The Routledge Handbook of English as a Lingua Franca

p.101

H.G. Widdowson

Introduction: a recurring theme

It has become a truism to say that the extent to which English is now used as a lingua franca is unprecedented in that it has become an all-pervasive feature of a globalized world. But although the extent of its use is unprecedented, it is, of course, by no means the only example of a language that, originally confined to and defined by a relatively small community of users, has spread beyond its borders to become an international means of communication. In this respect, it can be said that ELF is a new phenomenon in degree but not in kind. And as with other languages, the extension into lingua franca use quite naturally involves variation and change as the language is adapted to meet the needs of different communicative contexts and purposes. The study of ELF is essentially the study of adaptive variation and under what conditions this is activated. As such, it too has its precedents, for the study of linguistic variation has long been central to sociolinguistic enquiry.

A recurring theme here is the tension between variation as a pragmatic communicative expediency and variation as the expression of social identity. People invest their language with sociocultural values that they quite naturally seek to preserve and protect. If, as in the case of a lingua franca, a language is appropriated by outsiders and adapted to suit their own communicative requirements, the integrity of the language as representing these values would seem to be undermined. So it is that the intrinsic tendency of English, like any other language, to spread and mutate and continually adapt itself to different conditions has often caused alarm and despondency among linguistic preservationists who claim to have custody over the language and its ‘proper’ use. In their view, this tendency calls for intervention to ensure that if there is to be variation, it can be controlled so the communicative and communal integrity of the language is not compromised.

Prescriptions of appropriate language

This attitude to English is all too evident in current reactions to the uncontrolled spread of the language in its use as a lingua franca, as ELF researchers know all too well. But the attitude itself goes back a long way. In 1712, for example, Jonathan Swift published a pamphlet entitled ‘A proposal for correcting, improving and ascertaining the English tongue’, in which he inveighs against ways in which the language is corrupted, how it is full of what he calls ‘manglings and abbreviations’ (Swift 1712). In the previous century, we find Sir Thomas Sprat also complaining about the abuse of English. His objection, however, is not to the mangling of linguistic forms but to the extravagant eloquence of their use, what he refers to as ‘the luxury and redundance of Speech, this vicious abundance of phrase, this trick of metaphors, this volubility of tongue, that makes so great a noise in the world’ (1667: 111–113).

p.102

Sprat was one of the founders of the Royal Society, established for the ‘Improving of natural knowledge’ and his hostility to eloquence is because he sees it as a threat to its mission. As he puts it,

There is one thing more about which the Society has been most solicitous, and that is the manner of their discourse; which, unless they had been very watchful to keep in due temper, the whole spirit and vigor of their design had been soon eaten out by the luxury and redundance of speech.

(Sprat 1667: 111–113)

The necessary manner of discourse, he believed, was ‘a close, naked, natural way of speaking, positive expressions; clear senses; a native easiness: bringing all things as near the mathematical plainness as they can’ (Sprat 1667: 111–113). Sprat is effectively presenting the case for English for Specific Purposes (ESP), the specific purpose here being the advancement of science. This was, he believed, a variety of English used by ‘Artizans, Countrymen, and Merchants’ in contrast to the vicious verbiage of ‘Wits or Scholars’. It is of interest to observe that some 250 years later, Peter Strevens sounds a similar note in his article significantly entitled ‘Alternatives to Daffodils’ in which he speaks out against the predominance of literary uses of English in language teaching and in favour of what was then called the register of scientific English. He concludes by expressing a Sprat-like view in pedagogic terms:

the science and technology student is not at present adequately catered for: current English teaching practice is non-scientific and sometimes anti-scientific ... Perhaps ‘Fair daffodils we weep to see thee haste away so soon’… will give way in some overseas classes to language work more relevant to the science student’s eventual needs.

