Harmony as language policy in China: an Internet perspective

This paper provides an ethnographic understanding of harmony as language policy in China, grounded in a historical analysis of ‘harmony’ (和 he) as a distinct traditional Chinese (Confucian) ideal that gradually finds its new expressions through the policy of Harmonious Society (和谐社会hexie shehui) in contemporary China. The paper focuses on language practices surrounding ‘harmony’ emerging from the Internet, a discursive space and site that is both highly diverse and heavily contested with respect to policing processes, and notably so in the context of the PRC for its stringent measurement of censorship and sensitization of language use. It is shown that although the state is arguably the strongest stakeholder in implementing the policy of harmony, the actual processes of harmonization online develop in detailed, multidirectional and unpredictable rather than abstract, linear or monofocal ways. The paper offers a descriptive analysis of the field of Internet memes that critique and subvert the policies of harmonization. This includes new meanings for words (e.g. ‘harmony’ as euphemism for censorship) and puns around the acoustic image of hexie and other censorable words, resulting in the circulation and bricolage of myths and songs revolving around ‘river crab’ (hexie) and ‘grass mud horse’ (caonima) as placeholders of dissent, which feed back into offline popular (and critical) culture. The paper concludes with a discussion of the politics of harmony, pointing at the increasingly polycentric realities clashing with the modernist monocentric ideal of the state project of harmony, and with arguments for an ethnographically based understanding and inspection of language policy as an instrument for shaping sociolinguistic life.


Introduction
Whereas language policy at the state or institutional level generally aims for the ideal of harmony and social cohesion by emphasizing normativity and order, everyday language practices at the ground level demonstrate far more features of divergence, heterogeneity, and polycentricity. This is exceedingly so in view of the current stage of globalization and its outcome of 'superdiversity' (Blommaert and Rampton 2011). Digital mediascapes, for instance, open up a new, unprecedentedly complex, and less controllable space in which effects of formal policies are accompanied, de-centered, and transformed in a variety of sociolinguistic settings (e.g. Pietikäinen and Piirainen-Marsh 2009; Androutsopoulos and Juffermans 2014). This raises the question of how harmony, or any state-sponsored social project, is discursively negotiated and (co)constructed in and as social reality, as much as how language policy and perhaps society at large may be better conceptualized and understood when grounded in such reality.
The above observations are relevant to the current paper in two ways. On the one hand, they raise broad concerns about language policy research as theory and methodology, and they point us towards calls for a paradigmatic shift in this field (Ricento 2006;Shohamy 2006;McCarty 2011;Davis 2014): from a traditional focus on the formal policy (policy-as-text), often in the strict sense of the term, as a set of official documents, directives, and regulations produced by authorities such as the state, to a Hymesian (1980,1996) ethnographic perspective to policy (policy-asdiscourse) as dynamic, multifaceted, and situated social practices. Following this shift, and drawing on Foucault's notion of police, disciplinary power, and governmentality, Blommaert et al. (2009) suggest that language policy should be seen in terms of 'policing', i.e. processes of rational production and management of a normative structure that involves various sociopolitical actors and institutions with unequally distributed agency. This locates language policy in complexes of ideology and webs of cultural meanings and, as such, in constellations of microdiscursive practices that are anchored in different and often conflictual ideologies, indexical and constitutive of the macro-patterns of normativity and order. In such constellations, the state functions as but one of a range of possible centers of norms.
On the other hand, the questions of harmony and language policy draw attention to China as both a comparative context (to African and other contexts) and an interesting case in its own right. Harmony and language policy go hand in hand in China. Evolving from a well-entrenched classical Confucian ideal, 'harmony' has in recent years become a proper name that stands for an explicit discourse on the rationalization, maintenance, and enforcement of stability and order by the state in reaction to the rapid economic-political changes and sociocultural diversifications resulting from the country's modernization and globalization processes. This can be seen in the prevalent slogan of Harmonious Society championed by former President Hu Jintao. Not only the formal policies of language-which advocates a monoglot standard ideology (cf. Silverstein 1996)-but almost all recent official policies in China, have invariably adopted the state motif of 'harmony'. This has impacts on the way language and communication 'ought to' be and are actually practiced in (at least) the public sphere in China, including in its flourishing online environment. Harmony, therefore, is a crucial aspect and driving force of language policy and policing in the context of China.
This paper seeks an ethnographic understanding of harmony via and, thus, as language policy in China. As stated, we use ethnography in the Hymesian sense as a theoretical perspective rather than a mere method of social and linguistic inquiry. In this we follow Lillis's (2008) discussion of three levels of ethnography, i.e. a first and minimal level of ethnography conceived as talk around texts, a second level as full-fledged methodology comprising multiple data sources and a sustained and interactive engagement in the contexts of production and the communities of practice, and a third and deeper level of ethnography as interpretive theorizing (cf. Blommaert 2005;Rampton 2007). These three levels of ethnography progressively narrow the ontological gap between text and context in literacy research and discourse analysis more generally. Our use of ethnography needs to be appreciated at this third and deeper level, i.e. as an ethnographic-sociolinguistic study of harmony as a complex object of analysis.
