Keywords

A large proportion of the participants in the current study are first- or second-generation migrants, as is the case for other migrant-receiving big cities in multilingual China. The participants’ perspectives on places of origin, ethnolinguistic identities, and mother tongue proficiency were about changes, choices, flexibility, and struggles. The old certainty—the single linear relationship between language and identity—was constantly problematised . Their answers on language use, attitudes, and identities continually remind us that Guangzhou is no longer a predominantly monodialectal Cantonese city. It is a multidialectal city where Putonghua and Cantonese play the most important roles, while many other dialects coexist. Most of its current residents are at least bidialectal, and often tridialectal. The language of Guangzhou is not Cantonese, and Cantonese is not the language of Guangzhou. While it remains a stronghold of Cantonese besides Hong Kong, we are not even sure whether it is still Cantonese-dominant. However, most traditional cultural values (such as the perception of the ideal native speaker) and modern social institutions (such as the education system) are built on monolingual norms, which is how the problems arise. This is not only relevant to the “migrants”, but also to the “locals”, because they too are multidialectal and affected by the mismatch between the monolingual norms and the multidialectal realities.

1 Emergence of the Native Putonghua Speakers

In Chap. 4 we examined how the students perceived themselves as speakers of their “mother tongues” and other language varieties . What we see in these stories are how the participants discursively position themselves in the changing sociolinguistic landscape of Guangzhou in order to justify their language proficiency, language choices, and language attitudes.

The group of Grand-Estate students who claimed Putonghua to be their first dialect or even mother tongue is a conspicuous group. Researchers have noted this group in previous language-attitude research in China but glossed over the finding. For example, in Tang’s (2006) language-attitude study of Guangzhou secondary school students, the figure was 19 % (2005 sample); in Wang and Ladegaard’s study (2008) it was 15 % (2002 sample). In the Grand-Estate class that I studied, 46 % of the students identified themselves as Putonghua-D1 speakers, and the average of the two schools is 25 %. In short, the percentage of Putonghua-D1 speakers in Guangzhou has been on the rise over the past decade and in some communities, such as the Grand-Estate school community, the Putonghua-D1 speakers already outnumber the Cantonese-D1 speakers. Contrary to traditional sociolinguistic assumptions about the standard language, a considerable group of “native Putonghua speakers” is growing. We need to acknowledge their existence in order to study the sociolinguistic significance of the phenomenon.

So what will happen now that such a group has emerged in Guangzhou, China? Would they take the lead in transforming Guangzhou from the current Putonghua/Cantonese diglossia to Putonghua monolingualism, in which Cantonese and other regional dialects become endangered “minority languages”? These questions entail further and long-term studies. Here we are beginning the examination of the macro-, meso-, and microconditions from which such a group arises.

While the Chinese government has not attempted to eliminate regional dialects as explicitly as the Singaporean government, it nonetheless ensures that Putonghua is used in all public domains, and also encourages its use in private domains. There is no official statement that defines Putonghua as the mother tongue of all Han Chinese. However, by simply ignoring the enormous linguistic diversity within the Han Chinese population in all language policies and regulations, Putonghua monolingualism has become the norm in important domains in which language socialisation happens, such as at school, in the mass media, and in public institutions. Regional dialects are not explicitly banned in these domains, but neither are they “legitimate languages” (Heller 1996); thus they are implicitly banned (see Chap. 7).

On the other hand, the Putonghua monolingual norm in education has forced Putonghua into private domains, such as the family. As we have discussed, the school curriculum and many school teachers treat students as if they have the same satisfactory onset Putonghua competence. Some teachers even considered dialect-speaking students as naturally deficient language learners (Miss Pan’s “phonetic system theory” in Chap. 8) . Children who come to school without the necessary Putonghua proficiency could experience great difficulty (see Chen’s case in Chap. 5). Parents, who worry that their children may suffer academic disadvantage because their dialectal literacy is not valued at school and that school teachers are incapable of helping their children, have to prepare their children linguistically for such school requirements (Chap. 8). In some cases, this has resulted in a temporary or permanent language shift within the family; in other cases, the parents may decide to teach Putonghua instead of regional dialects as D1 to their children.