(Strevens 1971: 11)

What both Sprat and Strevens argue is that we need to recognize, that to be effective, English has to adapt to different requirements, to be appropriate to different communicative contexts and purposes. But in neither case is there any suggestion that this might involve any structural non-conformity from established linguistic rule. What they are talking about is not the use of a formally abnormal kind of English but the different functional use of what is formally a normal kind of English. The question arises, however, as to whether there are domains of use where conformity to rule does not necessarily meet communicative requirement and that call for some refashioning of the linguistic code itself.

p.103

Recodifications of English

Domains of use naturally multiplied with the expansion of empire but since the spread was a function of colonialism, English, like its colonized users, was under the jurisdiction of its British rulers. Whatever variable encoding occurred in its use, such non-conformity to rule was not sanctioned by authority as legitimate. But when the United States declared independence from colonial rule, a community of native speakers was created that was not bound by such authority and could change the language to suit the needs of the new nation. And Noah Webster set out to do just this:

Now is the time and this the country in which we may expect success in attempting changes to language, science, and government. Let us then seize the present moment and establish a national language as well as a national government.

(Webster 1789: 405)

For Webster, a recodification of English was called for so that it would serve the cause of national identity. The recodification was relatively minor, involving mainly orthographic and lexical innovation, but served its intended symbolic purpose. Over a century later, recodification was deemed necessary for an entirely opposite reason: not to support nationalism but to counter it. For, in Europe, nationalism, in the early part of the twentieth century, was taken to be a major cause of the First World War – Einstein called it ‘an infantile disease’. In consequence, in the aftermath of war, an anti-nationalist sentiment sought to eradicate this disease by the promotion of international understanding. It was in this socio-political context that Ogden devised his Basic English (Ogden 1930).

Unlike Webster’s intervention, the purpose here called for a radically recoded version of the language, a totally different kind of English, an elementary coding system, stripped of grammatical and idiomatic complexity so as to make it readily adaptable to any demands that might be made upon it as an international means of communication. But apart from this practical advantage, and as a corollary to it, Basic also had a symbolic advantage: this version of English was thereby de-nationalized in that it removed features of the language that made it distinctive as the national language of its native speakers, and who, therefore, could no longer claim ownership of it.

Communal and communicative functions

These attempts to refashion English so as to meet the needs of different domains of use are not only of passing historical interest. They raise controversial issues about the role and status of the language that are centrally relevant to an understanding of its international use and that remain unresolved to this day. One of these has to do with the relationship referred to earlier between the function of a language as an expression of communal identity, which is Webster’s concern, and its function as a means of wider communication, which is Ogden’s. How far are these functions compatible, how far in tension? In native speaker communities the two would seem to be naturally compatible, indeed symbiotically related in that effective communication is both cause and consequence of social cohesion. This is not usually the case with ELF, however, where the language is appropriated by other users and adapted expediently to meet their communicative needs, and where the identifying function may have little if any relevance. In this respect, it might be said that with such adaptive uses of English there is some reduction in function, which is why they are often taken to be defective – expressions of incompetence.

p.104

A second issue, which is made particularly prominent in Ogden’s proposal, is how far it is possible to identify what features of the standard language are actually essential for communication. Seen in this light, Ogden’s proposal is comparable to Jenkins’ Lingua Franca Core (LFC) (Jenkins 2000). One is concerned with lexico-grammatical and the other with phonological features of English, but both seek to separate out those linguistic features that have general communicative salience from those that only incidentally serve the dual communicative and identifying function of particular native speaking communities. In this respect, the design of Basic English can be seen as the specification of a lingua-franca core.

Where the two specifications differ, of course, is that Jenkins’ LFC is based on empirical findings of how people make variable use of an existing linguistic code, which therefore, has vitality (Stewart 1968). Basic, on the other hand, is an artificially constructed code: nobody actually speaks it and it can only become vital if it is actually put to use. So the de-nationalizing of English has the effect of de-naturalizing it. The problem posed by Basic was how far artificially abstracted encoding specifications can be vitalized to become a natural means of communication.

