Global English: From “Tyrannosaurus Rex” to “Red Herring”

The rise of English as the world’s main international language has prompted a social justice agenda underpinned by an assumption that English causes or exacerbates inequality and injustice in the world. In this position statement, I set out to problematise and complexify this assumption, suggesting that English is neither a “Tyrannosaurus Rex”, a “Cuckoo” nor a “Lingua Frankensteinia”, but a “Red Herring”, distracting attention away from the underlying causes of inequality. Within the theoretical framework of “verbal hygiene” (Cameron 1995, 2012a), and drawing on my own empirical work and that of others, I argue for widening the scope of global English and more broadly applied linguistics. I suggest that as socially committed applied linguists, we stand a better chance of solving “real-world problems” (Brumfit 1995: 27) if greater attention is accorded to systems of inequality that are not obviously language-based. I will suggest that a too narrow focus on linguistic injustice risks losing sight of the underlying non-linguistic conditions that produce this injustice. I conclude by suggesting some ways forward that centre on co-thinking language with political, social, economic, cultural and material conditions.


Introduction: Global English and Social Justice
No other language in history has grown as exponentially as English has in recent decades. Non-native users of English have been said to significantly outnumber native users for some time now. A third of the world's population are thought to have some degree of proficiency in English (Crystal 2008), an estimate that may have grown in the ten years since it was made. English has spread especially fast in transnational areas of life, such as business, science, popular culture and online communication (Haberland 2018;Crystal 2012;Lønsmann and Mortensen 2018;Gregersen et al. 2014;Hultgren et al. 2014;Dimova et al. 2015;Kuteeva 2013). English is the most widely taught foreign language in the world, and, in recognition of its ubiquity, governments across the world are lowering the age from which it is taught in schools; others are considering making it an official language (Lanvers and Hultgren 2018). Few would dispute that English has become the world's main international language and that it is likely to continue to grow (MacKenzie 2018).
Much discourse-academic and public alike-has been generated in the wake of the rise of English as an international language. Whilst some highlight the positive aspects, such as English enabling communication between speakers who would not otherwise have been able to interact (Van Parijs 2011;de Swaan 2001;Calvet 1998) and celebrate the range of ways in which English manifests itself locally (Pennycook 2007), many have been skeptical, as reflected in metaphors to describe English: "Tyrannosaurus Rex", "Hydra", "Trojan Horse", "Cuckoo", "Killer Language", "Lingua Frankensteinia" (Rapatahana and Bunce 2012;Skutnabb-Kangas 2003;Swales 1997;Cooke 1988;Phillipson 2006Phillipson , 2008. English has also been described as causing "linguistic imperialism", "linguicide" and "epistemicide" (Phillipson 1992;Skutnabb-Kangas and Phillipson 1995;Bennett 2007). Concerns have centred primarily on disadvantages and inequalities for non-native users of English as well as loss of cultural and linguistic diversity. Such concerns reflect a commitment of many applied linguists to issues of social justice (May 2003;De Costa 2015;Avineri et al. 2019) and "real-world problems" (Brumfit 1995: 27;Cook and Kasper 2005).
Of course, not all scholars who have contributed to our understanding of global English would necessarily see themselves as subscribing to a social justice agenda. The study of global English, as I here understand it, is a broad-ranging and interdisciplinary field of inquiry that has been approached by linguists, educationalists, economists, political philosophers and sociologists (Gazzola et al. 2018;Pennycook 2017;Crystal 2012;Van Parijs 2011;Mufwene 2010;Grin 1996;Brutt-Griffler 2002;de Swaan 2001;Canagarajah 1999;Graddol 1997Graddol , 2006Phillipson 1992). Political philosophers, economist and sociologists have tended to be more laissez-faire than linguists, highlighting the potential of English to enable communication between speakers who would not otherwise have been able to interact (Van Parijs 2011;de Swaan 2001;Calvet 1998). Some even argue that English should be actively promoted because of its potential to foster social justice and mobility (Van Parijs 2011). English as a lingua franca scholars have pointed out that English does not pose a threat to but actually enables multilingualism in that it allows people of different linguistic and cultural backgrounds to come together (Jenkins 2017). In general, linguists have not always engaged with the ideas of scholars in other disciplines possibly because they see them, rightly or wrongly, as devoid of a commitment to social justice, which many linguists hold in high esteem (May 2003(May , 2015. Applied linguists, in turn, have been accused of being "politically correct" and "sentimental" (Calvet 1998;de Swaan 2004). The laissez-faire approach to the global spread of English is detectible also among lay people who may adopt the pragmatic view that English is but a tool for communication, such as for instance scientists at Nordic universities who use English for teaching and research purposes . Most linguists, however, would take issue with such a utilitarian view of English (May 2015).
