (the pink cluster at the upper right corner). The network analysis reveals not only the social perceptions of Taiwanese Mandarin speakers but also those of mainlanders affecting a Taiwanese accent. Such perceptions are formulated largely via mediation instead of face-to-face interaction.
To make the emergent themes more distinct and understandable in the visualization, more contextually relevant descriptors were added to expand the explanatory power of the network. This resulted in the visualization shown in Figure 4. Key qualifiers from Peng (2018) are in red. Contextually relevant qualifiers are in green, and the words from the data that emerged as highly correlated with the key words are in beige. The emergence of linguistic features in Figure 4 visualizes the assignments of social meanings to Taiwanese Mandarin features. For example, “cute” is semiotically linked with you-mei-you3 and sentence-final particles. The lexical network pinpoints the features of Taiwanese Mandarin that invoke social meanings for Chinese mainlanders. In Figure 4, the cluster on the far left centers on “pretentious” (colored in red) with links to “celebrities,” “songs,” and “intentional.” “Pretentious” is further connected to two names of mainland celebrities. Building on the results of Figure 3, the model in Figure 4 extends the connections from the people who use gangtaiqiang to the linguistic features associated with it. By feeding the model with different types of data, this method uncovers dimensions of gangtaiqiang that the survey data fail to show.
Figure 4 shows that “pretentious” and “cute” are central to the findings. Essentially “pretentious” and “cute” are the perceptions of the same speech style. Inspired by the Japanese burikko performance, this type of youthful cuteness has been overrepresented by many celebrities stylizing with gangtaiqiang in the Mandarin-speaking media. The style is even considered emasculated when used by a man. Although nontraditional gender performances are being renegotiated across East Asia, whether versatile masculinity or youthful cuteness, such perceptions of gangtaiqiang in a way embodies persistent patriarchal norms where traditional gender ideologies are fortified.
In the past decade, gangtaiqiang has been styled and stylized in media in response to new waves of masculinity from across East Asia. The gendered representation of gangtaiqiang in media has instigated tensions and divergent opinions among audiences on the mainland, as Chinese masculinities are traditionally characterized by “filiality, brotherhood, and loyalty” (Kam 2002). For the younger generations, these androgynous styles are indicative of a young, nonconformist identity emulated by many millennials who are propelling forward more fluid and hybrid forms of Chinese masculinities. On the conservative side of the spectrum, many Chinese also perceive gangtaiqiang as soft, effete, and even effeminate, which manifests post-Mao heteronormative practices. The effete and effeminate perceptions of gangtaiqiang reflect Chinese mainlanders’ resistance to Taiwan’s newly acquired model of masculinity, which perhaps also mirrors the shifting power dynamics between the two political entities. In other words, the spread of changing gender norms and power dynamics through mass media is reshaping how language is performed, embodied, and perceived.
As shown in the visualization (Figure 4), “cute” and “soft” are the major perceptual themes of gangtaiqiang. This suggests that gangtaiqiang is increasingly linked to low status (defined as less intelligent and ambitious), as Chinese mainlanders turn to Beijing Mandarin for new models of power and masculinity. However, the fact that it also comes across as “pretentious” implies that it remains a useful stylistic resource for mainland celebrities to indicate a hip metrosexual identity on television. The implied solidarity (defined as socially attractive and benevolent) of gangtaiqiang may have contributed to its pretentious perception when a Taiwanese accent is being affected. The data point to a wave of resistance from the Chinese tradition of masculinity, or post-Mao heteronormative practices. The traditional notions of normative manhood is fortified by China’s ascent in its political and economic power. A large part of the Chinese national identity is also built upon its surge of global power. Gangtaiqiang juxtaposed against Beijing Mandarin seems to underscore the emerging masculine identity of the Chinese.
Such a perception is, of course, socially situated and may vary in different social contexts. The concept of gangtaiqiang, in fact, does not exist in Hong Kong and Taiwan. In a broader sense, the gendered perceptions of gangtaiqiang can perhaps be understood as China’s response to the growing trend of versatile masculinities across East Asia. Many women have turned away from traditional masculinity as a backlash against gender inequality. In a culture where manhood equates to power, perhaps the persistent patriarchal norms stem from men’s underlying fear of losing their privilege and hegemonic power. The acceptance of versatile masculinity might change these power dynamics and eventually decrease the privilege and hegemonic power of men.