In what follows, we begin with a historical analysis of 'harmony' as a distinct traditional Chinese ideal that gradually finds its new expressions through policy in contemporary China. We will then focus on language practices surrounding 'harmony' emerging from the Internet in China, a discursive space and site of policing that is highly diverse while also heavily contested and policed through stringent measurement of censorship and sensitization of communication (e.g. Tsui 2003;Yang 2009;MacKinnon 2011). From this perspective of the Internet, we will show empirically that although the state is arguably the strongest stakeholder in implementing the policy of harmony-or, better, harmonization-the actual processes of harmonization through policing online develop in detailed, multidirectional, and unpredictable rather than abstract, linear, or monofocal ways. The outcomes of such processes are, paradoxically, alternative ideologies of harmony as well as non-normative use of language. We will discuss the implications of these observations and our understanding of harmony as language policy-with reference to China and Chinese-in the final part of the paper.

Harmony as a Confucian ideal
'Harmony' originates from he (和), 1 a word celebrated by Chinese people as one of the core symbols of their cultural essence, alongside words such as 'fortune', 'longevity', and 'luck'. It is important to make clear at the onset that when we speak of harmony in a Chinese context, the use of the term conjures up a distinctive ideological load seated in over two millennia of Chinese history and Confucian traditions.
He is one of the central tenets of the Confucian system of ethical philosophy and political governmentality (cf. Yao 2000). The Confucian doctrines of he are incorporated by generations of Chinese in conceptualizing norms and orders that inform individual behaviors in relation to the moral self, the family, the state, and other levels of society. In this sense, he represents a specific set of historically enregistered and internalized discourses about what is meant by harmony, why harmony is important, and how to achieve it socially and politically. This is a crucial point for understanding the significance of harmony in terms of (language) policy in China today.
Although today he is invariably credited to Confucianism, its genesis predates Confucius (551-479 BCE). Li (2006), for instance, traces its existence back to as far as the earliest dynasties of Shang (sixteenth-eleventh century BCE) and Zhou (1066-256 BCE). He observes that the concept gradually evolved from its initial meaning of describing how different sounds or flavors respond to one another in ancient music and food rituals, to an aesthetical, ethical, philosophical, political, and metaphysical ideal that embraces harmony as the optimal way of constructing society and cosmos (Li 2006). It is believed that Confucius was the first to synthesize earlier thoughts about harmony and placed he at the centre of his philosophy. The previously variegated ideas were quoted and appropriated by Confucius and his followers to promote the social and political significance of harmony. He was held as 'the highest ideal' (Li 2006:588) of what was later to become Confucianism, one of the most influential thoughts and cultural traditions in China (and other Asian-Confucian societies).
What is interesting about harmony, according to Li (2006Li ( , 2008 and others, are the distinctions and dynamics between sameness and difference it defines. In the Confucian classics The Analects, he was a crucial criterion for junzi (the real gentleman)-junzi he er butong, xiaoren tong er buhe ('The junzi harmonizes but does not seek sameness, whereas [an unscrupulous man] seeks sameness but does not harmonize ', Li 2006:586). He er butong (harmony with distinction) is a popular saying people still use today to defend their stance and settle disputes. What is inscribed in these lines is the differentiation between harmony and sameness, between valid harmony based on the acknowledgement of difference versus sameness, and invalid harmony, based on the diminishing of difference; it also states the moral-ethical categorizations of harmony for which the order of good and bad, appropriate and inappropriate, is negotiated and established.
Such dynamics are crucial to the understanding of he. He does not equal tong (sameness), even though sameness is an important ingredient of harmony and must be valued and maintained 'at an appropriate level' (Li 2006:590). Not any kind of sameness leads to harmony. Li contends that the Confucian belief rejects the 'overpresence' of sameness and deems it as being in danger of imposing uniformity and disharmony. Difference, on the other hand, is a precondition and cornerstone of harmony because he is essentially about the harmonious, appropriate interplays of differences. Harmony presupposes the entailment of difference as tension, conflict, and 'strife': the process through which harmony is negotiated and sought. As Li (2006:592) argues, 'harmony is not only a state but, more importantly, a process, disharmony is necessarily present during the process of harmonization'.
Hence he entails hexie (harmony), heping (peace) and hejie (reconciliation), an equilibrium that is only acceptable and appropriate through strife and the harmonization of sameness and difference-managed diversity, so to speak. This includes the management of different roles and the knowledge of 'ought-ness' of behaviours based on the roles one assumes in society-what is called 'rites' or li (礼) in Confucianism. Li refers not only to ceremonial rituals performed on specific occasions, it is one of the five basic virtues (i.e. benevolence, righteousness, propriety, wisdom, and trustworthiness) in the Confucian ethics, and is deemed 'the way of humanity and the way of Heaven' (Li 2006:588), of behaving oneself as well as managing society.