The national language policies and educational policies affect every school, although different school communities can interpret and mediate the same policies in different ways. In the current study, an observable difference between the Grand-Estate and Sandwood School is that a significant group of native Putonghua speakers have emerged in the former school community but not in the latter. Why do migrant children in Sandwood acquire their parent’s dialectal D1s in addition to Cantonese, while their counterparts in Grand-Estate generally fail in both and speak only Putonghua?

1.1 The Impact of Social Network

Most of these Grand-Estate students’ parents are fluent speakers of their own dialectal mother tongues, while the students themselves are Putonghua-dominant, speaking no or very little of their parents’ mother tongues. The language shift has happened within two generations. David and Nambiar (2002, cited in K. K. Y. Cheng 2003) found that the push factor for intergenerational language shift was the high incidence of exogamous marriages and the many numbers relocating elsewhere. To apply that finding to the current study, “relocation” may be defined as moving away from their parents’ regional dialect region before they become 6 years old. In big cities of contemporary China, children usually start primary school around six. Before that, it can be assumed that the influences on their language learning mainly come from their family and the surrounding speech communities. Moreover, equivalence can be drawn between “exogamous marriage” and “interdialectal marriage” (“the mixed” children). That all the four “relocated and mixed” students in the Grand-Estate class had become native Putonghua speakers seems to confirm David and Nambiar’s hypothesis. However, a greater proportion of the Putonghua-D1 students were born in intradialectal families (11 out of 15). Moreover, over 90 % (10 out of 11) of the Sandwood students were “relocated”, “mixed” or both maintained their parents’ mother tongues as D1 which seems to provide counter-evidence. So we will look beyond the figures.

In discussing the case of New Zealand, Walker (2011) argues that access to domains, communities of practice, and social networks are necessary for creating contexts for language use so that linguistic diversity may be sustainable. Because of the implicit ban on dialects in public life in mainland China, the migrant students in the study often could find no domains other than the home in which to hear or use regional dialects. In this respect, native Cantonese speakers are “luckier”, because even if they migrate to a non-Cantonese-speaking region, it is still relatively easy to find Cantonese-speaking media and popular cultural products in Cantonese. Communities of practice (Eckert 2006) may be viable options but I found no preexisting communities of dialect communication in Grand-Estate. Comparing the native Putonghua speakers to the native dialectal speakers in the Grand-Estate groups, the social network factor seemed to be more relevant.

Lei, Yuan, Fan, and Qiu are Putonghua-D1 students among the key informants in the Grand-Estate School. They are either “relocated” or “mixed” and their stories have been discussed in Chap. 4. In contrast, Chen is “relocated” but speaks Teochew speech as his D1 and his case has been presented in Chap. 5. In identifying himself as a Teochew and claiming the centrality of Teochew-dialect proficiency to that identity, Chen made frequent reference to his Jiazu (family and clan) . Home town and Jiazu are more than symbolic, intangible concepts for Chen. While he no longer lives in the Teochew region, his parents have been sending him there during his school vacations, precisely for the purpose of enhancing his sense of being a Teochew and his Teochew-dialect proficiency.

The Jiazu settlements in Sandwood and in Chen’s home town Teochew are examples of the close-knit network that Milroy and Milroy (1985) talk about. Such a network “has an intrinsic capacity to function as a norm-enforcement mechanism, to the extent that it operates in opposition to larger scale institutional standardising pressures…the LOOSENING (original capitalisation) of such a network structure will be associated with linguistic change” (p. 359). A point of reference is the case of the rapid language shift within the Singaporean Chinese community from Chinese regional dialects to Standard Mandarin in home language use since the launch of the “Speak Mandarin Campaign” (discussed in Chap. 2). The percentage of Chinese households that use Chinese dialects as the dominant language has dramatically dropped from more than 80 % to less than 20 % (see Fig. 9.1 ; The figures are calculated based on the categories of “principal household language” (1980 census) and “language most frequently spoken at home” (1990, 2000, 2010 census), and are generalised to represent home language-use pattern), and the trends among the three major dialect groups (Hokkien, Teochew, Cantonese) are similar (Kuo 1985; Lee 2001; Singapore Department of Statistics 2000, 2010).

