The neoliberal coloniality of EMI in Hong Kong higher education: insights from online stancetaking

1 Introduction

In recent years English-medium instruction (hereafter EMI) has become an increasingly pervasive feature of higher education globally. Frequently discussed under the heading of internationalisation (Zhang 2018), in which ideas about the desirability of international curricular content, international research agendas and collaborations, as well as the free movement of students and scholars are central, EMI is often assumed to be the largely unproblematic means whereby the supposed benefits of internationalisation can be achieved. These hypothesised benefits are integral to the neoliberal marketisation of higher education, providing powerful academic publishers from the Global North with access to wider and more standardised markets (Thompson 2010), facilitating a largely one-way flow of students and scholars from South to North (Vavrus and Pekol 2015), and the privileging of Eurocentric form of knowledge (Heleta and Chasi 2023). Following Block (2022) and Sung (2022), we take the view that EMI, especially in postcolonial contexts, needs to be problematised in terms of linguistic and epistemic inequalities.

In this paper we focus on Hong Kong, and in particular on how members of the public familiar with local forms of implementation take up stances on EMI on an online platform. We begin with a short discussion of EMI in Hong Kong, before moving on to a discussion of stance (Du Bois 2007) and the specifics of online stancetaking as the most appropriate methodological lens for the analysis of the interactional data. This is followed by an analysis of the complex online stancetaking we encountered and concludes with a discussion of the findings.

2 EMI in Hong Kong higher education: a neoliberal coloniality perspective

Following scholars such as Wallerstein (1979) and Mignolo (2000), we take the view that capitalism as a world system could not have developed without European (and specifically British) colonialism. From this perspective, the spread of colonial languages such as English can be understood as having acted parasitically as ‘free riders’ on capital (O’Regan 2021), becoming in turn integral to neoliberalism as the currently dominant form of global capitalism. In post-colonial settings such as Hong Kong, English may be said to have been transformed from the language of colonial government to one of neoliberal coloniality. By this we refer to the way in which English in Hong Kong is doubly freighted – on the one hand, a residually enduring product of the hegemonic colonial order and the linguistic imperialism that accompanied it (Phillipson 1992), and on the other hand, a neoliberal site for the fashioning of the self as marketable in a highly competitive labour environment (Park 2016).

Scholars of coloniality (Grosfoguel 2007; Mignolo 2000; Quijano 1999, 2000), writing mainly about the Global South, argue that the end of colonialism has often meant the continued political, economic and cultural peripheralization of the ex-colony – what Quijano (2000) refers to as the ‘coloniality of power’. Clearly, the case of Hong Kong is altogether different. Unlike many ex-colonies there is no ongoing political and economic dependency on the ex-colonial power. However, following Maldonado-Torres (2007, p. 243) we agree that coloniality survives colonialism in multiple ways and that it ‘is maintained alive in books, in the criteria for academic performance, in cultural patterns, in common sense, in the self-image of peoples, in aspirations of self, and so many other aspects of our modern experience.’ From this perspective, it is the ‘coloniality of knowledge’ (Quijano 1999), and specifically the ‘coloniality of language’ (Veronelli 2015) for encoding that knowledge that is most relevant in the case of Hong Kong (see also Lin 2024; Lin and Luke 2006). The practices outlined by Maldonado-Torres entail the ongoing privileging of the colonial language over local languages in the post-colonial period, which amount to, as Veronelli (2015) notes, a racialisation and a dehumanisation of the expressive resources of those on the receiving end of such policies. Indeed, the Hong Kong government established the Native-speaking English Teacher (NET) scheme in 1997 (still recruiting today) to improve the quality of English being taught in public sector schools. Concretely then, what we are calling neoliberal coloniality refers to the way in which colonial values and attitudes (with regard to native speaker varieties of language and forms of knowledge and knowledge production) that emerged in the colonial period endure after the official end of colonialism and harmonise with contemporary neoliberal values, attitudes and forms of knowledge and knowledge production in which English is central. In the case of Hong Kong, we see this particularly in the role ascribed to English in the labour market and in the education system, along with the concomitant downgrading of Cantonese in the latter.

At the same time, we recognise that there is desire for English (and other colonial languages) across much of the world. Following Motha and Lin (2014), we understand desire as a situated and co-constructed phenomenon, shaped by social, historical, political, and economic contexts, with conscious and unconscious dimensions that can be leveraged for exploitative or liberatory purposes in language teaching and learning (p. 333). From this perspective, desire is socially produced. That desire for English exists in post-colonial settings such as Hong Kong is not to be wondered at, given the role that English plays in the world today. Institutions such as universities, increasingly desire to compete in the global educational marketplace through the implementation of EMI language policies. However, as we shall see below, the desires of local students do not always harmonise entirely with those of the institutions in which they study. Furthermore, studies in different contexts, especially in postcolonial ones, have shown that the pursuit of English often mediates linguistic and epistemic inequalities (Tupas 2015, 2022).