Based on the concept of “the mediatization of social life,” stylistic innovation as a meaning-making process is accelerated and intensified through mass media (Agha 2007; Mortensen et al. 2017). As television celebrities style and stylize in Taiwanese Mandarin to create new personae, new models of masculinity disseminated through mass media clash with Chinese mainlanders’ semiotic experience of normative masculinity. The mediatized representations of men and masculinity eventually transform the social perceptions of gangtaiqiang and the speakers thereof. This study also contributes broadly to the evolving gender ideologies in the Sinophone world by disentangling the intertwining relationships between the changing perceptions of gangtaiqiang, new models of masculinity, and shifting power dynamics between China and Taiwan.
This study draws upon text analytics to examine the perceptions of gangtaiqiang in relation to changing notions of masculinity. The text mining techniques provide evidence for Peng’s (2018) qualitative study regarding the social perceptions of gangtaiqiang by Chinese mainlanders. The lexical network analyses illustrate how qualifiers are connected by computing their frequency of co-occurrence. The network analyses reveal attitudes toward not only gangtaiqiang speakers but also Chinese celebrities who try to affect a Taiwanese accent. Our analyses point to an emerging realignment of attitudes among the Chinese toward gangtaiqiang. Chinese mainlanders, who once associated Taiwanese Mandarin with chic, urban television celebrities and young cosmopolitan types (Peng 2018), have increasingly come to associate gangtaiqiang with pretension, gentility, and even emasculation. The shift in attitudes can be, in part, ascribed to the way gangtaiqiang is stylized in televised media in response to the transcultural flow of hybrid masculinity. Building on the changing norms of masculinity in South Korea (Jung 2010), the results were further unpacked to discuss how this new type of soft masculinity is redefining the model of manhood in the Sinophone world. Finally, the study explores how mediatized conceptions of Taiwanese Mandarin (i.e., gangtaiqiang) are shifting to reflect changing gender ideologies and power dynamics between China and Taiwan.
Support for this project was provided by a PSC-CUNY Award (Award # 62502-00 50), jointly funded by The Professional Staff Congress and The City University of New York.
Agha, Asif. 2007. Language and Social Relations. Cambridge: Cambridge University Press.
Agha, Agha. 2011. “Meet mediatization.” Language & communication 31(3): 163–70.
Anderson, Eric. 2009. Inclusive Masculinity: The Changing Nature of Masculinities. Oxon, UK: Routledge.
Asher, Saira. 2018. “Flowerboys and the appeal of ‘soft masculinity’ in South Korea.” BBC News. https://www.bbc.com/news/world-asia-42499809 (accessed April 26, 2020).
Baker, Paul and Erez Levon. 2015. “Picking the right cherries? A comparison of corpus-based and qualitative analyses of news articles about masculinity.” Discourse & Communication 9(2): 221–36.
Baker, Paul and Erez Levon. 2016. “‘That’s what I call a man’: representations of racialized and classed masculinities in the UK print media.” Gender & Language 10(1): 106–39.
Baker, Paul and Balirano Giuseppe (eds). 2018. Queering Masculinities in Language and Culture. Basingstoke, UK: Palgrave Macmillan.
Beynon, John. 2002. Masculinities and culture. Milton Keynes: Open University Press.
Borgatti, Stephen P., Ajay Mehra, Daniel J. Brass and Giuseppe Labianca. 2009. “Network analysis in the social sciences.” Science 323(5916): 892–95.
Brezina, Vaclav, Tony McEnery and Stephen Wattam. 2015. “Collocations in context: a new perspective on collocation networks.” International Journal of Corpus Linguistics 20(2): 139–73.
Brezina, Vaclav, Pierre Weill-Tessier and Anthony McEnery. 2020. #LancsBox v. 5.x. [software]. Available at: http://corpora.lancs.ac.uk/lancsbox.
Brewer, Mick. 2018. “Good ol’ country boys playin’ on the farm: online articulations of rural masculinity by men who have sex with men.” Sexuality & Culture 22(2): 355–79.
Bucholtz, Mary and Kira Hall. 2003. “Language and identity.” In A Companion to Linguistic Anthropology, ed. Alessandro Duranti, 369–96. Oxford: Basil Blackwell.
