Reimagining educational linguistics: a post-competence perspective

1.1 Terminology

The initialism ‘SLA’ is used in this article to refer to the canon of cognitively and psycholinguistically oriented research and theory conducted within a specific Anglophone tradition in AL since c. 1970 for reasons that will become apparent below, rather than research and theory on additional language learning (ALL) or teaching (ALT) in general, terms which are more appropriate and more inclusive, both descriptively and ethically (Anderson 2022; Leung and Valdés 2019). In contrast to much writing in the SLA canon (see below), I choose here to recognise important distinctions in contexts within ALL and ALT, particularly between, foreign and second language learning/teaching, following Kramsch (2002). I also acknowledge Leung and Valdés’ (2019) four context types (p. 4), which, while more nuanced, retain the broader distinction between FL (their “Type 1” or “Type 2” contexts) and SL (their “Language Instruction Directed at Minority Learners”) contexts. I have avoided the recent preference in North America for the term “world languages” (see ACTFL 2017) to refer to (modern) foreign languages as problematic because no language has ever existed or could exist that does not fall under the scope of this term as currently defined – the terms “languages” and “language teaching” seem to suffice here.

2.1 Chomsky and Hymes

While the competence–performance distinction was, of course, first made by Chomsky (1965), it is important to note that both Hymes (1972) and Bernstein (2000) chose to use the two terms differently. Both were aware of, and rejecting, rather than misreading, Chomsky’s characterisation of them. In Chomsky’s work in the 1960s (e.g., 1965) the primary focus of interest was the underlying competence (implicit language knowledge) of the ideal native speaker of a language operating “in a completely homogenous speech-community” with a perfect knowledge of the language (p. 3). However, because this idealisation never exists in practice, competence cannot be observed directly. Performance, defined as “the actual use of language in concrete situations” (p. 4), offers evidence of the (frequently imperfect) realisation of competence, but “cannot constitute the actual subject matter of linguistics, if this is to be a serious discipline” (p. 4). It should also be noted at this point that Chomsky was “frankly, rather skeptical about the significance, for the teaching of languages, of such insights and understanding as have been attained in linguistics and psychology” (1970, p. 37), referring to his own work at the time.

Hymes (1972) was willing to accept Chomsky’s notion of an underlying, innate, early acquired competence that was shared by “normal” (p. 282) members of a community. However, he disagreed with Chomsky concerning the nature of this competence and the role and importance of performance. Hymes argued, firstly, that we need to envisage a competence that includes culture as well as language; indeed his choice of the term “communicative competence” was intended to emphasise all communication, not just linguistic (p. 281), even though this was not necessarily how others subsequently understood it (e.g., Canale and Swain 1980, Savignon 1972; discussed below). Hymes also reframed competence “as the most general term for the capabilities of a person”, noting that “competence is dependent upon both (tacit) knowledge and (ability for) use” (p. 282, italics in original); our understanding of not only what is possible, but also the extent to which it is feasible, appropriate and realised (his oft-quoted four conditions; p. 281). Also differently to Chomsky, Hymes argued that performance, which he often used interchangeably with “behavior” (e.g., pp. 280–281) to reflect his equal focus on culture, was important because analysis of its heterogeneity could reveal the commonalities of the underlying competence: “If one starts with analysis of the diversity, one can isolate the homogeneity that is truly there” (p. 276). While it is often overlooked, particularly by applied linguists, in the final section of his article, Hymes also briefly contrasts “a ‘long’ and a ‘short’ range view of competency” (p. 287), with the latter closer to Chomsky’s construct, and the former (long-range competency) recognising “continuing socialization” throughout life (here he was primarily discussing disadvantaged children from minority linguacultural backgrounds) that moved beyond early-acquired, tacit knowledge to recognise schooling and also Bernstein’s (1965) early work on restricted and elaborated codes.

2.2 Bernstein’s competence and performance models of education

As part of his wider research and theory on the sociology of education, Basil Bernstein (e.g., 2000) was interested in understanding the causes and impact of the increased prevalence of “competence models” – by which he meant all models that prioritised competence as a focus of pedagogy or an outcome of education – that occurred during the 1960s and 1970s. He identified notions of competence not only in Chomsky’s and Hymes’ work, but also in that of Jean Piaget, Claude Levi-Strauss and Harold Garfinkel. He noted similarities between these varied discussions of competence (albeit with some variation), which typically involved “practical accomplishments” that were generally “culture free”, often “biological” in their origins and “intrinsically creative and tacitly acquired” (p. 42). He observed that this notion of competence fitted well with the liberal progressive – even radical – zeitgeist of the 1960s before going on to make a critical observation about competence, echoing Chomsky’s concern (above):

What is at issue is how a concept which arose in the intellectual field, and whose authors had little or no initial connection with education, came to play such a central role in the theory and practice of education. (p. 44)

