Medium-of-Instruction Debate I: Mother Tongue Education and the Dual MoI Streaming Policy (1998–)

Abstract

In this chapter, we will examine the divergent social forces and competing ideological premises embodied in a few policy documents, notably the Llewellyn Report (Llewellyn SJ, Hancock G, Kirst M, Roeloffs KA, Perspective on education in Hong Kong: Report by a visiting panel. Government Printer, Hong Kong. Retrieved 28 May 2015, from http://www.edb.gov.hk/attachment/en/about-edb/publications-stat/major-reports/perspe_e.pdf, 1982) and the Education Commission Report No. 4 (ECR4, The curriculum and behavioural problems in schools. Government Printer, Hong Kong. Retrieved 27 May 2015, from http://www.edb.gov.hk/attachment/en/about-edb/publications-stat/major-reports/ecr4_e.pdf, 1990), the latter being instrumental in paving the way for the implementation of the dual MoI streaming policy from September 1998, whereby primary-school leavers (around age 12) are allocated to Chinese-medium or English-medium secondary schools. Drawing on a number of critical studies in the literature, we will give a critical appraisal of the educational merits of the streaming proposal in ECR4 by scrutinizing its questionable ideological premises. It will be argued that, as a justification of the streaming proposal, the scapegoating and stigmatization of ‘mixed code’ is ill-advised and unfounded. The monolingual classroom language ideology is outdated. To enhance the quality of classroom teaching and learning, there is a need for more rigorous classroom-based research on pedagogically sound and productive translanguaging practices, with a view to facilitating the effective teaching and learning of content subjects. A promising way forward is to investigate how and under what circumstances students’ L1 may be turned strategically into a useful pedagogical resource in the hands of properly trained plurilingual teachers.

5.1 Introduction

Given the marked linguistic distance between Cantonese and Modern Standard Chinese (MSC) on one hand (Chap. 3) and English on the other (Chap. 4), for Cantonese-L1 Hongkongers to come to grips with Chinese (SWC and spoken Putonghua) and English essentially through schooling under largely foreign-language-learning conditions is nothing short of a tall order (So 1992, 1998; cf. So 1984). But this is exactly the language-in-education policy goal of successive administrations of the Hong Kong government since the early 1990s, which came to be known as biliteracy and trilingualism¹ in the Special Administrative Region since the handover from 1 July 1997. Rather than enforcing the ‘mother tongue education’ policy across the board in all secondary schools from September 1998, the education authorities under the first Chief Executive of the SAR government, Mr. Tung Chee Hwa, allowed about 30% of the 400+ secondary schools to retain their much coveted English-medium status, provided a number of stringent conditions are met.² In other words, the mother tongue education policy was enforced mandatorily in about 70% of the schools, that is, CMI schools which resisted or did not merit the ‘EMI school’ label. Since 1998, as EMI schools tend to be more prestigious Band 1 schools,³ the 30% of EMI school places have become the prize of competition among keen primary school-leavers and their parents, especially those to whom ‘no English, no future’ reverberates like a relentless and haunting truism. In short, under the first SAR government, a well-intentioned mother tongue education policy was twisted and turned into a highly controversial, socially divisive dual MoI streaming policy.

That CMI schools and their students are routinely portrayed as ‘second rate at best’ in public discourse may be gauged by a mini critical discourse analysis (CDA) of a news story involving a retiring primary school principal surnamed Leung.⁴ According to that report, the principal received a marvelous “farewell gift” on the day when allocation of secondary school places were announced: it was reported that 20 out of 64 of the eligible pupils had been admitted to English-medium secondary schools – the best academic performance ever for that school located in Tai Kwok Tsui, a district inhabited by typically working class families. Of interest to us is what the exhilarated principal reportedly said, which was paraphrased towards the end of that news story:

Leung said pupils allocated to their favorite schools would find it a challenge as band-one schools are usually more demanding. He encouraged those going to Chinese-medium schools not to be disappointed, saying they will have more chances to stand out. (The Standard, emphasis added)

It doesn’t take a CDA expert to tell that the students’ disappointment was discursively constructed as a direct result of being allocated to Chinese-medium schools, an indelible label synonymous with ‘second rate at best’, which is destined to follow the CMI students for the rest of their lives. Regardless of whether such a denigration was intended by the school principal or the journalist, or both, the fact remains that in Hong Kong, being allocated to a CMI school is widely perceived as signifying ‘failure to secure a place in a first-rate EMI school’.

As the impact of education is far-reaching, affecting the life chances of future generations, it is understandable how the choice of medium of instruction is intimately tied up with a social concern about access to various forms of linguistic capital (Bourdieu 1991). Such a popular concern, in turn, helps explain why under the streaming policy, a school’s MoI label came to be perceived as indicative of its academic standards and standing, and how well its teaching staff and students ‘live up to’ the expectations of teaching and learning effectively through the medium of English. Before the mandatory segregation of schools by MoI effective from September 1998, it was this same perception which motivated secondary schools to label themselves as Anglo-Chinese schools rather than Chinese Middle Schools (So 1984, 1992). But the social dynamics involved in the MoI debate, within and beyond the realm of education, is much more complex than this. A clear understanding of the complexity of the intricate issues engrossed in this MoI debate is incomplete without clear answers to a number of questions. These include:

• 1. Why did the originally intended mother tongue education policy (100% Chinese-medium) eventually give way to the streaming policy (70% Chinese-medium; 30% English-medium)?

• 2. At the goal-setting stage, what divergent social forces and competing ideological premises were at work in the consultation and deliberation process, and how were they addressed by the education authorities?

• 3. Before and after the implementation of the dual MoI streaming proposal, what is it that makes streaming so controversial?

• 4. In the end, which social forces and ideological premises got the upper hand when the streaming proposal prevailed, and why?

• 5. Above all, who are the key stakeholders in the streaming policy debate, and what are their main concerns from their respective vantage points?

The last-mentioned question will be dealt with in Chap. 6. In this chapter, we will try to disentangle the main critical issues involved by addressing questions 1 to 4 above, which will necessarily require a fairly detailed retrospective account of the milestones and key issues arising at the policy goal-setting and implementation stages. Owing to its controversial nature, the MoI debate has generated a sizable body of critical works in the form of monographs (including PhD dissertations and departmental research reports), journal articles, chapters in edited books, and feature articles in local newspapers and magazines, in Chinese as well as in English. To understand how the SAR’s current language-in-education policy has evolved in the past decades since the colonial era, we will conduct a critical review of the relevant literature published mainly in the last two decades (Asker 1998; Choi 2003a, b; Evans 2000; Ho and Man 2007; Li 1998, 2008; Li and Tse 2002; Lin 1996, 1997, 2000, 2006, 2008; Lin and Man 2009; Luke 1992a, 1992b, 1998, 2005; Pennington 1998; Poon 2010, 2013; So 1984, 1992, 1998; Tang 2004; Tsui et al. 1999; Tung 1992, 1998; Wang and Kirkpatrick 2015). Below, we will first briefly recapitulate the key milestones since the 1970s. In particular, we will examine the ideological underpinnings embedded in one important policy paper – Education Commission Report No. 4 (ECR4 1990) – a 208-page document in which the rationale behind the controversial and socially divisive dual MoI streaming policy is spelled out. In the process, we will also make reference to other Education Commission Reports produced from the 1980s to the 1990s (ECR1–ECR7) where appropriate.

5.2 Language-in-Education Policy: From Goal-setting (1970s) to Implementation (1998)

The implementation of compulsory vernacular primary education in 1971 produced more and more English-knowing teenagers (So 1984, 1998), but the learning outcomes at secondary level, including English but also other content subjects, were disappointing. Like many home-grown Hong Kong academics and educators, successive British education panels (see Table 5.1) and experts were clearly aware of the obvious pitfalls – cognitive, linguistic and affective – of learning through an alien language. For instance, in the Llewellyn report (1982), reference is made to “quiescent pupils” (III.I.10, p. 26), while Lord (1987, p. 16) speaks of the “submerged majority” (cf. Tang 2004, p. 63). This highly unfavorable MoI-related learning condition prompted a group of young, home-grown intellectuals to openly query the socio-educational cost of learning through the medium of English (cf. title of the booklet At what cost, see Cheng et al. 1973/1979; Cheng 1979), and to question the inferior status of Chinese as a non-official language. Protests and rallies were organized, accompanied by critical commentaries in the mass media. These events gradually gathered strengths and culminated in a ‘Chinese as Official Language’ social movement.⁵ Yielding to massive pressure, in 1974, the government under Governor Sir Murray MacLehose agreed to recognize Chinese as a co-official language. Despite this significant landmark event and achievement, however, there was an overriding clause whereby English continued to reign supreme and its status remained unchallenged. In the legal domain, for instance, it was clearly stipulated that where diverging interpretations of different language versions should occur (various Ordinances, the Criminal Code, etc.), the English version would prevail.