Campbell-Kibler, Kathryn. 2011. “Intersecting variables and perceived sexual orientation in men.” American Speech 86(1): 52–68.
Carrigan, Tim, Bob Connell and John Lee. 1985. “Toward a new sociology of masculinity.” Theory and Society 14: 551–604.
Chapman, Rowena. 1988. “The great pretender: Variations on the new man theme.” In Male Order: Unwrapping Masculinity, ed. Rowena Chapman and Jonathan Rutherford, 225–48. London: Lawrence & Wishart.
Chen, Katherine H. Y. and M. Agnes Kang. 2015. “Demeanor indexicals, interpretive discourses and the ‘Kong Girl’ stereotype: Constructing gender ideologies in social media.” Journal of Language and Sexuality 4(2): 193–222.
Chua, Beng Huat. 2004. “Conceptualizing an east Asian popular culture.” Inter-Asia Cultural Studies 5: 200–21.
Connell, Raewyn W. 1995. Masculinities. Cambridge: Polity.
Coupland, Nikolas. 2003. “Sociolinguistic authenticities.” Journal of Sociolinguistics 7(3): 417–31.
Coupland, Nikolas. 2007. Style: Language Variation and Identity. Cambridge: Cambridge University Press.
Dashper, Katherine. 2012. “‘Dressage is full of queens!’ Masculinity, sexuality and equestrian sport.” Sociology 46(6): 1109–24.
Eckert, Penelope. 2019. “The limits of meaning: Social indexicality, variation, and the cline of interiority.” Language 95(4): 751–76.
Eckert, Penelope and Sally McConnell-Ginet. 1992. “Think practically and look locally: Language and gender as community-based practice.” Annual Review of Anthropology 21: 461–90.
Eckert, Penelope and Sally McConnell-Ginet. 1999. “New generalizations and explanations in language and gender research.” Language in Society 28: 185–201.
Ehrlich, Susan and Susan Levesque. 2011. “The strategic marginalization of working class masculinity in a batterers’ treatment programme.” Gender & Language 5(2): 271–301.
Elliott, Karla. 2016. “Caring masculinities: theorizing an emerging concept.” Men and Masculinities 19(3): 240–59.
Evans, James A. and Pedro Aceves. 2016. “Machine translation: mining text for social theory.” Annual Review of Sociology 42: 21–50.
Fiske, John. 2011. Understanding Popular Culture. London: Routledge.
Gao, Helen. 2019. “‘Little fresh meat’ and the changing face of masculinity in China.” The New York Times, 12 June 2019. https://www.nytimes.com/2019/06/12/opinion/little-fresh-meat-china.html (accessed April 26, 2020).
Glaser, Barney and Anselm Strauss. 1967. The Discovery of Grounded Theory. London: Weidenfield & Nicolson.
Gold, Thomas B. 1993. “Go with your feelings: Hong Kong and Taiwan popular culture in Greater China.” The China Quarterly 136: 907–25.
Gough, Brendan. 2001. “‘Biting your tongue’: negotiating masculinities in contemporary Britain.” Journal of Gender Studies 10: 169–85.
Hall, Matthew. 2015. Metrosexual Masculinities. UK: Palgrave Macmillan.
Hall, Matthew and Brendan Gough. 2011. “Magazine and reader constructions of ‘metrosexuality’ and masculinity: a membership categorisation analysis.” Journal of Gender Studies 20(1): 67–86.
Hanlon, Niall. 2012. Masculinities, Care and Equality: Identity and Nurture in Men’s Lives. Basingstoke, UK: Palgrave Macmillan.
Hiramoto, Mie and Cherise Teo. 2014. “I am the invincible sword goddess: mediatization of Chinese gender ideology through female kung-fu practitioners in films.” Societies 4 (3): 477–505.
Hunter, Sarah C., Damien W. Riggs and Martha Augoustinos. 2017. “Hegemonic masculinity versus a caring masculinity: implications for understanding primary caregiving fathers.” Social & Personality Psychology Compass 11(3): e12307.
Irvine, Judith T. and Susan Gal. 2000. “Language ideology and linguistic differentiation.” In Regimes of Language: Ideologies, Polities, and Identities, ed. Paul V. Kroskrity, 35–84. Santa Fe, NM: School of American Research Press.
