Towards Critical ELF pedagogy

It is easy to think of English as a ‘thing’: a language with a structure, a history, and a set of norms to be learned. The problem, however, is that these assumptions do not sit comfortably with how English is actually used. English (indeed any language) is more than just an inventory of rules and words we learn and apply: it’s something people do with whatever linguistic resources they have available, often with strangers, often under conditions of uncertainty.

In an article published in January 2026 in World Englishes,¹ Henry Widdowson discusses this tension by juxtaposing World Englishes (WE) and English as a Lingua Franca (ELF), two pluralistic approaches to English language teaching (sometimes hastily conflated under the broader Global Englishes label). In the article, Widdowson argues that the two are not different labels for the same phenomenon, but two different ontologies, i.e., fundamentally different ways of understanding what language is. What I want to do in this post is explore what such a distinction means for language teaching and language teacher education.

To do this, I read Widdowson’s article alongside Ruqaiya Hasan’s work on literacy² and Norman Fairclough’s critical discourse framework.³ I believe that when read together, these perspectives show that debates about English –even when they seem very theoretical and disconnected from practice– are very practical, in that they suggest what kinds of linguistic agency learners develop and what futures this makes possible for them.

So, if you find yourself wondering whether the English you teach is really the English your learners need, this discussion aims to help you think that question through.

What are World Englishes?

Perhaps the most natural starting point for this detour into practice-facing theory might be to remind ourselves what World Englishes are. In a seminal paper written in 1985,⁴ Braj Kachru challenged hierarchical conceptualisations of English, which positioned ‘native speaker’⁵ models as the only legitimate variety, and other varieties as deficient approximations (or ‘broken English’).

Kachru’s ‘Three-circle’ model

In place of a binary with ‘correct’ and ‘deficient’ English (the varieties of middle- and upper-class native speakers and global learners, respectively), Kachru proposed a model with three concentric circles (Figure 1):

• an inner circle, representing the varieties in the UK, Australia and New Zealand and North America. This set of varieties projected norms outwards.
• an outer circle,⁶ which encompassed varieties where norms were still developing. Typically, these were post-colonial settings, in the Indian subcontinent and Africa, where local varieties of English were gradually gaining legitimacy.
• an expanding circle, including all the other ‘World Englishes’ around the world, that still looked to the inner circle for norms.

This re-conceptualisation of English(es) was profoundly important. Firstly, it challenged the idea that English was a single language. In Widdowson’s words,

[t]his designation represented English worldwide in terms of distinct varieties, but these were conceived not as subordinate lectal variants of the same language but as having superordinate lingual status as different kinds of language. 

Widdowson (2026). [original emphasis]

In language education, it also encouraged tolerance of accents and hybrid forms, and it proved fundamental in raising the professional status of teachers who were not native speakers of inner-circle varieties. By doing so, it paved the way for more democratic practices in English Language Teaching.

Limitations of the WE approach

Yet the World Englishes perspective, for all its transformative impact, remained bounded by several important theoretical constraints. The most important of these was that it seemed to use national communities as the primary unit of analysis, and it described fairly standard norms within each community. In other words, it assumed that everyone in the USA spoke the same kind of English, and everyone in Mexico spoke another, undifferentiated, variety.

“Edit[ing] out the socio-cultural or ethnic diversity” within communities (to use Widdowson’s phrasing) was, in some ways, an understandable simplification, and it may have made some sense in the mid-eighties. But such a description, which assumes stable, settled communities within nation-state borders, is much less tenable today. Here’s what Widdowson says:

Here what we so frequently find is internecine conflict within national borders, alongside the massive disruption and dispersion of established communities, with multitudes of migrants and refugees driven from their settled communal lives to seek to escape from poverty and persecution to more privileged lands—and with the communities in these privileged lands striving to keep them out as posing a threat to their own communal security. So although at a very general level, one can describe varieties of the language associated with national communities as WE, this can only be a very partial description of English in the world, which cannot be just pluralized as nationally named varieties confined within the Outer Circle.

Widdowson (2026)

In writing this, I don’t wish to ignore Kachru’s contribution. The World Englishes paradigm can and should be credited with challenging past colonialism. But it’s fair to say that it struggles with present transnational mobility and fluid, hybrid identities.

The English as a Lingua Franca perspective

In place of the norm-focused descriptions that World Englishes scholarship produces, Widdowson advocates a resource-focused description: English as a Lingua Franca (ELF).

In the ELF paradigm, emphasis is on what language users do, i.e., on how they negotiate meaning without relying on shared variety conventions. When speakers who have different linguistic backgrounds need to communicate, they make expedient, adaptive use of the English resources at their disposal – they make do with what English they have. This kind of meaning negotiation is, arguably, the predominant reality of English encounters today.

