Studies in Variation, Contacts and Change in English
Please cite this article as:
Mackenzie, Ian. 2015. “Will English as a lingua franca impact on native English?”. Can We Predict Linguistic Change? (Studies in Variation, Contacts and Change in English 16), ed. by Christina Sanchez-Stockhammer. Helsinki: VARIENG.
author = "Ian MacKenzie",
title = "Will English as a lingua franca impact on native English?",
series = "Studies in Variation, Contact and Change in English",
year = 2015,
booktitle = "Can We Predict Linguistic Change?",
number = "16",
editor = "Christina Sanchez-Stockhammer",
publisher = "VARIENG",
address = "Helsinki",
url = "https://urn.fi/URN:NBN:fi:varieng:series-16-2",
issn = "1797-4453"
English used as a lingua franca (ELF) by non-native speakers contains a huge amount of variation and usually differs quite considerably from English as a native language, although there is some evidence that innovations or hybrid forms shared by speakers of many L1s are diffusing into common ELF usage. In this article I describe some of these forms and consider the psycholinguistic and sociolinguistic mechanisms that might lead to – or prevent – such usages being propagated among native English speakers. Although English has a long history of language contact and language shift, with, e.g., Celts (Britons) and Scandinavians adding to and subtracting from the language, this happened long before there was widespread literacy, education, mass media, and language standardization. Today, native speakers could imitate ELF forms if they found them particularly expressive, or for reasons of (overt or covert) prestige – imitating the speech of a culturally dominant or socially attractive group, although this currently appears improbable. Moreover, non-native speakers who relocate to Anglophone countries are more likely to accommodate to local speakers than vice versa, and their children will acquire the variety of the wider local community. In parts of larger cities this could be a ‘multi-ethnolect’ containing various L2-derived features, but such forms are unlikely to be widely diffused. At present, there is no evidence that native English speakers are accommodating to ELF usages, and my (tentative) prediction is that this is unlikely to happen.
English used as a lingua franca (ELF) by non-native speakers contains a huge amount of variation and usually differs quite considerably from English as a native language (ENL), either as a consequence of imperfect learning, or because ELF speakers choose to disregard forms they consider to be communicatively redundant, and instead seek to enhance communicative effectiveness. There is some evidence that innovations or hybrid forms shared by speakers of many L1s are diffusing into common ELF usage, and some ELF researchers argue that these usages will inevitably have an effect on native English. In this article I will consider the psycholinguistic and sociolinguistic mechanisms that might lead to – or prevent – such usages being propagated among native English speakers. My argument will take into account the fact that English in Britain has a long history of language contact and language shift, with (among others) Celts (Britons) and Scandinavians adding to and subtracting from the language, as predictions about the future generally stand to gain from a knowledge of the past. 
English as a lingua franca is a term that reifies the kind (or kinds) of English used around the world as an additional or international language. ELF can be defined 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). This definition does not exclude the usages of native English speakers in ELF interactions, or those of speakers of nativized or indigenized ‘outer circle’ varieties, but this article concerns usages that are typical of non-native ELF speakers.  Spoken (and to a lesser extent written) ELF usually differs from ENL, and contains a lot of non-standard forms. As Ferguson (2009: 129) puts it, it should be viewed “as a fluid cluster of communicative practices where speakers draw on a wide, not clearly bounded range of linguistic features – some standard, some non-standard, and others not English at all (at least according to the conventional view).”
ELF clearly does not need to be the same as ENL, as it is not part of any native target culture in which particular ways of speaking and behaving are appropriate. Rather than imitating the norms of native English speakers, ELF speakers are said to (or recommended to) adopt ways of speaking (with their bi- or multilingual interlocutors) which aid mutual intelligibility and successful communication. Hence ELF must be “functionally not formally defined; it is not a variety of English but a variable way of using it” (Seidlhofer 2011: 77). The proponents of ELF argue that it is different from English as a native language but not deficient or inferior (e.g. Seidlhofer 2011: 120). Thus the difference between ELF and the older concept of EFL (English as a foreign language) is an outlook or an attitude: while EFL learners make mistakes (or errors), ELF users are said to show a lot of variety: instead of restricting themselves to the realizations of native English speakers, they exploit unused latent possibilities of English morphology, syntax and phraseology. 
ELF speakers greatly outnumber native speakers of English, and many researchers expect ELF to impinge on ENL.  For example, Seidlhofer (2003: 7) asserts that “it is the non-native speakers of English who will be the main agents in the ways English is used, is maintained, and changes.” More recently, Mauranen (2012: 33) has argued that “[i]t is reasonable to expect the sheer scale of ELF use to have an effect on the English language,” and that “it would be surprising” if ELF did not “have a significant impact on ENL communities” (Mauranen 2012: 57).
Prime candidates for diffusion are forms that are widely shared among the world’s L2 English speakers, because they are common in many other languages and therefore acquired comparatively easily in English as an additional language. As Mauranen (2012: 29) puts it, ELF is a site of “second-order language contact” – a contact between hybrids: “a large number of languages are each in contact with English, and it is these contact varieties (similects) that are, in turn, in contact with each other. Their special features, resulting from crosslinguistic transfer, come together much like dialects in contact” (Mauranen 2012: 30).  Mauranen suggests, very plausibly, that innovations or hybrid forms shared by speakers of many L1s “will diffuse into common usage, so that ELF will become more stable” (Mauranen 2012: 32) and focused. Speakers tend to accommodate to each other’s linguistic uses, adhering to what Keller (1994: 100) calls a fundamental maxim of communication: “Talk like the people around you.” Mauranen (2012: 33) describes current ELF speakers as the “first generation” following the “explosive expansion” of English use that coincided with the internet in the mid-1990s, and suggests that “by the time the third generation learns English, we may expect English already to show clear traces of lingua franca influence.” Mauranen’s “English” here must include ENL.
As one would expect, many innovative or hybrid forms that have diffused into common ELF usage are simplifications of grammatical features of ENL that differ from those of the majority of natural languages, and are therefore ‘difficult’ for L2 learners. These include:
These usages can easily be explained in terms of L2 transfer or grammatical replication, analogical levelling, simplification, generalization, reconceptualization, the reduction of tense and aspect distinctions, the reduction of redundancy, and its opposite – increased explicitness by the exploitation of redundancy, and so on.  All of them except (2) are also found in many nativized New or World English varieties (Kortmann & Lunkenheimer 2013). But while some ELF speakers may systematically use such forms, others are much less consistent in their linguistic choices, and use a seemingly random range of variants, with different degrees of conformity to the preferred expressions of the native language, and to those of their supposed L2 contact variety or similect, as a result of what Mauranen (2012: 41, 217) calls the fuzzy processing and shaky entrenchment or – from another perspective – the imperfect learning to be expected in a second language.