While Bernstein identified a number of potentially attractive features in competence models – features that we would today often recognise as components of “learner-centred education” in its various guises (see Anderson 2023b; Sriprakash 2012) – his discussion is also critical of such models, including their high cost of implementation, their “invisible pedagogy”, the tacit modes of control involved, and the difficulty, particularly for disadvantaged learners, in acquiring competence (also a concern of Hymes 1972), anticipating the more recent criticism in AL of (neo-Chomskyan notions of) native-speaker competence as a goal for additional language learners:

… this idealism of competence, a celebration of what we are in contrast to what we have become, is bought at a price; that is, the price of abstracting the individual from the analysis of distributions of power and principles of control which selectively specialise modes of acquisition and realisations. (Bernstein 2000, p. 43)

In contrast to competence models, Bernstein adopted the term “performance models” to refer to those that are more consistent with the historical norms of mainstream education, in which learners are socialised into the discourses of the subject or institution (“acquirer” means learner in Bernstein’s framework):

… a performance model of pedagogic practice and context places the emphasis upon a specific output of the acquirer, upon a particular text the acquirer is expected to construct and upon the specialised skills necessary to the production of this specific output, text or product. (p. 44)

Bernstein also noted that while performance models are more controlling in their practices, because of their higher visibility, lower levels of expected learner autonomy and explicit transmission of social norms, they can be seen to be more teachable than competence models, involving explicit evaluation criteria and more directive control (stronger “classification” and “framing” in Bernstein’s terms; e.g., 1973) that enable learners, particularly those that are more disadvantaged, to acquire the required discourses, codes and realisation rules (i.e., literacies) of education. As Ivinson (2011, p. 16) notes, in performance models “the criteria for achieving legitimate texts are made relatively explicit to children within local classroom cultures. Success relies less on producing personal or original texts and more on reproducing specialist-subject criteria.” While Bernstein’s discussion of these models is restricted to Western educational contexts, where increasing educational funding throughout the twentieth century allowed competence models to have a widespread influence on private and national education systems, lower cost performance models have remained the norm in low- and middle-income national contexts worldwide (Barrett 2007; Sriprakash 2012) where per child investment in education may be only a fraction of that in OECD countries (Anderson 2023b).

Thus, while Bernstein’s discussion of competence models identifies key features of Chomsky’s competence that were shared with other writers at the time, his use of the term “performance” is very different to that of Chomsky, positing it as the second of two dichotomous approaches (competence and performance models), rather than as one of the two inter-related (not dichotomous) aspects of cognition and behaviour that were of interest to both Chomsky and Hymes.

2.3 The competence model of applied linguistics

The origins of the competence model that came to dominate applied linguistics are closely intertwined with the origins and early development of both applied linguistics as an academic discipline and SLA studies as its primary area of research focus, both of which emerged in the 1960s and largely in North America and the UK (Howatt and Widdowson 2004; Hudson 2020; Hult 2008). The single and most powerful influence in these origins was Chomsky’s work (e.g., 1965), which centralised competence as the construct of interest (as discussed above), but also, critically, his vision of a homogonous (i.e., monolingual, monoglossic), native-speaker, target language community as the wider social context for learning (Flores and Rosa 2022).

It is in Chomsky’s work, not Hymes’, where we first find many of the core constructs and assumptions that went on to dominate the competence model of ALL in AL and SLA discourse communities. Here, language learning is first popularised as a process of acquisition; Chomsky’s “acquisition model” (1965, p. 30) provided the basis for theorising and describing early childhood language learning yet was also embraced unproblematically by the then nascent SLA community as an appropriate construct for additional language learning (see Anderson 2022). In a work acknowledged as seminal to the establishment of SLA as a field, Corder (1967) accepts Lambert’s (1966, as cited in Corder 1967) distinction between acquisition and learning (subsequently popularised in the work of Stephen Krashen), arguing for the primacy of the former and assuming that “the processes that take place in the learning of first and second language … apply in both circumstances” (Corder 1967, p. 163, italics in original). Following Chomsky, early SLA researchers frequently focused on the acquisition of grammar, particularly orders of acquisition, perceiving this to be their primary area of interest (e.g., Bailey et al. 1974). Further, because Chomsky’s area of interest was early childhood language learning, his focus was on the acquisition of the spoken language. This phonocentrism was imported uncritically into early SLA research and theory (Pennycook identifies it as foundational to the AL movement; 1994, p. 135) along with the assumption that “acquisition” of the spoken language was primary. As Kramsch observes (2002, p. 61), “SL research … was mainly psycholinguistic in nature. It sought to explain and predict the acquisition of English by immigrants to English-speaking countries”. In this sense the AL community envisaged, characterised and researched “SLA” as a process of assimilating or integrating the other to our norm; what might be called the integrationist objective of ALT dominant at this time.