There are also academic activities that cannot be positioned clearly as either for or against global English. The world Englishes paradigm has sought to redefine the terms on which global English operates by pointing to the legitimacy of varieties of English spoken outside of its "mainland", such as Indian, Nigerian and Singaporean English (Kachru 1990;Bolton 2006). In a similar vein, the English as a lingua franca paradigm has also made a significant contribution to shifting away from an assumption that there is one standard way of using English (Jenkins 2007;Seidlhofer 2011;Mortensen 2017). Consequently, there have been numerous and wellmotived calls for a complete overhaul in the way in which English is taught and tested (Kiczkowiak and Lowe 2019;Rose et al. 2020;Rose and Galloway 2019;Jenkins and Leung 2019;Brutt-Griffler 2017;Leung et al. 2016). There are also scholars, mainly non-linguists, who have taken a more explanatory approach, proposing models designed to account for how English spreads, by highlighting its "communicative potential" (de Swaan's 2001), and how it operates according to the "maximin principle" to minimize exclusion in a conversation (Van Parijs 2004: 115). Economists have modelled the costs involved in learning languages and translating between them (Grin 1996;Hogan-Brun 2017;Gazzola et al. 2018), whilst others have predicted the spread of English based on demographic changes (Graddol 2006).
In this position statement, I will argue that branches of applied linguistics and sociolinguistics committed to a social justice agenda would gain from an explicit acknowledgement that language is mostly a contingent, secondary factor and not a cause of inequality. I will suggest that language needs to be decentred from our analyses. To clarify my argument, it may be useful to contrast it with a counter argument. One work which embodies current thinking in applied linguistics is the book Linguistic Diversity and Social Justice. In this book, Ingrid Piller argues that applied linguists must "put linguistic diversity on the map of contemporary social justice debates" (2016: 5). While I share Piller's commitment to social justice and wholeheartedly agree with her intentions, in this position statement, I experiment with reversing her argument. In contrast to Piller, I would suggest that language has always been a key trope in applied linguistics, and that the time has come not to "put it on the map of contemporary social justice debates", but to tone it down. From this it follows that any intervention focused solely on matters of language is unlikely to succeed in putting things right.
I accept that it is useful for applied linguists to invoke language, and language-related concepts, as a shorthand for other more material sorting mechanisms. I also believe that few, if any, would assign a causal link between language and inequality. As Piller puts it: "linguistic diversity intersects with social justice" (2016: 5, my emphasis). This intersection is conceptualized neatly in coinages such as "raciolinguistics", "accent bias" and "linguistic racism" and evidenced in the many ways in which linguistic minorities may suffer multiple forces of oppression and discrimination not only through their race but through their language or way of speaking (Alim et al. 2016;Sharma et al. 2019;Dovchin 2019). It must also be acknowledged that stereotyping and discrimination can happen solely on the basis of how a person speaks, as experiments in social psychology and perceptual dialectology have shown.
This piece, however, is an invitation to unpick what more precisely the intersection between global English and social justice might consist of. What do we mean, more specifically, when we say that language intersects with inequality? I would suggest that such an "unpicking" is interesting not only as an academic exercise, but that it will put us in a better position to rectify any injustice in the world. I suggest that imprecision carries the risk of mistaking correlation for cause and of conflating language with other systems of inequality. With a lack of precision, we risk misdiagnosing the problem and proposing the wrong solutions. In essence, I will argue that if as applied linguists we want to stand a chance with the social justice agenda, we need to redirect our attention away from language and towards economic, social and political inequalities, or at least keep them all in view.