Comparing WE and ELF perspectives

Drawing on Widdowson’s discussion and other sources on ELF, I have put together Table 1, below, which contrasts the World Englishes and English as a Lingua Franca perspectives.

Aspect	World Englishes (WE)	English as a Lingua Franca (ELF)
Assumptions	Addresses past colonization & post-colonial identity Motivated by symbolic recognition of national identity and linguistic legitimacy Assumes stable communities and nation-states	Addresses present global realities (international mobility, displacement) Motivated by practical communication needs across diverse global contexts Recognizes global instability
Linguistic perspective	Views language as a variety Norm-focused: emphasis on describing how language is conventionally used Disregards linguistic variation that doesn’t represent national norms	Views language as communication Resource-focused: emphasis on how people exploit their language potential Embraces linguistic flexibility and unpredictability
Sociocultural perspective	Intra-communal: focuses on nationally-defined, settled communities, especially within the Outer Circle (typically post-colonial nation-states) Describes geographically distributed but locally bounded usage	Inter-communal: focuses in fluid, international users globally, who do not share community conventions of English usage Describes globally pervasive and transnational usage

As Table 1 shows, the World Englishes and ELF perspectives are fundamentally different. While Widdowson acknowledges that both approaches stem from a motivation to promote social justice, he argues –not unconvincingly– that ELF is much more relevant to the current geopolitical and educational realities.

More often than not, ELF is the only feasible means available to people who have to cope with problems of oppression, displacement and deprivation: refugees seeking to escape from these effects, for example, or officials seeking to counter and alleviate them. The problems that ELF users have in communicating are inevitably bound up with these real-world problems they need to communicate about. It is no exaggeration to say that the future well-being and very survival of most people on the planet are often dependent on a capability to make effective lingua franca use of whatever English is available to them as communicative resource.

Widdowson (2026)

Beyond WE and ELF

Up to this point, I have largely followed the terms of Widdowson’s (2026) argument. I note that the distinction he draws, between World Englishes and English as a Lingua Franca, is ontological: that is to say, it is about what kind of entity we believe language to be, and what kinds of communication become visible (or invisible) depending on the perspective we adopt. Accepting this distinction already has significant consequences for how we think about English language teaching.

To make these implications more explicit, I will try to place Widdowson’s argument in dialogue with Ruqaiya Hasan’s work on literacy and Norman Fairclough’s critical discourse framework. Read in this way, the contrast between WE and ELF is not only about alternative understandings of language, but about the kinds of linguistic agency that language education enables or constrains.

Language education
through the lens of literacy

The first of the building blocks in the argument I am developing is the distinction, by Ruqaiya Hasan (a prominent theorist in the Systemic Functional Linguistics tradition), between three qualitatively different ways of engaging with language and meaning:

• Recognition literacy is about mastering the formal features of language (sounds, grammar, vocabulary etc.) and using them with accuracy.
• Action (or functional) literacy refers to the ability to use language for practical purposes; this involves understanding existing genres and practices and using them effectively.
• Reflection (or critical) literacy treats language as a social practice embedded in power relations, and involves understanding how language constructs knowledge, identities and social relations.

What matters here is that these aspects of literacy are not simply cumulative skills. Rather, they represent different modes of engagement with language, and each one makes certain forms of agency possible while constraining others. Hasan’s perspective recognises developmental dimensions (in the sense that more sophisticated literacies build upon foundational ones). But crucially, she draws attention to the fact that educational systems systematically provide students with differential access to these literacies based on social class and other forms of stratification. For working-class students, these are often the recognition and action literacy levels; privileged students, on the other hand, gain access to reflection literacy. In this way, pedagogical choices reproduce social inequality.

Language education
through the lens of discourse

The second strand in my thinking stems from Norman Fairclough’s writings. If you are not familiar with his work on Critical Discourse Analysis, this builds on a three-level understanding of language:

• Treating language as text (or an object for description) involves a strong focus on the linguistic form (e.g., vocabulary, grammar, cohesion, and rhetorical choices).
• Treating language as discursive practice requires understanding how people produce and understand texts. It is about knowing genres, norms, modes of communication and communicative context (who speaks to whom, and how they position themselves), and about understanding how all of the above shape the activity of communication.
• Treating language as social practice is about connecting text to broader social structures and to the power relations that permeate them. This connects language to implicit questions of ideology, inequality, and institutional authority, and foregrounds questions about whose interests particular discourses serve and whose they marginalise.

While Hasan’s literacies and Fairclough’s CDA dimensions emerge from different theoretical traditions and serve different primary purposes, I find that share some structural parallels that make them analytically powerful when read together. Both draw attention to the consequences of foregrounding one dimension of language at the expense of others. Taken together, they help us understand explain how English language teaching can systematically privilege particular norms, forms of competence, and social interests, and while denying certain learners access to critical awareness of how language relates to power.