Variability in a native language is now generally assumed to be a form of what Weinreich et al. (1968: 101) describe as “orderly” or “structured heterogeneity,” in which variants are conventionally associated with social values that determine their appropriate use. Hearers perceive or recognize particular forms as having social as well as semantic values (or as encoding sociolinguistic as well as linguistic information). Particular forms can index a speaker’s age, sex, status, regional origin, and so on. Native-like command is taken to include the evaluation and use of heterogeneous structures, and sensitivity to contexts of use. Yet this is often not the case in ELF, in which speakers tend not to replicate native speakers’ contextually determined style shifting. Given the wide range of L1-influenced similects, as well as the existence of speakers whose usages are less consistent than the logic of similects would suggest, native English speakers often experience ENL-ELF interactions as a form of multiple dialect contact.
Thus there are two competing ways of describing widely shared ELF usages – as accidental or intentional. We can describe them as errors resulting from imperfect learning, given “the lousy language learning abilities of the human adult” (Trudgill 2001: 372), and invoke the many plausible mechanisms described by SLA researchers, such as transfer and interference (Weinreich 1953), interlanguage and fossilization (Selinker 1972), approximative systems (Nemser 1971), crosslinguistic influence (Kellerman & Sharwood Smith 1986), imposition and source language agentivity (Van Coetsem 1988), and so on. Alternatively, we can describe them as the product of the conscious and deliberate disregard of forms that are known to be part of the native language, but which ELF speakers consider incongruous or idiosyncratic or afunctional or communicatively redundant, in order to increase regularity, minimize redundancy, reduce ambiguity, and generally enhance explicitness, clarity and communicative effectiveness, and also, perhaps, to symbolize non-native-English-speaking-identity. This might be described as the “lingua franca outlook” on language, a natural corollary of the “lingua franca factor,” which consists of “the inherent interactional and linguistic variability that lingua franca interactions entail” (Firth 2009: 150).
This functional and indeed teleological account flies in the face of many standard accounts of the nature of language change, such as Lass’s (1997: xviii) insistence that language users “simply have to make do with what’s historically presented to them, and cope with it when it changes,” and Keller’s (1994: 13) apodictic statement that “speakers change their language neither intentionally, nor to a plan, nor consciously. This is generally true, and there is nothing more to it.” For Keller, language change (certainly that of a native language) is an epiphenomenon and a non-intended consequence or by-product of people’s aggregated linguistic choices. For many ELF researchers (e.g. Erling & Bartlett 2006; Jenkins 2007; Seidlhofer 2011; Cogo & Dewey 2012), by contrast, changes in the use of English by non-native speakers can be intentional, deliberate and functional. For the purposes of this article, however, we can remain wholly agnostic as to whether the imperfect learning or the deliberate change argument is the more likely; what is important is that either way, these widely-attested usages exist.  Innovations have occurred, among greater or lesser numbers of ELF speakers, and what remains to be seen is whether they will be more widely diffused, and propagated among native English speakers.
Importantly, the attested innovations in ELF corpora circumvent what Weinreich et al. (1968: 102) call the actuation problem inherent in functional accounts of language change: if a change in a structural feature that takes place in a particular language at a given time has a functional motivation, why does it not occur simultaneously in other languages with the same feature, or indeed why was it not actuated sooner?   It can safely be assumed that speakers of ELF similects are speaking English in much the same way as speakers of their L1 did in the past, although there used to be fewer L2 English speakers. The difference is that in the days when ELF was known as EFL (English as a foreign language), they would most likely not have had the insouciance and self-confidence that come with the “lingua franca outlook.”
But the question in my title remains to be answered: will English as a lingua franca impact on native English? If so, how will this happen? There are indubitably a large number of non-native speakers of English, but it is also widely argued that most ELF usage takes place in the absence of speakers of ENL. For instance, Thomason (2001: 21) states that most users of English as a lingua franca “have no opportunity (and often no desire) to practise by talking to native speakers of English.” Jenkins (2006: 161) describes native English speakers as a “small minority” in intercultural communication, so presumably most of them must be spending most of their time at home in the inner circle countries. Back in the days when Jenkins (2000: 227–228) used the term English as an International Language (EIL) rather than ELF, she even suggested that in inner circle countries, this variety would need to be incorporated into the secondary school curriculum, while “[f]or those who have already reached adulthood, it will be necessary to attend adult EIL classes in the same way that ‘NNS’ [non-native speaker, in scare quotes] adults do.” So if native speakers rarely hear ELF variants, how or why will they imitate them? Language change via accommodation and diffusion is generally held to require wide-scale, protracted, intensive, naturalistic, face-to-face interaction, and there is a critical threshold for the awareness of new features.  Both Bloomfield (1933: 476) and Labov (2001: 19) posit a principle of density, such that the diffusion of change can be explained mechanically by the amount and frequency of communication. I will argue below that today, non-native (ELF or ESL) speakers who find themselves for any length of time in inner circle countries are more likely to accommodate to the local majority than vice versa, even if this was not always the case in centuries past. 
There is a long history of contact-based change in English (and in all the other languages spoken natively in the British Isles for the past 2000 years), with Celts (Britons) and Scandinavians both adding to and subtracting from the language as a result of transfer and imperfect learning following language contact and language shift.
Leaving aside the question of what proportion of the British population shifted from a Celtic language (Brittonic or Brythonic) to Latin (Late Spoken Latin) during four centuries of Roman occupation, there are many reasons to believe that speakers of Brythonic Celtic languages (Cornish and Welsh), who gradually assimilated to Anglo-Saxon culture, transferred Celtic syntactic constructions into Old English over several generations, sometime after the 6th century, especially in the western and northern regions furthest from the Anglo-Saxon heartlands.  These notably include periphrastic do or do-support, and the progressive form (Filppula et al. 2008).