Having said this, there is perhaps one way in which language might be said to play a more pivotal role in creating unjust social relations. This is if we think of "language" in the sense of "discourse" and assume that larger social structures are both reflected in, and emanate from, everyday and institutional uses of language. Certainly, as Foucault and others have taught us, discourse is potentially material in effect, producing "practices that systematically form the objects of which they speak" (Foucault 1972: 49). In other words, social structures and categories can be "talked into existence" with real-world consequences for justice. The argument pursued here, however, is slightly different in that the critique is targeted, not at the idea of discourse as socially constitutive, but at the idea of language as a reified and essentialised entity that is somehow to blame for social injustice. Applied linguists are increasingly seeking to find ways of incorporating materiality into their theoretical and analytical frameworks (Block 2014(Block , 2018Canagarajah 2017Canagarajah , 2018Pennycook 2016Pennycook , 2018Ricento 2015). Thus, in what follows, I continue work in this vein by making a case for applied linguistics to decentre language.

Theoretical Framework: Verbal Hygiene
My argument is framed within the theory of "verbal hygiene" as expounded by the British sociolinguist Deborah Cameron (2012aCameron ( [1995). With echoes to American linguistic anthropology and the works of the French sociologist Pierre Bourdieu, verbal hygiene centres on the idea that people have an irresistible urge to debate, discuss and sometimes regulate and intervene in matters of language. Verbal hygiene refers to a "motley collection of discourses and practices through which people attempt to 'clean up' language and make its structure or its use conform more closely to their ideals of beauty, truth, efficiency, logic, correctness and civility" (Cameron 2012a: vii). Whilst most of Cameron's examples of verbal hygiene come from intra-linguistic discourses, i.e. talk about what is considered correct and appropriate (mainly) within the English language, the concept is arguably applicable to inter-linguistic discourses too, i.e. talk about the global spread of English and about how to manage it vis-àvis other languages.
A central point in the verbal hygiene framework is that languagerelated debates and interventions are rarely only about language. As Cameron puts it: "complaints about language changes are usually symbolic expressions of anxieties about larger social changes" (Cameron 2012a: 238). Cameron believes that, in contemporary society, it is commonplace to make "a mountain out of a mole-hill" where language is concerned (Cameron 2013: np). She explains: In any given time and place, the most salient forms of verbal hygiene will tend to be linked to other preoccupations which are not primarily linguistic, but are rather social, political and moral. The logic behind verbal hygiene depends on a common-sense analogy between the order of language and the larger social order, or the order of the world. The rules of language stand in for the rules that govern social or moral conduct and putting language to right becomes a sort of symbolic surrogate for putting the world to right. (Cameron 2012b: transcribed from an oral presentation) If a preoccupation with language really is a cover preoccupation for some underlying anxieties, then, as Cameron suggests, this would explain why opinions about language are often expressed with such passion and fervour. Because, as she says, in most cases they are "not just debates about language" (Cameron 2012b: np), but debates about the current state of the world and about how to put it right.
In the next section, I will scrutinise three key assumptions to build my argument that language is rarely the be-all and end-all in matters of social justice. These assumptions are 1) non-native speakers are disadvantaged by the spread of English; 2) English threatens other languages and 3) language policy will curb the spread of English. These three assumptions are widespread not only in applied linguistics, where they have risen to assume status of unchallenged orthodoxies, but also in lay circles. This is why, in the following, I will consider lay and specialist discourses together. What makes Cameron's verbal hygiene framework stand out is that no one is exempt from it. Even us linguists, who normally pride ourselves on our descriptivist and impartial stance, are guilty of it. It is, according to Cameron, impossible for linguists to adhere to their own ideal of descriptivism because opinions about language are never neutral (see also Jaspers 2017). I hope that scrutinising these three assumptions in applied linguistics will at the very least spur people in the field to come up with counterarguments.