Spaces in the language education landscape

Both the literacies and the CDA framework move through three levels of engagement with language, progressing from surface forms to deeper questions of power and social practice. These provide a starting point for developing three unique spaces⁷ in the landscape of language education, each one of which involves different affordances and constraints.

Teaching language as a code

The first space foregrounds the material, formal features of language. At this level, learners approach language as a relatively autonomous system, and engage with its rules and patterns more or less independently of use. Pedagogical traditions, such as the grammar-translation method, audiolingualism and –more recently– many language learning apps, operate in this space.

When lodged in this space, language education effectively fosters linguistic skills (the ability to read and write), while at the same time encouraging conformity to norms defined elsewhere. In this sense, it lends itself well to producing a linguistically functional, compliant workforce, even as it limits learners’ scope for critical agency.

Teaching language as communication

In the second space, attention shifts to language-in-use. This involves practising communicative skills, while raising awareness of genre expectations (e.g., writing an effective letter of complaint), the role of communicative context (e.g., selecting appropriate levels of formality) and issues of positioning (e.g., thinking about which linguistic choices make us appear cooperative or assertive).

In social terms, such pedagogy produces socially calibrated language users. Applied to English language teaching, it is about helping learners participate effectively in established practices without questioning them.

Teaching language as power

The third space involves viewing language as social practice embedded in relations of power. Operating at this level, we might encourage learners to examine how language constructs social realities, how particular ways of speaking and writing become authorised, and how communicative norms connect with institutional interests and social hierarchies. Pedagogical work in this space foregrounds questions of ideology, inequality, and authority, and invites learners to reflect on whose voices are legitimised, whose are marginalised, and what the consequences of such decisions are.

Language education in this space is less concerned with smooth participation in existing practices than it is with developing critical awareness and transformative agency. It is about making it possible for learners to contest the discursive conditions under which communication takes place. In social terms, this orientation does not aim to produce a compliant or merely well-aligned workforce, but citizens capable of recognising how language serves particular interests and of acting upon that recognition.

Overview

We can now distinguish three fundamentally different configurations of language education, as shown in Table 2:

Space	Language is	Learners are	What we teach learners	Affordances	Constraints
Teaching language as code	An inventory to be mastered: vocabulary, grammar, syntax, cohesion	Reproducers of fixed rules	Decode, encode, and reproduce linguistic structures accurately	Basic communication, conformity to norms	Maintains norm conformity; limits agency to unquestioning reproduction
Teaching language as communication	A resource for interaction: Genres, communicative contexts, positioning, production and consumption of texts	Meaning-makers, flexible users	Use language for practical purposes, navigate genres effectively, and position oneself appropriately	Effective participation in established practices	Genres/norms accepted as givens, limited ability to transform practices
Teaching language as power	A site of struggle and transformation: Power relations, ideology, interests, social structures, inequality	Critical agents, social actors	Interrogate how language constructs reality, recognize whose interests are served, produce new knowledge	Critical awareness, transformative agency, resistance to dominant norms	Requires institutional courage

WE, ELF and spaces in the language education landscape

After this somewhat extended diversion into how language education might be configured, I would now like to return to the question of World Englishes, English as a Lingua Franca and the way they connect to English Language teaching.

From code to communication

A lot of mainstream English Language Teaching — the kind of teaching that Kachru challenged when putting forward the World English paradigm — operates firmly in what I describe as the language-as-code space. Despite the advent of communicative language teaching, learners need to master a fixed set of rules and words and occasionally genres and communicative patterns, as defined by native speakers. The goal of language learning is accuracy, as measured in standardised assessment, and deviation equals linguistic deficiency.

World Englishes appears more inclusive, in that it broadens the range of what counts as ‘acceptable’ or ‘correct’ language, to include local varieties of English. But even so, it remains firmly embedded in the language as code space. The norms are different (e.g., Indian English rather than British English), but their role in the education dynamics is unchanged.

What the ELF paradigm offers, as Widdowson rightly points out, is an alternative perspective — one which replaces emphasis on code with emphasis on communication. This is not an entirely straightforward move: Early ELF scholarship aimed to identify new norms, such as a set of core pronunciation features (Jenkins, 2000).⁸ However, recent work (including Widdowson, 2026, which prompted this post) has re-oriented to a more interaction-based understanding.

Communication under conditions of inequality

The shift from code to communication has been pedagogically necessary, but it is far from sufficient. Widdowson powerfully emphasises that many people globally use ELF in “high-stakes encounters”, where survival and well-being are at stake (e.g., refugees communicating with officials, workers in exploitative labour conditions, people addressing displacement and inequality).