Thus, as a result, partly of Chomsky’s influence, but also of the primarily Anglophone-community origins and the integrationist objective of ALT within the early AL movement, a large number of assumptions concerning the nature of language and language learning were imported uncritically into foundational AL theory and research, despite Chomsky’s (1970, see above) concerns regarding their relevance for language teaching:

1. assumptions concerning the “target language” (TL) community as broadly homogenous and monolingual;

2. a characterisation of the “native speaker” as a person born and/or socialised unproblematically into the target language community at an early age;

3. depictions of the teacher as a “native speaker”, and their standardised variety of the target language as the norm, model and goal (these three were often conflated) of SLA;

4. a vision of the classroom as located in the wider TL community, where “exposure” and “authentic” texts are abundant;

5. the othering of learners as alien and/or disadvantaged, requiring integration;

6. assumptions that the only valid form of learning is acquisition of the spoken language and the related assumption that therefore only implicit learning (Krashen’s “acquisition”) is valid;

7. a primary interest in the systemic acquisition of grammar and comparative neglect of other areas of learning (e.g., lexis, genre conventions, literacy, interculturality);

8. the primacy (both sequentially and hierarchically) of spoken over written language;

9. assumptions that “exposure” to “authentic language” constitutes the primary, and most appropriate, input for learning;

10. models in which learner language is characterised as deficient (e.g., “interlanguage”) until and unless it approaches the (typically monolingual, white, middle class) spoken variety of the NS.

Many, if not all, of these assumptions would today be critiqued, if not rejected as flawed by many writers who identify as ALs (see, e.g., Douglas Fir Group 2016; Flores and Rosa 2022; Leung and Valdés 2019; Ortega 2014). And while mainstream AL itself, and much SLA research, particularly since the turn of the century, can be seen to have moved towards more sophisticated, more socioculturally aware, more multilingual and more inclusive models of ALL (e.g., Douglas Fir Group 2016), these same assumptions have continued to exert a strong influence (see Anderson 2022), particularly because they became foundational to the new pedagogy that emerged in the 1970s and remains dominant in notions of best practice in AL and SLA communities today: Communicative language teaching (CLT).

2.4 Competence in CLT

There is a pervasive misunderstanding, if not myth, in applied linguistics that it was Hymes’ notion of communicative competence that provided the ‘big idea’, and primary goal of CLT. In reality, however, what was embraced was the term itself, used simply to mean “linguistic interaction in the target language”, and not Hymes’ theorisation of it (Paulston 1974, p. 347). For example, Sandra Savignon’s early and highly seminal research and scholarship on the acquisition of “communicative competence” (1971, 1972, 1975) never once mentions Hymes, nor his four conditions (see above); neither does Wilga M. Rivers’ writings in TESOL Quarterly that built on Savignon’s work (e.g., Rivers 1973). Even Canale and Swain (1980), who proposed the three, then four (Canale 1983) strands of communicative competence (grammatical, sociolinguistic, strategic and discourse) that are now familiar, and did discuss Hymes’ construct, nonetheless depicted it very differently to Hymes, instead locating it firmly within the grammar-focused, integrationist, native-speakerist Chomskyan tradition of AL as discussed above:

[t]he communicative approach that we envisage is thus an integrative one in which emphasis is on preparing second language learners to exploit … those grammatical features of the second language that are selected on the basis of, among other criteria, their grammatical and cognitive complexity, transparency with respect to communicative function, probability of use by native speakers, generalizability to different communicative functions and contexts, and relevance to the learners’ communicative needs in the second language. (Canale and Swain 1980, p. 29)

In its strongest subsequent realisations, such as task-based language teaching (TBLT), CLT became the epitome of a competence model, as characterised in Bernstein’s framework (see above). It prioritises the implicit “acquisition” of spoken language through weakly classified, learner-independent, democratically framed use of “authentic texts” and interaction in “naturalistic” and autonomous TL communities, and is typically assessed through task achievement in primarily transactional language use (e.g., Ellis 2020; Long 2015; Nunan 2004).^[1] At the time that CLT emerged, this was an ambitious (radical, even) goal that proved useful, particularly for adult professional use, ESP and preparation for academic study (see Allwright 1976; Howatt 1988; Johnson and Morrow 1976). Importantly, early discourse on CLT was emerging primarily in discussion of English language teaching (ELT/TESOL) and was inextricably linked to Anglophone SL communities (mainly UK, North America, Australasia) and predominantly in HE and adult contexts, where many of the above assumptions must have seemed tenable (Holliday 1994; Phillipson 1992). As ELT grew into a lucrative business, funding university AL departments, teacher education programs and vast publishing houses, all of the above became embedded in a new industry that benefited those who were privileged enough to board the (ostensibly white, native-speaker, male-dominated) bandwagon (Pennycook 1994; Phillipson 1992).