Table 5.1 A selected list of education panel reports during the colonial era

In 1978, a nine-year compulsory education policy up to Secondary 3 (Grade 9) was implemented. In 1981, a panel of four experts led by Sir John Llewellyn from the UK was invited to review the language situation and educational policy provisions in the colony. After extensive investigation and meeting with representatives from various groups of stakeholders, the panel submitted a report to the government (the Llewellyn Report 1982), where the dilemmas were clearly articulated and the recommended policy options spelt out unambiguously. While it was widely recognized that learning through one’s mother tongue was the most effective (UNESCO 1953, cf. UNESCO 2003), the panel had no doubt that the economic well-being of Hong Kong hinged on a significant part of its workforce being conversant in English. Failing this, internally the government would be short of English-knowing skilled workers to staff its civil service at different departments, offices and ranks, in which case effective governance would be adversely affected. Externally, employers of transnational consortiums and local companies would find it difficult to hire employable English-speaking staff to engage and meet the needs of non-Cantonese-speaking employers and clients.

In view of such “a classic public policy dilemma”, a difficult choice between, on one hand, prioritizing the lingua-cultural needs of Chinese Hongkongers but with “a possible decline in the economic prosperity” as a consequence, and, on the other hand, ensuring sufficient numbers of speakers conversant in English at the expense of “the educational progress of the majority” (Llewellyn et al. 1982, p. 29), the Llewellyn Report made a compromise recommendation as follows:⁶

The dilemma lends itself to a typically Hong Kong solution, that of compromise. This would involve, in the long term, a shift towards complete mother tongue education in the early compulsory years through abandoning the fiction that the Anglo-Chinese and Chinese middle schools use only one language as the medium of instruction. Such a solution would support a wholehearted push towards genuine bilingualism after P6 [Primary 6], including the tertiary level. From F1 [Secondary 1] there should be a progressive shift to genuinely bilingual programmes so that by the end of FIII [Secondary 3] students are receiving approximately half of their instruction in each language, with putonghua continuing to be an option which can be built into the secondary school timetable as well as being offered on an extra-curricula [sic.] basis at public expense. (Llewellyn et al. 1982, p. 29, para. III.1.20)

From the point of view of safeguarding the best interests of Hong Kong society, this recommendation appears to have the merit of reconciling the dilemma rooted in Cantonese-L1 students’ difficulties of learning through an unfamiliar language (Chap. 4), and the need to foster and facilitate the development of plurilinguality among those students who manage to survive an EMI teaching and learning environment. More importantly, this report probably sowed the seeds of dual MoI streaming some 15 years later. Three paragraphs earlier, it states that:

An obvious way out (…) is for the Government to impose Cantonese as the medium of instruction in FI-III [Secondary 1 – Secondary 3] of all secondary schools so that the first nine years of schooling (PI-FIII) [Primary 1 – Secondary 3] would be in the ‘language of the heart’. A pragmatic variant on this would be to leave alone the small number of schools which have been genuinely successful in using English as a medium of instruction. (1982, para. III.1.17, emphasis added)

In the domain of employment, however, after the Second World War English gradually became more and more relevant to Hongkongers’ education and work life, largely because many Chinese parents in an emerging middle class were attracted by the symbolic value of English in terms of its strong potential for facilitating “upward and outward mobility” (So 1992). This helps explain why Chinese Middle schools were eclipsed by the immensely more popular Anglo-Chinese schools since 1950s, as So (1992) remarks (cf. Li 2002a):

For somebody who possesses tertiary education qualifications or more, he will be assured of either an upward passage and become a member of the local, expanding bourgeoisie; or an outward passage and become a member of the Overseas Chinese communities in one of the advanced, English-speaking nations of the world.

In short, a successful English-medium secondary education has become the principal determinant of upward and outward mobility for the people of Hong Kong. Many, if not most, aspire to both. (So 1992, p. 78)

As a result, many Hongkongers no longer felt so strongly that English was imposed on them; rather, English was gradually seen society-wide as a form of symbolic capital (Bourdieu 1991) that is worth harnessing through hard work and investment (Poon 2010; cf. Norton 2013a, b). This perception, in turn, fuels the prestige of English-medium schools, which explains the general preference for schools to label themselves as Anglo-Chinese before the policy of mandatory segregation of MoI-based schools was implemented in 1998. Such a perception manifestly remains pervasive some 19 years after the handover. According to a recent corporate survey conducted by a Singaporean company on behalf of the credit card consortium Visa (Kwong 2015), of all the middle class parents polled in Asia, mainland Chinese respondents topped the list, with 51 per cent expressing a strong desire for their children to be sent abroad for education. Hong Kong and Indian respondents, at 39% and 34% respectively, were second and third on the list. Their most favored destination was the US, followed by Britain and Australia, for “it affords a broader outlook later in life and widens career options” (Kwong 2015).

Another lasting impact of the Llewellyn Report (1982) is the proposal that an Education Commission be set up to oversee the language-in-education policy provisions by deliberating long-term and short-term priorities, scrutinizing their resource implications, and rolling out a road map for the government’s consideration. Subsequent to the formation of the Education Commission in 1983, three Reports (ECR1 1984; ECR2 1986; ECR3 1988) were issued. All this culminated in the concretization of the dual MoI streaming policy in ECR4 (1990):

Needs were defined in ECR1 [1984]. Research findings were selectively provided in ECR2 [1986] to substantiate the views on those needs. Assuming that the substantiation was not problematic, ECR4 [1990] proposed a framework hoping that the process by which those needs were to be achieved could be well managed and monitored. (Tang 2004, p. 157)

Drawing on research findings suggesting that “only around 30% of students may be able to learn effectively through English” (ECR4 1990, p. 102), ECR4 proposed a Medium-of-Instruction Grouping Assessment (MIGA) framework, whereby primary school-leavers would be assigned to one of three types in terms of their academic ability to learn through the medium of Chinese or English:

C – Students who would learn best through the Chinese medium

B – Students who would probably learn better through the Chinese medium but who are possibly able also to learn in English

E – Students who are able to learn effectively in English many of whom could equally well learn in Chinese should they so wish (ECR4 1990, p. 107)

ECR4 (1990) further proposes that schools be classified into three types: Chinese-medium, two-medium, and English-medium. By providing parents and schools of students’ MIGA results, it was believed that individual schools would be able to make an informed and responsible decision regarding the pedagogically most reasonable and productive MoI for their students.⁷

By 1996, mechanisms for streaming Hong Kong students into Chinese-medium and English-medium schools were progressively concretized for implementation shortly after the return of sovereignty to the People’s Republic. As noted by Poon (2010), several months before the first Chief Executive, Mr. Tung Chee Hwa, took office on 1 July 1997, a draft decree to introduce comprehensive Chinese-medium education across the board in all secondary schools was floated, only to be aborted after being severely criticized in the media and strongly resisted by various stakeholders in the education sector. Indeed, so overwhelming was the popular outcry and the craving for some space for English-medium education in the media that the first SAR government had no choice but to give in and replace that comprehensive ‘mother tongue education’ policy with the dual MoI streaming policy. Based on public opinions amplified in local newspapers, Poon (2010) identified support for the “compulsory Chinese medium instruction policy” on educational grounds (e.g., the Professional Teachers’ Union and the Hong Kong Government Secondary Schools Principals Association) and patriotic grounds (e.g., the Hong Kong Federation of Education Workers). In general, however, “the policy was poorly received territorially by students, parents and schools” (Poon 2010, p. 38). With regard to the nature and extent of the medium-of-instruction dilemma, the results of a survey conducted in July-August 1997 by the Hong Kong Federation of Youth Groups said it all: whereas 55% of the students and parents agreed that CMI was more effective, 73% expressed concerns about lower English standards, while half of the respondents were convinced that CMI students’ life chances in terms of access to university education and job opportunities would be unduly compromised (South China Morning Post, 19 September 1997). There is also evidence that the latter concern was widely shared. In May 1997, some schools and Parent-Teacher Associations put up adverts in several newspapers stating their firm support for English-medium education (Poon 2010, p. 38).