Harmony operates on five hierarchically interrelated scale-levels (Li 2008). On the elementary level is the individual-personal awareness and desire to self-cultivate one's internal harmonization as a moral duty to the keeping of order in society; this is the foundation of a moral society. The second level concerns 'a nexus of human relationship' (Li 2008:429), namely, the five major interpersonal/ethical relationships in which the individual self exists: between ruler and minister, parent and child, husband and wife, or between siblings, and between friends. The next level of harmony is to do with the governance politics of the state. Harmonious governance is to bring about order in society through the virtuous functioning of government officials rather than penal laws. The fourth level involves promoting harmony as peaceful coexistence beyond the state borders, in the world. And finally, at the most fundamental level, harmony is applicable as a universal law and a cosmological order generated by the interactive process of balancing human, nature, and society. The ultimate goal is to realize taihe (grand harmony) throughout the cosmos, which derives from harmony at the lower levels.
Harmony, thus, is a carefully constructed normative complex in Confucianism that relies on the cultivation of prescribed virtues and maintenance of morality and ethics. It provides an early model of humanist ideal of organizing life at multiple levels and achieving desired balance and order. To this end, differences and conflicts are regulated and controlled, through strife between individuality and collectiveness, by means of stipulated norms and rules.
Returning to the issue of policy and policing, it is not difficult to see that the Confucian ideal of harmony, especially with a long historical trajectory of being the state-sponsored political and ethical system since Emperor Han Wudi's rule (156-87 BCE), can function as a coercive force on people's perceptions about 'how to be' or 'how to act' and the normative organization of society. This ideal, as we will see next, is continuously intertextualized, updated, and reinvented-eventually, established as policy-in tune with the development of a modern China.

The reinvention of harmony
As already mentioned, 'harmony' or hexie (和谐) has recently become a prominent discourse pattern in China, embodied in pervasive expressions like Harmonious Society initiated by the government. If one travels to China nowadays, one would find an overwhelming presence of the word hexie in the public sphere: on the TV, in newspapers, on public transport, in classrooms and offices, and on street billboards and banners. Blommaert's (2010:142ff) insightful account of the 'harmonious golf' sign in a Beijing street, about how the national political slogan of harmony is superimposed onto the global corporate discourse of golf, offers a poignant example.
This is an arduous comeback of Confucianism after its marginalization in China since the early twentieth century. This return signals a discursive shift centering upon the Confucian 'jargon' of harmony, the revamping and redeploying of which reflect the state attempts to establish new orders as its engagement with globalization processes deepens. Harmony, in this sense, serves as a strong contemporary rhetoric that dominates the order of meaning making in China, as seen e.g. in the official policy of 'language harmony' (yuyan hexie).
Understanding the discourse of harmony in China's language policy certainly benefits from insights into the philosophical-epistemic dimension of he embedded in the Confucian traditions (as examined above). Meanwhile, this understanding needs to be situated in the present framework of talking and behaving that is emerging under/in the name of hexie in response to processes of globalization. To establish this framework is to further investigate the social-historical dimension of hexie in which harmony is a discourse of cultural and philosophical tradition as well as an indicator of wider social and political changes in the light of China's modernization and globalization. We will see that hexie involves considerable discursive shifts, not just a reactivation of he.

New Confucianism
The discursive shifts of hexie are by no means random. This becomes clear when we consider it as part of the successive discursive shifts about Confucianism unfolded over the course of the last century. Prior to that, Confucianism stayed more or less as a stable, mainstream value system in a largely enclosed Chinese society since the Han Dynasty. The downfall of the last monarchy Qing and the incoming of Western ideologies, such as capitalism, liberalism, and communism in the early 1900s, effectively ended the orthodoxy of Confucianism in China. A new generation of intellectuals emerging from the May Fourth movement (ca. 1919) denounced the hegemony centuries' political manipulations of Confucianism had exerted, especially over the equality, freedom, and creativity of women and youth. Having said that, a total breakaway and disregard of Confucianism, at least of its emblematic and intellectual values, also provoked devastating identity dilemmas for a newborn nation-state. In fact, Confucianism was never far off the scene. As Dirlik (1995:234) asserts, 'These same intellectuals [who decried Confucianism] would, in ensuing years, engage in efforts to find some reconciliation between "Western" and "Chinese" values, out of which would emerge what has come to be called "New Confucianism"'.
New Confucianism is an umbrella term that captures the ongoing neoconservative transformation and reinvention of the Confucian traditions over the past decades, of which three generations of exponents have developed (Makeham 2003;Fan 2011). Over these generations, the status, focus, and impact of Confucianism shifted with the ebb and flow of economic and political conditions (Dirlik 1995;Bell 2006;Louie 2011). In a nutshell, the first generation  of the three returned to Confucianism in search for a sense of self from within a crisis-ridden China, amidst desires to depart from its condemned 'feudal' past. The second one  was led by scholars outside Mainland China which was then occupied by the Cultural Revolution and the consequent ideological distancing from Confucianism in favour of Maoist communism and nationalism. This period was engaged mainly in dialogues between Oriental-Chinese Confucianism and Western-Kantian philosophy. The most recent decades (1980-present) saw a pan-China 'rediscovery' of Confucianism (Bell 2006) accompanying the East Asian economic booms in the 1980s, as-not so much its content, but-'the evaluation of that content with respect to the question of modernity' had changed (Dirlik 1995:236). In this newest wave of Confucianism, China reemerged as its center and leading advocate, while embarking on the post-Mao 'reform and opening-up' course and reengaging with wider processes of globalization. It is this third phase of Confucianist 'renaissance' (Fan 2011) that gives rise to the current political discourse of harmony.