Although periphrastic do – the use of do as a semantically empty syntactic place-filler indicating tense – only appears in Middle English texts, it may have existed in vernacular registers for several centuries, but have been considered too colloquial or inelegant to be suitable for the written language of the Old English period. While it is now only used in affirmative sentences as an intensifier, periphrastic do is used to this day in negatives and questions, and it is argued that this aided the change from Old English’s SOV order in subordinate clauses (and frequent VS order elsewhere) to the SVO order that became standard in Middle English (McWhorter 2011: 279). It is also suggested that the (periphrastic) English progressive to be + the-ing verb form (which can also function as a gerund) derives from the Brythonic construction of the substantive verb ‘be’ + preposition yn + verbal noun (Filppula et al. 2008: 27, 60).
Tristram (1999: 30) points out that early language contact and the shift from Brythonic to English have been supplemented by continuous contact with Welsh speakers ever since, with loose-knit open network ties close to the Welsh-English border, and geographical mobility, so that present-day English can be described as “a ‘Celticised’ West-Germanic language.” Tristram also insists that it was speakers of Celtic languages who triggered what must be seen as a typological change from a predominantly synthetic language to a predominantly analytic one, with the result that today both the Celtic languages and English have periphrastic constructions including do-support and the progressive, a preference for a nominal style with light verbs, analytically formed prepositional and phrasal verbs, relative clauses with stranded prepositions, and the internal possessor construction (the use of possessive pronouns rather than the article for body parts, etc. – e.g. she’s washing her hair, rather than the hair). Filppula et al. (2008: 24) point out that the situation involving extensive societal bilingualism with considerable L2 transfer that existed before the Celts in England shifted to English sometime before the 10th century or so was replicated more recently when there was a shift from Celtic languages to English in most of Wales, Scotland and Ireland.
Alongside these additions to the language, there are many other differences between Modern English and Old English (or between English and all the other Germanic languages), that can be thought of as simplifications. These include the absence of grammatical gender marking on articles, pronouns, nouns and adjectives; of inflectional case marking; of inherent reflexivity marking (e.g. English to remember vs. German sich erinnern, ‘to remember yourself’); of directionally marked adverbs (such as the former hither and thither); of the second-person thou pronoun; and of V2 word order (placing finite verbs in second position of a declarative main clause) (McWhorter 2007: Chapter 4). It has often been suggested that the simplification of these relatively complex and overspecified features was the result of non-native acquisition (the same process as that said to operate in ELF today).
Post-adolescent and adult learners prefer regularity and transparency to irregularity and opacity. They find things like inflections and grammatical gender and agreement hard to learn, so they tend to simplify, or indeed pidginize (Trudgill 2011: 35) the language they are learning, by regularizing irregularities, increasing morphological transparency (replacing synthetic structures with analytic ones), and reducing redundancy, while also transferring elements from their L1.  In cases of language shift, this leads to children learning their parents’ imperfectly acquired L2 as their L1, and subsequently transmitting this simplified grammar to the next generation. If there is a large enough proportion of non-native speakers, native speakers begin to accommodate to simplified aspects of the non-native variety which are thereby transferred to the native variety. Hence English can be described as a regularized and simplified high-contact koiné (Trudgill 2011: 67), or a semi-creole (McWhorter 2011: 295), resulting from dialect levelling, the loss of minority forms, and perhaps the creation of interdialect forms, or alternatively from feature competition and selection (Mufwene 2001).
So far, so good. Disagreement arises, however, concerning the respective roles of the Celts and Scandinavians. Trudgill (2011: 34) argues that simplification occurs when adults learn a language, while long-term contact involving territorial co-habitation, intermarriage and child bilingualism is likely to lead to complexification. Hence the simplification of Old English into Middle English must have resulted from contact with the Brittonic Celtic language (or languages), because Anglo-Saxon/Norse bilingualism should have complexified English rather than simplified it. McWhorter (2011), on the contrary, argues that whereas the Celts maintained their languages for generations and acquired English gradually, infusing it with Celtic features along the way, the Scandinavian Vikings – largely males who practised exogamy – needed to function in Old English almost as soon as they invaded, but acquired (and passed on) a functional but approximate and incomplete version of the language. Although Old English and Old Norse had similar phonetic and morphological patterns, parallel grammatical structures, and many cognate words, specific forms (such as verb inflections, and gender attributions in the noun phrase) often conflicted. It may have been the case that omitting certain grammatical features did not impede communication. Hence mutual accommodation between speakers of two languages with an adstratal rather than a superstratal/substratal relation led to convergence and koineization, and a linguacultural fusion of Angles and Danes in the north of England, in which English lost various grammatical features, including some that it shared with Old Norse.
For present purposes, it does not matter whether the contact-induced simplifications in English during this period were the work of the Celts, or the Vikings, or both; the point is that they occurred.  The question here is whether the processes that simplified English in the 7th to 9th centuries are still in operation, and whether post-adolescent and adult acquisition will influence native English in the future. The answer to this question would appear to be no, however: as McWhorter (2011: 17) stresses, the period of simplification by way of widespread second-language acquisition preceded “widespread literacy, education, and language standardization.” Languages that were simplified when their transmission was “interrupted” by incomplete non-native acquisition were largely used orally, at a time when reading and writing were “elite activities” (McWhorter 2011: 17).  Today, however, the native version of the English language “is always available and influential in print, the media, and schools,” and consequently “a highly complex language can survive endless waves of second-language acquisition intact” (McWhorter 2011: 17). 
Furthermore, even though non-native speakers of English (and speakers of nativized New English varieties) greatly outnumber native ones in the world today, most of the former are (logically, if not by definition) not living in ‘inner circle’ English-speaking countries. Just like most (younger) immigrants to inner circle countries today, ELF users who spend any length of time in an ENL environment are far more likely to accommodate to the local community’s ways of speaking than vice versa. Most native English speakers probably are exposed to non-native usages, but the average Briton or American or Australian today probably hears far less non-native English than, say, the Anglo-Saxons in the 5th–9th centuries who were exposed to the L2 English of speakers of various Celtic and Norse languages. Today, the children of non-native English speakers who relocate to Anglophone countries, just like children whose parents speak a markedly different native dialect from the local one, will almost invariably acquire the variety of the wider local community rather than their parents’ ELF (Weinreich et al. 1968: 145; Labov 1982: 47).