2.5 The calving of educational linguistics

It is notable that the foundation of educational linguistics occurred in the mid-1970s, when both SLA and CLT were on the ascendence and the ELT “industry” was rapidly expanding, disseminating its competence model of learning around the world: “the predominant notion of AL is that it serves the needs of language teaching, particularly ELT” (Hult 2008; p. 14; also Phillipson 1992). It is on this backdrop that Spolsky (1974, 1978 proposed the term “educational linguistics” to refer specifically to aspects of AL that had relevance to language education (Hornberger 2022; Hult 2008; Spolsky 1978). The agenda he proposed, in part out of frustration of the “application” model of AL at the time (Spolsky 1978, p. 2), was well-intended, focussing on the needs of education and educators (rather than linguists), and offering a vision of EL that emerged from questions of practice rather than academic theory (Hornberger and Hult 2008). And while Spolsky’s first definition and description of EL occur in a paper on the language development of Navajo children that aimed to enable the Navajo “to make informed decisions about language education policy” (Spolsky 1974, p. 3), his vision of EL was firmly rooted in the integrationist objective of AL, as discussed above:

The scope of educational linguistics is the interaction of formal education with language. It is concerned with describing and analysing language education in all its aspects. This involves the assessment of a child’s communicative competence at the time he enters school at subsequent stages in his education, and the whole range of activities undertaken by the educational system to bring about changes in its pupils’ linguistic repertoires. (Spolsky 1974, p. 2)

Like many in the North American AL community at this time, Spolsky (1974, 1978 makes extensive use of the term “communicative competence”, without once referencing Hymes’ (1972) original paper or wider theory; yet his referencing of Chomskyan theory was extensive and meticulous (e.g., 1978, Ch. 8).

As his opinion settled, Spolsky sometimes adopted broader definitions of EL (e.g., 1999, p. 1; “the intersection of language and education”). Nonetheless, he continued to view its primary goal being to “offer information relevant to the formulation of language education policy and to its implementation” (1974, p. 3; also, 2008, p. 2), rather than, say, pedagogy, curriculum or teacher education. In this sense, Spolsky’s attempts to establish the new field on the most appropriate footing were undermined by the bias in his own interests and concerns. Indeed, without reason to question the foundational Chomskyan assumptions of AL, and with a strong belief in the competence model inherent in CLT, Spolsky and his collaborators, in essence, attempted to ringfence a large part of AL (including much of SLA) as EL territory, and this has continued largely unchallenged (albeit often ignored or simply unnoticed) ever since. For example, the authorship of the first Handbook of Educational Linguistics (Spolsky and Hult 2008) includes a strong SLA focus and reveals a strong Anglophone bias, with 83 % of its contributors with first affiliations to institutions in Anglophone-dominant countries, particularly the USA (56 %).

Thus, as a result of the above foundational assumptions, the Anglophone origins of both AL and EL, and the unstoppable behemoth of ELT, educational linguistics matured into applied linguistics business as usual (AL = EL = ESL = CLT), promoting an integrationist view of language, language learning and teaching that was predominantly neo-Chomskyan, Western, Northern and Anglophone – a learner-centred, democratic, liberal/progressive mode “competence model” in Bernstein’s terms (2000).

3.1 Kramsch’s view from foreign language research

Google Ngram data indicates that the term “foreign language teaching” increased gradually in printed use from the start of the 20th century, and then rapidly in the post-war era, during the expansion of ELT (see Figure 1). In contrast, the term “second language teaching” came into common use in the 1960s, supporting the narrative presented above. It is notable also that the former term has declined in printed use more recently, corresponding with the rise of SLA research, which has consistently attempted to de-emphasise the difference between the two terms, perceiving that SLA findings generally apply to both context types (e.g., Ellis 2008; Larsen-Freeman and Long 1991; see Leung and Valdés 2019). In contrast, Kramsch’s (2002) article Standard, norm, and variability in language learning: A view from foreign language research makes a sharp distinction between the two – not only the context and nature of FL and SL learning environments (which SLA research sometimes acknowledges, albeit briefly), but also concerning important differences in the teaching environments involved, particularly the extent to which curricula, curricular outcomes, and norms typically differ between the two. Consistent with the argument presented above, Kramsch notes:

Since the seventies, the findings of SLA research have progressively subverted FL methodology, both liberating and enriching, but at the same time shortchanging FL learning. By focussing on an authentic native speaker norm, it has highlighted the rich social context of language use and measured learners’ progress according to how well it conforms to the economy of verbal exchanges on the native speaker linguistic market. But in so doing, it has tended to reduce the FL to its currently realized communicative uses in situations of everyday life, in approximation to and conformity with stereotypical native speakers, thus often silencing other, FL-specific forms of language potential. (2002, p. 60)

Figure 1:

Frequency of terms: “foreign” and “second language teaching”. © Google Books Ngram Viewer, 2023.