The harmonious society
China's reengagement with globalization since 1978 has hugely enhanced the country's economic-political power and, consequently, its social diversification and restratification. China joined the World Trade Organization in 2001, and in merely 10 years, its GDP had increased almost five times and it became the second-largest economy in the world. Notwithstanding, the inequality of wealth in China is also growing at an alarming speed, with its Gini coefficient index reaching 0.73 in 2012. 2 Reclaiming and revaluing Confucianism under such conditions becomes once again necessary and all the more important. For one thing, it reasserts the part of the cultural and national identity that China had been alienated from under the dominance of imported ideologies and self-inflicted disruptions. For another, this reforges a domestic political-ideological framework that can appeal to nationalistic nostalgia and, at the same time, reestablish order in a rapidly changing and restructuring society. It is in such a context that the Chinese government initiated the latest wave of New Confucianism by reintroducing hexie as the spearhead jargon, calling into question the need of a new order legitimatized by redeploying Confucianism.
Harmonious Society (和谐社会 hexie shehui) was first put forth in September 2004, when former President Hu Jintao gave a speech on 'building a harmonious socialist society' at the fourth plenary session of the Sixteenth Central Committee of the Communist Party of China. This was propounded in the following year when 2 The Gini coefficient index measures income inequality, with \0.25 generally considered the most equal and [0.6 the most unequal; 0.4 is the internationally recognized warning line. See People's Daily article 'Report Suggests Top 1 % Families Own Over 1/3 of The Country's Wealth' (http://society.people. com.cn/n/2014/0725/c1008-25345140.html).
CPC's sixth plenary session passed the strategic document of 'Chinese Communist Party Central Committee's Resolution on Major Issues of Building a Socialist Harmonious Society'. According to President Hu, a harmonious society is 'a scientific development concept' consisting of six elements: democratic rule of law, fairness and justness, honesty and friendliness, vitality and liveliness, stability and orderliness, and coexistence of man and nature. It is urged that a harmonious, i.e. orderly, political environment and social structure with regained morale are needed in China to address the deepening social divide, discontents, and tensions, as well as to fill the perceived ideological and 'moral vacuum' left by the Cultural Revolution (Louie 2011).
Harmonious Society bespeaks an effort in building a culturalist-humanist image of the state by re-cherishing the core Chinese values encoded in the Confucian concept of hexie while turning it into a new rhetoric 'to react and redress an increasingly less balanced and less fair domestic landscape' (Yu 2008:123) and, ultimately, to reassert the state's authority. In this context, hexie becomes a metonym for a self-defendable form of power and coercion that imposes certain order and normativity. This recentering of harmony, as we will see next, is also reflected in the way language policy in China has been (re)formulated and expanded, incorporating hexie as a major trope and motif.

Language harmony
The monoglot standard of Putonghua and its hegemonic dominance over other Chinese varieties has evolved out of sociohistorical practices (Dong 2010) and can thus be seen as a continual process of harmonization in the sense of Confucianism. The hierarchical order as harmony in the domain of language use is made more transparent and justifiable as it merges with the political discourse of social harmony in recent years. This merge is illustrated in the official poster used in 2009 for the twelfth annual National Putonghua Promotion Week organized by the Chinese National Commission on Language and Script Work (see Figure 1).
In this poster, the layers of significance of harmony are semiotized in: (a) hexie as an oversized word placed in the top-center, announcing the theme of the state-led language campaign; (b) the recursive pattern of hexie as the background, inscribed in the calligraphic font 'seal' (zhuan, originated from the Confucian period over 2000 years ago) and written vertically-an aesthetic and archaic organization of semiotic features indexing hexie's historical and cultural roots-which frames the poster as well as, symbolically, the language campaign; and (c) two sentences that spell out the updated (sociolinguistic) meaning of hexie in modern-day China: 'love the motherland language and script; build a harmonious language life'.
Government propaganda like this has been in practice ever since the founding of the PRC to install a monoglossic order in society. This order is not only about which language varieties are more prestigious than others (e.g. Putonghua compared to dialect or other languages), but also about the different and unequal degrees of legitimacy and authenticity certain language variety as social capital may have (or not) to afford people voice (X. . Regulation of language is regulation of voice (Blommaert 2005;Juffermans and Van der Aa 2013), thus, a form of policing through formal policy. The first language law passed in China in 2000, the Law on the Standard Spoken and Written Chinese, is an example and also a pretext of the Harmonious Society discourse that followed a few years later.
What can be seen in the above poster is that the policy of a monoglossic order is being relabelled with hexie, an ideal that emphasizes order and normativity as appropriateness by virtue, and, thus, becomes a bone fide voice. Such a voice is further sanctioned by the nationalist sentiment (the call to love the motherland language and script) aroused by Confucianism from which hexie originates, and is seen as iconic of Chinese history and culture (the archaic styling of the word hexie). Any voice implying an alternative, heteroglossic order is, hence, against harmony and morally inappropriate, and may be regarded as an act of disharmonization and subversion.