While the innovations and variants of non-native speakers are easily explicable – either as the accidental outcomes of imperfect learning, or as the deliberate disregard of structural features deemed to be communicatively redundant – their replication and dissemination by native speakers would have to involve different reasons. There would have to be reasons (if speakers did not perceive some kind of advantage in new forms, they would not use them) and the most plausible candidates are the perception of heightened expressiveness, and considerations of prestige.
Enhanced expressivity is widely evoked as a cause of language change. As well as the static maxim quoted above – “Talk like the people around you” – Keller (1994: 101) posits dynamic maxims such as “Talk in such a way that you are noticed” and “Talk in an amusing, funny, etc. way.” Harris & Campbell (1995: 54) discuss what they call “exploratory expressions” used for emphasis, reinforcement, clarity, avoidance of ambiguity, or just for fun (or simply by mistake) – constructions that “may be judged ungrammatical, stylistically odd, or foreign, but will nevertheless be understood.”  Some ELF usages could easily fall into this category, and potentially be picked up by native speakers “as a poetic expression, as a periphrasis motivated by the desperation of not finding a more appropriate means of expression, as a way of deliberately producing stylistic oddity or foreign flavour, or for other stylistic reasons or communicative needs” (Harris & Campbell 1995: 54). An expression widely adopted in this way would be reanalyzed as an acceptable part of the grammar.
Even Lightfoot (1991: 60), who does not accept that adults can continue to change their ‘grammar’ throughout the lifespan, proposes expressiveness as the causal mechanism for changes in use. Haspelmath (1999) objects that expressiveness just means saying very clearly what you think, which is not very different from another of Keller’s maxims (which Haspelmath labels the maxim of clarity), “talk in such a way that you are understood,” and suggests that there is rather a maxim of extravagance, or better still, since the intention is to be noticed and to impress the hearer, a maxim of impressiveness. Although native speakers are very unlikely to see any functional advantage to abandoning, say, delicate nuances of meaning provided by an aspectual system that was many centuries in the making, they might easily pick up on particular non-native variants – different ways of saying the ‘same thing’ – or particular tokens of particular constructions, if they sound more expressive (or impressive), or appear to have a certain stylistic or poetic effect or discourse function.
The other main reason to replicate common non-native usages would involve prestige. This could be what Labov (2006 ) called overt prestige: imitating the speech of a culturally dominant group, but also the contrary – covert or negative prestige: imitating the speech of a socially attractive but non-dominant group. As Milroy (1992: 129) points out, prestige is not the same thing as standardization: although many people seeking socioeconomic advancement consider it to be in their interests to use standard forms, prestige can also “be subjectively attached by speakers to forms or varieties which are very distant from, and in conflict with, the codified norms of the standard.”
Thus some native speakers could identify with particular groups of ELF speakers and imitate or accommodate to their usages (Giles & Smith 1979), in what might be called acts of identity (Le Page & Tabouret-Keller 1985).  For example, professional people could regularly encounter ELF speakers with the same type of job, background and educational qualifications as themselves, form relatively strong ties with them, and imitate some of their ways of speaking English. However, the diffusion of linguistic innovations among native speakers does not necessarily require large-scale identification with the innovators. According to the Milroys’ (1985) account of language change (and its contrary, deliberate language maintenance), it is only necessary that a suitably respected or admired member of a social network also has ties (probably weak ones) with another group or network, and begins to imitate one of their usages in interactions with his or her own primary network, and is imitated in turn. As Milroy (1992: 180–181) puts it, “mobile individuals who have contracted many weak ties … are in a particularly strong position to carry information across social boundaries and to diffuse innovations of all kinds.” If a few members of a (largely or wholly native speaking) social network take up an innovation (in this instance originating with ELF speakers), the innovation can then diffuse to the group as a whole, thereby assuming a certain social significance, and symbolizing or indexing some sort of social value associated with the group, but without any conscious aspiration to sound like or identify with the originators of the innovation. As long as there are a few early adopters, the standard S-curve is likely to appear: the change, which started slowly, will pick up speed and proceed rapidly, before slowing down again when most speakers have adopted the change in question. 
If enough people adopt new variants (probably individual expressions – particular tokens of certain constructions), they will progressively become more automatized or routinized or more deeply entrenched in most hearers’ minds (Langacker 1987: 58–59), thereby becoming primed and more likely to be activated, merely as a function of exposure to use, and independently of any intentional goals (Croft 2000: 32). Diffusion of innovations – leading to what Croft (2000) calls differential variation, or a shift in the relative frequencies of variants – can proceed by way of what Thomason (2001: Chapter 6) calls “passive familiarity” rather than “deliberate decision.”
Yet native English speakers are likely to encounter a wide range of ELF speakers, from service industry employees to (perhaps) professional ELF-using colleagues, using the language very differently, and for very different purposes. Not all of these people will be perceived by native speakers as belonging to a prestigious group with a social status to identify with. Consequently, many of their variants are more likely to be perceived as usages lacking any sociolinguistic value or stylistic salience, than to be reanalyzed as indicators of social prestige that should perhaps be imitated. Extensive weak ties in loose networks are necessary conditions for change, but they are clearly not sufficient ones.
As argued above, changes or simplifications arising from second-language acquisition today have to compete with a highly standardized language widely used in schools, in the media, and in print. Native-like command of heterogeneous (and socially marked) structures also implies an awareness of what is not native-like. All non-native usages run up against a community’s linguistic regularities or conventions, and the nature of conventions is that almost everyone in a speech community conforms to them, expects almost everyone else to conform to them, and would prefer any additional member of the community to conform to them (Croft 2000: 98).
If adult speakers could not change their language, however, ELF forms would have almost zero chance of being diffused among native speakers. The nature of most ELF interactions – professional contexts as well as social life and tourism – is such that it tends to be adults who are exposed to ELF variants. Children are marginal in the social networks that could conceivably lead to the propagation of linguistic changes, so change would have to come from accommodation by adults and subsequent propagation to children in the input for L1 acquisition. Yet whether they imitate non-native usages or not, it does seem that “people of all ages can (and do) modify and restructure their language” (Kerswill 1996: 178), and continue to acquire additional social and stylistic registers.  Harris & Campbell (1995: 49) insist that “[t]he grammar of an adult can change” and is “best viewed, not as an inflexible completed object, but as an adaptable, constantly growing set of generalizations.” Construction grammarians (like sociolinguists) tend to emphasize the existence of variation, arguing that languages and their component parts are creatively stretched with use and change even when they are not being widely learned and used as L2s. For example, Hopper (1988: 120) states that grammatical regularities “are always provisional and are continually subject to negotiation, renovation and abandonment,” so that grammar is forever emergent. Hilpert (2013: 2) argues that “with every new text genre and every additional speaker, more variation enters the picture, and less of an invariant core system remains,” and certainly ELF (along with World Englishes) has led to an expansion of the feature pool, bringing about what Meyerhoff & Niedzielski (2003: 547) call “a broadening of the set of forms available to English speakers.” 