The specific implementation of EMI in Hong Kong can be traced back to the nineteenth century, when English was established as the official language for the colonial administration as well as a common language for business and education (Evans 2016). This process was underpinned by the establishment of the University of Hong Kong (hereafter HKU) in 1911, in which English played a central role. Frederick Lugard (1910), the then governor of Hong Kong who helped found what was the city’s first university, presented this as means to ‘civilise’ the locals, stating:

[…] I would emphasise the value of English as the medium of instruction. If we believe that British interests will be thus promoted, we believe equally firmly that graduates, by the mastery of English, will acquire the key to a great literature and the passport to a great trade. (p.4)

Thus, education in English was constructed officially from its inception as beneficial to the colonial power and the personal, cultural and material gain of the colonised. In this way, the currently pervasive neoliberal discourse of English as an essential skill in the labour market is also intimately bound up with a politically and economically defined hierarchy of languages, in which English was historically more valued than Cantonese (Lin 1997). Today, EMI is used as a key selling point by HKU and other Hong Kong universities. The official HKU website states:

We are the first university in Hong Kong to adopt English as the main medium of instruction, and we enjoy a distinctive position as an English-medium university in Asia. HKU had formed academic links with more than 400 overseas universities and hosts over 16,000 non-local students on campus.

Thus, early adoption in the colonial period and HKU’s contemporary role as a distinctive provider in the Asian and international educational marketplace are clearly signalled. In such a scenario, Cantonese takes second place to English as the language of global capitalism, education and, by implication (echoing Lugard), of ‘civilised’ Hong Kongers.

3 Stancetaking and online stancetaking

As the use of EMI continues to exert an impact on teachers and students in Hong Kong, it is important to understand how EMI is made sense of by the public and specifically by those with experience of this educational policy. As the current study sets out to trace the connections between EMI and neoliberal coloniality in online interactions, we require a framework that can attend to both situated positionings and wider ideological implications. In this regard, the concept of stance and the process of stancetaking are particularly useful.

Although definitions of stance abound, Du Bois’s (2007) conceptualisation has proved influential (see Jaffe 2009; Kiesling et al. 2018). This is described as entailing:

a public act by a social actor, achieved dialogically through overt communicative means (language, gesture, and other symbolic forms), through which social actors simultaneously evaluate objects, position subjects (themselves and others), and align with other subjects, with respect to any salient dimensions of the sociocultural field. (Du Bois 2007, p. 163)

Du Bois differentiates between three types of stance functions, namely evaluation, positioning and alignment. Stancetakers evaluate what is called an object of stance (or stance object) and position themselves on an epistemic scale (expressing knowledge – e.g. ‘I know’) or an affective one (expressing feelings – e.g. ‘I am glad’), or both (e.g., ‘I am amazed to discover’). They also display alignment or disalignment when interacting with other stancetakers. Following Jaworski and Thurlow (2009), we take the view that (1) evaluation sheds light on whether an object is taken as desirable/undesirable or good/bad; and that (2) stancetaking establishes interpersonal positionings while instantiating ideology or indexing an ideological position. From this perspective, stancetaking is useful in bringing out the ideological inflections in the voices of online stancetakers, as well as the inherent instability of ideology (Holborow 2012), as posters discuss the merits and demerits of EMI in Hong Kong.

In its original formulation Du Bois’s tripartite model refers to social interaction between two speakers. Clearly, stancetaking online is more complex, partly because it entails more than two participants and partly because of the way in which the structural design and arrangements of online platforms impact on interactions. First, unless a discussant directly quotes a particular post when contributing to a discussion, it is often unclear to which contribution the stancetaker is responding. Second, online interactions are qualitatively different from face-to-face interactions taking place in real time. For instance, the asynchronous nature of many online discussions can result in temporal discontinuities in the flow of communication, as participants join the discussion at different points in time and are not necessarily aware of the content of all previous contributions. This can lead to digressions from the original topic requiring the need for reminders of the topic from the original or other posters. Third, interactions are not always mediated through the same attention to face work that characterises much face-to-face interaction. The anonymity of online posting can entail (as in the case of this online discussion) a high degree of face threatening and insulting language.