Sociolinguists in China (e.g. Feng 2007;Zhou 2006;Zhang and Xie 2010) also argue that maintaining harmony in language use is an indispensable aspect of constructing a harmonious society. They reason that the realization of language harmony relies on people's awareness of the norms of conduct and willingness and 'sensibility' in conforming to such norms (Feng and Zhang 2006). It is suggested that nonstandard, non-normative, and innovative uses of language across domains, such as commercial language, literary works, and online communication, all risk violating and harming linguistic and social harmony (Feng 2007). Following this logic, harmony has to do as much with self-compliance of normalization as with top-down policing and active interventions of state power. So far, we have discussed the notion of 'harmony' with respect to he, in the classical Confucian sense, and hexie, in its evolvement into a political discourse through the reinvention of Confucianism in China. Taken together, these provide ethnographic contextualizations necessary for us to dissect the meanings of harmony as observable texts and practices in China. In the following section, we examine harmony as processes of policing in China's virtual space. We will demonstrate that, alongside the state policing, there are considerable non-stateoriented interactions and influences from grassroots users of the Internet. Such practices imply that 'language harmony' is not only about policies and legislations of language per se, but more about the policing of voice and the validity of using certain linguistic features to express oneself. The Internet has profoundly transformed the way people access information and communicate. It offers unprecedented potential of freedom and democracy to authoritarian states and exposes its citizens to alternative norms, values, and resources that were unavailable before. With the new technology, the speed and velocity of such changes are extraordinary, posing new challenges to the existing social order. One of the main challenges is related to superdiversity-new forms of diversity that make use of the Internet either as a space and medium of production, or as a tool for inventing new resources of meaning making (e.g. Varis and Wang 2011;X. Wang 2012;Velghe 2014). What's more, the Internet allows wider, more active and democratic participation in economic and sociopolitical discussions and public civic life at the grassroots level (Zhou 2005;Yang 2009). In the case of China, however, all of these may interfere or even endanger the building of a harmonious society in the eyes of the state. Not only may online communication dispute the state prerogative of defining practices of meaning making, it also decreases the exclusive power of state control and opens various aspects of social and political issues for negotiation and debate.

Online policing and harmonization
Structuring and maintaining virtual order is, therefore, on the top of the agenda for constructing and reinforcing language harmony in China. Devising and implementing Internet censorship policies have been a vigorous and sometimes aggressive way of policing and controlling online behaviors (Varis et al. 2010). For instance, in addition to language rules, such as the Language Law of 2000, it is common practice to use automatic screen-masking to block 'disharmonious' language useranging from profanity to politically sensitive words or topics-by substituting with asterisks or deleting it altogether. Sometimes an entire webpage or website is removed. The government also contrives a system that inflicts self-monitoring online. CNNIC issues new legislations almost every year regarding the management of Internet Protocol addresses in order to accurately track the activities of individual end users online. The panoptic surveillance measures are conjointly carried out by the Internet police (see Figure 2) who inspect and enforce judicial punishment against 'disharmonious' behaviors. The law enforcement and policing online began in 2003 when the Ministry of Public Security launched the massive Golden Shield censorship project, known as the Great Firewall of China. In a blog entitled '25 Shocking Facts about Chinese Censorship', Wilkins (2009) lists all of the above and other measurements, including the use of spyware and the ban of transnational social networks such as Twitter, YouTube and Facebook, in order to censor and control what is happening on the Internet in China. It is widely reported in the international media, such as the New York Times, the Guardian, and alike, that the Internet censorship in China is the most stringent in the world. In China, more people are employed by the government to monitor and 'guide' online conducts, such as the notorious 50 Cents Party (wumao dang) who are allegedly paid at the rate of half a yuan per post to write as so-called grassroots netizens, and to steer public opinions into a 'harmonious' direction. When Christine Lagarde, the head of the International Monetary Fund, decided to charm China by appearing on its most popular social media Sina Weibo during her official visit to China in November 2011, she was instantly cautioned by a netizen named Damo Duhang, 'Please be careful to write! Here is not France. If your word is sensitive, someone would hexie you' (Chin 2011).
While the state doctrine of Harmonious Society is used as a mandate to justify the control of communication and the quashing of 'disharmonious' speech online, the word hexie has turned into a satirical placeholder for the domineering maintenance of social stability and political order. Netizens started using hexie as a euphemism for Internet censorship. When they say that a user is 'harmonized', the suggestion is that the person has somehow been brought into compliance by government agency, whether by physical force or by losing access to his/her account. By appropriating this word, netizens voice criticism of claims that state-imposed censorship is the means to build a 'harmonious society'. This attitude is illustrated in a widely circulated picture online, which shows the word 和諧 (hexie), in traditional characters, with the radicals 口 (mouth) all being plastered over (see Figure 3). Through this image, netizens argue that harmony is in fact a policing strategy adopted by the authorities to muzzle them, to silence their voices.