English as an international language

As it turned out, the problem disappeared as Ogden’s proposal was overtaken by events. Basic was designed to meet the needs of international communication, but as the century progressed, English was becoming increasingly established as an international language without the need for artificial intervention. The fledgling domain of scientific enquiry, for example, which Sprat believed needed to be linguistically nurtured, took flight and found its own global manner of discourse in English under its own impetus. Attention now was focused not on how the language might be refashioned for international use, but on how this use affected the status of the language – how ‘English’ was to be defined now that it had become what Larry Smith called an ‘International auxiliary language (EIAL)’ (Smith 1980). As Strevens puts it:

In the case of the language called ‘English’ the sheer numbers of English users whose individual performances (and competences) are summated within the fiction of ‘English’, their worldwide geographical distribution, the great range of social needs and purposes they serve, and the resulting myriad of identifiably different versions of English – all these factors combine to produce a paradox: as English becomes ever more widely used, so it becomes ever more difficult to characterize in ways that support the fiction of a simple, single language.

(Strevens 1980: 79)

Over 30 years later Seidlhofer points out that this fiction that English is one language and ‘evermore shall be so’ still prevails and is a major obstacle in understanding the nature of ELF (Seidlhofer 2011). What then has happened over the intervening years? How far has the changed status of English been recognized and what characterization has been proposed as a factual alternative?

Strevens seeks to resolve the paradox he refers to by making a distinction between intranational and international uses of the language. By the former term he means the English adopted as a community language in ex-colonial countries, those in what Kachru later called the outer circle (Kachru 1985). Other countries, he says, need the language ‘for contact with the external world, for communication with other individuals and communities, for access to science, and other uses for which English is the vehicle’ (Strevens 1980: 81).

p.105

The other countries that Strevens specifically mentions are Japan, Turkey and Brazil – countries located in Kachru’s expanding circle. Interestingly, however, Strevens makes the point that communities who use English intranationally also need to use the language for international purposes as well, so that his concept of international cuts across the Kachru distinctions and in effect refers to the use of English as a lingua franca. However, subsequent developments in the study of EIAL did not focus on international English in this sense but on the intranational English, that of outer-circle users. And the main concern was to describe the recurrent linguistic features that defined distinct so-called nativized varieties associated with particular communities of users. These were accordingly assigned independent status as World Englishes, thereby providing symbolic confirmation of the political independence of these ex-colonial countries. The English of the inner circle was thereby de-nationalized and then re-nationalized as the language of outer-circle communities. The primary focus was, then, on language as a means of expressing social identity, and in this respect the World Englishes movement (the WE paradigm as it is usually called) can be seen as the empirical analogue of Webster’s proposals for modifying English so as to make it a symbol of independence and separate national identity – proposals that, when taken up and vitalized in use, became established in the United States as, in effect, the first of the World Englishes.

World Englishes and English as a lingua franca

The extensive work on World Englishes, initiated and inspired by Kachru and Smith, and published in the journal of that name, was, then, principally concerned with the description of varieties, linguistically distinct versions of intranational English that had their own communal identity and integrity. The variable manifestations of ELF, of international English as Strevens defines it, are not accounted for. Indeed, there is a tendency among WE proponents to dismiss such variation as unworthy of study precisely because it does not display the formal regularities that can be systematically identified as constituting a separate variety (for further discussion see Seidlhofer 2015).

Interestingly, Strevens himself takes up a position that is in line with an ELF rather than a WE perspective on variation. Referring to Smith’s label EIAL (English as an international auxiliary language) he makes the following comment:

[T]he locution ‘English as …’ presupposed the existence, to the minds of some people, of ‘an English’ i.e. of a degree of reification, of a more or less finite, describable entity different in some definable respects from other forms of English.

(Strevens 1980: 94)

But in the minds of WE people it is presupposed that there are indeed such describable entities, forms of English that are different from each other in some definable respects. This is precisely what is held to justify the plural form World Englishes.