Assumption 1: Non-Native Speakers Are Disadvantaged by the Spread of English
Despite challenges to the notion of "native speaker" (Davies 2003), there is a prevailing assumption in many fields in applied linguistics that nonnative speakers of English are disadvantaged by the spread of English as a global language. Indeed, entire research fields have been partly or wholly founded on this premise, including minority language research, language maintenance and revitalization, linguistic rights, English as a medium of instruction, English for research publication purposes, and others. The idea is intuitive. Anyone who speaks more than one language is likely to have found themselves in a situation where they would have felt more comfortable speaking the language they know best. Mother-tongue instruction and, more recently, translanguaging, has long been advocated by UNESCO (1953UNESCO ( , 2007 in compulsory schooling in ex-colonial contexts on the grounds that children learn better if they are taught in a language they already know. However, aside from the fact that this may not be a practical possibility in highly multilingual contexts where there is a shortage of local-language teachers and teaching material, more pertinent to the point made here is that even if everyone were able to conduct their business in the language they know best, this would be unlikely to do away with the source of inequality. Of course, advocates of bilingual or mother-tongue education might rightly argue that even if medium of instruction is rarely the only factor associated with disadvantage, seeking to eradicate language-related injustice is still better than not doing anything at all. This is true of course. However, we might like to ask if a focus on language on its own is sufficiently effective and/or the extent to which it draws attention away from more fundamental causes of inequality. Jaspers (2019) questions the role of medium of instruction for improving educational attainment (see also Block 2018). He queries the tendency of many applied linguists to present the advantages of using a child's home language in school (including translanguaging) as an evidence-based fact rather than for what he sees it, i.e. ideology. Writing about the Dutch situation, Blommaert (2017) notes that although good school attainment has often been attributed to using specific languages at home or in school, the critical role of parents' income and educational background has been overlooked.
In the context of higher education, a systematic review of the literature was unable to conclude that medium of instruction has a bearing on learning outcome (Macaro et al. 2018). The inconclusiveness of Macaro et al.'s review might be interpreted as evidence that learning is an immensely complex phenomenon. The degree to which it takes place can in all likelihood not be assigned to a single factor, including language of instruction. Rather, it is likely to depend on a myriad of factors, extrinsic and intrinsic, psychological and sociological (Coleman et al. 2018), and language is likely to be only one, perhaps even rather insignificant, factor in the grander scheme of things.
Another field which is at least partly premised on the idea of nonnative speaker disadvantage is English for research publication purposes. Although the field has sought to move away from a deficit view of nonnative English research writers, Hyland argues that the field is still characterized by a "pervasive view which asserts that EAL [English as an additional language] scholars are disadvantaged in the cut-throat competitive world of academic publishing by virtue of their status as second language writers" (2016: 66). Feelings of linguistic disadvantage are empirically evidenced in studies revealing how both manuscript authors and journal gatekeepers point to non-nativeness as being in one way or another problematic (Lillis and Curry 2015;Ferguson et al., 2011;Perez-Llantada 2018). Interpreted within a framework of verbal hygiene, such topicalization of "non-native speakerness" and its problematic nature is not surprising. It points to how language intersects with and sometimes acts as a proxy for phenomena in the wider world. However, as has interestingly been shown, although negative comments on non-native English usage are given in abundance by journal reviewers and editors, they don't necessarily lead to an article being rejected on those grounds (Lillis and Curry 2015;Ferguson et al. 2011). Another study (discussed below), which adopts a more macro-level perspective, suggests that the key factor in determining publishing success may not be the first language of the author.
Drawing on bibliometric analyses, O'Neil (2018) found that whilst the English-dominant countries Australia, United Kingdom, Canada and the United States feature in the top ten most productive countries in terms of scientific output, so too do the non-English-dominant countries Switzerland, Sweden, Netherlands, Germany, Spain and France. Another study, based on an analysis of articles published in the top five journals by impact factor in six subject areas, shows that authors with English as an additional language produced 57% of the output compared to 43% by native speakers of English (Hyland 2016). With the proviso that there is no error-free method to establish the first language of an author, and both studies cited above use proxies for this indicator, the evidence suggests that native-speaker status may not be the only or even the most important factor determining publishing success (see also Kuteeva 2015).