Hexie, river crab, and caonima
The parody of 'harmony' has, ironically, turned the word itself into a so-called sensitive word, an object of policing. When the word hexie begot censored and 'harmonized' online, netizens adopted a new word, 'river crab' (河蟹), to replace the direct use of 'harmony', since these two words are tonologically different homophones: héxié for 'harmony', and héxiè for 'river crab'. Images of river crabs are shared online to express discontent with the state censorship and suppression of free speech (see Figure 4). Soon, 'river crab' became a 'meme' (an Internet buzzword) that symbolizes, euphemistically, the ideological battle between 'harmonization' and 'counter-harmonization' in China's cyberspace.
The move from héxié 和谐 to héxiè 河蟹 indicates an extraordinary effect of policing. Rather than uniformity and loss of voice, the enforcement of language harmony online has stimulated and facilitated new forms of (super)diversity and new opportunities and ways of self-articulation. This is important to our understanding of language policy as ethnographically informed processes of policing. Even though harmony and Internet censorship are forcefully implemented as top-down policy, this policy is being negotiated and resisted by the subordinate group and their individual agency, leading to oppositional responses and unexpected outcomes.
To illustrate this point, we turn to another well-known Internet phenomenon since 2009: a 'modern myth' (Hopkins 2011) about 'river crab' fighting 'grass mud horse'-another Internet meme created by netizens. 'Grass-mud horse' comes from căonímă 草泥马, a seemingly innocent nonsense word. However, it is a carefully invented homophone (again with different tones) of another harmonizable expression, càonĭmā 肏你妈, which means 'fuck your mother'. Although the Chinese censorship system aims to curb obscene use of language, the pun effect of caonima enables netizens to transgress while satirizing the policy with impunity. This eventually makes the word an icon of grassroots aspirations for freedom of speech. Netizens even designed a written form for this three-character-phrase, by combining elements of each of the three characters 草, 泥 and 马 (see Figure 5). As a netizen named Kenneth Tan explains, 'The 艹 radical refers to 'grass' (草), 尼 resembles 泥 and both are homophones, while 马 is the character for 'horse'. The new character even has a recommended pronunciation jiàyú'. Initially, caonima became popularized for being a clever euphemism of a swearing word that can escape the touch of harmonization/censorship. But, over time, it took on a whole new life beyond this function. A mythical animal depicted as a furry, amiable-looking alpaca was created to give a physical embodiment to the 'grass mud horse', a previously nonexistent creature, and it started roaming on the Internet. Furthermore, a story was invented and circulated, telling that the magical beast lives in the 'Ma-le Gobi' desert (mălègēbì 马勒戈壁) and feeds on 'fertile grass' (wòcăo 沃草). Although the environment in Ma-le Gobi is extremely harsh, the grass mud horse lives a happy life there. But one day, the river crab (symbolizing harmonization/censorship) moves into Ma-le Gobi. The grass mud horse and the river crab have a fierce fight and, finally, the grass mud horse wins the battle and goes on living in the fantasy land of Ma-le Gobi thereafter (see Figure 6).
The story (with several slightly varied versions) is a dramatic elaboration of resistance against Internet policing by Chinese netizens. One might argue that the protagonist, the grass mud horse, represents the repressed, and the river crab represents the repressor. The use of stories becomes here the 'hidden transcript' (Scott 1990) of public political discourse, by developing euphemistic lexicons, images, and narratives through which language use and meaning making are coded in such a way that they are recognized and shared by subordinate groups, but lie beyond or beneath the patrol and surveillance of the authority. These can be understood in terms of 'metro-practices' (Arnaut 2012), acts of communication or identity that travel underground, below the radar of panoptical governmentality. Both 'Ma-le Gobi' and 'fertile grass' are such examples: the former is the homophone of the Chinese vulgar expression mālegèbī 妈了个屄 (your mother's vagina), and the latter is that of wŏcào 我肏 (I fuck). In using extreme profanity, subversive puns, as well as the metaphorical plot of the grass mud horse defeating the river crab, netizens are able to utter deep resentment and symbolic defiance of China's Internet censorship and figuratively enact the struggles through a fantasy drama of war. The triumph was celebrated across the Internet and spawned reproductions in more vivid forms of language online.