This is all in contrast to the Chomskyan theory that, rather than ‘simply’ building up a set of constructions from the utterances they hear, children are endowed with an innate Universal Grammar (UG) and a Language Acquisition Device that scans the ambient linguistic input of external language or E-language to which the child is exposed and sets the predefined parameters of UG to match it, thereby constructing a grammar or internal language or I-language. According to this theory, language change is not so much the result of innovations invented and disseminated by adults as of children internalizing new grammars by setting the parameters of UG differently from their caregivers (or the previous generation as a whole) as a result of changes, possibly minute ones, in the primary linguistic data or E-language, which can trigger a new system that generates a very different set of sentences and structures (see Lightfoot 1991, 2006).
Yet a great deal of evidence (e.g. Weinreich et al. 1968: 184–185; Labov 1982: 46–47; Kerswill 1996) suggests that pre-adolescent children and adolescents turn away from the norms of their caregivers and accommodate to the norms of their peers and slightly older children (probably after an interim ‘bidialectal’ period during which they have two heterogeneous systems characterized by orderly differentiation), several years after the original setting of UG parameters. Where children acquire their particular dialect pattern from a relatively homogeneous peer group (or from slightly older children), they are likely to perceive occasional or infrequent “ELF usages” in the speech of both children and adults as irrelevant ungrammatical oddities. In fact Chambers (2002: 123) argues that the children of immigrants seem to be wholly unaware that their parents speak differently than the wider community until they reach the critical period around puberty. They have an “innate accent filter” that leads them to pick up native patterns of phonology and speak like their peers; it is as if they had “a device which allows individuals to disengage themselves from certain eccentric or at least diversionary communities of practice that fall within their worlds, presumably in order to allow them to participate fully in the communities that will play more integral roles in forming their identity.” This disengagement seems to apply to lexicogrammar as well as phonology, so that community norms are perpetuated by the offspring of immigrants. 
On the other hand, Meisel (2011: 138) argues that if children acquire a language in an environment without established community norms in which non-native speakers are in the majority, e.g. in ‘heteroglossic’ working-class neighbourhoods in major cities with many heritage speakers, the variable “input for the next generation … may very well contain the kind of conflicting evidence which is likely to trigger reanalysis in subsequent first language acquisition,” leading to language change on the part of monolingual native speakers as well as bilinguals. A construction-based account, on the contrary, would describe children as having acquired multiple constructions (probably including transitory ones), used in different settings, even if generative grammarians would describe them as requiring different UG parameter settings. The same might well apply to children simultaneously acquiring both a local and a standard dialect.
What appears to be far more frequent than the resetting of (monolingual) ‘parameters’ by children acquiring their first language is teenagers in multi-ethnic urban peer groups choosing to mix forms from languages spoken by their friends. This can be a way of negotiating individual and group identities, which Rampton (2005) describes as “crossing.” In multilingual immigrant communities in which younger children acquire the target language largely by way of unguided informal group second language acquisition in multiethnic friendship groups, instead of from their caregivers, the result can be “a new ‘multi-ethnolect’ containing some features of minority origin” (Kerswill et al. 2013: 267), or if not a ‘lect’ as such, at least a smallish repertoire of shared innovations. A number of linguists have given examples of such “polylingual languaging” (Jørgensen 2008) in cities in north-west Europe. Many of these usages, however (certainly morphosyntactic ones, as opposed to phonological ones), are likely to recede in adulthood, and the majority of native English speakers in Britain and the USA are no more likely to accommodate to such innovations, or to similarly innovative hybrid forms from ELF, than they have accommodated, over the past two centuries, to major immigrant varieties (e.g. Irish, Italian, Yiddish, Hispanic, or indeed African-American). 
However, despite the reasons and mechanisms that might lead native speakers to imitate non-native uses, the question remains as to whether most ELF speakers who relocate to inner circle countries continue to “express both their L1 identity and membership of the international ELF community” (Jenkins 2007: 25), or rather try to adhere to Keller’s maxim of talking like the people around them, in this case the native-speaking majority. The same applies to migrating speakers of nativized World English varieties. It is for this reason that Meierkord (2012: 2) argues that ELF should be subsumed into a broader category, “Interactions across Englishes,” involving speakers of all three circles.
Cogo & Dewey (2012: 43) have analysed and based a number of conclusions on a corpus recorded in a “community of practice” of foreign language teachers living and working in London. This community does indeed appear to constitute an ELF setting in an inner circle country in which native speakers are “no longer the norm providers” (Cogo & Jenkins 2010: 275), but Cogo & Dewey (2006: 78) also suggest that non-native speakers tend to accommodate much more to native uses when (perhaps due to some “unplanned and unavoidable occurrence”) an L1 English speaker is present. Many other professionals do feel the need to conform to ENL norms (and ask their ENL speaking colleagues to proof-read their written communications). As Ferguson (2012: 179) puts it, “what counts as an effective use of linguistic resources in an ELF context in Bratislava, say, may index something rather different, and attract different, less favourable evaluations, when used, say, in a formal presentation to a mixed American/international audience in New York.” Consequently, irrespective of any willingness to imitate non-native usages, native speakers may not in fact be exposed to all that many of them.
I have suggested above that grammar is inherently variable and probabilistic and usage-based and forever emergent, and that adolescents and adults might accommodate to new usages if they consider them to be more expressive or impressive than conventional usages, or saw them as somehow carrying social prestige. So will native English speakers accommodate to the constructional (and phraseological and lexical) variation of ELF and depart from the entrenched constructions of their native speech community? For the moment, corpora show no evidence of any of the widely diffused ELF usages listed above spreading into ENL:
Thus overall, few of these widespread ELF features seem to offer particular expressive gains to native speakers of English, and they seem unlikely to appeal to speakers who want to be noticed and to impress by finding new ways to say old things. These features are certainly salient to native speakers: just as much as the 3rd person singular -s inflection, the logic of uncountable nouns, and the constraints on using stative verbs in the progressive form have low salience and communicative value for non-native speakers, the absence of the -s inflection, the use of the plural with uncountable nouns, and the use of the progressive with stative verbs tend to be noticed by native speakers. The last of these features in particular is frequently used by comedians and script-writers to imitate the English of non-native speakers. On the other hand, usage-based theory holds that speakers are intuitive statisticians who keep a mental record of the relative frequency of use of all the constructions they hear (Bybee 2007), so the more native speakers are exposed to such usages, the more they will become entrenched, and potentially primed for use by way of passive familiarity.