Both qualitative and quantitative methods have been applied to examine stancetaking in online conversations. In Kang and Chen’s (2014) qualitative study examining the construction of the Hong Kong gender stereotype known as the ‘Kong Girl’, stancetaking analysis focuses on data about a woman named Jenny, the only stance object of interest in their study. Given their large database, they illustrate what they consider to be telling examples of stances only. In contrast, Kiesling et al. (2018), adopting a quantitative computational approach, annotate all interactional stances, with a view to analysing thread structures and keywords that they argue are both indicative of and predictive of stance dimensions. However, our database is larger and messier than that of Keng and Chen’s, and the predictive aspect of Kiesling et al.’s study is not applicable. Instead, our approach is informed by those studies that seek to analyse large datasets (e.g. Cutting 2020; Kang 2023; Strange 2023) through a combination of thematic analysis and stancetaking. From this perspective, a thematic analysis is initially conducted to organise the data practically, making its manageable for subsequent stancetaking analysis. This procedure is outlined in the following section.

4 Methods

The study focuses on a discussion thread on LIHKG (see https://lihkg.com/thread/2458603), a popular Hong Kong-based Reddit-like online forum. The discussion topics are wide-ranging and mainly organised by ‘channel’, which stipulates the expected theme (e.g., current affairs, computer hardware, relationships, academic, etc.) for contributors. The forum has a conventional design that allows users to start threads, post replies, and ‘like’ and ‘dislike’ posts.

The selected thread for this paper is a noticeably heated one on EMI in the Master of Social Work (MSW) programme at the University of Hong Kong (HKU), titled ‘HKU MSW] 可能HKU會係最離地既大學’ (Translation: ‘HKU MSW] HKU could be the university most detached from reality’). Interestingly this was initiated on the politics channel, rather than the academic channel. Beginning on 23 March 2021, the thread attracted 480 replies from 112 different users, creating a corpus of 22,171 words, with the total posts attracting 565 likes and 160 dislikes (as of April 2024; see https://osf.io/sny9v for our translation). The originator of the thread identified themself as a student on the programme and criticised the use of EMI in programme delivery. This generated a discussion on various related and unrelated topics. Relevant topics included EMI, MSW, HKU, social work programmes in general, both at HKU and at other universities, social work as a profession, as well as personal attacks by posters on other posters, and meta-comments about the discussion itself.

Before deciding on how to conduct the stancetaking analysis, we had to decide how best to approach the database. As suggested above, the flow of online interactions is somewhat diffuse and unpredictable, with comments not necessarily following preceding posts sequentially. In fact, the original poster had to place a pinned message at the top of the page as a reminder to all discussants (regardless of when they joined the discussion) that the focus of the thread was the impact of language on learning and employment and not the nature of social work itself. Given the inevitable messiness of the data, we approached the analysis in stages.

We first read and re-read all the posts to develop our analytical sensitivity and understanding of key issues. To start with, we focused on the originator’s long and complex first post because of its importance in establishing the basis for the discussion. After re-reading it, we confirmed that EMI is the main stance object, which in turn is linked to three key issues we have labelled themes, namely practicality, flexibility and fairness (see Table 1 below for description and summary of the number of coded references under each theme). Before addressing the stancetaking, we opted for a grounded thematic analysis of all the posts made in response to the initial post to see whether and how issues related to these themes were taken up by the discussants. Accordingly, all the contributions were first coded through a series of open and axial coding (Strauss and Corbin 1990) and constant comparison (Glaser 1965) processes on NVivo 20. We established that, as with the initial post, the stance object of EMI also permeates the whole thread and that relevant responses also fall predominantly under the three themes of practicality, flexibility and fairness. All these themes refer directly to the language-in-education policy on the programme, while the theme of practicality also encompasses the use of English in the field of social work. As there are copious amounts of data under each theme, our stancetaking analysis entails selected examples to reflect the overall focus of the thread (see Appendix for translation of the first post).

We now present an analysis of the first post, followed by a stancetaking analysis of the responses and the ways in which coloniality and neoliberalism are indexed.

5 Analysis: the first post by 你呀媽係彈彈波

The originator of the thread starts with a very long post (see https://osf.io/sny9v for our translation), where their dissatisfaction with the use of EMI within the MSW programme of HKU is expressed at length. In line with the title of the thread, they first set the scene by evaluating the university and the programme as generally ‘disconnected from reality’, contrasting it with the practical nature of social work. The poster complains:

Actually, HKU being disconnected from reality is nothing new. I just never thought that the MSW (Master of Social Work), a supposedly down-to-earth programme, could be so fucking disconnected from reality. (lines 3–5)

They then mention a range of issues regarding the orientation of the programme design (i.e., the focus on professional writing rather than practical social work skills) (lines 6–9) and assessment (i.e., group work versus individual work) (lines 11–6). They stress that they want to ‘write something more practical, something actually helpful to future clients’ (line 20–1). Subsequently, the poster moves on to the main issue of the use of EMI on the programme and details the problems associated with the aforementioned themes of practicality, flexibility and fairness (lines 31–66).