In their critical review of the MoI debate during the past 100 years from colonial days to 1 year after Hong Kong’s return of sovereignty to China, Tsui et al. (1999) observe that successive panels of British education inspectors and experts were unanimous in recommending the use of the local students’ mother tongue as the MoI (see Table 5.1). However, up until the early 1980s, the expert recommendation for some form of vernacular education was consistently disregarded by the colonial government.⁸ This led Tsui et al. (1999) to conclude that, despite its obvious pedagogical merit, the educational agenda (of providing vernacular education) was consistently neglected so long as the political agenda (of promoting English through education under the colonial government) prevailed. To justify its policy choice, however, the colonial government never failed to point to societal needs for English, as evidenced by its arguably indispensable role in international trade and commerce, in addition to strong parental preference of English for their children. The social and economic agendas, therefore, were used as a convenient pretext for privileging the political agenda at the expense of the educational agenda. Such a stance appeared to have changed, however, after the signing of the Sino-British Joint Declaration in 1984, when the political future of Hong Kong was sealed. From then on, successive colonial administrations began to embrace a pro-mother-tongue-education stance, as reflected in the policy documents, ECR1 (1984) – ECR7 (1997). As Lee Kwok-Sung, Principal Education Officer of the Education Department (1998) put it, until the time of the 1997 language-in-education conference commemorating the renationalization of Hong Kong (Asker 1998), the colonial government had been promoting mother-tongue education for over a decade (Lee 1998, p. 111).

In April 1997, a consultation document entitled Arrangements for Firm Guidance on Secondary Schools’ Medium of Instruction was issued. After undergoing minor revision, the official ‘firm guidance’ directive was formally rolled out several months later, in September, as detailed in the Medium of Instruction Guidance for Secondary Schools (Lee 1998, p. 113; see also Poon 1999). In view of its timing, Tsui et al. (1999, p. 205) believe that it is “by no means a coincidence that the implementation of ‘firm guidance’ should have been in 1998, a year after the handover in 1997”. Then, on the basis of detailed comparison and analysis of the MoI policy (changes) in Malaysia, Singapore and India after these former British colonies attained independence, Tsui et al. (1999) conclude that, in Hong Kong as in other postcolonial societies, the pedagogical merits of mother tongue education were foregrounded and vindicated only after the educational agenda converged with the political agenda (pp. 205–206), and that the “social, economic or educational [agenda] will come to the fore if they converge with the political agenda. If they do not, then they get pushed into the background” (p. 210).

5.3 Dual MoI Streaming Proposal (ECR4 1990): Questionable Premises

In view of its tone-setting function, Education Commission Report No. 4 (ECR4 1990) is in many ways “the watershed marking a new beginning in the language policy evolving process of Hong Kong” (Tang 2004, p. 114), paving the way for the important ‘firm guidance’ consultation document issued in April 1997. As with any important policy document, ECR4 relies on a number of premises to buttress its policy line – the dual MoI streaming proposal in this case. These premises include: mother tongue education, the threshold hypothesis and the linguistic interdependence hypothesis (Cummins 1979), the maximum exposure argument, societal needs for English in Hong Kong, and local parents’ preference for English-medium education for their children. Most of these premises have come under critique by various scholars either at the ideological or the implementation level, as shown in the critical review below.⁹

5.3.1 Threshold Hypothesis and Interdependence Hypothesis

According to ECR4 (1990), the dual MoI streaming proposal is guided by Cummins’s (1979) threshold hypothesis and interdependence hypothesis. Tung (1992) examines the theoretical grounding and support of the threshold hypothesis and observes that:

It is clear that Cummins’s Threshold Hypothesis specifies two different thresholds. However, in the Report of the working group set up to review language improvement measures [Education Department 1989], there is reference to only one threshold. The problem is, it is unclear that the Education Department’s threshold coincides with either of the two thresholds. (Tung 1992, p. 121)

Tung (1992) then examines the projected percentage of students seen as capable of following English-medium (30%) and Chinese-medium education (70%) in ECR4 (1990) respectively, and points out that neither accords with Cummins’ (1979) higher or lower threshold. Tung (1992) further queries the quality and quantity of ECR4’s empirical evidence as follows:

So far, the threshold levels have typically been indicated by children’s scores on vocabulary measures or reading comprehension tests. This is acceptable for research purposes but not for applications where we wish to determine whether a particular child can benefit from instruction in a second language. (Tung 1992, p. 122)

All this leads Tung (1992) to conclude that “the attempt by the Education Department to apply the Threshold Hypothesis in Hong Kong is clearly an example of misapplication of Western ideas” (p. 123).

On the other hand, Tung (1992) considers Cummins’s (1979) interdependence hypothesis entirely worth supporting. He reviews a number of empirical studies in other multilingual contexts in which solid evidence was obtained regarding the linguistic and cognitive advantages of a threshold level of competence in the students’ first language on their learning of content subjects in a second language, suggesting that positive transfer is at work. Tung (1992) believes that linguistic interdependence (e.g., between Chinese L1 and English L2) is especially crucial for decontextualized learning of content subjects through reading and writing in L2 (Cummins 1983). On this basis, Tung (1992) pleads for stronger support for more local research into the linguistic interdependence between Chinese and English in the education domain.

5.3.2 Maximum Exposure Hypothesis

According to ECR4 (1990), the streaming proposal was conceived largely to facilitate students’ English proficiency development by maximizing their exposure to English in school. The idea is to produce proficient users of a target language by maximizing students’ exposure through using it as a medium of instruction exclusively. Such a premise may be traced back to one of the six recommendations in ECR1 (1984), as follows:

Secondary schools which use Chinese as the instructing medium should be given additional resources to strengthen the teaching of English to avert any conseqential drop in the standard of English due to reduced exposure. (Lee 1998, p. 111, emphasis added)

The ‘maximum exposure’ hypothesis looms large in ECR4, with 30% of students being assigned to EMI secondary schools to receive English-only instruction (ECR4 1990, pp. 103–104). The ‘maximum exposure’ argument is also used to condemn the use of mixed code in what is supposed to be English-medium classes (pp. 100–101). Tung (1992, p. 128) suspects that with ‘maximum exposure’ being hailed as a premise, the government was trying, on one hand, to please the business sector by acceding to their demand for English-speaking or English-knowing workers through schooling, and on the other hand, to satisfy the wish of many parents to whom English-medium education is equated with their children’s career development and success. As Tung (1992, p. 129) further points out, drumming up the ‘maximum exposure’ argument for support is misguided for it has been widely discredited in earlier research, as Cummins and Swain (1986) have noted:

Although intuitively appealing, there is a considerable amount of research evidence that refutes a simplistic ‘maximum exposure’ hypothesis. Clearly, sufficient exposure to the school language is essential for the development of academic skills; however, equally or more important, is the extent to which students are capable of understanding the academic input to which they are exposed. (Cummins and Swain 1986, p. 80)

Pedagogically speaking, therefore, privileging exposure to a target language at the expense of students’ understanding borders on being unethical (Choi 2003b; Tang 2004). This view is rightfully stated in Tang’s (2004) critique, who draws attention to the sacrifice, cognitive and intellectual, that comes with learning content subjects through an alien tongue:

[W]hen the purpose of having more exposure to a second language, i.e. English in our case, is to enhance that language as the ultimate goal at the expense of learning more effectively in one’s own tongue, the benefit of exposure in such context cannot be justified in either ethical or educational ground. (Tang 2004, p. 139)

Tang (2004) conducts a critical discourse analysis of the key language-in-education policy documents from ECR1 to ECR7. In his eloquent critique, he reveals an unmistakable positivist orientation in their theoretical grounding, which may be characterized as “a technocratic form of policy analysis that emphasizes efficiency and effectiveness” (p. 66). He examines the findings of one government-commissioned classroom-based study in ECR2 (1986), and points out that:

Overall, only a tiny percentage (2–3%) [of students] preferred monolingual English presentation. About a half preferred monolingual Chinese oral presentation and a third monolingual Chinese written presentation. The remainder preferred bilingual modes of spoken or written presentation. (...) In other words, students preferred understanding more in the class through either monolingual Chinese or bilingual modes of spoken or written presentation to learn through English. The priority of the students’ need was clearly evidenced. (Tang 2004, p. 135, emphasis added).