She observes that, as a result of the influence of the AL/SLA movement, “FL teaching started to espouse the same norms as SL teaching”, that “the globalization of the textbook publishing industry and of teacher training programs further served to extend the NS norm from SL to FL learning” and that even “FL research [from the early 1980s] has contributed to the SLA research agenda” (p. 62), mischaracterising FL learners as impoverished SL learners who are, or should be, striving towards the same outcomes, albeit in more challenging contexts. She points out that, of ten SLA research generalisations presented in Lightbown’s seminal article (2000), eight are predicated on SL linguistic norms, and the remaining two (items 9 and 10) describe challenges that many FL learners, with limited exposure, input, motivation and appropriate support, are unlikely to overcome.

Kramsch then goes on to document how FL norms in three mainstream educational systems (those of USA, France and Germany), while varied, posit curricular outcomes that are simultaneously much wider and richer (e.g., including developing intercultural awareness, social and emotional learning, metacognitive language analysis skills, etc.) than those of SL contexts, and exhibit no need to make reference to NS norms; indeed, in Germany it is the “native speaker [who] is viewed as “the Other, the Foreign”” (p. 67) and FL learning as “a primary component of the learners’ search for their own identity” (p. 67).

Thus, prevalent assumptions in AL and EL concerning what language is, its manifestations in society, how it is learnt, and even how it should be taught all originate in the strong competence model adopted by SLA research and SL instruction, as discussed above. These have only subsequently been applied to FL contexts. In this sense, FL contexts become an add-on, rather than a fundamental starting point for educational linguistics. As a result, approaches to ALT that originate in SL contexts, especially CLT and TBLT, have been “exported” to FL contexts worldwide, often uncritically (Rabbidge 2019). When these have frequently proven to have less than the expected impacts on learning (see, e.g., Canagarajah 1999; Holliday 1994; Hu 2002), rather than questioning these fundamental assumptions, there has been a tendency to blame the context itself, the curriculum or assessment practices, the teachers, the learners or even the education system as a whole (the “large class” problem, for example; Coleman 2018), rather than recognising these as common constituent elements of primary and secondary education worldwide (Ekembe 2016) and, instead, critically evaluating the validity of the approaches being exported (Anderson 2023b; Canagarajah 1999; Holliday 1994; Tabulawa 2003).

3.2 Key characteristics of additional language learning in basic education

Bringing together the above discussed changes in contexts for language learning and teaching, from BANA to TESEP, and Kramsch’s observations on FL instruction as by far the most common type of ALT in the world today, I here identify a number of important characteristics common to many ALT contexts, particularly for FL instruction of English, but also often for many SL and bilingual education programmes around the world that are often overlooked by applied and educational linguists:

1. Class characteristics: The vast majority of learners in basic education worldwide study in classes of over 25, often over 30 pupils (Villoutreix 2012). Such classes are often not differentiated by ability level for language instruction, meaning that some secondary classes commonly include learners spanning one or even two proficiency bands (e.g., A2 to B2). As a result, partly of class sizes and ability differences, but also of learner characteristics discussed below, behaviour management challenges are an everyday reality for many teachers, particularly in secondary grades, and often lead to lower time on task as a result.

2. Learner characteristics: For many language learners in basic education, the subject language plays comparatively little role in their lives outside the classroom (e.g., Zein 2022), even in the case of English in countries where it may have official SL status, such as India or Kenya (see, e.g., Annamalai 2005; Dhillon and Wanjiru 2013). As a result, learner opportunities for exposure to and use of the subject language are often limited to their studies, particularly for more disadvantaged learners, and those in rural areas (Zein 2022). Further, because language learning is compulsory in basic education, learner interest, engagement and motivation, both to study additional languages, and to use them in class, varies considerably in both type and quantity, particularly for teenage learners, who are undergoing complex, basal identity development at such ages that interplay with ALL and ALT in complex ways (see, e.g., Coleman et al. 2007; Lamb 2012).

3. Teacher characteristics: Low teacher proficiency in one or more of the languages of the classroom, particularly the subject language, is an everyday reality in basic education around the world (e.g., Hayes and Raman 2015; Kaslan et al. 2020; Mody 2013; Zein 2022).^[3] In the case of English, this challenge is exacerbated by global shortages of English teachers (Patel et al. 2023; Zein 2022). Depending on the country, some or many such teachers may receive general educational training and little or no subject-specific pedagogy, particularly at primary level (e.g., British Council 2018; Patel et al. 2023; Zein 2022), meaning that their pedagogic practices, beliefs and interests may often be those of generalist teachers. Nonetheless, many language teachers also share more-enabled languages (first and/or community languages) with their learners – an important affordance discussed further below.

4. Curricular and assessment characteristics: The vast majority of formal language learning is bound and regulated by national, state or private educational curricula and syllabi, compulsory (often high-stakes) assessments and important child welfare, social and emotional development and pastoral support guidelines which provide the primary content and conditions for learning (Kramsch 2002). As such the development of learner language proficiency is typically only one of several prescribed curricular outcomes.