What does seem to matter rather more is whether you find yourself in a well-resourced and well-networked context. O'Neil's study reveals that a mere twenty nations produce 80.4% of the world's global scientific output while the remaining 211 produce 19.6%. Tellingly, the twenty most scientifically productive nations are also the ones who possess 78.4% of the global wealth, measured in GDP. In other words, the richest nations produce the vast majority of scientific output. Just like the world's economic resources in general, scientific output is concentrated in a very small minority of countries, irrespective of whether these countries are English-dominant. This confirms that there are certainly inequities in global academic publishing and that a cause for social justice could be pursued. However, it would appear that economic factors are more important than linguistic ones in explaining these inequalities. In essence, then, if applied linguists are committed to a social justice agenda then we need to ponder the significance of other more material factors alongside language.

Assumption 2: English Threatens other Languages
Another assumption that is widespread in lay and academic circles alike is that the spread of English engenders a loss of languages and cultures around the world. What I seek to challenge here is not that language loss happens-it clearly does. Rather, I want to draw attention to the imprecision in attributing language loss and language endangerment to English, which in my view erroneously locates the problem in the linguistic sphere rather than in the underlying cultural, technological and social developments that lead to it. In the following, I'll try to explain what I mean, illustrating it with the Nordic debate about "domain loss".
In the Nordic countries, English has been said to cause "domain loss", which refers to the idea that the national Nordic languages (Swedish, Danish, Norwegian, Finnish and Icelandic) may lose functionality or be marginalised in transnational areas of life which are felt to be of key importance, such as science, higher education and business (see references in Hultgren 2016a; Hultgren et al. 2014). This has at times been framed as an issue of social justice on the grounds that if scientific output is communicated in English, it will be inaccessible to those segments of the population whose English proficiency is not sufficiently high, and that if university students have been trained in English, their ability to undertake professional functions upon graduation (e.g. as vets, farmers and doctors) will be hampered. In both cases, democracy is said to suffer (see references in Hultgren et al. 2014;Hultgren 2016a).
Within the framework of verbal hygiene, we should ask what underlying ideologies such concerns may index; what they may stand as proxies for and the extent to which they are underpinned by evidence (see, e.g., Kuteeva et al. 2020). Elsewhere, I have identified the ideologies underlying domain loss as ranging from romantic nationalism, antiimmigration, anti-Americanism and anti-bureaucratisation (Hultgren 2014a). It is revealing that, in the Danish debate, those who have been particularly evocative about the spread of English can be found on opposing sides of the political spectrum. The far right populist party (the Danish People's Party) have argued for a strengthening of Danish on the grounds of protecting Danish heritage and those "left-behind" segments of the population with comparatively low levels of English proficiency, whereas members of the left wing parties have argued the same but for very different reasons: anti-globalization and anti-corporation. 1 The fact that the same linguistic argument (Danish must be strengthened) can be used for promoting what are diametrically opposed political ideologies illustrates how language debates are essentially vehicles for political and ideological debates.
Even if we accept the idea that discourses about language loss and death are symptomatic expressions of some underlying ideologies, as empirically committed scholars, we could still explore if these concerns are justified. Certainly, as already mentioned, concerns about language loss are not plucked from thin air. As is the case with claims about nonnative speaker disadvantage, they are empirically documented in a voluminous and established body of work on language death, language loss, language endangerment, language maintenance and language revitalization (Fishman 1991;Nettle and Romaine 2000;Evans 2009). This body of work, dubbed "salvage linguistics" (O'Rourke and Pujolar 2013), is based partly on an underlying concern with "saving" threatened languages from extinction. But, as already noted, in the interest of precision, it is worth asking if the threat is posed by English or by underlying cultural, technological and social developments. Futhermore, it is worth asking how English or indeed any language can actually be empirically operationalized beyond being an ideological construction. I explore these issues below.