Native speakers of English do modify their speech when talking to non-native speakers with limited proficiency, usually by syntactic simplifications, recasts and repetition; this is often called “foreigner talk” (Ferguson 1971), and parallels the way parents modify their speech for young children (caretaker talk). In fact interactional adjustments by native speakers and more competent language learners are at the heart of the “interactionist approach” to foreign language teaching (Gass & Mackey 2015). But of course such dynamic modifications tend not to diffuse into everyday native English usage.
There may be many other frequent ELF usages that are less salient (and less likely to be stigmatized by English language teachers), as well as individual speakers’ “exploratory expressions,” that are slowly finding their way into ENL (into the inherent variability of some individual speakers’ heterogeneous repertoires) through dialect contact. Language contact is inherently unpredictable. Given intense and/or long-lasting contact, anything can happen: Thomason & Kaufman’s (1988: 14) statement that “as far as the strictly linguistic possibilities go, any linguistic feature can be transferred from any language to any other language” is widely quoted. Moreover, as Heine & Kuteva (2005: 5) point out, “[c]ontact-induced language change is a complex process that not infrequently extends over centuries, or even millennia,” as with some of the Celtic examples above. Therefore, any predictions we make should be suitably modest or cautious.
I am less convinced than Mauranen (2012: 57) that “it would be surprising” if ELF did not “have a significant impact on ENL communities,” but Peter Schrijver (personal communication) suggests that given time, ENL is likely to be “changed beyond recognition,” and that my position resembles that of “a first-century AD upper-class Roman from Rome” stating that ‘New Latin’ would not succeed in taking over “because in the 200 years or so that Rome had been a world power nothing of the sort had happened yet.”
In short, I believe that we can safely predict the following: that post-adolescents and adults acquiring L2s will continue to simplify and reduce and regularize them, and transfer elements from their L1s, but also that many expanding circle speakers who relocate to inner circle countries will strive for something resembling native-like usage; that some (or many, or most) people will continue to innovate linguistically, trying either to express themselves better or just to impress other people; and that most people will continue intermittently to imitate the speech of culturally dominant or socially attractive groups. What we cannot state with any degree of certainty is that native English speakers will find non-native English speakers either culturally dominant or sufficiently socially attractive to imitate. We can also predict that while most younger and second generation immigrants will acquire or adopt the local community’s ways of speaking, parts of larger cities with multi-cultural communities and ongoing immigration could see the development of multi-ethnolects containing various L2-derived features. However, such forms are unlikely to be widely diffused beyond these communities. As McWhorter (2011: 17, quoted above) points out, unlike in the 7th to 9th centuries, when it appears highly probable that native English speakers were influenced by the imperfect L2 English of Celts and Scandinavians, today there is widespread literacy, and standardized native varieties of English remain dominant in education and the media, and will probably resist the effects of non-native acquisition. At present, there is little evidence that native English speakers are accommodating to usages deriving from nativized World Englishes or from ELF, or that this is going to happen in the future. But of course, as always, time alone will tell.
My thanks go to the editor of this volume and to anonymous reviewers whose comments and suggestions have – as always – greatly benefitted this article.
 Historical sociolinguists regularly invoke the uniformitarian principle, such that natural processes occurring today almost certainly also occurred in the past. As Labov (1972: 161) puts it, “the same mechanisms which operated to produce the large-scale changes of the past may be observed operating in the current changes taking place around us.” We may also expect them to operate in the future. [Go back up]
 Following Kachru (1985: 12), it has become conventional to talk about three circles of English. The countries in which English is used (by native speakers) as a primary language make up the inner circle, while countries in which English has an official second-language role in a multilingual setting (mostly former British colonies in Africa and Asia) constitute the outer circle. That part of the world where English is learnt and used as a foreign (or additional) language, for communication with speakers from all three circles, is categorized as the expanding circle. [Go back up]
 E.g. Cogo & Dewey (2012: 78) suggest that “[d]eciding what constitutes an error” is “not a particularly ELF-compatible way of thinking about language” and “probably the wrong kind of question to ask.” [Go back up]
 According to Crystal (2003: 65–67) there are about 400 million native English speakers, 430 million L2 speakers of (outer circle) World Englishes or New Englishes, and a global figure (including the expanding circle) of 1.5 billion English speakers. Crystal (2008: 5) revises this figure “in the direction of two billion.” [Go back up]
 Trudgill (2004) outlines a similar process in the formation of new colonial dialects (in Australia, New Zealand, South Africa, etc.) by way of dialect mixing. Children were faced with a hugely variable set of models, with no single peer-group model to accommodate to. Eventually, a new dialect appeared consisting of those variants that were most widely shared in the mixture. [Go back up]
 See MacKenzie (2014, Chapter 4) and the references therein. Other ‘unfocused’ aspects of ‘ELF grammar,’ not always attributable to specific similects, include the very varied use of prepositions, articles, plurals, and verb complementation patterns. [Go back up]
 There is no contradiction here: reduced redundancy involves the omission of inflections and distinctions that are not salient to non-native speakers; increased explicitness (such as adding prepositions by analogy with other verbs) is only redundant from a native speaker viewpoint. [Go back up]