Notwithstanding students’ clear MoI preference (Cantonese or bilingual), ECR4 (1990) focuses on how well the students were “coping well” (learning through English), and opposes that construct with “performing better” (learning through Chinese), whereby the meaning and goal of education is defined as students’ ‘understanding’ when learning through the medium of English. Apart from ‘understanding’, Tang (2004) also problematizes other constructs such as ‘advantage’ (p. 136), and ‘well-educated’ (p. 139). Then, using the dual MoI streaming proposal in ECR4 as an illustration, by “making things clean, calculating, and homogenizing” (p. 156), Tang (2004) shows how thorough the designers of that proposal are in their positivist way of thinking:

[S]tudents were categorized into three streams¹⁰ for the convenience of mapping a ‘clean’ and manageable plan. A timetable was designed to make sure that everything would be processed according to schedule and students were properly channeled to different ‘homogenizing’ groups. Students’ ability was turned into numbers so that ‘calculation’ could be processed based on which streaming or grouping could be made possible and manageable. (Tang 2004, p. 156)

Further, for the sake of ‘effectiveness’ and ‘efficiency’, and to minimize ‘contingency’ and ‘uncertainty’, bridging programs were prescribed to ensure that the maximum number of students could attain the privileged goal of English competence when learning through the medium of English (Tang 2004, p. 156), even though “many schools see the bridging courses as adding to rather than solving the problems which teachers and students face” (Johnson 1998, p. 268). Finally, after laying bare the ideological premises in successive language-in-education policy documents, Tang (2004) concludes that as an institutional framework, the dual MoI streaming proposal approximates an “input-process-output-quality-assurance factory model, (...) where participants were treated as passive agents serving the functional needs of a system” (Tang 2004, p. 159), with the assumption that people would comply once the targets and criteria were set. Tang (2004) goes on to raise a rhetorical question: Granted, that research findings strongly suggested that some 30% of students were linguistically capable and fit to learn through the medium of English, the problem is: “if this [mastering English as L2] is the aim of education, then it is the aim for just 30% of the student population. What about the rest?” (p. 165).

5.3.3 Economic Forces: Societal Need for English

Pervasive in every single language-in-education policy document from ECR1 (1984) to ECR7 (1997) is a claim or premise that there is great demand for English-speaking or English-knowing workers. Where does this demand come from, and who exactly perceives a demand for it? From the point of view of Hong Kong’s demographic composition since the Second World War, it is clear that English has been widely perceived by Chinese Hongkongers as economically a valuable asset to have, but socially and affectively an alien language to learn or use. During the colonial era, despite being the vernacular and the principal medium of written communication among the absolute majority, Cantonese and Hong Kong Written Chinese (HKWC) were only secondary in importance given that until 1974, English was the only official language in the colony. At the same time, from secondary education onwards, owing to the need for more and more people with adequate knowledge of spoken and written English in the multilingual workplace, there was increasing pressure for the colonial government to produce more and more English-knowing school-leavers by expanding the scale of English-medium education. For largely demographic reasons, however, the exclusive use of English for intra-ethnic communication among Chinese Hongkongers has been relatively rare, as So (1998) explains:

[W]hen over 95 percent of the population in Hong Kong speak the language [Cantonese], its use comes naturally and often is taken as a given except for the few occasions when a bilingual Chinese wants to make a symbolic statement by switching from Cantonese to English or Putonghua. Actually, nowadays, Chaozhou-hua, Kejia-hua, Minnan-hua, Siyi-hua, Putonghua or the English-in-Cantonese Mix (the native tongue of the local educated class) are used more often as group/solidarity markers in Hong Kong than Cantonese. (So 1998, p. 160)

Up until the 1970s, therefore, despite its utilitarian and instrumental value, English was widely perceived by Chinese Hongkongers as minimally relevant to their lifeworld, English was felt by many to be an anomaly imposed by the colonial government on the schooling population (Poon 2010, 2013; cf. Cheng 1979).¹¹ Such a popular perception, however, is in stark contrast with the expectations of employers in the business sector, who consider English to be essential for sustaining Hong Kong’s economic well-being. As noted by a high-ranking executive of “the largest and politically most influential bank in Hong Kong” (Tsui et al. 1999, p. 205) at a conference commemorating the return of Hong Kong’s sovereignty to China:

English, which some have wrongly associated with colonialism, is today the most widely used language in the world of business. It is the common link and the language of trade in the global village. If Hong Kong is to remain the great economic success that it is in the competitive global economy, it is vital for its voice to be heard and its products to be promoted. A good command of English is essential for that, especially among the territory’s leaders. (Au 1998, p. 180)

Au then goes on to lament Hong Kong students’ “unsatisfactory level of language standards”, university graduates included, and regrets that many employers have to “organize language training to improve the effectiveness of their communications” by offering “remedial as well as vocational training, often to clerks and managers alike”, thus adding to the companies’ “undesirable business cost” (Au 1998, p. 183).

The bank that Au represented was part of a consortium of big firms which would periodically lament Hong Kong’s declining English standards (see, e.g., Evans 2000, pp. 192–194; cf. ‘the complaint tradition’, Bolton 2003), a voice that was often and continues to be amplified in mass media from time to time. In 1990, in addition to making an “outcry” (Tsui et al. 1999, p. 205), the business sector also launched the Language Campaign, with the explicit goal of helping improve English standards in schools. All this points to the influential role of the business sector in shaping the direction of the government’s language-in-education policy. Embedded in the policy’s premise, namely ‘Hong Kong’s strong demand for English’, is primarily the voice and interest of the local and transnational business sector. Johnson (1998) speaks of “the English lobby”, consisting of the business community and the tertiary institutions, which brought their influence to bear on the policy deliberation process throughout the 1980s, their main argument being, to sustain Hong Kong’s future economic development and to assure its status as an international business and financial center, larger numbers of bilinguals with high standards of English are needed (cf. Tang 2004, p. 156). How influential the business sector has been in shaping the local language-in-education policy agenda may be gauged by Lin’s (1997) remark, “It seems that what the business interests in Hong Kong want is cheap but good foreign-language-speaking labor, ready made from the school system” (Lin 1997, p. 430).

5.3.4 Social Forces: Local Parents’ Preference for English

A similar dilemma was faced by Cantonese-L1 (especially working class) parents who are in favor of an EMI education for their children. Relative to the allegation, that most Cantonese-L1 parents do not understand what is in their children’s best educational interest, So (1992) gives a succinct defense from the vantage point of working class parents as follows:

[M]ost parents somehow know that on the one hand, the educational consequences of English-medium secondary education are not as catastrophic as some pundits would have them believe. On the other hand, the education offered by Chinese Middle Schools is not as easy and effective as their advocates say it is. (...) After all, they know a local Cantonese student will not be able to make his grade in a Chinese Middle School with his Cantonese alone. (...) What matters is really the student – Anglo-Chinese school or English-medium school – could master the two standard languages. (So 1992, p. 80)

And, with regard to the allegation that Cantonese-L1 parents are ‘lemmings’ who had no idea which language of instruction works best for their children, So (1992) reassures us that these parents know very well what they want:

[Hong Kong parents] would like their children to have access to English-medium education, and may, as a result attain a level of English proficiency that would enable them to progress in the local society. In fact, what the parents are shunning are Chinese Middle Schools, not instruction in Chinese. These parents may have very high expectations of their children, but they are not lemmings. (So 1992, p. 82)

The pro-EMI position was eloquently argued for by T.-L. Tsim, a business leader of grassroots origin who made it to the English-medium University of Hong Kong through hard work. Back in 1978, when the MoI debate flared up again,¹² Tsim wrote in the South China Morning Post that whereas EMI secondary education was admittedly a pain for the majority of Cantonese-L1 students, he questioned the wisdom of denying EMI education to bright and linguistically gifted students from modest families, given that access to good English-medium education was a springboard to social mobility (see Fig. 5.1).

Fig 5.1

T.-L. Tsim’s plea that English-medium education should not be barred from bright and linguistically gifted students from a working class background

Having made a strong case for EMI, Tsim then concluded that “it is children who come from less well-to-do families but who have the potential to succeed that will be losing out” if a blanket mother-tongue education policy were to be implemented regardless of students’ and their parents’ choice (Tsim 1978/1979, p. 157). That Tsim’s views were taken seriously by the education authorities under the colonial government may be gauged by his membership in the influential Education Commission set up in 1983. A few years later, in his commissioned report on English proficiency in Hong Kong submitted to the Hong Kong Language Campaign, Tsim (1989, para. 1.11) states the status and function of this group very clearly:

The Hong Kong Chinese who can act as a bridge between East and West, between the expatriates who speak no Cantonese and the locals who speak little English, belong to perhaps the top ten to twenty per cent of the class in our Anglo-Chinese schools. Every effort should be made to ensure that they will be able to learn English, to learn in English if they want to, and use English in their adult life without fear of social ostracism. (Also cited in Choi 2003b, p. 687)

5.4 Discussion

5.4.1 Dual MoI Streaming: A Controversial and Socially Divisive Policy

Without a doubt, English occupies an important place in the Hong Kong SAR government’s language-in-education policy, which is largely dictated by its woman- and man-power development needs as the former British colony gradually evolved from a manufacturing base in the 1960s into a knowledge-based economy since the 1980s. English is regarded by language policy makers of the SAR as important linguistic capital which is crucial for the continued well-being of “Asia’s World City”, and widely perceived by Hongkongers as an indispensable language for upward and outward mobility (So 1992). This is why, notwithstanding the restoration of Chinese sovereignty and the logical move to valorize the Chinese language (vernacular Cantonese and SWC) through the ‘mother tongue education’ policy, English-medium education continues to have a place in the local secondary-school curriculum. This is also the background to the controversial and socially divisive dual MoI streaming policy, enforced from September 1998, to maintain 100 officially sanctioned English-medium secondary schools, which later expanded to 114 after 14 of the 20 schools complained and subsequently attained the EMI status through appeal. The rest of the 300+ secondary schools have remained Chinese-medium, with the CMI label being received by some schools with pride, but seen as an eyesore by many others.