Strevens continues with a disclaimer: ‘This has not been the intention of “auxiliary” in EIAL. On the contrary, the original concept had been of differing purposes for using English rather than different versions of the language’ (Strevens 1980: 94; emphasis in the original). If this was indeed the original concept of EIAL, it would seem to correspond more closely to an ELF rather than a WE perspective, one that recognizes that there is English in the world other than World Englishes, and one that looks at the process of variation itself, at how English takes various forms to satisfy differing communicative purposes rather than at different formal versions of the language (for further discussion see Widdowson 2015).

p.106

Intentionally or not, these remarks by Strevens can be seen as indicating these two separate future developments in the study of international English. But these are perhaps more clearly signaled 10 years earlier in research that makes no reference to international English at all. I refer here to the work of William Labov. In a celebrated paper (Labov 1969a) he challenges the privileged status of Standard English and the idea prevalent at the time that failure to conform to it on the part of speakers of what was then called Black English vernacular (BEV) was evidence of diminished communicative and cognitive capabilities and that its speakers had a verbal deficiency that needed to be remedied by education. This idea that non-conformity to Standard English results in verbal deprivation and communicative impairment is one that ELF scholars are all too familiar with. They too have had to contend with the idea of the essential superiority of the standard language, an idea as institutionally entrenched as ever, and the assumption that users of the language who do not conform to it must necessarily be communicatively defective and in need of remedial treatment by educational intervention – witness the widespread view of ELF that it is really only another name for learner English. In this respect, Labov anticipates a basic tenet of ELF study that people are perfectly capable of communicating effectively without conforming to the norms of the standard language and that such conformity will even often be contextually inappropriate and may well result in ineffective communication.

But we need to note that the people Labov is talking about in this paper are members of a particular community of speakers, and his aim is to demonstrate that the language they use is contextually appropropriate for their own intra-community purposes. In this case, therefore, the communicative and identifying functions of language use converge. Labov is concerned to show not only that the use of features of non-standard English can be pragmatically effective on particular contextual occasions, but that these features are consistent markers of social identity. Though non-conformist in relation to the norms of the standard language, they conform to other norms that define a different version of the language with its own system of rules and its own social conventions of appropriate use. As Labov puts it ,‘All linguists who work with non-standard Negro English recognize that it is a separate system, closely related to Standard English but separate from the surrounding white dialects by a number of persistent and systematic differences’ (Labov 1969a: 32).

Labov’s demonstration of the logic of non-standard English is designed to present a case for the recognition of African American English as having its own independent status and integrity as a variety in its own right. There is a clear correspondence here with the work of WE scholars, similarly concerned with establishing versions of English that are dialectal in that they represent the means of communication and the expression of identity of separate speech communities, defined as groups of people who share the same primary socio-cultural space. They too focus attention on intra-community domains of use where the communicative and identifying functions naturally converge.

Language variety and variation

But of course, what we have with ELF is the use of English between people who do not belong to the same speech communities, as these are traditionally defined, who therefore do not share the same primary socio-cultural or lingua-cultural space. The question arises as to what communicative demands are made on the language in ELF domains of use and how these demands can be provided for. We can perhaps best approach this question by first considering domains of use where the demands are small and relatively easy to meet. I refer to the use of so called pidginized forms of English. These have been a subject of linguistic study for over 50 years, and again, raise issues that bear directly on current thinking about ELF. In reference to Smith’s EIAL, pidgin domains of use are local and not international but like EIAL the language can be described in the two different ways already discussed: in ELF terms as a variable process of pidginization or in WE terms as a pidgin, a language variety. This is a representative definition:

p.107

The process of pidginization is usually assumed to begin when a language is used only for very limited communication between groups who speak different native languages. Sharply restricted in domains of use, it undergoes varying degrees of simplification and admixture. If a new stable variety of the language emerges from this process, it might be described as a pidgin.

(Rickford 1977: 191–192)

Since pidgin English is used for ‘communication between groups who speak different native languages’ it conforms to Seidlhofer’s definition of ELF as ‘any use of English among speakers of different first languages for whom English is the communicative medium of choice, and often the only option’ (Seidlhofer 2011: 7).

And as with ELF what is emphasized in Rickford’s definition is use as a process. Just as a pidgin may emerge as a stable variety so it is conceivable that variation in ELF interaction in some contexts may over time also lead to relative stability, when, for example, regularities in ELF use emerge in what has been called a particular discourse community (Swales 1990) or community of practice (Wenger 1998). And of course if the language becomes communal in this way, so it acquires an identifying function accordingly.