In a study of new words, I set out to compare the proportion of lexical borrowings from English in the science domain with what had been found in a non-science domain. The aim was to establish whether claims about "domain loss" could be empirically substantiated (on the reasons for operationalizing "domain loss" as "lexical borrowing", see Hultgren 2013). I recorded, transcribed and analysed the talk produced by lecturers with Danish as their L1 in 11 undergraduate teaching sessions in physics, chemistry and computer science at the University of Copenhagen (Hultgren 2013). Although English is increasingly used as a medium of instruction at Danish universities, most programmes at undergraduate level are still delivered in Danish. Calculating the number of lexical borrowings from English, I found that at 0.6%, the proportion was exactly the same as in the non-science domain (Rathje 2010), debunking any myths that Danish local terminology was on the verge of disappearing. Moreover, I found that the number of lexical borrowings was significantly higher in computer science than in physics and chemistry, reflecting the significant developments this area of life has undergone in recent decades.
It is also relevant to mention that I encountered significant methodological challenges in calculating the proportion of lexical borrowings. This is because of well-known complexities in deciding at what point a word ceases to be a borrowing and starts to be an integral unit of the recipient language. This was further complicated by questions over whether acronyms and propernames, sometimes in hybrid form, e.g. "Fortran", "ASCII-definitionen" (the ASCII definition) and "Javadockommentarer" (Javadoc comments) should be considered lexical borrowings at all. These turned out to be particularly frequent in computer science and were the main reason why computer science exhibited a higher proportion of lexical borrowings than chemistry and physics. Moreover, on closer inspection, many "new" words turn out not to be "new" at all, but are created by combining already existing linguistic resources, often of Latin and Greek derivation, such as, e.g. "dissociative electron attachment", "solid state ionics" and "orthogonal synthon paradigm" . In other words, much as they have done throughout history, speakers combine existing linguistic resources to denote new social phenomena. This is unsurprising and entirely in accordance with well-documented sociolinguistic processes.
These findings pose a challenge to "domain loss" and a key assumption that has underpinned it, that of "complete and society-bearing language" [komplet og samfundsbaerende sprog] (Nordic Council 2007). The notion of a "complete and society-bearing language" assumes that the national language is a priori endowed with the required linguistic resources it needs to be functional. This, of course, is an illusion when you redirect attention from "language" to "register" (Agha 2007;Blommaert 2013). Indeed, as the LSP [language for specific purposes] scholars, Laurén et al., put it "[i]t is a fact that no language covers all possible domains at all levels" (2002: 25), thereby acknowledging that no language is ever at any one time "complete and society-bearing". As is well-known, most languages, English included, are made up of a mix of linguistic and communicative resources reflecting the influence from different cultures and different developments throughout history, and it is only when certain people, notably the powerful, decide to call that particular concoction of linguistic and communicative resources a "language" that it becomes one.
The findings reported above should not be interpreted as a denial of language loss if this is understood as "the loss of a particular language named as X". Nor should they be seen as justification for not taking seriously the plights and rights of minitory language speakers. There is a lot of work to do in granting marginalised groups the cultural, political and linguistic rights that are enjoyed by more powerful groups. What they do illustrate, I think, is the need to co-think language with the cultural, political and social world. Language changes because the social world changes. By drawing attention to some of the empirical complexities in establishing what a language really is, they also illustrate the axiom that languages are ideological constructions. Together, they highlight the intangible, elusive and secondary nature of language and reinforces the point that the remit of linguists needs to be broader than language. As I will argue in the next section, no social justice agenda centred on language alone can do anything to reverse the spread of global English.

Assumption 3: Language Policy Will Curb the Spread of English
To address the perceived injustice and encroachment posed by English, applied linguists (and lay people) sometimes propose language policy as a solution. Whilst language policy can be understood in numerous ways involving creation, interpretation and appropriation (Barakos and Unger 2016;Hult and Johnson 2015;McCarty 2011), it usually entails some way of interfering in language to obtain a desired linguistic outcome (Hult and Johnson 2015). The Nordic countries have been particularly pro-active in seeking language policy solutions to the perceived threat from English from a top-down level. This has centred on the notion of "parallel language use", defined as "the concurrent use of several languages within one or more areas. None of the languages abolishes or replaces the other; they are used in parallel" (Nordic Council 2007: 93;Hultgren 2016b). As the other two assumptions in applied linguistics discussed above, the concern with language policy is also underpinned by its own well-established field in applied linguistics, language policy and planning (LPP). Although LPP researchers have argued for the need to incorporate more material, nonlinguistic elements into their analyses, this angle remains underdeveloped and undertheorized (Ricento 2015).