Above was the backdrop to the implementation of the dual MoI streaming policy which, as warned by critics, proved to be highly controversial. Perhaps the most widespread educational concern with the two-tier secondary school allocation system was the unintended but unavoidable labeling effect on CMI students, who have to put up with a popular perception of having ‘failed’ to make it to one of the EMI schools, which are without exception more prestigious. How damaging such a perception is to CMI students’ self-esteem may be gauged by pictures and TV footages of primary school-leavers in tears shortly after results of the allocation of secondary school places were released in July 1999 – the first time when CMI/EMI secondary school places were allocated after the streaming policy was implemented. Those CMI students who could not hold back their emotions were embittered not only by a shattered dream to enter an EMI school of their choice and wish, but also the harsh, socially constructed label of being academically ‘inferior’ or ‘second rate at best’.

5.4.2 Scapegoating CCS and ‘Mixed Code’: Misguided Justification of Segregated Monolingual Instruction

If the ‘language quandary’ (Lord 1979) of Hong Kong students’ low attainment level in English is metaphorically likened to a disease, then there is no question that classroom code-switching (CCS) has been socially constructed as and popularly held to be the symptom, if not the pathogen, both requiring treatment. CCS, often equated with ‘mixed code’, refers to “the alternating use of more than one linguistic code in the classroom by any of the classroom participants (e.g., teacher, students, teacher aide)” (Lin 2008, p. 273). Whereas the term CCS places a stronger emphasis on the process of switching between conventionally discrete languages, ‘mixed code’ commonly refers to the language output resulting from CCS. As exemplified below, such a negative view not only pervades policy documents like the Llewellyn Report (1982), Education Commission (1994, 2005); Education Commission Reports 1–7 (ECR1 1984 – ECR7 1997), it is also commonly found in academic and public discourse.

The panel led by Sir John Llewellyn (Llewellyn et al. 1982) was clearly aware of Hong Kong teachers’ use of ‘mixed code’ as a common practice in their teaching, regardless of the stipulated MoI. This is clearly borne out by classroom-based research. In his study of “bilingual switching strategies” in the teacher talk of Anglo-Chinese secondary schools, for example, Johnson (1983) found that bilingual teachers code-switched every 18 seconds on average (cf. Ho and Man 2007, p. 13). Such a common bilingual interaction practice is presented in the ‘official MoI discourse’ as undesirable, suggesting, implicitly or explicitly, that compared with monolingual instruction, ‘mixed code’ is pedagogically not conducive or even detrimental to students’ learning. For example:

“teaching and learning are generally more effective if the medium of instruction is either the mother tongue or English” (paragraph 6.4.3, ECR4 1990; also cited in Tang 2004, p. 147)

ECR4 endorsed the principles for MoI and recommended regular reviews to monitor progress and stronger measures to encourage Chinese-medium instruction and minimize mixed-code teaching (Education Department 1997, para. 1.2, emphasis added)

[ECR4 recommends] that regular reviews be conducted to monitor progress and to consider whether stronger measures might be needed to achieve the objectives of encouraging Chinese-medium instruction and minimizing mixed-code teaching. (Principal Education Officer, Education Department, Lee 1998, p. 112, emphasis added)

Such an ‘anti-mixed code’ stance is also shared by quite a few academics working in language-related disciplines and other areas within the humanities. For example, an English language teacher educator said:

[U]ntil very recently, more than 90% [of local secondary schools] advertised themselves as English-medium schools. In effect, what this meant was that textbooks, some writing on the board, and examinations were in English but everything else was mainly in the mother tongue, Cantonese, with some loan words (mainly technical) in English. This form of Cantonese and English use is described as ‘mixed code’ and is generally thought to be the worst of all modes of instruction. (Falvey 1998, p. 76)

In a (1997) feature article in the South China Morning Post, Laurence Goldstein, a professor of philosophy at HKU, made two points in support of the impending streaming policy: to give EMI students partial immersion and maximum exposure in English, “but not a mixing of the two”, and to get rid of the “curious mixture of languages”. A similarly hostile stance against ‘mixed code’ may be found in two separate articles authored by a renowned professor of (Chinese-English) translation, (Lau 1997, 1999). In his (1999) article, for instance, he said:

The problem that Hong Kong is facing now is not whether mother-tongue education should be implemented, but how to root out the mixing of languages of instruction (i.e., sometimes Cantonese, sometimes English)….¹³ (Lau 1999, p. 35, my translation)

The ‘avoid mixing’ advice is also shared by many Chinese language education experts. For instance, Tse et al. (2014) advise bilingual parents who are keen on giving their children the best of two (or more) home languages, as follows:

Father may wish to communicate with the child only in English, mother only in Cantonese; different people may use specific languages to create a bilingual environment. But it is important to avoid mixing languages in the same sentence, for that would likely lead to misunderstanding when [your child tries to] make sense [of your language input] and get confused when using [the target languages].¹⁴

Similarly, in her critical analysis of the effectiveness of the SAR government’s fine-tuning policy of the MoI policy in 2009, and the extent to which it helps mitigate the dominance of English in the education domain, Poon (2013) characterizes ‘mixed code’ consistently as a “language problem” on a par with “declining English standards”. In her view, any use of classroom translanguaging by the EMI teacher is a pedagogical problem to be resolved and a classroom language use pattern to be eradicated:

Prior to the 1990s, the Hong Kong government adopted a laissez-faire attitude towards MOI, and the language problem of using mixed code in teaching was not addressed. The problem was deepened throughout the 1980s as evident in some studies (e.g., Johnson 1983). The Hong Kong government then proposed the streaming policy in 1990, hoping to address the problem of using mixed code in EMI schools and at the same time solve the age-old problem of declining English standards. (Poon 2013, p. 45, emphasis added)

In some cases, the teacher’s use of mixed code in class is explained as proficiency-related, which may be true, as Tsui et al. (1999) point out when summarizing the empirical evidence in support of the dual MoI streaming proposal:

[T]he prevalent use of mixed code in English medium schools was a result of the lack of an adequate command of English not only of students but also teachers. (Tsui et al. 1999, p. 199).

Such an observation, however, helps reinforce the popular perception and social stigma against translingual practice, not only in society but also in the education domain, even though elsewhere ‘mixed code’ is demonstrably more successful in engaging students when they are trying to make sense of EMI subject content. For instance, in their review of Ip and Chan’s (1985) two-year longitudinal study involving 7,500 junior secondary students on the amount of English spoken in class, Tsui et al. (1999) point out that:

Students with a high level of English proficiency coped well in English medium education whereas those who had low English proficiency suffered. This study further showed that more and more Cantonese was used in instruction in Anglo-Chinese schools. Teachers often resorted to Cantonese to explain complex concepts as Cantonese or mixed code was more effective in promoting classroom interaction. (Tsui et al. 1999, p. 198)

There is no question that such an ‘anti-mixed code’ attitude is pervasive in society, where it is widely perceived as a linguistic anomaly, reflecting a popular perception that ‘mixed code’ is indicative of the code-mixer’s inability to adhere to linguistic purity by failing to use ‘pure’ language. What is interesting is that some critics and opponents of ‘mixed code’ may themselves be among the heavy code-mixers themselves, even though when made aware of that common social practice, they may feel the need to apologize, seriously or light-heartedly in passing depending on the context (see Chap. 2).