Fig. 9.1
figure 1

Trends in dominant household language use among Singaporean Chinese

Full size image

However, we should not overlook the social changes that took place preceding and throughout the language shift —the change from monoethnic to multiethnic living and the diminishing of the “clan associations” (Gupta 1994; Li Wei et al. 1997)—another example of a close-knit social network . Dialect-medium education used to be provided by the clan associations. With the end of the clan associations and the strong official propaganda against using dialects, most of the younger Singaporean Chinese are unable to or consider it unnecessary to speak regional dialects .

In contemporary Chinese society, the rapid modernisation and massive migration has greatly loosened the traditional close-knit networks. By sending Chen to his Jiazu in the Teochew region during vacations, Chen’s parents have deliberately and proactively preserved access to such networks for Chen even after migrating to Guangzhou. However, similar bonds were either unattainable or had been given up by most other migrant students in the Grand-Estate school community. “Lone” migrant students like Yuan, Fan, and Lei found their “home towns” (that is, their parent’s home towns) geographically too distant to be tangible. The “home-town speech ”, which they could no longer speak or understand, had also become emotionally too distant to be consequential.

A question remains. The migrant students in Sandwood are just as removed from their dialectal home town as those in Grand-Estate. Why have they not become native Putonghua speakers ? It is possible that the close-knit network of “the locals” has also had an effect on the “outcomers” . While the “outcomers” are not part of the locals’ network, people in the village nonetheless live together and have much more frequent, close contact with each other than people living in modern high-rise residential blocks. While the network serves as a norm-enforcement mechanism for the locals, the norm may appear to the “outcomers” as “speaking dialects to your family” and “speaking dialects in public places”. Cantonese plays an important role in local business, factories, and everyday life. Most of the migrant children’s parents can speak Cantonese and they have encouraged their children to learn it. The children have learned their “home town speech” at home. When playing with other children in the neighbourhood or staying in their parents’ little shops, they have acquired Putonghua, Cantonese, and other dialects, as well as the habit of using them in a hybrid way. At school, they have also had much more opportunity to hear Cantonese from their teachers and classmates, even in class and in the school offices, as the native Cantonese students and teachers are more used to speaking Cantonese among themselves (compared with Grand-Estate).

Moreover, the migrant students in Sandwood mostly fall into the category of “floating population” , a very deprived, and marginalised (Zhang 2002; Mackenzie 2002) group of dwellers in the city because they are usually both non-Hukou and rural migrants. Cantonese as a locally important language probably provides more convenience, social mobility and access to symbolic resources than Putonghua at the township or even district level. The Cantonese-speaking norm and the overt prestige of Cantonese might have jointly resulted in the Sandwood migrant students’ higher proficiency in Cantonese, while the institutional segregation (Hukou) and the local social network distinction may have made them more determined to keep their “roots” by maintaining their “home-town speech” . In contrast, the migrant families in Grand-Estate school community are more well-off than their Sandwood counterparts. Living in a region where the dominant language is clearly Putonghua, migrant families not bound to any dialect-enforcing network seem to have accepted the economically-dominant and pragmatic-minded belief that regional dialects have little relevance in contemporary society (Li Wei et al. 1997). Such a belief ignores the need of young migrants to find a sense of belonging and identification through language, and denies the benefit and importance of becoming multilingual, while maintaining societal multilingualism. To be more exact, almost all participants thought it important to be multilingual, but multilingual competence in Putonghua and other Chinese dialects was not the form of multilingualism they valued .

1.2 The Search for “Roots” and a “Sense of Belonging”

It is often said in the local prodialect media reporting that Putonghua is a language without roots and that it has limited historical or cultural value—interestingly, this is contrary to the Singaporean official language ideology that only Putonghua provides access to Chinese cultural heritage. Yet the language ideology underlying the pro-Cantonese discourse and the “dialect-bashing” discourse of the Singaporean government is the same—historicisation is key to the legitimisation of language varieties (Milroy 2001) and speakers must find their roots through the legitimate language .