Variation, then, can be studied in two ways. One way is to relate language variation to language change, to trace its developmental trajectory by identifying interim stages of its emergent progress towards stabilization. So one might identify certain rudimentary forms of English as a basilect and then trace how through variable and extended use it develops into a mesolect and subsequently into an acrolect, an established variety (Bickerton 1975). Similarly, Schneider (2012) argues that in some of its manifestations at least, ELF variability can be taken as an interim stage of variety development. Such an approach focuses attention on the linguistic properties that are manifested in different ways of using English and so is essentially concerned with variation as an indication of language change.

But a second way of studying variation is to consider the process itself and the communicative conditions that give rise to it. Here the focus of attention is not on linguistic forms as such but on their pragmatic function, that is to say, to quote Strevens again, on ‘differing purposes for using English rather than different versions of the language’ (1980: 94; emphasis in the original). To do this is to move from considering not the linguistic features of the texts that are manifested in language use but the discourse that they realize – what users mean by the linguistic forms they use. Some of these discoursal purposes are satisfied by a very rudimentary use of linguistic forms and here there are uses of ELF that can indeed be characterized as pidginized English. But the point is that such uses are appropriate to purpose and as and when purposes make greater communicative demands, so the language quite naturally complexifies to meet them for, as Halliday puts it: ‘The particular form taken by the grammatical systems of language is closely related to the social and personal needs that language is required to serve’ (1970: 142).

p.108

In the study of pidginization, this complexification is generally discussed as a feature of creolization, the process whereby the language extends into a wider range of domains of use and eventually gets stabilized as a creole: a variety, a version of the language associated with a particular community of speakers, in short, a World English. And as such, as mentioned earlier, the communicative and identifying function of the language converge. But the two functions do not have to converge. This process of complexification does not have to be uniquely associated with creolization. As the use of English extends into the domains of international communication, it will naturally complexify to meet a wide and heterogeneous range of contexts and purposes. The use will sometimes approximate to the linguistic norms that characterize native-speaking varieties. But this approximation is incidental, a function of communicative expediency, and carries no necessary identifying significance. What brings about complexity in ELF usage, and the extent to which it is or is not in conformity to established encoded forms of the language is the complexity of the contextual demands that it has to satisfy. These demands may be intra-community in a conventional sense, as is the case with creolization, and, as we have seen with BEV, but they can also be, and in ELF are, inter-community – international in the Strevens sense. With ELF we have the same motivation that furthers the development process of creolization – the increasing functional demands that lead to formal complexity – but these are associated not with intra- but with inter-communal contexts and purposes. With ELF we have variation that is not necessarily an interim stage on the way to a variety but a continually adaptive pragmatic use of linguistic resources (see Widdowson 2015).

One can of course describe the actual linguistic forms that this variation can take without subscribing to the view that they are embryonic features of a variety. This involves focusing attention on the variants that are textually produced rather than on the pragmatic discoursal process that produced them. This is generally done by referring variant forms to the conventional norms of standard usage. Thus, with some ELF interactions, one can point to such textual features as the ‘absence’ of definite articles or the third person s, or the non-standard use of prepositions or the plural morpheme. To do this is to give privileged benchmark status to these conventional norms. Where variants are so regular as to constitute a variety, as in the case of World Englishes, they are then, of course no longer considered as errors but are legitimatized as sociolinguistic markers of identity and pragmatically appropriate to context in intra-community domains of use. But where these variants are used inter-communally, that is to say when people from different primary communities communicate, as is the case with ELF, the error stigma tends to remain, in spite of the evidence that they serve to textualize their users’ discourse intentions in pragmatically effective ways.