One such example is an online video called 'the song of caonima' which went viral after its release in 2009. The song (again with a few versions) features a digital There is a herd of Grass Mud Horses (fuck your mom) Who live in the MaLe Desert (your mother's cunt) They are lively and intelligent They are fun-loving and nimble They live freely in the MaLe desert (your mother's cunt) They are courageous, tenacious, and overcome the difficult environment Oh, lying down Grass Mud Horse (Oh, fuck your mother!!!) Oh, running wild Grass Mud Horse (Oh, fuck your mother, hard!!!!) They defeated the River Crabs (censorship) in order to protect their grassland (free speech) River Crabs (censorship) disappeared from the MaLe Gobi Desert forever!!! The infectiously funny yet perplexingly distorted digital productions, such as this, suggest powerful yet humorous attacks against the harmonization force. The central narrative about caonima as feisty survivors and warriors who fight bravely against invaders to protect their scarce resources is, in fact, a hidden political dissent and activism through language violence against harmony-as-hegemony. This kind of struggles extends even beyond the Internet, as the image of caonima goes offline, enters the corporeal world, and is turned into consumable goods and identity statements in popular culture (see Figure 7). Not only has caonima been transformed into a new cultural product of online spoofs (Meng 2011) and symbolic interactions (S. Wang 2012) for mass consumption, it goes on to expand deeper into Chinese society and becomes an exploitable material with multiple meanings that inspires and provokes a more explicitly 'disharmonious' democratic movement, notably by public intellectuals. For instance, the Chinese artist and dissident Ai Weiwei makes himself a leading actor of this movement by posting photo images of himself, posing naked with only a small furry caonima doll blocking or 'harmonizing' his genitals. This is highly controversial not least because of the public display of nudity (which has led to criminal accusations of 'pornography' against Ai Weiwei): the composition of these images comes with a highly offensive caption which rests on the pun between 'grass mud horse covering the middle' (căonímă dăngzhōngyāng 草泥马挡中央) and 'fuck your mother the Central Party' (càonĭmā dăngzhōngyāng 肏你妈党中央). This visual and semiotic reframing of caonima further broadens the sociolinguistic repertoire of the word and makes it a transparent symbol of ridicule and contempt over the control of Internet communication imposed from the above-as depicted in a cartoon impression of Ai Weiwei's act of art by the Chicago artist Tom Tian (see Figure 8). By leaping over the heads of the police, a naked Ai Weiwei, with one hand shielding himself with the grass mud horse and the other raised in a fist high in the air, shows ultimate condemnation, rebellion, and subversion against the hegemony of harmony exercised by the authority. Figure 7 Caonima T-shirt and toys (http://www.gxyin.com/ShowNews.aspx?id=569)

Discussion
From hexie to caonima, what we have is a complex story about harmony and harmonization, Internet censorship, hegemony and suppression, resistance and struggle, semiotic innovation and digital creativity from the Chinese cyberspace (cf. Nie 2009). In the name of traditionalism and nationalism represented (selectively) by the Confucian ideal of harmony, the state is keenly restructuring order and rebalancing social disparity while maintaining an authoritarian system. This ambition of harmony inevitably leads to the 'harmonization', namely, coercion and even denial, of diversity and individual voice which are enhanced by the Internet as a new social arena and new package of resources for constructing alternative identities. The Chinese Internet censorship is a new and overt form of policing in response to this phenomenon of globalization. Nevertheless, it is challenged from below.
The above discussions illustrate how innovative manipulations of linguistic, semiotic, and literacy resources via computer-mediated communication creates a new genre of protest and contention through which the process of state policing is sabotaged and challenged. The invention of 'river crab' and 'grass mud horse', together with their associated lexicons, images, puns, and stories, relies on the sophisticated interplay of visual, verbal, and symbolic texts capable of expressing multiple meanings through the same form. This transformation of language function is made possible by the infrastructure of the Internet and is propagated and transmitted from online to offline, making these words a socially recognizable and 'enregistered' (Agha 2005) set of codes that offer semiotic, aesthetic, symbolic, and political capitals, thus, allowing Chinese netizens to develop their own voices in the presence of tough policy.
Subversive Internet memes like the ones examined here are an important and distinctive part of micro-politics that 'takes advantage of unique possibilities of the Chinese language, as well as the technological possibilities of the Internet' (Hopkins 2011). Although memes tend to be contingent, unstable, and temporary-also depending on the extent of policing-their instant usability and trendiness can appeal to mass audiences and can therefore potentially generate mass campaigns against censorship (or other forms of) policing through fast, informal, micro language transgression. The word hexie offers a good example of how 'harmony' is turned on its own head and changed from a symbol of policing and homogenization to that of contention and counter-homogenization, totally opposite to what it was intended to mean by the authorities. It spawns a string of new memes, all of which are developed into codes with multiple functions that can be used and appropriated in various settings and environments as anti-policing instruments.
Thus, ironically, in the process of harmonization, 'harmony' has caused a wave of 'disharmonious' behavior and noise. Such politics of (dis)harmony on the Internet can be taken as an indication of super-diversification of voices. Nevertheless, what seems a semiotic carnival drawing wide participation in a range of formats can only momentarily escape the control and inhibition of the state power. According to Global Voices (a multilingual community of bloggers who report about citizen media stories from around the world), as a consequence of its popularity, the online appearance of hexie and caonima is officially suspended, and a notification to the Chinese forum managers about the policy banishing of these words is repeatedly tweeted (see Figure 9).
Policing seems omnipresent, but so do phenomena of hexie and caonima. The banality of power (Mbembe 1992) in the name of harmony already presupposes the existence of disharmony. The question to ask, then, is 'not whether the Internet will democratize China, but rather in what ways the Internet is democratizing (or will democratize) communication in China' (Tai 2006:184). Conclusions This paper has reviewed the genesis of 'harmony', a state-political term in China from its early philosophical sources to its contemporary deployment as a tool for social ordering. Harmony, as we have seen, was never an unambiguous concept and has always been contested, remodelled, and challenged, by means of shifts in the intertextual links of the concept. In this sense, harmony joins political core terms such as 'freedom', 'democracy', 'development' and so forth: semantic floaters that, when used skilfully, can stand for entirely different realities. As Bolinger (1971) argued long time ago, the 'pure' or 'original' semantics of a term can never stand for the total array of its actual forms of usage. Word meanings per se are poor indicators of the actual life of words in human social and cultural practice.