Without denying that language policy initiatives can serve important symbolic functions for minority language speakers, and thereby fulfil an important social justice function, the position in this paper is that, unless language policies are co-thought with more material aspects, they stand little chance of having a real effect. In a study on policies in higher education (Hultgren 2014a), I compared language policies with other policies not centred on language, hereunder research and educational policies. What I found was that despite all eight of Denmark's universities having some form of language policy in place that advocates parallel language use, i.e. concurrent use of Danish and English, this contrasted markedly with policies in the research and educational domain, whether operating at institutional, national and supranational level. For instance, one of the drivers that pushed scholars to publish in English-medium journals was the introduction of bibliometric evaluation regimes, akin to those already in place in many Anglophone countries. Although these evaluation regimes do not explicitly direct researchers to publish in English, they do award more points (and hence more governmental funding) to those institutions whose researchers publish in higher-ranking journals, which, as we know, tend to be in English. Thus, it is a research policy, not a language policy, that drives the shift to English. Consequently, the only thing that could possibly reverse this trend would be to put into place another research policy with a different aim or indeed not to have any research evaluation regime at all. What will not revert the trend is a policy centred on language alone, even if this policy declares the institution as being committed to parallel language use (see also Fabricius et al. 2017;Saarinen 2017).
Many more policies could be mentioned that indirectly engender an increased use of English. For instance, Danish universities have targets for the recruitment of international staff. This was a directive by the Danish government to make Denmark more competitive in the global knowledge economy. This, of course, encourages the recruitment of non-Danish speaking staff and is likely to expand the use of English as a lingua franca.
Another key driver of English is the creation of a European Higher Education Area which is meant to standardize degrees and enable mobility. As has been observed, not a single word in the Bologna Declaration is devoted to language (Philipson 2006). This is because it is not a language policy; it is an educational policy with unforeseen or ignored consequences for language. Policy initiatives like these indirectly lead to an increased use of English and can be said to constitute a type of "covert language policy" (Piller and Cho 2013;Soler-Carbonell 2017). However, crucially for the argument put forth here is that these are all policies that originate in a domain other than the linguistic one. They fall variously within the realm of research policy, EU policy, educational policy and institutional policy. Such policies will continue to promote and increase the use of English-often indirectly and implicitly. This is why there are increasing calls for adopting a broader approach to the rise of English in non-Anglphone universities, and more generally, in any domain (Dafouz and Smit 2020;Doiz and Lasagabaster 2020).

Why Do We Place So Much Emphasis on Language?
Before I conclude, it is worth reflecting on why we are so blinded by language. Firstly, in some ways, it is understandable that applied linguists should be so, given our disciplinary affiliation. Given that we are linguists, it is not surprising that language and language-related matters take centre stage in our analytic and theoretical frameworks. More broadly, perhaps, there is something to be said about the inextricability of language and culture (Risager 2006). It is, in some ways, impossible to talk and think about language without also talking and thinking about culture and vice versa. When minority language groups fight for their linguistic rights, they often do so because they see them as a symbolic instantiation of their cultural and political rights. However, the idea that language and culture are to some extent inextricable does not mean that by putting the linguistic world to right, we will automatically put the world to right.
Perhaps another reason why language has come to take such a front seat in contemporary society may be that the social, cultural, political, economic and technological changes are currently felt by many to be intense. It has been suggested that in times of perceived destabilization of norms, circumstances in which we arguably find ourselves today given increased physical and virtual contact between users of different languages, debates about language tend to intensify and norms become more explicitly negotiated (Duchêne and Heller 2008;Cameron 2012a). Globalization has brought with it significant changes in the political, economic, cultural and technological domain, and what we may be witnessing is people airing their underlying anxieties about these breakneck changes by reference to language.