Iwabuchi, Koichi. 2002. Reentering Globalization: Popular Culture and Japanese Transnationalism. Durham and London: Duke University Press.
Jung, Sun. 2010. Korean Masculinities and Transcultural Consumption: Yonsama, Rain, Oldboy, K-Pop Idols. Hong Kong: Hong Kong University Press.
Kam, Louie. 2002. Theorising Chinese Masculinity: Society and Gender in China. Hong Kong: Cambridge University Press.
Kam, Louie. 2014. Chinese Masculinities in a Globalizing World. Hong Kong: Routledge.
Kawa, Nicholas C, Jose Michelangeli, Jessica L. Clark, Daniel Ginsberg and Christopher MaCarty. 2018. “The social network of US academic anthropology and its inequalities.” American Anthropologist 1(121): 14–29.
Kiesling, Scott F. 2004. “Dude.” American Speech 79: 281–305.
Kiesling, Scott F. 2007. “Men, masculinities, and language.” Language and Linguistics Compass 1: 653–73.
Kiesling, Scott F. 2018. “Masculine stances and the linguistics of affect: on masculine ease.” NORMA 13: 191–212.
Lawson Robert. 2011. “Patterns of linguistic variation among Glaswegian adolescent males.” Journal of Sociolinguistics 15: 226–55.
Lawson Robert. 2013. “The construction of ‘tough’ masculinity: negotiation, alignment and rejection.” Gender and Language 7: 369–95.
Lawson, Robert. 2020. “Language and masculinities: history, development, and future.” Annual Review of Linguistics 6: 409–34.
Levon, Erez. 2006. “Hearing ‘gay’: Prosody, interpretation and the affective judgments of men’s speech.” American Speech 81(1): 56–78.
Lin, Angel and Avin Tong. 2008. “Re-imagining a cosmopolitan ‘Asian us’: Korean media flows and imaginaries of Asian modern femininities.” In East Asian Pop Culture: Analysing the Korean Wave, ed. Chua Beng Huat and Koichi Iwabuchi, 91–125. Hong Kong: Hong Kong University Press.
Lippi-Green, Rosina. 1997. English with an Accent. New York: Routledge.
Milani, Tommaso. 2013. “Are ‘queers’ really ‘queer’? Language, identity and same-sex desire in a South African online community.” Discourse & Society 24: 615–33.
Milani, Tommaso. 2015. “Theorizing language and masculinities.” In Language and Masculinities: Performances, Intersections, Dislocations, ed. Tommaso Milani, 8–33. London: Routledge.
Miller, Laura. 2004. “You’re doing burikko! Censoring/scrutinising artificers of cute femininity in Japanese.” In Japanese Language, Gender, and Ideology: Cultural Models and Real People, ed. Shigeko Okamoto and Janet S. Smith, 148–65. New York: Oxford University Press.
Mortensen, Janus, Nikolas Coupland and Jacob Thogersen (eds.). 2017. Style, Mediation and Change: Sociolinguistic Perspectives on Talking Media. Oxford: Oxford University Press.
Moskowitz, Marc L. 2009. Cries of Joy, Songs of Sorrow: Chinese Pop Music and Its Cultural Connotations. Honolulu: University of Hawaii Press.
Nocaj, Arlind, Mark Ortmann and Ulrik Brandes. 2014. “Untangling hairballs.” In Graph Drawing: 22nd International Symposium, GD 2014, Würzburg, Germany, September 24-26, 2014, Revised Selected Papers, ed. Christian Duncan and Antonios Symvonis, 101–12. Berlin/Heidelberg: Springer.
Palla, G., Albert-László Barabási and Tamás Vicsek. 2007. “Quantifying social group evolution.” Nature 446(7136): 664–7.
Peng, Chun-Yi. 2018. “Mediatized Taiwan Mandarin: social perceptions and language ideologies.” Chinese Language and Discourse 9(2): 162–83.
Peng, Chun-Yi. 2020. “Effects of media exposure on regional associations: a case study of Mandarin aspectual you.” Chinese Semiotic Studies 16(3): 329–43.
Preston, Dennis R. 1993. “Folk dialectology.” In American Dialect Research, ed. Daniel R. Preston, 333–78. Philadelphia, PA: John Benjamins.