With regard to practicality, the poster emphasises that the language of social work in Hong Kong is predominantly Cantonese, which makes EMI impractical. They explain:

Although there’re a lot more English-speaking clients these days, in most cases, we use Cantonese to communicate with our co-workers and clients. In this sense, I really think we should use Cantonese, not English, for counselling [practice]. (lines 25–7)

While acknowledging the presence of English-speaking clients, here we see a questioning of the institutional language policy and the linguistic hierarchy it upholds. As the most commonly used language of communication between colleagues and with clients, Cantonese is evaluated as the most suitable language for learning on the programme, at least as far as counselling practice is concerned. This evaluation is maintained (lines 31–36) and the practical nature of the social work profession itself is highlighted (lines 32–33), thereby framing the teaching of social work in an ‘impractical’ language as unrealistic. The poster then says:

If we’re doing it in English, how can the vast majority of the students (99 %) apply what they’ve learnt in their job in the future? They’re also not able to learn what they should learn. (lines 34–6)

The use of English is evaluated here as something that hinders the application of professional knowledge in the future, as well as students’ ability to learn in class in the present. Thus, the impracticality of the language policy is compounded by the educational disadvantage it presupposes for those students who are also unable to learn course content.

This is immediately followed by a comment about flexibility. While acknowledging the official MOI is English, the poster points out that this is not always rigidly observed in practice, thus hinting at its inherent unworkability:

Some students pointed out that the University states explicitly that the medium of instruction is English, but this proved not to be the case […]. (lines 38–9)

Noting that both Chinese and English were used by previous cohorts (lines 39–41) and by the poster themself in casework practice, they then pose the rhetorical question ‘If we’re sticking to this principle, why were there exceptional cases?’ (line 43). The case for flexibility regarding language use is thus made by referencing instructors’ attested failure to follow the institutional monolingual policy.

The poster then turns to fairness (lines 46–66). This issue has already been implicitly raised in the assertion that the language policy disadvantages students from acquiring much-needed practical skills. They return to this explicitly by referring to the presence of ethnic minority and mainland Chinese students taking the course, given that these students may not speak Cantonese. Anticipating the argument that not using EMI is a form of ‘discrimination’ (line 46), the poster positions non-Cantonese speaking students as unreasonable:

They [referring to ethnic minority and mainland Chinese students] asked why there’re only English and Cantonese for us to choose from [referring to the attested use of Cantonese and English mentioned above], but not Hindi or Mandarin. If not using English in classes means discrimination, should we, the Cantonese speakers, learn their mother languages to accommodate them? (lines 47–50)

Here the poster disaligns in advance with those who will argue that using EMI is appropriate in multilingual classrooms and positions such students as demanding unreasonable accommodation, as well as being guilty of moral blackmail (lines 53, 65–66). In adopting this stance, the poster is in fact elaborating on their earlier positioning of two ethnic minority students as problematic (lines 28–9). The poster questions why ethnic minority students ‘always feel like [they are] being discriminated against’ (line 54) and unfairly ‘play[ing] the race card when any issue arises’ (lines 58–9), and why future social workers in Hong Kong do not learn ‘the local culture and language’ (lines 51–2). What then, we might ask, is the poster’s own position on an appropriate language policy for the Social Work degree? Although not clearly articulated, the following comment provides an indication of the nature of the stance being taken:

Some said using entirely Chinese or English is being fair, which I disagree with. When it comes to equity, if we’re all using English, it’ll be really unfair to those with poor [English], and a great advantage to those with fluent English. (lines 62–4)

Significantly the poster disaligns here with those who think English-only or Chinese-only pedagogy is fair (line 62), but again they make the point that an EMI only approach benefits those with ‘fluent English’ (lines 63–4). So the case being made is for some kind of bilingual approach, but with the blame for the policy being directed at the non-Cantonese-speaking students: ‘There’s no absolute fairness in this world, and I really don’t understand why just the two students are favoured over the vast majority’ (lines 71–2).

In the following section, we turn to a stancetaking analysis of the selected posts based on the aforementioned three themes. It should be noted that these necessarily overlap in some cases. A range of examples under each theme has been selected to reflect stances taken.