We are now in a position to formulate two sets of concluding reflections. First, we will extract some general points from our earlier analysis and look at what this tells us about harmony as larger patterns of political and ideological struggles. Secondly, some general observations can be made regarding the nature of language policy and language policy research.
We have seen that the intensive use of the term 'harmony' in contemporary China is guided by a desire, or demand, for social order. This demand appears to be spurred by the accelerating social differentiation in the PRC in the wake of its rise to global economic prominence. With the emergence of a sizeable professional middle class and a smaller (nevertheless important) class of super-rich people (e.g. Tomba 2009), China is rapidly becoming a class-stratified society characterized by inequalities between rich and poor. This is accompanied by the availability of new information and ideas and new opportunities and resources for identity making provided by the Internet.
It is in the context of such escalating social and political divide that 'harmony' must be seen and understood: it is a slogan that responds to the rapid fragmentation and diversification in society by putting some 'spin' on it: in spite of such growing diversities, the Chinese must have a common focus and invest themselves into a project of social cohesion and 'harmony'. This concern with 'harmony' is, thus, an attempt towards re-emphasizing the modernist monocentric ideal inscribed in the state structure of China. Harmony should produce, legitimatize, and enforce centripetal forces in society and politics, and prevent society from spinning out of control.
This, as we have seen, does not always work according to plans. The monocentric orientation of 'harmony' clashes with the increasing polycentricity in Chinese society, with escalating social, cultural, and political fragmentation-an increasing divergence of values, opinions, and other objects of 'ideology'. The Internet is a carrier for such accelerating forms of polycentricity, and we have shown some of the many ways in which Chinese netizens address, in practice, the state's and their own understandings of harmony. The Internet, obviously, is a platform not just for centripetal forces in society but also (and perhaps more so) for centrifugal forces, forces that take subjects out of the monocentric orbit of the state. This tension between a centripetal and monocentric social politics, and a centrifugal and polycentric potential afforded through the Internet, is well understood by the Chinese authorities, and could be at the core of the state's attempts to monitor and constrain Internet use. Similar reactions against the 'chaotic' dimensions of the Internet by the state can be observed elsewhere too; think of the knee-jerk reactions by several Western states when WikiLeaks started publishing previously confidential documents.
The future of 'harmony' as a useful concept in Chinese politics will depend on the way in which it can be deployed as a 'niched', non-totalizing concept targeted at the policing and regulation of certain aspects of social life. If it is applied to the totality of social life, it will backfire, because it is an inadequate descriptor of social processes and, consequently, can only be used against specific social processes, as a means to repress and eliminate certain forms of social processes. That is, it can only be used successfully as a potentially repressive policing instrument. It will then share the fate of many other concepts deployed by central authorities in attempts to 'control' and 'reduce' escalating social diversity. 'Integration', 'social stability', 'social cohesion', and other widely used terms will almost inevitably become (or have already become) targets of contestation and conflict, since they are irrelevant as descriptors of the social realities. A monocentric understanding of legitimate identities is likely to lead to coercive and excluding practices in the age of globalization and superdiversity.
So how do we understand language policy in view of the evidence presented here? It is clear that language policy, any language policy, is not a singular object, the features of which can simply be 'read off' core documents and semantic analysis of the core terms in the language-political vocabulary. It is best to see it as a highly complex and non-linear set of practices that are lodged in specific sociolinguistic contexts. The forces that create language-political effects are not unified either, perhaps not even readily identifiable or entirely unpredictable. A more ethnographically-based analysis would bring out the specific factors influencing the direction of these processes, and show us why sometimes coercion will prevail, and why sometimes resistance and transformation occur.
It is also unwise to see language-political statements and key terms as descriptors of sociolinguistic realities. This 'fallacy of internalism' (Thompson 1990) assumes that political realities are contained in political texts, an assumption that has inspired many scholars in language policy. Texts and terms, however, do not predict their own uptake and implementation. In fact, uptake and implementation are fields of research in their own right and require entirely different approaches than the critical textual analysis of language policies. They demand ethnographic inspection; and when such ethnographic inspection is performed, researchers will often encounter unexpected outcomes (cf. McCarty 2011). We can then see formal language policies -texts and their concepts-as flexible and unstable instruments; they may be in design for dogmatic deployment but in reality receive defiant interpretations and adverse consequences, as illustrated in this paper. We can also see formal language policies as just one instrument for shaping the sociolinguistic lives of people; it rarely occurs as the only instrument. Societies and their sociolinguistic environments are polycentric and become increasingly so. Language policies such as that of harmony will therefore have to share their space of manoeuvering with other sets of prescriptions and normative expectations.