Text and discourse: analysis and interpretation

There are then two approaches to the description of ELF variation, and they correspond to two historical traditions in linguistics. One treats instances of language use, what Saussure refers to as actes de parole, as text, the overt linguistic forms that language users produce. The second approach considers what users intend to mean by the texts they produce, and how their texts are variously interpreted, in other words what discourse a text is designed to realize, and what discourse is derived from it. To illustrate the distinction we can return again to Labov. When he is pointing out the distinctive linguistic features of a particular instance of BEV: for example, negative concord (as in ‘you aint goin’ to no heaven’) or invariant be (as in ‘when they be sayin’) he is talking about text. When he says that these features are used in a contextually appropriate way to express a complex argument, then he is interpreting the text as discourse. When he demonstrates that BEV is systematic, he is treating it as text but when he argues it is logical, he is treating it as discourse.

p.109

Over recent decades, with the development of corpus linguistics, textual description has become increasingly precise and detailed. This has revealed that there are recurring patterns of usage within texts, co-textual inter-dependencies that combine linguistic items into phraseological units in accordance with what Sinclair calls the ‘idiom principle’ (Sinclair 1991). Since text is obviously produced as a sequence of linear units, the recently proposed ‘linear unit grammar’ (Sinclair and Mauranen 2006) would therefore seem to be particularly well suited to its description. And it has indeed been applied to the description of ELF text by Mauranen and her colleagues (Mauranen 2012).

Text description is a matter of tracing the occurrence of different linguistic forms, and, as indicated earlier, where they vary from conventional usage, these variants can be described by taking some established version of the language as a norm of reference. This is what Labov does when pointing out the distinctive features of BEV, and what Schneider does when he makes lists of how variants in ELF compare with those in WE varieties of English. In both of these cases the variants that are identified are isolated grammatical features. Where the linear unit approach to the description of ELF text differs, and breaks new ground, is that the variants identified are within phrasal sequences where linguistic elements are compounded into larger linear units. What is of particular interest here is not the occurrence of particular grammatical or lexical features in ELF text, but the extent to which their occurrence in these phrasal units differs from ‘normal’ usage.

This is the approach that is applied to those data in the ELFA corpus of spoken English that are produced by non-native speakers in academic domains of use. The non-conformist variants that occur are described as approximations: phrasal units that resemble but do not fully replicate normal English. In producing such variants ELF users are said to approximate to normal English but do not quite get it right. In some cases, the approximations are said to be ‘formal’ in that it is the linguistic form that users do not get right, but in other cases the approximation is said to be ‘semantic’ in that the linguistic form is ‘correct’ but not used in accordance with the conventions of idiomatic usage. Approximations, then, are ‘expressions that are close to the target, but not entirely precise’ (Mauranen 2015: 40)

The linear unit approach to text description, though innovative in many ways, is not entirely unprecedented. In the middle of the last century, Zellig Harris expressed the belief, very much in tune with current thinking, that ‘language does not occur in stray words or sentences but in connected discourse’ (Harris 1952: 3). Harris made no distinction between discourse and text and so he proceeded to consider the nature of this connection by proposing procedures for analyzing texts by identifying what he calls morphological combinations as sequential patterns of structural and semantic equivalences. At a time well before the advent of corpus linguistics, his approach to text analysis was of course necessarily somewhat rudimentary and very different from that based on linear unit grammar, but it raises a fundamental issue about the nature of such analysis – an issue that is directly relevant to the study of ELF. After demonstrating how his analysis works, Harris concedes its limitations:

All this, however, is still distinct from an interpretation of the findings, which must take the meaning of morphemes into consideration and ask what the author was about when he produced the text. Such interpretation is obviously quite separate from the formal findings, although it may follow closely in the directions which the formal findings indicate.

(Harris 1952: 29; emphasis in the original)

p.110

To ‘ask what the author was about when he produced the text’ is to raise the pragmatic questions of what discourse the author intended to textualize, and what discourse might be interpretatively derived from it. Although Harris recognizes that text analysis alone cannot deal with such questions, he suggests that discourse interpretation may follow closely from its findings. He does not explain, however, how it might follow, just how the formal features of texts are indicative of their pragmatic significance. The issue of the relationship between text analysis and discourse significance is left unresolved.