Therefore for native Putonghua speakers in Guangzhou, Putonghua does not provide access to regional identities and thus the sense of roots. They have to search for their roots in other directions. Many of them symbolically chose their parents’ home-town dialects for access to their roots, even though they can barely speak those dialects. This loosening connection between language and identity is a symbolic one. It may be argued that “precisely because of its intangibility, symbolic ethnicity can be a very enduring matter” (Edwards 2010, p. 230) , but it also seems to accelerate the loss of regional dialects. This phenomenon has also been noted in previous language-attitude studies on the Singaporean Chinese community in which rapid language shift has happened over three generations. The loosening of ties occurred first between regional dialects and dialectal identities in the parents’ generation and currently between Mandarin proficiency and the ethnic Chinese identity in the younger generation (Li Wei and Zhu Hua 2010; Li Wei 2011). In this way, it is possible to be an ethnic Chinese in Singapore with little knowledge of Mandarin.

On the other hand, the place that seems more consequential for the native Putonghua speakers’ regional identity is Guangzhou. Qiu, Yuan, and Fan, the native Putonghua-speaking girls, are remembered for their “hammering on the table” during focus group sessions when I was asking a series of questions about their places of origin and their “home-town speech” . They protested and questioned the legitimacy of these questions. While they have places of origins defined by official categorisation, they talked as if that was of no significance to them. However, in their individual interviews or in the “off-stage” conversations during the focus group discussions, they had said that they were Guangzhouers . Qiu’s statement was a strong declaration: “From where? From here!” They had grown up in Guangzhou with their friends and family around, and they even spoke or understood more Cantonese than their parents’ “home-town speech”. Yet above all, they said they did not care whether they could speak Cantonese like “natives”. As they said in the focus group, they are Guangzhouers who speak Putonghua as their D1, and some find it unproblematic to be Guangzhouers who do not speak Cantonese (Chap. 4). They questioned the rigid link between language and identity, and they refused to cooperate with the request that potentially forged such a link and questioned their self-categorisation. They managed to position themselves against such a one-identity-one-language norm by discursively problematising and rejecting it during the focus group sessions. Outside the focus group sessions, they were still faced with such a norm, and were sometimes even regarded as a “failure” by their family and others (Chap. 4) .

It has been noted that for young migrants who have or prefer dual/multiple identities, the multiplicity could be as enriching and complementary as it is isolating and dislocating. They may feel a sense of being “both here and there”, or “neither here nor there” (Walker 2011). They are challenged and pulled in different directions in their search for and construction of new hybrid, fluid multiple identities. Putonghua, due to its translocal and nongeographic nature, does not provide a ready-made “sense of place and belonging” in this contextFootnote 1 (Ethnic Voice New Zealand 2005). Cantonese is both the dominant local dialect of the “receiving” city and a prestigious regional dialect per se, which makes it a likely candidate for providing that sense of local belonging.

However, the parents of these students typically did not facilitate their children’s acquisition of Cantonese at home. Neither of Lei’s parents, nor the housemaid or grandmother could speak Cantonese and he had mainly picked it up at kindergarten before he moved to the Putonghua-dominant Grand Estate. Fan’s parents, on the other hand, are fluent Cantonese speakers, but they claimed to have deliberately left the question of regional identity open, and therefore did not insist on Fan’s learning or using any language other than Putonghua. In the case of Qiu, her father, who is multidialectal in Teochew dialect, Cantonese and Putonghua, expressed regret that Qiu could not speak either dialects, but also insisted that the general language environment was more important than family influence. Hence he would not take responsibility for teaching dialects at home.

While most parents shared the view that language environment is crucial, they differed with regard as to what the language environment is composed of. Some, such as Chen’s parents, considered home as an integral part of the larger language environment and saw nothing “unnatural” about teaching dialects to the children at home. In the immediate language environment of Grand-Estate School, where Cantonese has virtually become a “minority language”, the parents’ decision about whether to facilitate the learning of Cantonese at home might crucially determine the children’s competence in, and identification with, it. Moreover, for families like Lei’s in which none of the adults could speak the local dialect, school (including kindergarten) is a vital site where migrant children can encounter and pick up the dialect. By not taking responsibility for facilitating the students’ learning of local dialects or even banning their use, the school and the teachers are collectively responsible for language loss in the local dialects. By not attending to the linguistic and psychological challenges that the children are facing, both school and parents should be held accountable for the linguistic and cultural discontinuities (Walker 2011) experienced by the students.