In these contexts, communication is never politically neutral: Refugees must communicate, but in interactions where power is radically asymmetrical; workers must communicate, but within economic structures of exploitation; displaced people must communicate, but from a position of extreme vulnerability. For learners facing these realities, developing communicative capability without political awareness means that they can express themselves clearly while still remaining exploited, negotiate meaning while still being denied justice, and communicate effectively while still facing systematic marginalisation.

The question we need to ask, then, is how learners can move from adaptive functioning within existing structures to critical transformation of those structures?

Space	Mainstream ELT	World Englishes	English as a Lingua Franca
Teaching language as code	full alignment	full alignment	alignment with early work
Teaching language as communication	compatible with communicative methods	not compatible	full alignment
Teaching language as power	–	–	compatible with a critically reframed version of ELF

From communication to empowerment

The point I want to make is that language education cannot remain at the level of communication that ELF-oriented pedagogy enables, any more than it can remain at the level of code associated with mainstream ELT and WE.

The critical dimension encoded in the ‘language as power’ space is not an optional add-on or “critical thinking bonus”. It is pedagogically necessary because learners cannot develop full agency as language users without political awareness. It is ethically necessary because in contexts of high-stakes encounters, global inequality, and systemic oppression, providing communication skills without critical awareness is insufficient and even complicit. And it is practically necessary because learners who understand language as political practice are better equipped to navigate, negotiate, and transform the actual power-laden contexts they face.

Such concerns, I think, run through Widdowson’s critique of mainstream ELT and WE, and his advocacy for ELF pedagogy. Positioning Widdowson in a politically explicit vantage point on language education may seem like a bold move, especially given his long-standing resistance to overtly political framings.⁹ But the expressed concern with norm imposition, communicative asymmetry, and the conditions under which English is used in situations of inequality make for an argument that is political by implication if not by declaration.

Conclusion

In the conclusion of his paper, Widdowson draws attention to the call, originally in Kachru’s (1985) paper, for “new paradigms and perspectives for linguistic and pedagogical research” that are responsive to the needs of a changing socio-political reality. That call is still timely, and Widdowson’s article makes a convincing argument why ELF is fundamentally better suited than mainstream ELT and WE for this role. The argument I am making, though, is that ELF needs to more overtly embrace its political implications in order to respond to this call. A critical ELF is not one that abandons the achievements of code description (WE) or communication focus (ELF); rather it situates them within an analysis that recognises language education as fundamentally about power, justice, and transformation.

What would this mean in practice? It would mean that ELF pedagogy moves beyond teaching learners to negotiate meaning flexibly, to also teaching them to recognize when communicative breakdowns stem from power asymmetries rather than linguistic differences. It would mean examining not just how people communicate across lingua franca contexts, but whose communicative labor is valorised and whose is rendered invisible, whose accommodations are expected and whose are optional. The question for teachers, then, is not simply what English to teach, but what kind of agency we want our teaching to enable.

---

Footnotes

1. Widdowson, H. (2026). WE, ELF and ELT: Perspectives on English and applied linguistics. Advance Access. World Englishes. https://doi.org/10.1111/weng.70010 ↩︎
2. Hasan, R. (1996). Literacy in society. Longman. ↩︎
3. Fairclough, N. (2001). Language and power (2nd ed.). Longman. ↩︎
4. Kachru, B. (1985). Standards, codification and sociolinguistic realism: The English language in the outer circle. In R. Quirk, & H. Widdowson (Eds.), English in the world: Teaching and learning the language and literatures (pp. 11–30). Cambridge University Press. ↩︎
5. Of course, ‘native speaker’ is just short-hand for ‘visibly western’, as many people who grew up with English as their first language but didn’t have the expected skin colour found out. For more on the experiences of Visible Ethnic Minority Native English Speakers, you may want to read Eljee Javier’s (2015) work ↩︎
6. Am I the only one who is troubled by the fact that the ‘outer circle’ is actually in the middle of the diagram? ↩︎
7. If you happen to be familiar with complex dynamics systems theory, you will recognise that what I am describing here are attractors. ↩︎
8. Jenkins, J. (2000). The phonology of English as an international language. Oxford University Press. ↩︎
9. As some readers may remember, Widdowson (1995) challenged early approaches to Critical Discourse Analysis (CDA), as outlined in Fairclough’s Discourse and Social Change (1992). To summarise his reservations, he was concerned that CDA conflated text interpretation with ideological critique, lacked systematic linguistic evidence, and prioritised researcher bias over objective pragmatics. Fairclough (1996) responded by accusing Widdowson of misrepresenting CDA’s three-dimensional framework (text, discursive practice, sociocultural practice) and clinging to an apolitical, positivist view of language. ↩︎

---