6 Analysis of interaction

7 Discussion and concluding remarks

Examining the range of stances taken up with regard to EMI in the HKU social work programme has enabled us to shed light on the complexities that neoliberal coloniality presupposes in this multilingual educational and professional setting. Despite a majority of stances in favour of a more flexible language policy (i.e. one that entails the use of Cantonese alongside Chinese, at least in some parts of the course), our analysis shows that EMI in this higher education setting and beyond is generally taken for granted. The most critical posters to the platform do not call for the complete removal of English, rather for a reconsideration of its use in some aspects of the course. This consent to the use of English overall, we suggest, reveals an acceptance of the geopolitics of knowledge production and consumption that privileges English as the language of global capitalism (O’Regan 2021), as well as its role as a local colonial inheritance. From an institutional perspective, HKU has little reason not to maintain its EMI policy due to the hegemonic status of English in academia, the pressure to engage in internationalisation and the neoliberalisation of higher education globally (Block 2022). However, the data reveal a significant degree of resistance to the exclusive use of English and point to a perceived need for change, as well as revealing the particular value of a stancetaking approach – and it is to this that we now turn.

As stated above, we follow Jaworski and Thurlow (2009) in viewing stancetaking – certainly as far as language choice in education is concerned – as having a profoundly ideological aspect. ‘Stance’ they argue, is ‘an act of self-identification and social identification by which I say something not only about myself and my view of the world, but by which I also make a judgement about you and about others as being like me or unlike me’ (p. 220). Clearly the attitudes to EMI expressed in the database are indexical of posters’ wider views of the world – the specifically colonial inheritance that English represents in Hong Kong and the role it plays in global edu-capitalism. However, and in our view equally significantly, stancetaking also tells us something important about the nature of ideology itself. Perspectives on ideology are notoriously divergent, with many scholars taking the view that it can be understood straightforwardly as a collectively shared social imaginary (van Dijk 1998). Here, however, we align with Holborow (2012, p. 29), who argues that ideology is most usefully understood as a one-sided representation of the world articulated from the perspective of the dominant or capitalist class, and one which is always ‘part-believed and part-rejected’. It is a perspective, she writes, which simultaneously ‘starts from the real but sees the real through a certain lens’ (p. 29). On this view, ideology is inherently unstable and contradictory. This, we suggest, is evidenced in the stancetaking we have analysed, and underscores the ideological dimensions of neoliberal coloniality. Consistent with the existing literature (e.g. Block 2022; Phyak and Sah 2024), EMI is found to be sustained by neoliberal ideology at both the institutional and individual levels in present-day Hong Kong and, consistent with its variegated nature (Brenner et al. 2010), local language-in-education policies and practices are shown to be shaped by colonial values and attitudes that are deeply entrenched.

Regarding Holborow’s point about ideology’s link to the real world, it is important to recognise that attempts to shape consciousness as to the nature of reality (one of ideology’s key functions) run into trouble when the lived experience of those on the receiving end suggests that reality is somewhat different. The initial poster begins with the charge that HKU is ‘disconnected from reality’ - a view at odds with how the university sees itself. From HKU’s perspective on the reality of the higher education landscape, EMI is framed positively and is integral to the university’s unique selling point as a global player and as a destination for international staff and students. However, the sustained and acrimonious challenge to this view articulated throughout the thread shows that reality from the perspective of a ‘down-to-earth programme’ (in the words of the initial poster) can be understood very differently. As Holborow (2012, p. 30) puts it, ‘[m]arket maxims may seem to make sense at one level but not at another […] Contradiction lies at the heart of both the construction of an ideology and how it is received’. Overall, we suggest that the thread reveals how neoliberal consent to an all-pervasive EMI policy (if not as yet, the socially constructed neoliberal desire for English itself), is far from guaranteed.

It is beyond the remit of this paper to explore in depth the practical problems posed by the language policy. However, one key issue we see as needing to be addressed is how those students who do not speak the local language can have their educational and linguistic needs addressed on such a programme. The data suggest that the current policy has the potential to fuel linguistic and racial tension between students – not only between Cantonese speaking students and international/ethnic minority students but with mainland Mandarin-speaking students as well. As the institutional language policy is partially mediated by its internationalisation efforts, administrators may have to re-think what ‘internationalisation’ means. Should international students, for example, be offered language classes in Cantonese to facilitate their integration into a more flexible EMI pedagogy? A truly internationalised space, we suggest, would embrace the linguistic resources of all parties involved by acknowledging the heteroglossic nature of language, while seeking to address all students’ linguistic needs. This would not be without problems, but it would have the value of recognising the complexity of teaching and learning in an equitable manner in an internationalised space by seeking to address the problems associated with the neoliberal coloniality of EMI.