1.3 Questions of Integration in the Era of Multilingualism

The Grand-Estate migrants and their Sandwood counterparts differed in their perceptions of the importance of knowing Cantonese for living in Guangzhou. The differences between their socioeconomic status and community language environments may account for their more or less instrumental motivation (Gardner and Lambert 1972) for learning Cantonese. Yet there are more symbolic and essential issues underlying the integrative motive. How important is Cantonese/or should Cantonese be for immigrants to be culturally integrated into Guangzhou society, assuming that integration is their goal and the goal of the city? Guangzhou has historically been a city receiving immigrants. Many of its current “local” population are descendants of immigrants who came here decades or centuries earlier. In the past, it was taken for granted that immigrants should integrate with the local community through learning and using Cantonese, whereas now there is open debate as to how we deal with social diversity and unity, and what kind of integration we are aiming for (Edwards 2010).

There is a continuing debate concerning multiculturalism. Yet in the case of Guangzhou, the language situation further complicates the issue of integration. The local language of the host community is Cantonese and many believe that it is the best way of expressing Guangzhou culture. However, Cantonese is a prestigious language and, at the same time, a regional dialect that has no official status because of the monolingual language policy for the entire Han Chinese population in China. The outstanding economic achievement and cultural influence of the Cantonese-speaking region has provided a counterbalance in the unfair competition between Cantonese and the official variety of Putonghua. Yet as other regions are catching up economically and the status of Putonghua is rising to an unprecedented height, Cantonese seems to be losing ground in the competition. In national discourses, Putonghua is the language of integration and unity, while regional dialects are considered to be divisive; this resembles the official discourse in Singapore (Bokhorst-Heng 1999). Among the many language conflicts (or indeed conflicts that are disguised as language issues) online and in the media, one need not go far to find arguments such as: Should immigrants integrate with the local Guangzhouers , or should the locals assimilate with “the rest of the country”? Are the native Cantonese so unique and arrogant that they may have the privilege to speak a different language while “everyone else” speaks Putonghua? The constant struggle that the migrant students experience in negotiating their ethnolinguistic identity and their search for a sense of belonging reminds us that these issues call for more urgent attention .

Our positions on these issues are based on our attitudes towards the relationship between language and culture, language and identity, language and community, and, fundamentally, our assumptions about what language is. Such positions are, in turn, premises of our decision concerning what kind of multilingual education (at home and at school) is considered suitable for Guangzhou citizens or for any citizens in times of mobility and globalisation. Pennycook (2006) warns that we should be careful in demanding that “heritage language learners” connect with their “home culture” and should not take it for granted that the learners will renew contact with their country of origin (referring to Kramsch 2006). In fact, renewing “contract” may be a proper metaphor for renewing “contact”. To what extent should they keep renewing the ethnolinguistic contract with their place of origin? What is the relationship between that contract and the new contract with the current place of residence?

Answers to the two questions are largely ideological and the debate is open. My position is based not so much on the learners’ obligation as on the extent to which they are inhibited or empowered. The individual’s ability to integrate with a community is reciprocal with his/her ability to participate in that community (Touraine 1994, p. 209, translated and cited in Mick 2011): “For integration to exist, an individual or collective subject must be able to modify a social or cultural entity, which means that the importance lies as much on identity as on participation”. In other words, the participants become integrated when they are able to problematise, negotiate and (re)construct what it means to be members of certain communities—the form they would like their contracts to take.

A third question is what kinds of language practice mediate the renewal or making of contracts. A discursive approach to language attitudes has enabled us to examine language practices in detail. Language attitudes—evaluative practices regarding languages, language communities, and other language-related issues—are the means for positioning oneself in the process of participation. The participants need to draw on various languages, or semiotic tools, to accomplish the participation. However, monolingual language ideologies result in different semiotic means being discounted as resources for participation. Putonghua may not be regarded as part of a Guangzhouer identity, but speaking Cantonese is not considered helpful for constructing a “Chinese” identity either. Regional dialects as a whole are “forbidden” languages in socially valued domains, such as education, media, and other formal and institutional activities, domains that would positively valorise a language. In other words, native dialectal speakers are not able to participate in such domains through their native languages, so that the dialect voices are excluded from constructing social realities and identities related to these domains.