5. Modality characteristics: Written language, particularly in the learning of English, is likely to be as important as spoken language, if not more so. Not only are exams predominantly (or wholly) delivered and taken in written modalities in primary and secondary education in many national contexts (Patel et al. 2023; Zein 2022), future studies (e.g., HE and FE, online courses) and assessments for and during these are likely to involve significant amounts of written language, frequently in (e.g., EMI or CLIL programmes), or with, globally dominant languages, such as English.

6. Language characteristics: While all classroom communities engaged in ALL are, by definition, either bilingual (including the subject language) or multilingual, the learners and teacher often (not always) share, and frequently use, a “more-enabled” (Durairajan 2017) alternative language to the subject language (e.g., Copland et al. 2022). This community language constitutes a key affordance for so-called “L1 use” in the additional language classroom. Recent discussion of the complex, flexible languaging practices involved characterise it as translanguaging and emphasise the importance of the learners’ full repertoire in facilitating learning (Anderson 2018; García and Li 2014).

These realities are seen as common characteristics of ALT in large-scale (especially state-sponsored) systems, even in Europe (British Council 2018), which themselves are both embedded within, and reflective of social practices of the countries and cultures in question (Bourdieu and Passeron 1990; Bruner 1996). They are at variance with the majority of the foundational assumptions discussed above, and are frequently overlooked or ignored by AL and EL commentators and researchers on issues of methodology or curriculum, particularly by those who conduct research in high-income countries and those who investigate only elective tertiary education. To provide just one example of this neglect, four out of five institutional characteristics discussed by Long (2015, pp. 371–372) as necessary for the successful implementation of TBLT do not hold true in the contexts here described.

4.1 Competence and literacy: a continuum framework

At this point in the discussion I would like to step back from Bernstein’s (2000) characterisation of competence and performance models as simply dichotomous. Any critical historical perspective must recognise that to escape the influence of a sociohistorically-dominant construct such as competence since the 1960s, a reimagined EL must also avoid any reactionary pendulum swing to the opposite polarity and, instead, attempt to reconcile, or at least relate, apparently conflicting models of education. As Sfard (1998) notes in her discussion of two conflicting metaphors for learning, “metaphorical pluralism embraces a promise of a better research and a more satisfactory practice” (p. 10), also observing that “an adequate combination of [both] metaphors would bring to the fore the advantages of each of them, while keeping their respective drawbacks at bay” (p. 11). And while many scholars working in the competence tradition of SLA may perceive their theories of learning incompatible with others in education, there is evidence that individual institutions (see, e.g., Ivinson and Duveen 2006) and teachers (e.g., Anderson 2023b; Barrett 2007) can balance, or even move flexibly between aspects of performance and competence models depending on context-specific factors. Indeed, as both Barrett’s (2007) study in Tanzania and Sriprakash’s (2012) in India find, suitably identified aspects of both models can be combined in an eclectic approach, something that has long been argued for in the literature on language teaching pedagogy (e.g., Bax 2003; Kumaravadivelu 1994; Ur 2013). Providing appropriate outcome measures are selected (those that are valued by key stakeholders in education in a given context, rather than SLA researchers), research can and does indicate that varied, eclectic approaches can be effective (e.g., Anderson 2023b). As such, a post-competence educational linguistics does not mean abandoning the recognition of the importance of developing learner proficiency in additional languages; rather it recognises that this will often be part of a wider range of educational outcomes, including the ability to communicate appropriately across one’s full range of contexts of use, discourse types, modalities and repertoires, and the ability to recognise, understand and contextualise languages and language use within wider understandings of the world, as Kramsch (2002) argues. Numerous other cross-curricular outcomes that education systems, or specific stakeholders, may choose to prioritise should also be recognised (e.g., social and emotional learning; core competencies; see, e.g., Wang and Luo 2019).

With these observations in mind, Figure 2 offers a proposed continuum between two polarities for educational models (termed “orientations”) based on Bernstein’s competence and performance models, but with consideration specifically of additional language education in mind. As such, and given the potential confusion that Bernstein’s choice of the term “performance model” may cause, particularly in our field, I have chosen to reframe the alternative to “competence orientation” as “literacy orientation”, a term which, as far as possible, encapsulates the values of this orientation, particularly in relation to language pedagogy, while still retaining the polarity of Bernstein’s performance models. The two poles are characterised as follows:

• Competence orientation (language as innate faculty): models that view language education as the enabling of the faculties of natural language learning. Language is a system to which learners need to be exposed, given opportunities to communicate in, and to learn through trial and error.

• Literacy orientation (language as social construct): models that view language education as the development of the means of cultural interaction. Language is a system which learners need to be introduced to, and taught the rules, discourses and evaluation criteria of, both in formal education and wider society.