There is thus strong evidence, at the time when the government’s language-in-education policy was being formulated, that ‘mixed code’ was socially constructed as “the prime cause of educational and language problems in 1990 in the tone-setting Education Commission Report No. 4” (Lin 2000, p. 181). It is thus not surprising that, in accordance with such government-led ‘anti-mixed code’ rhetoric, the language-in-education policy measures, in ECR4 as well as subsequent policy documents, are so designed as to ensure that ‘mixed code’ be eradicated, hoping that it would give way to ‘pure’ language instruction. This is why ECR4 (1990), an important policy paper in which the rationale of the dual MoI streaming policy is spelled out unambiguously, should make the reduction of CCS or ‘mixed code’ its primary target. Section 6.4.2 warns against “mixed-code teaching, as a result of which children may not become proficient in the full range of language skills in Chinese or English” (ECR4 1990, p. 99). In the overall design of the dual MoI streaming policy, elimination of mixed code is axiomatically a corollary of the ‘pure’ or ‘no mixing allowed’ language of instruction philosophy, as stipulated in section 6.4.4:

6.4.4 Given our view that it would be better if one clear medium of instruction for teaching, textbooks and examinations were used, it follows that the use of mixed-code should be reduced as far as possible. The corollary to this is that it is necessary for students to be grouped according to which medium of instruction is most appropriate for them. Students will need to be placed in Chinese-medium classes or English-medium classes on the basis of their ability to learn effectively in that medium. (ECR4 1990, pp. 100–101)

The rationale or justification for a CMI/EMI divide is stated in section 6.4.3, as follows:

6.4.3 We recognise that teaching and learning are generally more effective if the medium of instruction is either the mother tongue or English (for those who are able to learn effectively through this medium). Unfortunately, however, the use of mixed-code is quite common in many of our classrooms. In English-medium schools, while the textbooks, written work and examinations are in English, teachers often use Cantonese to explain the lesson material to students and to conduct discussions with students. In some cases, this can lead to time being wasted on translation of English texts in class and, worse still, learning being reduced to rote memorisation of facts in English. (ECR4 1990, p. 100)

Choi (2003b, p. 678) notes that “commissioned academic research played a prominent part in the development of the mechanism of selection”. Regarding research-based evidence, a few government-commissioned studies are cited, in ECR4 as well as in other Education Commission Reports (e.g., Brimer 1985; Johnson 1985),¹⁵ but in terms of offering the empirical evidence needed to justify the dual MoI streaming policy, the findings of those studies failed to make a convincing case for any causal link between the use of CCS in class and students’ weak academic performance. Quite the contrary, there is some indication that it is not at all an impediment to learning as claimed. For instance, in Ho and Man’s (2007, p. 16) review of Brimer’s (1985) data and findings, they conclude that “mixed code may not be handicapping but it was the requirement to perform in English (tests) that hinders students’ performance”. Further, relative to the 30/70 split between EMI and CMI in the streaming proposal, both the quality and amount of putative evidence in support of that threshold (30% EMI) are open to doubt. As Tung (1992) observes:

[I]t is not clear whether any threshold level could be described in such detail as to allow an observer to tell with confidence when a pupil’s language ability has reached a threshold level. (...) There needs to be a clear description of a full range of linguistic tasks that a child has to be able to do in order to be judged to have reached a threshold level and to be able to benefit from a certain type of education. (Tung 1992, p. 122)

Ho and Man (2007, p. 17) similarly query: “How can we determine that it is the top 30% and not 20% or less? How can we define and prove whether the students can benefit from English as an MoI?” A more sweeping conclusion is reached by Tung (1992) as follows:

Arguing from the needs of industry and commerce for more workers capable of functioning in English and concluding that a sizable proportion of the student population should be identified and taught only through the medium of English cannot be supported by research on the development of bilingual proficiency. (Tung 1992, p. 128)

As shown in section 6.4.3 of ECR4 (1990) cited above, rather than reassuring the reader with sound empirical evidence that using ‘mixed code’ is bad, it is simply reiterated that ‘pure’ language instruction is not possible due to ‘mixed code’. That this claim borders on being a fallacy is evidenced by two types of evidence. First, the more educated Hong Kong Chinese people are, the more difficult it is for them to resist sprinkling English expressions of various lengths into their otherwise ‘pure’ Cantonese when interacting with fellow Cantonese-English plurilinguals like themselves. So (1998, p. 160), it will be recalled (see above), regards “the English-in-Cantonese Mix” as “the native tongue of the local educated class”. That this is the case has been clearly demonstrated by the ‘One day with only English’ experiment (Li and Tse 2002; cf. Li 2011; see Chap. 2). Plurilingual Chinese readers who have any doubt about this may give it a try; by consciously avoiding the use of any English while interacting with others in Cantonese, it will become clear where the needs for English in plurilingual interaction lie, plus a good chance of first-hand experience why Cantonese-English ‘mixed code’ (or Chinese-English ‘code-mixing’ in written mode) is so difficult to avoid.

Second, the claim that no meaningful learning takes place through ‘mixed code’ or CCS sounds preposterous and simply does not stand up to reason. Perhaps the clearest counter-evidence comes from highly successful Hong Kong Chinese academics whose outstanding achievement would not have been possible without the mediation of ‘mixed code’ during the formative stage of their education at secondary level. This is reminiscent of Prof. Daniel C. Tsui (崔琦教授), recipient of the Nobel Prize in physics in 1998. Inspired by this exciting news story in October 1998, one fellow alumnus of Chinese-medium Pui Ching Middle School wrote a feature article in Hong Kong Economic Journal (Anonymous 1998) lamenting the inflexible dual-language streaming policy which had just been announced for about 2 months. Apart from lauding and congratulating Prof. Tsui’s crowning achievement for a natural scientist, the writer pointed out that his shining academic performance was due in no small measure to the use of both English and Chinese at Pui Ching Middle School, where teachers would teach in English first, before explaining the main points again in Chinese:

‘At that time the teaching methods at Pui Ching Middle School emphasized Chinese and English equally, whatever the mode of bilingual teaching. The purpose was to ensure that students understand completely. Even in English lessons, after something was taught entirely in English, often the main points would be reiterated and explained one more time. That was so different from the present system, where English is forbidden by the mother tongue education policy, while Chinese is so rigidly banned in EMI lessons.’ (my translation, cited in Li 2008, p. 26)¹⁶

What this anonymous alumnus of Pui Ching Middle School said here gives us much food for thought as we ponder and weigh the desirability of two MOI policy options: (a) to rigidly adhere to ‘pure’ language use by cleansing ‘mixed code’ despite pedagogically sound reasons for translanguaging to students’ more familiar language, inside or outside the classroom; or (b) to conduct serious research into productive translanguaging practices and, in so doing, better understand how and under what circumstances such practices could be turned into pedagogical resources to facilitate content-subject teaching and learning more productively and effectively. The choice seems very clear: while the advantage of exposure to ‘pure’ English is indisputable, it should not come at the cost of clarity and understanding of whatever the students are learning at hand. With educational merits as the yardstick for measuring academic success, it seems unthinkable how a language-in-education policy would value the medium (of teaching and learning) at the expense of the message (the content to be learned).

The ideology of linguistic purism has come under vehement critique. It has been dismissed by many scholars as unduly biased and, given its pervasiveness in plurilingual interaction, amounts to unrealistic wishful thinking (e.g., Choi 2003a, b; Lin 2000; Luke 1992b; So 1992). Luke (1992b, p. 111) found it paradoxical why the mother tongue education policy, if theoretically and pedagogically so well founded, was not applied across the board to all lower secondary students. According to Luke (1992b), the streaming of 30% of primary school-leavers to EMI schools appeared to be a strategic compromise on the part of the government in an attempt to please both parents (cf. ‘social agenda’, Tsui et al. 1999) and employers in the business sector (cf. ‘economic agenda’, Tsui et al. 1999) by meeting their demand half way. To justify the provision for EMI schools while rolling out the mother tongue education policy, ‘mixed code’, which is widely perceived as linguistically ‘impure’, is foregrounded and presented as a scapegoat.¹⁷ Such an argument, however, is built only on the flimsiest of empirical evidence and support (Luke 1992b, p. 112).¹⁸

Both Luke (1992b) and So (1992) indicate that what is generally referred to as ‘code-mixing’ is a natural mode of bilingual interaction, which is commonly found in multilingual societies and is entirely consistent with the government’s language-in-education goal of developing students’ English proficiency and using Cantonese to facilitate the learning of English-medium content subjects effectively. Such a common practice is more recently referred to as translanguaging (e.g., Creese and Blackledge 2010; García and W. Li 2014; García and Lin, in press; see Chap. 2); being sociolinguistically conditioned, translanguaging cannot be wished away by any top-down policy (Luke 1992b, p. 116). More recently, there is also some evidence of translanguaging taking place in English and Putonghua lessons in some local primary schools (Wang and Kirkpatrick 2015, p. 20). So (1992, p. 83) affirms the constructive role of ‘mixed code’ on the grounds that “in varying degree of effectiveness, [it] helps the student adapt to the English-medium environment”. He goes on to dispel ECR4’s (1990) stigmatization of ‘mixed code’ as follows:

This claim [of ‘mixed code’ being the culprit of poor learning outcomes] is misplaced, and indicates a lack of understanding of the language dynamics in local classrooms, and of the development of bilinguality under local conditions on the part of the architects of the Streaming Proposal. The fact is the use of the mixed-code is itself a reflection of the reality of students’ needs. (So 1992, p. 87)

According to So (1992), rather than being a “form of corrupted speech”, ‘mixed code’ is “a mark of bilingual behaviour”, and so any “application of monolingual norms, on the part of language purists” is not only inappropriate, but it also reflects the misguided value judgment of a parochial “monolingual, inward-looking society” (So 1992, p. 87). Rather than pursuing a socially divisive dual MoI streaming policy, therefore, both Luke (1992b) and So (1992) call for more classroom-based research, with a view to identifying pedagogically sound translanguaging practices and productive bilingual teaching and learning strategies.