The institutional constraints on regulating linguistic access to participation are rigid and compartmentalising, but the participants can transgress them in moment-to-moment discursive negotiation (see Chap. 6). Language learners could reject the stifling “either/or” choice of linguistic and cultural membership, even if only temporarily, and inhabit a third space (cf. Mick 2011; Pennycook 2006) or translanguaging space (Li Wei 2011). Such dynamic, inhabited space is most evident in the discursive construction of self, language, culture, and community among the migrants or the “diasporic” population .

In the current study, the migrant students, especially the native Putonghua speakers , exemplify how such space may be created and inhabited. The native Putonghua students questioned the “linguistic contracts” predefined for them—the connection between “home town speeches” and “home towns”, or the connection between Cantonese and being “Guangzhouers” . They foregrounded the indeterminacy of the relationship between language, culture, and membership, and demanded the recognition of a new type of ethnolinguistic identity. Recognising such indeterminacy and multiplicity is a means of resisting the symbolic domination of monolingual language ideologies that privilege some versions of meaning and language practice as natural, inevitable and incontestable (Gee 2008; Makoni and Pennycook 2005). Pennycook (2006) sees language education not as a functional enterprise that serves other pragmatic agendas, but as a practice of translingual activism . Educators need to understand the translingual and transcultural spaces inhabited by students, and to open up important space “to oppose the incursion of homogenous discourses and to look to multiple sources of cultural renewal” (p. 114).

2 Multilingual Native Cantonese Speakers vis-à-vis Monolingual Language Ideologies

The migrant students are not the only group affected by the conflicts between the “monolingual mindsets” and multilingual realities (Mick 2011; Walker 2011) in the rapidly changing city of Guangzhou. The native Cantonese students are also linguistically hybrid. Unlike their grandparents (and even their parents) who are monolingual Cantonese speakers, they are bilingual in Putonghua and Cantonese (or even multilingual). They are surrounded by a multidialectal sea and began dealing with the challenges from when they were very young—perhaps as soon as they were born and started listening to the world.

Monolingual language ideologies in the form of “self-evident common sense” are the most common form of everyday symbolic domination imposed on multilinguals. The study demonstrated that the one-on-one analogy between language and identity immediately created a dilemma for children born into the increasingly commonplace interdialectal marriage families. In Chap. 4, we have discussed the story of the Sandwood “Cantonese-Hunannese” girl, Ying. She was aware that the local-outcomer marriage of her parents may be frowned upon by some people in the linguistically and culturally conservative Sandwood community. This belief had made her feel suspicious about, and reluctant to answer, questions on place of origin. Because of her father, she is accepted as a local but the fact that she grew up with three languages (Cantonese, the Hunan dialect and Putonghua) was often overlooked. She knew that her spoken Cantonese was somehow not as “good” as that of the other locals, and that when she spoke Cantonese, people might judge her and make her feel inadequate. That was why she felt hesitant about using Cantonese with strangers and preferred Putonghua. Ying’s Cantonese pronunciation errors were singled out each time, whether the activity was casual conversation or language-oriented. Min’s insensitive questioning of Ying’s Cantonese competence was the last straw, and Ying pointed out that she was not deficient, only different. By using a rhetorical “what if” question rather than a straightforward statement, Ying suggested that other people did not understand the challenges that she had been facing and certainly did not put themselves in her shoes.

Ying was right in pointing out that she is not a monolingual and it is not fair to judge her by monolingual standards. Yet she is not “half-Hunannese-half-Cantonese” but both Hunannese and Cantonese. The point is she is a complete person who has an integrated linguistic repertoire comprising the Hunan dialect, Cantonese and Putonghua, rather than being three inadequate monolinguals in one (Cummins 2008; Walker 2011). This should be the basis of our understanding of all aspects of multilinguals’ language life, including their language competence, language attitudes and language identities. The social pressure on Ying comes from a language ideology that fails to recognise this premise.