 I am inclined to believe that a certain proportion of ELF speakers are aware that their usages differ from those of native English speakers, and that they purposely use their ‘ELF variants’ because they believe them to have functional (communicative, discursive) advantages, but this will be the subject of another article – or, like several parts of this paper, a chapter of a longer work. For a contrary view, see Mollin (2006). [Go back up]
 Analysis of attested, innovative (or at least non-native) usages of ELF speakers also sidesteps the problem that explanations of language change often account for dissemination rather than innovation, taking it for granted that novel variants already exist somewhere, perhaps in “a restricted subgroup of the speech community” (Labov 1972: 178), or in another social network with which a speaker has weak ties (Milroy & Milroy 1985), so that as Croft (2000: 55) points out, the Milroys’ “innovator” is in fact merely a transmitter. [Go back up]
 Dixon (1997: 106) suggests that as people watch television and movies from other countries, British, American and Australian dialects are converging: people “pick up speech habits from each other – words and grammatical constructions, and even some habits of pronunciation.” See also Meyerhoff & Niedzielski (2003). It used to be argued (e.g. Trudgill 1986: 40) that electronic media are not very instrumental in the diffusion of linguistic innovations, but this was before people (especially young people) began to spend a large proportion of their time using the internet, watching YouTube, etc. [Go back up]
 Schrijver (2014: 31–33, 48, 82) argues that Latin-speaking Britons (or possibly Latin-Celtic bilinguals) from the higher social classes fled west and north from the Anglo-Saxon warrior-settlers. As they greatly outnumbered the speakers of Highland British Celtic, they Latinized its sound system, and consequently that of its present-day survivors Welsh and Breton. [Go back up]
 It is widely argued that adolescents are past a fixed, neurologically-given ‘critical threshold’ (Lenneberg 1967) for language acquisition, but a usage-based approach provides an alternative explanation: the difficulties of adult L2 acquisition are simply the result of prior L1 learning – the entrenchment of particular patterns and constructions – and transfer, an inescapable and pervasive component of language learning (Ellis 2006). [Go back up]
 Unless they didn’t, of course. Thomason & Kaufman (1988: 264) deny that “the degree of change exhibited in English” in this period was “anything other than normal,” and insist that the simplifications seen in Middle English “probably were taking place in OE before Norse influence became relevant” (Thomason & Kaufman 1988: 303). The “Norsification” of English was less a case of accommodation to imperfect learning than “a fad whereby an English speaker would parade his knowledge of Norse while speaking English” (Thomason & Kaufman 1988: 298), a kind of prestige borrowing because Norse added “a few subtleties of meaning and a large number of new ways of saying old things” (Thomason & Kaufman 1988: 303). Thomason & Kaufman (1988) disregard the possible influence of Celtic speakers. Meanwhile Lass (1980: 139) rejects external or contact explanations on principle, describing language contact as a “rather gross” contingency, and bemoaning those “contact romantics” who seek to derive “the maximal number of characters in a given language from contact sources” (Lass 1997: 201). He insists that language-internal explanations must be preferred to exogenous ones for “considerations of parsimony,” and that “endogenous change is … part of the nature of the beast” (Lass 1997: 207–208), to which Milroy (2003: 145) retorts, “it is in the nature of the beast [language] to resist satisfactory explanations on how it can change within itself.” It will be clear that the present article is written by an unabashed contact romantic. [Go back up]
 Some examples of grammatically anomalous, expressive ways of saying things, some from known authors: a rose is a rose is a rose; been there, done that; curiouser and curiouser; I could care less (meaning ‘couldn’t’); long time no see; needless to say; no can do. [Go back up]
 There is a huge literature on the way speakers imitate each other. See also Bock (1986) on syntactic priming; Niederhoffer & Pennebaker (2002) on linguistic style matching; Pickering & Garrod (2004) on subconscious interactive alignment; Szmrecsanyi (2005) on ‘persistence’ and ‘micro-entrenchment’; etc. [Go back up]
 In fact, as Aitchison (2013: 95) suggests, S-curves tend to be made up of lots of little overlapping S-curves, each corresponding to a different register or genre or linguistic environment. [Go back up]
 It could be argued – and quite possibly demonstrated – that this is not in fact a broadening of the set of available forms, because every conceivable variant used by an ELF speaker has at some point been attested in traditional rural dialects in the inner circle and/or somewhere in outer circle varieties (see e.g. Kortmann & Lunkenheimer 2013), but many usages may appear to be new to inner circle speakers not conversant with traditional dialects or World Englishes. [Go back up]
 Cheshire et al. (2011: 167–168) suggest instead, perhaps more plausibly, that such children are aware of their parents’ variants as part of the “feature pool” (Mufwene 2001: 4), but choose to orient to community norms. The claim that infrequent “ELF data” are largely irrelevant to the child is just as valid for usage-based or constructionist accounts of language acquisition (e.g. Clark 2003; Tomasello 2003) as it is for Lightfoot’s Chomskyan account. [Go back up]
 Trudgill (2010: Chapter 5) gives some examples from American English of what seem to be the results of contact with German, Yiddish and other European languages, including the grammatical constructions I like to skate (as opposed to I like skating) and Are you coming with? (as opposed to Are you coming?), and restricted collocations of verbs like have and take compared with British English. However, these constitute only a few very minor features. [Go back up]
 Except (although decreasingly so) in traditional East Anglian dialects in England. Trudgill (2002: 97–99) gives a speculative but convincing explanation of the origin of this feature that depends on the high proportion of non-native speakers in the area in the 16th century (see also Trudgill 2010: Chapter 2). This critical mass of non-native speakers is not found in inner circle countries today. [Go back up]
 E.g. of the regularized past tenses quoted above from ELFA, the BNC (British National Corpus) gives just 3 hits for feeled, 3 for bringed (but all in a metalinguistic context, talking about wrong usages), and zero hits for fighted, heared, losed and teached. [Go back up]