5.4.3 Outdated Monolingual Classroom Language Ideology

That the training of proficient speakers/writers of English should be factored into the educational outcomes of the SAR’s curricula from primary to tertiary levels is beyond dispute. By extension, even though the road map towards biliteracy and trilingualism since the inception of this language-in-education policy in the mid-1990s has been riddled with problems and queries regarding students’ learning effectiveness, deemphasizing English or eliminating Putonghua in the curriculum has never been seen as an option. The key question is whether the current policy measures give us the greatest mileage, and in what ways students’ learning effectiveness – in content subjects as well as the target additional languages – could be enhanced without unduly complicating an already crowded curriculum and aggravating students’ learning burden. Below, I will problematize one tacitly followed axiom or principle that has informed the SAR’s language-in-education policy provisions to date, namely, an outdated monolingual classroom language ideology.

In Chap. 2, we saw that in informal social interaction where no monolingual norms prevail, plurilinguals would naturally draw on all the linguistic resources within their repertoire to make meaning. As Canagarajah (2013) has argued convincingly, for centuries in many multilingual societies, notably those in European nation states and their former colonies, that unmarked translingual practice is clearly at odds with the monolingual ideology propagated by national governments which were/are guided politically by the ‘one people one language’ dogma. Language labels such as Dutch, Flemish, German and Luxembourgish were created, their differences played up while similarities de-emphasized, in order that discrete boundaries between language varieties could be clearly demarcated. Such a reality is subsequently enforced through standardization and codification, and perpetuated through the nation’s language policy, which may be monolingual or multilingual. Not only is the choice of national language(s) written into the national constitution, used in mass media, and made visible in society, but such a belief is also hammered into schoolchildren’s minds through education. Seen in this light, it comes as no surprise that the linguistic pattern of communication characterized by translanguaging (W. Li 2011; W. Li and Zhu 2013), which is normal, usual, unmarked in social interactional terms, should be viewed with disdain and shame by the populace at large, especially in the education domain. One consequence of the naturalized use of language labels from Arabic to Zulu is that few non-language experts would query the received wisdom of ‘pure language’, let alone querying the value-loaded judgment of all kinds of identity-driven motivations of translanguaging (including ‘crossing’, Rampton 1995), which is implicit in such pejorative terms as ‘code-mixing’, ‘mixed code’, ‘hybridization’, or even ‘(language) bastardization’. We should bear in mind this critical perspective when deliberating issues related to language-in-education policy measures and their implementation in Hong Kong.

Since the colonial era, the language-in-education policy in Hong Kong has been dominated by a monolingual classroom language ideology, as Low and Lu (2006) observe in their survey of ‘code-mixing’ among teachers and students in the home and classroom:

Generally, opposition to the use of mixed code is based on the belief that mixed code communications will not only hinder L2 learning but also retard the development of L1 learning. Mixed code has been described as the leading factor contributing to the general decline of students’ language proficiency. Such an assertion was found in some documents that support the recent changes and adjustments in educational policies of Hong Kong (...). Conversely, a discussion of the detrimental effects of mixed code was omitted from other reports and research publications. Importantly, there were little data or empirical evidence to show that codemixing was responsible for, or might lead to, low proficiency in L1 and L2 if it was used extensively. Nor was data available to support why mixed code caused undesirable results in language learning. (Low and Lu 2006, p. 183)

Another justification of the ‘no code-mixing allowed’ MoI policy is premised on the argument that class time taken up by ‘mixed code’ would be time wasted to the extent that students’ exposure to ‘pure’ English would be reduced. Such a stance is clearly evidenced in, for example, the Education Bureau’s (2009) Legislative Council brief on the need to implement the fine-tuning of the medium of instruction policy:

Although mother-tongue teaching can remove the language barriers for students, effectively stimulate their interest in learning and encourage greater involvement in the learning process, students learning in their mother tongue have limited exposure to English during lesson time and this may affect their bridging over from junior secondary levels to senior secondary and/or post-secondary levels at which EMI teaching may be adopted to a comparatively greater extent. (Education Bureau 2009, ‘Justifications’, p. 2, emphasis added)

These two concerns – eliminating classroom code-switching and ensuring maximum exposure to English – are like both sides of the same coin. At the policy implementation level, their combined effect is that, where English is used as the medium of instruction, be it content subjects or English lessons, no Cantonese is allowed; on the other hand, where the MoI is Cantonese, teachers are advised to refrain from using any English. Failing this, the teacher risks being reprimanded by an inspector (‘language police’, So 1992, p. 88) dispatched routinely by the education authorities on surprise visits, who would issue (sometimes unfriendly) reminders to all teachers that any mixing of Cantonese into their supposedly pure English-medium lessons would be inappropriate. Worse, such reprimanding may sometimes take place in front of their students. Over the years, in my capacity as lecturer of various courses in different MA in TESOL programs (e.g., a course like ‘Social context of language education’), I have heard anecdotes how frustrated in-service teachers were humiliated by rigid and indifferent school principals and/or inspectors. After sharing their emotional outpouring during the break or after class, however, the same in-service teachers made it clear that they saw nothing wrong, for translanguaging to their students’ more familiar language at specific junctures of their lessons was, in their professional judgment, pedagogically the most productive and appropriate decision relative to the objective of meeting the teaching and learning goals at hand. Many of the embittered front-line teachers of English also felt that the unsympathetic school inspectors tended to miss or simply chose to ignore that point when repeating their ‘no mixing allowed’ verdict and reminder regarding the teachers’ use of ‘mixed code’ during the EMI lessons they observed.

How widespread is this shaming experience among front-line EMI teachers, and what impact does it have on the quality of teaching and learning in their lessons? While to my knowledge there has been no research into these two related questions, the extent to which many EMI teachers feel unsure about whether it is right to translanguage to their students’ L1 may be gauged by the title of a booklet: How to have a guilt-free life using Cantonese in the English class: A handbook for the English language teacher in Hong Kong (Swain et al. 2011). There, the circumstances under which Cantonese may be put to use, by the teacher and/or students, in English-medium content lessons are clearly spelled out. The authors explain why and how Cantonese-dominant students should be allowed to use their L1 in EMI lessons, among other reasons to seek quick clarification, from their peers or teacher, or to process abstract or intellectually challenging information which is already available in their L1 before packaging that information in idiomatic and lexico-grammatically correct English. If teachers have to self-monitor and be constantly on guard against sporadic surveillance occasioned by some school inspector’s surprise visit, it is difficult to imagine how they could maintain high morale and have the peace of mind to exercise their professional judgment regarding the most productive pedagogy, which in context may include translanguaging to students’ L1 to cater for the weaker students’ learning needs. In short, such MoI flexibility is lost, or deprived from resourceful bilingual teachers’ inventory of pedagogical options, just to meet the higher-order objective of maximizing students’ exposure to English. It is doubtful whether such a guideline, which has been enforced with rigor for nearly two decades, serves the best interests of our students and teachers from the pedagogical point of view. That ‘mixed code’ is both a cause and result of the students’ poor English is a view also shared by some local academics (see above). Such a view suggests that no meaningful learning takes place where ‘mixed code’ surfaces. It remains unclear, however, to what extent this view is matched by empirical evidence (Low and Lu 2006).¹⁹

5.4.4 Policy Implications: De-stigmatizing Translanguaging and Researching Pedagogically Sound Translanguaging Practices in the Classroom

Given that English is not often used for intra-ethnic communication among Hong Kong Chinese, who make up the absolute majority of the local population (see Chap. 1), classroom teaching is an important site and context for the majority of Cantonese-L1 students from primary to tertiary levels to learn and be exposed to this international language systematically. Class time being precious, the current language-in-education policy is so designed as to maximize students’ exposure to English. Toward this end, however, the empirically discredited ‘maximum exposure’ hypothesis (Cummins and Swain 1986, p. 80; see above) has been hailed like a dogma, a straitjacket that prevents bilingual EMI teachers from turning to their students’ more familiar home and community language to facilitate the give-and-take in classroom teaching and learning. This is so because translanguaging is seen as undermining students’ exposure to English and thus prohibited. Punitive measures for EMI teachers who are ‘caught’ violating the ‘no mixed code allowed’ guideline are not uncommon, even though when school inspectors or ‘language police’ (So 1992) are out of sight, EMI teachers often have no choice but to resort to translanguaging to meet the teaching and learning goal at hand – an unapologetic practice in EMI classrooms and an open secret among EMI teaching professionals in Hong Kong.