Figure 2:

The competence–literacy continuum.

I have also included an intermediate orientation between these polarities, contingently labelled “pluralist”, incorporating the ability to locate features partway between the two polarities and/or the ability to integrate elements of both polarities alongside one another. The layout of Figure 2 recognises that while specific programmes or educational systems may be eclectic in their combination of features, it will be possible to locate most (on balance) at specific points on the continuum; as such, it includes options for both weaker and stronger orientations on both sides. And while Figure 2 may be seen to present the pluralist orientation as a best-of-both-worlds option, it could also be argued that it runs the risk of diluting either or both alternatives unnecessarily for some contexts. In this sense the competence–literacy continuum should be seen as a tool that raises awareness of the full range of language education orientations without necessarily prescribing any one over the others.

4.2 Implications of a reimagined educational linguistics

By highlighting the bias of the competence orientation, identifying its often-neglected (in AL and EL) alternative (the literacy orientation) and the possibility for a pluralist orientation to exist between these, the above proposal brings with it a number of implications that are briefly explored here.

Firstly, the above continuum recognises a range of teacher–student relationships and interactions in the classroom appropriate to varying context types and goals. While competence orientations typically stress the role of teachers as low-intervention facilitators of acquisition, this contrasts with the more directive role that teachers in basic (K9 or K12, depending on the system) education may need to adopt when supporting learner cognitive, social and emotional development, for which guiding and scaffolding roles may be more appropriate. Likewise learners may assume roles varying from passive subjects – when mastering specific discourse types and genres – to increasingly autonomous agents of their own learning, as wider goals, individual needs, cultural norms and curriculum foci change. Related to this, the continuum may help to counter the often oversimplistic “learner-centred” narrative of competence orientations, which typically prioritises learner–independent activities over teacher-led instruction, despite the fact that both are frequently found to be essential to basic education (Anderson and Taner 2023; Campbell et al. 2004; Hattie 2009).

Secondly, a post-competence perspective enables us to move beyond the oversimplistic, often damaging and frequently unexamined distinction between classroom and “real world” that has dominated competence models in EL for the last 50 years. The frequent allusions to authenticity (e.g., authentic texts), naturalness (e.g., natural English) and reality (e.g., real communication) as features of a desired world of competence beyond the classroom (and often beyond the reach of so many language learners worldwide) can be rejected (see Cook 1997; Widdowson 1998). Classroom communities can themselves be viewed, researched and represented as real communities, valid communities and valued communities – their practices and culture seen as no less important than those of other loci of social practice (Bruner 1996). Textbooks are authentic texts (for learning), teachers and learners are also users (and vice versa), and their idiolects (including repertoires, grammars, accents and discourse features) are seen not as defective, but as dynamic features of their current identities and emergent selves.

Thirdly, the phonocentric bias of the competence era may be rejected along with the bias of SLA discourse that prioritised implicit “acquisition” (as typically measured through spontaneous speech) over explicit “learning”. Writing is no longer seen as necessarily inferior, or secondary to speech, either as a classroom practice or measure of learning. Both modalities may be important (depending on the programme orientation) and valued. Explicit language knowledge, so often valued by FL teachers around the world, becomes not just a resource of limited use towards a more important end (as often characterised by SLA researchers; e.g., Ellis 2008), but itself may be a primary goal of language education, useful for developing more complex literacies and metalinguistic skills that will be of greater use across varied subjects and practices in lifelong education. Intermediary and hybrid modalities may also be valued (informal messaging registers; voice to text and text to voice, etc.).

Fourthly, the multi-faceted, externally-imposed syllabi and curricular goals that invariably shape education in FL contexts can be recognised as valid norms of formal ALT, rather than as obstructions to “real learning” or “needs-based” curricula. The outcomes of such curricula are frequently valued by many learners, parents and other stakeholders in education today (e.g., governments as sponsors of basic education). Such goals and stakeholder perspectives may become a primary focus of interest and understanding for EL, rather than simply critiqued as misguided or replaced with pared-down, simplistic (and ultimately problematic; see Leung 2022) constructs of proficiency or competence as the only valid goal.

Fifthly, assessment in EL, as a means to measure learning, may also become more varied, potentially recognising diverse curricular goals such as the development of mediation, translation and metalinguistic skills, intercultural and translingual awareness and literary analysis to counter the invariably monolingual, proficiency-obsessed focus of much language testing today (see Leung 2022). Because many of the skills in question can be assessed through written texts, valid assessments become easier to design and administer, and more cost-efficient as a result – a key concern in many public educational systems.