 I have had a number of native-English-speaking graduate translation students (i.e. people with a certain interest in and feel for languages) who claimed to be wholly unaware of these distinctions. [Go back up]
 Moreover, such a change would go against the standard path of grammaticalization. Bybee et al. (1994) describe a common path in which resultative constructions in a language (expressing a past action with a resulting current state) develop into perfects (expressing the current relevance of a past action), which then evolve into pasts or perfectives (expressing temporarily bounded situations); there are no cases of perfects giving way to the present. [Go back up]
 E.g. of the verbs from VOICE cited in Seidlhofer (2011: 145–146), COCA (the Corpus of Contemporary American English) gives 9 hits for discuss about (compared with 12 in VOICE), zero for reject against, and 64 for return back. [Go back up]
BNC = Davies, Mark. 2004–. BYU-BNC. (Based on the British National Corpus from Oxford University Press). Available online at http://corpus.byu.edu/bnc
COCA = Davies, Mark. 2008–. The Corpus of Contemporary American English: 450 million words, 1990-present. Available online at http://corpus.byu.edu/coca
ELFA = The Corpus of English as a Lingua Franca in Academic Settings. 2008. Director: Anna Mauranen. http://www.helsinki.fi/elfa/elfacorpus
VOICE = The Vienna-Oxford International Corpus of English (version 1.0 online). 2009. Director: Barbara Seidlhofer; Researchers: Angelika Breiteneder, Theresa Klimpfinger, Stefan Majewski, Marie-Luise Pitzl. http://www.univie.ac.at/voice
Bock, J. Kathryn. 1986. “Syntactic persistence in language production”. Cognitive Psychology 18: 355–387. doi:10.1016/0010-0285(86)90004-6
Chambers, J.K. 2002. “Dynamics of dialect convergence”. Journal of Sociolinguistics 6(1): 117–130. doi:10.1111/1467-9481.00180
Cheshire, Jenny, Paul Kerswill, Sue Fox & Eivind Torgersen. 2011. “Contact, the feature pool and the speech community: The emergence of Multicultural London English”. Journal of Sociolinguistics 15(2): 151–196. doi:10.1111/j.1467-9841.2011.00478.x
Cogo, Alessia & Jennifer Jenkins. 2010. “English as a lingua franca in Europe: A mismatch between policy and practice”. European Journal of Language Policy 2(2): 271–294. doi:10.3828/ejlp.2010.16
Cogo, Alessia & Martin Dewey. 2006. “Efficiency in ELF communication: From pragmatic motives to lexico-grammatical innovation”. Nordic Journal of English Studies 5(2): 59–93. http://ojs.ub.gu.se/ojs/index.php/njes/article/view/65/69
Crystal, David. 2008. “Two thousand million?” English Today 24(1): 3–6. doi:10.1017/S0266078408000023
Ellis, Nick C. 2006. “Selective attention and transfer phenomena in L2 acquisition: Contingency, cue competition, salience, interference, overshadowing, blocking, and perceptual learning”. Applied Linguistics 27(2): 164–194. doi:10.1093/applin/aml015
Erling, Elizabeth J. & Tom Bartlett. 2006. “Making English their own: The use of ELF among students of English at the FUB”. Nordic Journal of English Studies 5(2): 9–40. http://ojs.ub.gu.se/ojs/index.php/njes/article/view/69/73
Ferguson, Charles A. 1971. “Absence of copula and the notion of simplicity: a study of normal speech, baby talk, foreigner talk and pidgins”. Pidginization and Creolization of Languages, ed. by Dell Hymes, 141–150. Cambridge: Cambridge University Press.
Ferguson, Gibson. 2009. “Issues in researching English as a lingua franca: A conceptual enquiry”. International Journal of Applied Linguistics 19(2): 117–135. doi:10.1111/j.1473-4192.2009.00225.x
Ferguson, Gibson. 2012. “The practice of ELF”. Journal of English as a Lingua Franca 1(1): 177–180. doi:10.1515/jelf-2012-0009
Firth, Alan. 2009. “The lingua franca factor”. Intercultural Pragmatics 6(2): 147–170. doi:10.1515/IPRG.2009.009
Gass, Susan M. & Alison Mackey. 2015. “Input, interaction and output in second language acquisition”. Theories in Second Language Acquisition, 2nd edn, ed. by Bill VanPatten & Jessica Williams, 180–206. New York: Routledge.
Haspelmath, Martin. 1999. “Why is grammaticalization irreversible?” Linguistics 37(6): 1043–1068. doi:10.1515/ling.37.6.1043
Jenkins, Jennifer. 2006. “Current perspectives on teaching World Englishes and English as a Lingua Franca”. TESOL Quarterly 40(1): 157–181. doi:10.2307/40264515
Jørgensen, Normann J. 2008. “Polylingual languaging around and among children and adolescents”. International Journal of Multilingualism 5(3): 161–176. doi:10.1080/14790710802387562
Kachru, Braj B. 1985. “Standards, codification, and sociolinguistic realism: The English language in the outer circle”. English in the World: Teaching and Learning the Language and Literatures, ed. by Randolph Quirk & H.G. Widdowson, 11–30. Cambridge: Cambridge University Press.
Kerswill, Paul. 1996. “Children, adolescents and language change”. Language Variation and Change 8(2): 177–202. doi:10.1017/S0954394500001137
Lass, Roger. 1990. “How to do things with junk: Exaptation in language evolution”. Journal of Linguistics. 26(1): 79–102. doi:10.1017/S0022226700014432
Meisel, Jürgen M. 2011. “Bilingual language acquisition and theories of diachronic change: Bilingualism as cause and effect of grammatical change”. Bilingualism: Language and Cognition 14(2): 121–145. doi:10.1017/S1366728910000143
Meyerhoff, Miriam & Nancy Niedzielski. 2003. “The globalisation of vernacular variation”. Journal of Sociolinguistics 7(4): 534–555. doi:10.1111/j.1467-9841.2003.00241.x
Milroy, James & Lesley Milroy. 1985. “Linguistic change, social network and speaker innovation”. Journal of Linguistics 21(2): 339–384. doi:10.1017/S0022226700010306
Niederhoffer, Kate G., & James W. Pennebaker. 2002. “Linguistic style matching in social interaction”. Journal of Language and Social Psychology 21(4): 337–360. doi:10.1177/026192702237953
Pickering, Martin J. & Simon Garrod. 2004. “Toward a mechanistic psychology of dialogue”. Behavioral and Brain Sciences 27: 169–225. doi:10.1017/S0140525X04000056
Seidlhofer, Barbara. 2003. “A concept of ‘international English’ and related issues: From ‘real English’ to ‘realistic English’?” Strasbourg: Council of Europe. http://www.coe.int/t/dg4/linguistic/source/seidlhoferen.pdf (PDF)
Szmrecsanyi, Benedikt. 2005. “Creatures of habit: A corpus-linguistic analysis of persistence in spoken English”. Corpus Linguistics and Linguistic Theory 1(1):113–149. doi:10.1515/cllt.2005.1.1.113
Trudgill, Peter. 2001. “Contact and simplification: Historical baggage and directionality in linguistic change”. Linguistic Typology 5(2–3): 371–374. doi:10.1515/lity.2001.003
Weinreich, Uriel, William Labov & Mervin Herzog. 1968. “Empirical foundations for a theory of language change”. Directions for Historical Linguistics, ed. by Winfred P. Lehmann & Yakow Malkiel, 95–189. Austin, TX: University of Texas Press.