With effect from September 2012, the 12-year compulsory education policy and new 334 curriculum structure were implemented. Before that, huge amounts of funding were channeled through Language Fund to support various language enhancement initiatives at different levels (Miller and Li 2008). Now that nine-year compulsory education has been extended to 12-year, still more resources are needed. One crucial question is: How efficient is the language enhancement funding used? The current language-in-education policy prioritizes maximum classroom exposure to English and, to ensure that all students assigned to EMI schools have the aptitude to study through the medium of English, a 30/70 split was imposed, such that English-medium education is reserved for the minority. Such a design is meant to simulate teaching and learning conditions akin to those that are characteristic of immersion in English-L1 countries. The research evidence to justify that 30/70 split, however, is obscure to say the least (Low and Lu 2006; Tung 1992). It is not clear that the primary school-leavers assigned to EMI schools, despite being Band 1 students, are all capable of studying through the medium of English effectively (Choi 2003b, p. 675). This point is echoed by the following rhetorical question:

[A]nybody who knows the local situation will wonder: Where are we going to find that thirty per cent secondary students who will be able to have their education exclusively in English (...)? (So 1992, p. 88, emphasis in original).

So’s (1992) skepticism is corroborated by anecdotal evidence of individual EMI students being obliged to repeat Secondary 1 (Grade 7), for according to their academic results, some students could not cope with studying through English and needed more time to adapt to the EMI environment, with no guarantee for survival beyond the repeated school year.

5.5 Conclusion

Towards the end of his critique, So (1992) asks: ‘Do we really need linguistic segregation in our schools?’ (subheading, pp. 86–88). Recall the three main design features of the dual MoI streaming proposal (cf. So 1992, p. 86):

• 1. Mother-tongue hypothesis

• 2. “better if one clear medium of instruction for teaching, textbooks and examinations are (sic.) used” (ECR4 1990, p. 101)

• 3. “students should be grouped by reference to a medium in which they could learn effectively” (ECR4 1990, p. 101)

As we have seen above, apart from limited educational merits due to its ill-advised and questionable premises, the streaming proposal has proved to be socially divisive largely due to its labeling effect, even though this may not be intended:

[W]hat streaming will do to the schools is not so much provide them with new information, but put a medium of instruction label on them, as well as on their students; and in so doing, effect linguistic segregation in the secondary sector. The Streaming Model is our linguistic ‘Berlin Wall’ of the 1990’s. (So 1992, p. 86)

In sum, the streaming policy arguably suffers from two inadequacies. Firstly, instead of ensuring students’ effective understanding and efficient learning regardless of their choice of MoI, the policy places high priority on maximizing EMI students’ exposure to English, the advantage of which is remote compared with the immediate pedagogical concern of the EMI teacher who is under obligation to cover the syllabus timely and whose priority is to ensure that the teaching point at hand is accurately understood. Often there is no choice but to flout the ‘no mixed code allowed’ guideline by switching to students’ more familiar language, an irresistible classroom practice that unleashes a sense of guilt (Swain et al. 2011), which is pedagogically counterproductive to the detriment of the quality of teaching and learning. To overcome this problem, nothing short of a fundamental U-turn in mindset is needed, namely, to think of the language shared by both teacher and students as a pedagogic resource for learning the target language, English. As a prerequisite, the negative or even hostile attitude toward translanguaging (i.e., CCS or ‘mixed code’) should give way to creative thinking, in particular, how the students’ L1 could be mobilized as a teaching strategy and turned into a learning resource. There are plenty of empirical research findings across different multilingual contexts showing how this can be done strategically and productively, for example, translanguaging tasks that facilitate child migrants’ cognitive academic language proficiency (CALP) development in Canada (Cummins 2013cf. Lewis et al. 2012a, b). Similar classroom MoI studies have been conducted in UK (Creese and Blackledge 2010) and the Hong Kong context (Lin and Wu 2015; Lo 2015). Rather than being a cause or result of poor English, translanguaging is commonplace in social interaction between plurilinguals (W. Li 2011; W. Li and Zhu 2013). And, far from being a symptom or disease, that plurilingual mode of interaction is natural, independent of the levels of language proficiency within the plurilinguals’ repertoire. Put differently, translanguaging involving Cantonese and English is very common among Cantonese-L1 speakers in their informal speech and writing, regardless of their proficiency level in English, especially when the topic is related to things that they have learned or come across through the medium of English, that is, the ‘medium-of-learning effect’ (MOLE in short, see Li and Tse 2002; Li 2011; Chap. 2). What is needed is rigorous classroom research to identify pedagogically sound and productive translanguaging strategies along the lines proposed by Lo and Lin (2015) and their colleagues (e.g., engaging students in think-pair-share in Cantonese as a means to enhance their understanding and confidence, before helping them produce and package the same information in English in accordance with lexico-grammatical accuracy norms in EAP, see Lin and Wu 2015; Lo 2015; Tavares 2015). In this way, rather than being an unwelcomed classroom intrusion to be avoided at all cost, students’ L1 has good potential for playing an instrumental role toward better and clearer understanding of conceptual learning, and their quality of learning will more likely be enhanced (cf. flexible education, Weber 2014). This proposed coping strategy is consistent with Choi’s (2003b) plea to resolving the dilemma, namely, maximizing pedagogical soundness on one hand, while ensuring students’ access to that valuable symbolic capital called English on the other:

[T]he elitist official policy of language streaming and enforced monolingual mode of learning, based on the ideology of language ‘purism’, has to be abandoned, or undermined. Various bilingual modes of teaching as well as classroom communication should be explored so that the first language could be used constructively both for content learning and for supporting the development of the second language, for the majority of the students. (Choi 2003b, pp. 690–691)

The second problem is related to the funding formula. Currently, the secondary and tertiary sectors claim the lion’s share of funding support for various language enhancement initiatives. As is well-known, however, the language learning outcomes leave much to be desired. According to one statistic widely shared and circulated among ELT professionals, where English is taught and learned as a foreign language, it takes at least 200 class hours to bring a tertiary student’s English proficiency level up by half a band score (e.g., an increase from 5.5 to 6.0) on the nine-band IELTS scale.²⁰ This is just a mean figure, with no guarantee that the 200 class hours would actually produce that result for every individual student. The return is grossly disproportional to the investment. In light of the psycholinguistic and neurolinguistic evidence in the last two decades (see Chap. 7), it would appear that the current funding policy is lopsided, in that the bulk of language enhancement resources is directed to key learning stages at the educational hierarchy where learners’ language learning sensitivity is the lowest. To capitalize on the “time-delimited window in early life” (Mayberry and Lock 2003, p. 382), therefore, in accordance with the Chinese adage ‘yielding twice the result with half the effort’²¹ rather than the opposite, it would make more sense for stronger support and more resources to be directed at a life stage which, from the point of view of language learning effectiveness, has been shown to be most productive, namely, at the preschool (kindergarten) and lower primary levels (age 4–8, see Chap. 7). As things stand at present, however, these two key stages are relegated to a lower priority, both in terms of regulatory measures and government funding. As of 2016, there is some indication that free compulsory education will likely be further extended from 12 to 15 years to include the preschool years at kindergarten (age 4–6). If that is the case, it would be opportune time for the education authorities to review the existing policy governing preschool education. In Chap. 7, we will examine the empirical evidence in support of the above-mentioned “time-delimited window” and explore its relevance to the early introduction of Putonghua at the preschool and lower primary levels.