A third possible reason why we, and here I refer specifically to us as scholarly linguists rather than to lay people, are predisposed to foreground language may be that scholars in all fields have a tendency to accept and reproduce the axiomatic assumptions in their fields. Drawing on Bourdieu's (1990Bourdieu's ( , 1991 notion of field and Bernstein's (2000) understanding of disciplinary formation, May (2019) argues that academic disciplines from the 19th century onward and their subsequent organization into self-regulating communities amount to what Bernstein terms "singulars". Academic singulars are characterized by strong boundary maintenance, which are supported culturally (via professional associations, networks, and writing) and psychologically (in students, teachers, and researchers). As a result, "singulars develop strong autonomous self-sealing and narcissistic identities" (Bernstein 2000: 54), preventing them from seeing the trees for the wood. In combination, singulars militate against wider interdisciplinary-let alone, transdisciplinary-engagement. If this is so, then how can we break through this impasse? I'll consider this below.

Conclusion: A Roadmap for Global English
In this position statement, I have set out to question some Panglossian orthodoxies about global English, namely that the global spread of English engenders injustice and that language policy is the solution to addressing this injustice. To an already long list of zoological and mythological creatures used to describe English, e.g. "Tyrannosaurus Rex", "Hydra", "Trojan Horse", "Cuckoo", "Killer Language" and "Lingua Frankensteinia", I will add "Red Herring". I see global English as a red herring because it is purported to be significant, but it actually detracts the focus away from the real issue. There is, I would suggest, an urgent need to provincialize language and to engage to a much greater extent with the underlying factors that cause English to spread.
If it is accepted that English, and language in general, is not the sole or even the most important reason for inequality in the world, then an obvious follow-on question is what we as applied linguists concerned with social justice should preoccupy ourselves with instead. Should we put our feet up and bury the discipline? On the contrary. I would suggest that applied linguistics has a key role to fill, and to conclude this paper, I propose two ways forward.
Firstly, we could keep the notion of verbal hygiene at the forefront of our thinking and be mindful of the possibility that discourses about language, although often very strongly felt, will often just serve as a starting point for a deeper, non-linguistic examination. As Salö (2017, drawing on Bourdieu) suggests, we should exercise "epistemic reflexivity", in order to raise our critical awareness, query our own epistemological baggage, and reflect on the presuppositions in our field. To Salö sociolinguistic research seems to end up showing and saying exactly what one would have expected it to show and say, based on the position-social, academic or otherwise-from which the research was produced. Often, this is because scholars embody the values of the group they investigate and, all too often, fail to create a rupture with their inherited view of the problem they investigate. (2017: 2) As empirically committed scholars, we should acknowledge that the discourses produced in our field, even though seemingly centred on language, may not in actual fact be linguistic. However, we should also recognise that language offers an important window into social structures and change that may not have been as obvious to someone not habituated to focusing on language.
Secondly, and alongside a continued attention to language, applied linguists could widen their lens and acknowledge that a focus on language alone is not enough neither for understanding a social problem, nor for seeking a solution to it. As is well-known, the reason for the dominance of the English language is inextricably linked with power and imperialism, as Philipson (1992) has importantly shown. The historical junctures at which the English language has spread are believed to have coincided with those at which English-speaking peoples and nations have engaged in imperialist expansion. This goes back to the arrivals of the Saxons, the Angles and the Jutes on the British Isles in 449 AD, and in more recent history, to three key historical events. The first is British imperial expansion from the 17th century onwards which took the language from its island birthplace to settlements around the globe; the second is the industrial revolution from the 18th century, which solidified the English language as one of scientific progress; and the third and most recent is associated with the rise of the US as a superpower and the global spread of an economic model based on capitalism. Consequently, methodologies and theories need to be expanded in order to allow us to understand the political, economic, social and cultural systems and processes that undergird the current world order and cause English-or any languageto expand. This calls for greater interdisciplinarity (see further in the epilogue). I started this paper by suggesting that assigning injustice and inequality to the linguistic sphere risks misdiagnosing the problem and proposing the wrong solutions. As I have argued, the real cause of disadvantage and injustice lies not in global English, but in political, economic, social and cultural structures.