And it is an issue that is raised in the linear unit analysis of ELF text. Here too the question arises as to how its formal findings are to be interpreted. When approximations are said to be ‘not entirely precise’ this refers to their linguistic form and not to their pragmatic function: obviously being pragmatically precise does not depend on how closely one gets to a target linguistic norm. So the question arises as to what motivates these approximations, what these ELF users ‘were about’ when they produced their texts? As with any pragmatic use of language, the linguistic forms that occur are those that users have reason to suppose are appropriate to context and purpose. In the domains of academic use that ELFA is concerned with, what these ELF users assume to be appropriate is presumably the English that is established as conventionally normal in academic discourse. Not surprisingly, then, it is these norms to which they seek to conform. Approximations are of their nature norm-referenced and if the norm is known, meanings can be inferred by reference to it: ‘Approximations that are close enough to their target may not pose too much difficulty for a hearer to construct the meaning from the elements that are there’ (Mauranen 2015: 42). The target referred to here is presumably the English that is conventionally used in academic texts, that is to say, the standard language, and if the hearers are familiar with this, then it is indeed likely that they would be able to normalize the approximation. This, of course, presupposes that meanings are inscribed in their particular textualized form and so are not directly recoverable from the non-conformist variant but only via reference to the standard norm. Effective communication, therefore, is assumed to depend on conformity.

But of course these conditions that encourage deference to conventional norms do not by any means obtain in all domains of ELF use. As in any other use of language, linguistic variation in ELF is pragmatically determined as users draw expediently and adaptively on whatever formal resources they have at their disposal to get their meanings across, and these resources will, as has been pointed out in the ELF literature, be multilingual in that they will naturally include the users’ knowledge of their own first languages (Klimpfinger 2009; Hülmbauer and Seidlhofer 2013; Jenkins 2015; Hülmbauer 2016). So in many, perhaps most cases, ELF users cannot rely on a mutual knowledge of conventionalized norms. As, for example, research based on the VOICE corpus has shown (e.g. Pitzl forthcoming; Seidlhofer 2009), they have to negotiate conditions for understanding as they go along and what is meant by the variants they use, whatever their linguistic provenance, has to be inferred in the discourse process by reference to context and purpose. In these cases where there is no conventionalized norm to refer or defer to, the interpretation of what users are about does not, as Harris puts it ‘follow closely’ from the textual forms they produce. This, of course, underlines the importance of recognising that the linguistic features of ELF are pragmatically determined by local contextual factors and cannot be generalized. As Harris indicates, the relationship between the formal findings of text analysis and their interpretation as discourse remains problematic.

And not only problematic for the study of ELF. As Labov points out, the understanding of all language use involves an enquiry into ‘how things are done with words and how one interprets these utterances as actions: in other words, relating what is done to what is said and what is said to what is done’ (Labov 1969b: 54–55).

p.111

Conclusion

And it is this that links ELF study with the past. For all the work that I have referred to, in one way or another, has been concerned with this relationship between what is said and what is done, with what forms of English are needed to appropriately textualize the discourses of different domains of use. For Sprat, the forms appropriate to the discourse of scientific enquiry had to be of the ‘mathematical plainness’ he supposed to be characteristic of the speech of artisans; for Webster and Ogden the language code itself had to be refashioned to suit changed national and international discourse needs respectively. Subsequent sociolinguistic study has shown how users fashion existing linguistic resources for themselves to suit their communicative needs. What informs all of these developments is the recognition that different discourses, whether these are transient or become stabilized as varieties, will make different textualizing demands on linguistic resources, and it is the same recognition that informs the study of ELF.

With digital communication and the vastly extended networks of interaction, discourses increase in their diversity and so the forms of English that textualize them will quite naturally vary in accordingly diverse ways. The extent of this diversity is unprecedented, but not the adaptive process of variable discourse textualization that drives it. In this respect ELF as a phenomenon is not new. And, since, as we have seen, different approaches to dealing with this adaptive process have a long history, ELF study is not in its essentials new either. It too has its precedents, and tracing them can perhaps reveal more clearly just what these essentials are.