While Ying’s interdialectal family background makes her a more obvious multilingual, the fact that the majority of the participants, the local “pure” students, also grew up bilingually through Cantonese and Putonghua is too often neglected. The parents of these students often take it for granted that with Cantonese spoken at home and accessible local networks, their children will just naturally become perfect native speakers of Cantonese. To their disappointment, especially in the Grand-Estate School where there are few Cantonese-speaking opportunities, they find that their children are speaking Cantonese with a strange Putonghua accent, lexis and structures. To quote their parents, their Cantonese is “half a bucket of water” (半桶水) or “neither salty nor bland” (不咸不淡). All indicate the sense of under-qualification and hybridity. The tie between Cantonese proficiency and the Guangzhouer identity is the strongest for them, because they cannot “blame” relocation or family background for not being able to speak (1) “standard” and (2) “pure” Cantonese. This is another “common sense” language ideology shared by many that makes their own life difficult.

Firstly, the concept of “standard” is an ideology-ridden social construction (Milroy 2001). Speakers of certain languages come to believe that their language exists in standardised, uniform forms, so that variations in language use are just deviant, incorrect, accented or “dialects” (Deumert 2004). The second sense of “standard” refers to a measure of achievement—how well someone has achieved in comparison with the standardised norms just mentioned. Nonlinguists often do not realise that standardisation and standards are imposed rather than natural. Compared to the official standardised variety Putonghua, there are neither official Cantonese standardisation measures, nor guidelines for spoken or written Cantonese used in the press, nor Cantonese school lessons in Guangzhou. It is not even a legitimate medium of instruction as it is in Hong Kong. It means that the “standard” of Cantonese is considerably vaguer and the Cantonese speakers have little institutional assistance to help them meet the standard or standards.

The language ideologies that there must be a standard form of Cantonese and that native Cantonese speakers must speak “standard Cantonese” are contradicted by the reality of the vague standard and lack of institutional support. However, native speakers who fail to live up to the standard are likely to be punished and may not be able to defend themselves if they too are subscribers to such language ideologies. In Chap. 4, Hay was verbally attacked by Du because Du thought that Hay did not speak the kind of Cantonese spoken by Guangzhouers , and Hay was excluded from being one of “us” because of that. Hay, of course is a Guangzhouer by all “conventional” criteria. Du was not really denying that but discrediting Hay as a “knower” because of his accent. This phenomenon has been found in other multilingual contexts and theorised as “epistemic injustice” (Esch 2010, referring to Fricker 2007). It happens when the semiotic means (whether a style, an accent, a dialect or a language, and the distinctions between these are not clear-cut) used by the speaker is stigmatised.

However, I would argue that in addition to the stigma associated with Hay’s particular accent, misconceptions and lack of language awareness about variation within languages also played an important role in reinforcing the epistemic injustice. It has been pointed out (Agha 2003) that accents become recognisable to different hearers in different ways. A Liverpool working-class accent may sound like British to an American, like English to a Scottish person, like a northerner to an English southerner, like a Liverpudlian to a northerner, like working-class to a Liverpudlian. To my ear, Hay sounded like a typical Huangpu person (we were both born in Huangpu District); Du sounded like someone from the older but not central part of Guangzhou; whereas Su sounded as if he were from an outer suburban district or a county-level town. They all have accents that are “standard” in their own right. However, lacking awareness of the variations within Cantonese and believing that there is only one standard form of (Guangzhou) Cantonese, the students, including Hay himself, reasoned that there must be something wrong with the way Hay speaks.