Queen, Robin. 2012. “Days of our lives: language on a daytime drama.” Gender and Language 6(1): 151–78.
Queen, Robin. 2015. Vox Popular: The Surprising Life of Language in the Media. Oxford: Wiley-Blackwell.
Rapp, Jessica. 2020. “South Korean men lead the world’s male beauty market. Will the West ever follow suit?” CNN Style. https://www.cnn.com/style/article/south-korea-male-beauty-market-chanel/index.html (accessed August 22, 2020).
Reyes, Angela. 2014. “Linguistic anthropology in 2013: Super-New-Big.” American Anthropologist 116(2): 366–78.
Rule, Alix, Jean-Philippe Cointet and Peter S. Bearman. 2015. “Lexical shifts, substantive changes, and continuity in State of the Union discourse, 1790–2014.” Proceedings of the National Academy of Sciences 112(35): 10837–44.
Rutherford, Johnathan. 1988. “Who’s that man?” In Male order: Unwrapping Masculinity, ed. Rowena Chapman and Jonathan Rutherford, 21–67. London: Lawrence & Wishart.
Ryan, Ellen B., Howard Giles and Miles Hewstone. 1988. “The measurement of language attitudes.” In Sociolinguistics: An International Handbook of the Science of Language and Society, vol. II, ed. Ulrich Ammon, Norbert Dittmar and Klaus J. Mattheier, 1068–81. Berlin: de Gruyter.
Schuelke-Leech, Beth-Anne and Betsy Barry. 2017. “Philosophical and methodological foundations of text data analytics.” In Frontiers in Data Science, ed. Matthias Dehmer and Frank Emmert-Streib, 145–68. London: Taylor & Francis Group.
Seidler, V. 2006. Transforming Masculinities: Men, Cultures, Bodies, Power, Sex and Love. London: Routledge.
Sinclair, John. 1991. Corpus, Concordance, Collocation. Oxford: Oxford University Press.
Song, Geng and Derek Hird. 2013. Men and Masculinities in Contemporary China. Leiden, The Netherlands: Brill.
SturtzSreetharan, Cindi L. 2017. “Language and masculinity: the role of Osaka dialect in contemporary ideals of fatherhood.” Gender and Language 11: 552–74.
van Dijk, Teun A. 1980. Macrostructures: An Interdisciplinary Study of Global Structures in Discourse, Interaction, and Cognition. Hillsdale, NJ: Lawrence Erlbaum.
van Dijk, Teun A. 2001. “Critical discourse analysis.” In Handbook of Discourse Analysis, ed. Deborah Tannen, Heidi E. Hamilton and Deborah Schiffrin, 352–71. Oxford: Blackwell.
van Dijk, Teun A. 2009. Society and Discourse: How Social Contexts Influence Text and Talk. Cambridge: Cambridge University Press.
Wong, Heung-wah and Hoi-yan Yau. 2016. “The anthropology of Chinese masculinity in Taiwan and Hong Kong.” In Changing Chinese Masculinities: From Imperial Pillars of State to Global Real Men, ed. Kam Louie, 220–43. Hong Kong: Hong Kong University Press.
Yuasa, Ikuko P. 2010. “Creaky voice: a new feminine voice quality for young urban-oriented upwardly mobile American women?” American Speech 85(3): 315–37.
Yueh, Hsin-I Sydney. 2017. Identity Politics and Popular Culture in Taiwan: A Sajiao Generation. Lanham, MD: Lexington Books.
Zhang, Qing. 2005. “A Chinese yuppie in Beijing.” Language in Society 34: 431–66.
Zhang, Qing. 2017. Language and Social Change in China: Undoing Commonness through Cosmopolitan Mandarin. New York: Routledge.
Burikko is derived from buru and ko to mean “fake child” or “phony girl.” In Japanese, the term buru means “to pose, pretend, or act,” and the suffix ko means “child” or “girl.” Burikko performance is a stylized and gendered vocal style that indexes a type of “youthful cuteness” or, for some people, feigned naiveté, which downplays and masks the adult femininity of the speaker (Miller 2004).
The name of a popular Taiwanese talk show.
You-mei-you is a common pattern for yes–no questions in Mandarin.