And finally, while the so-called “target language” (i.e., the named additional language that typically constitutes the substantive curricular focus; e.g., English) will inevitably be central to the goals and assessment of much ALT, the learners’ more-enabled prior language(s), as the repertoire to which additional languages are added, is/are also recognised as the foundation for learning, particularly in models with strong literacy orientations. Adaptive repertoire deployment, as typically advocated in translanguaging models (e.g., Cenoz and Gorter 2021; García et al. 2016; García and Li 2014), is also recognised in the framework, reflecting language use practices in societies across the planet (Canagarajah 2013). In this sense, the full discursive repertoire of the classroom community may be valued and leveraged appropriately, enabling learners and teachers – as translingual practitioners (Anderson 2018) – to move between and gain awareness of language varieties, lects and registers, and to appropriate these for different uses, such as achieving curricular outcomes, engaging in authentic communication and exploring individual and shared identities as required.

4.3 A research agenda for a reimagined EL

The arguments presented here also have potential implications for future research in EL, including, but not limited to the following, much of which recognises the importance of theorising from classroom communities (as Spolsky intended), rather than towards them from linguistics:

• Reimagined contexts for research: The bias in AL for researching language learning in tertiary contexts, which are inevitably more accessible to most academics, is unlikely to change in this era of continued cuts to research budgets and available time in many countries. Nonetheless, research on hitherto comparatively neglected primary and secondary ALT, particularly in middle- and low-income contexts may warrant increased emphasis, particularly research which seeks to understand these contexts and their literacy or pluralist orientations by investigating aspects of curricula, the practices of teachers and the experiences of learners. Such research is likely to provide key foundational evidence on which other EL research can build.

• Reimagining collaboration with researchers of other school subjects: While the pre-service education of teachers of languages often happens alongside teachers of other school subjects (e.g., through modern foreign language teacher certification), it is notable that, due in part to the origins of AL, and in part to the subsequent dominance of the competence model in AL, many educational linguists often interact little, or not at all, with educational researchers of other subjects. This is influenced in part by departmental structure in some universities (AL departments may be separate from educational ones), conference-going practices, journal preferences and other publication choices. By recognising literacy and pluralist orientations as no less valid than competence orientations, a post-competence perspective opens up a wide range of useful opportunities for collaboration with researchers of other subjects that may not have been seen as valid from the strong competence perspectives of the past. For example, research on similarities and differences in teacher roles and development, learner identities and behaviour, cross-curricular themes or assessment practices across different subjects may offer insights that enable educational linguists to better understand language classrooms as part of wider school and societal communities and – importantly – also to increase interest in researching languaging practices in classrooms of other subjects (see, e.g., Nikula and Moore 2019; Smit 2018).

• Reimagining research into written literacy development: Likely as a result of the phonocentric bias of AL, childhood written literacy development has rarely interested EL researchers. Yet this is clearly an important goal of all formal education, particularly, but not only, at primary levels (e.g., Indian learners may need to develop literacy in three or more scripts, which may continue into secondary education). Furthermore, it overlaps in complex ways with other aspects of additional language development, through, for example, phonics instruction – an important area of research almost entirely neglected by applied linguists, despite Bialystok’s observation that “what is beyond dispute is the crucial role that phonological awareness plays in learning to read an alphabetic script” (2001, p. 168). Among potential issues for a research agenda relevant to EL is the task of identifying appropriate phonics syllabi for additional language learners whose first written script is non-Latin, or simplified lexical content for initial literacy instruction for learners in highly multilingual low-income contexts (e.g., in parts of Cameroon, Nigeria or Papua New Guinea) who may be forced to study English as their first written language, despite it being an additional language for them (see Anderson 2015).

• Reimagining multilingual practices in ALT: The recent increase of interest in translanguaging has, to date, been led by research in bilingual education, predominantly in high-income countries (see García and Li 2014). A reimagined EL is better able to see the relevance of this research for other types of language classrooms, particularly in FL contexts in highly multilingual low- and middle-income countries in the global South, where, despite extensive evidence of translanguaging in wider society, research in additional language classrooms has been fairly limited until recently (e.g., Makalela 2015; Vaish and Subhan 2015), likely because such practices have been over-simplistically characterised as either “use of L1” (e.g., Lee and Macaro 2013) or “grammar-translation” (e.g., Yu 2001) by researchers adopting competence orientations.

• Reimagining understandings of quality in language teaching: Research investigating issues of educational quality in AL, particularly within the SLA canon, has tended to focus on specific interventions designed to demonstrate impacts on learning narrowly conceived as improvements in underlying proficiency (i.e., competence) in a specific “target language”. A reimagined EL offers a wider understanding, both of education and language learning, enabling it to adopt a wider range of outcome measures as potentially valid, such as enhanced multilingual literacies or expanded repertoires, improved intercultural awareness and learner exam performance, which has typically only been accepted as a measure of quality if exam tasks are competence-oriented (and often they are not in basic education). More holistic constructs of teacher-embodied quality, such as teacher expertise, may also be of increased interest in a reimagined AL for its high ecological validity and context specificity (see Anderson 2023a, 2023b).