The consequences of such “epistemic injustice” soon became apparent in the same focus group session when Hay joined Du and Su in teasing Chen for not reading the poem fluently in Cantonese. Chen lost no time in fighting back but targeted Hay only by saying, “You can’t read it either! You are a Guangzhouer but you can’t read it!”, whereas only a moment previously, Chen was the one backing Hay up as a Guangzhouer. Hay’s accent, or rather others’ misconception about that accent, had turned him into an easy target for violence. The violence was symbolic (cf. Herr and Anderson 2003) in this case (but might have been physical in others): people were judged according to a linguistic standard that is mistaken and oppressive, and punished for not meeting the standard. The “inadequate” native speakers become vulnerable to blame, insults and contempt not only from their fellow native speakers, but also from speakers of other languages. Those blaming others at one moment may be subject to the same kind of violence at another moment. Consequently, such symbolic domination and violence indeed victimises everyone. Even though in these sessions the violence seemed relatively mild and fleeting, the way it cumulated and suddenly broke out was startling and worrying.

The second issue to be examined besides “standard Cantonese” is “pure Cantonese”, or why the students “fail” to speak like their parents and grandparents. Having grown up amid day-to-day language contacts between at least two languages, the students make sense of their worlds in heteroglossic/polyglossic rather than monoglossic terms, and their multilingual competence is a reflection of such a way of sense-making. Here we can borrow Bakhtin’s notion of “simultaneity” (Woolard 1998, referring to Bakhtin 1981) to understand these students’ ways with words (borrowed from Heath 1983). Bakhtin sees language forms as “both/and” with a real simultaneity of contrasting elements in tension rather than “a mere wavering between two mutually exclusive possibilities” (p. 4) which is often implied by code-switching approaches. To contextualise “simultaneity” in the study of social language practices, Woolard (1998) highlights the under-researched phenomenon of “bivalency”, which is defined as “a simultaneous membership of an element in more than one linguistic system” (p. 6). To determine the linguistic membership of bivalent elements is more an ideological than “objective” undertaking. So far as the distinction is ideological, it is as much the ideology of the speaker as that of the hearer and the analyst.

In the current study, there were many examples in the students’ language practice that indicated bivalency or simultaneity in general (such as the interrelated translingual phenomena of interference, borrowing, and code-switching). For example, some students directly used the Putonghua pronunciation of the originally Cantonese word when negotiating meanings (see also Liang in press). It is also common for Guangzhou residents to use Cantonese sentence-final particles when speaking Putonghua. A similar phenomenon can be found when Singaporean Chinese speak Mandarin or English (Li Wei et al. 1997; Gupta 1994) .

Such bivalent language uses reflect the fact that the different languages are not kept strictly apart in the participant’s brain. What an individual has is not several discrete repertoires of single languages, but one organic, hybrid, heteroglossic and polyglossic repertoire at his/her disposal . This is the kind of language practice disliked and disapproved of by the students’ purist parents, teachers, and sometimes even themselves. The assumption is that one can and should speak languages in their “pure” forms and not to speak in such a “contaminated” way. They fail to see that in the current era characterised by mobility, multiplicity, and heterogeneity, such a way of speaking is only “natural” and very adaptable to the challenges of everyday life. A study conducted in Hong Kong shows how counterproductive and unrealistic it is to require people to talk as if they lived in an isolated, homogeneous, monolingual world (Li and Elly 2002). Twelve university students were asked to use Cantonese only for one day. It was found that such a purist way of speaking compromised their communicative skills and often caused communication breakdown or interpersonal discomfort. That study considers only code-switching and code mixing practices. If taking into consideration other forms of simultaneous language practices, the attempt to eliminate all these “impure” expressions in speech, if it were ever possible, would do even greater damage to communication and productivity. The purist norm imposed on the multilinguals inhibits their multilingual development and communication, ignores or stigmatises their multilingual competence, and is indeed another form of symbolic domination (Heller 1995) .

Social changes in the multidialectal city of Guangzhou, due to migration, have served “to denaturalise and problematise boundaries and essentialised unities” (Bailey 2007, p. 259) , including: the meaning of mother tongue, the one-on-one analogy between language and identity, the standard language ideology, and linguistic purism. The study showed that the participants were able to discursively undermine the monolingual norms using multiple semiotic resources and ethnolinguistic identities. However, the students’ struggles also foregrounded the extent to which our society and education has failed to recognise multiple and flexible ethnolinguistic identities, to raise awareness of language variation and diversity, and to respond to the learning needs, capacity, and strategies of the new multilingual and heteroglossic generation.