English as a Lingua Franca : A New Variety in the New Expanding Circle ?

The traditional tripartite model of English world-wide has so far ignored one of the most important functions of English today, namely that of a lingua franca between non-native speakers. In integrating ELF into models such as Kachru’s, the question that arises is whether it can count as a variety. Evidence from a corpus-based study of lingua franca English as it is used between European speakers indicates that it is not a non-native variety in the traditional sense. It is suggested here that ELF could be conceptualised as a register, which can be integrated into the varietyand nation-based model only on a functional level. 1. Three circles and modern realities Of all the models describing English world-wide, Kachru’s (1985) threecircle model (which is a continuation of Quirk et al.’s 1972: 3 tripartite model) has proven the most influential, dividing English use into three categories: the Inner Circle, Outer Circle and Expanding Circle. However, as Bruthiaux (2003) has pointed out, Kachru’s model is problematic in several respects. Its most serious drawback is that it does not spell out clearly what it wants to categorize, since nations, types of speakers, functions of English as well as types of variety are all referred to. For instance, the model distinguishes native speakers in the Inner Circle from the non-native speakers in the Outer and Expanding Circles. In addition, it treats the functions that English performs in the communities in question, emphasizing that English is used in all domains and for all communication purposes in the nations of the Inner Circle, and for many important intranational functions in the Outer Circle. In the Expanding Circle, English is learnt as a foreign language for the purpose of communication with the Inner and Outer Circles. Finally, and importantly, the model considers the standard-orientation of the English used. The Inner Circle communities, which are termed norm-providing, possess their own varieties of English, while Outer Circle communities, which Kachru sees

as norm-developing, are in the process of developing their own varieties, the New Englishes.As learners, the speakers in the Expanding Circle are not given the right to their own variety-development, but are seen as norm-dependent.Kachru's three-circle model thus fuses several levels of analysis: it is nation-based, grouping different nations into the circles on the basis of a combination of nativeness, functions of English, and variety status.
A further problem, adding to the confusion of the model, however, has gained attention in recent years.The model does not appear to take into account the fact that English has acquired a new dominant function world-wide: that of a lingua franca between all three circles, but especially within the Expanding Circle.Research on English as a lingua franca has multiplied, and has indeed become the newest trend in the study of English world-wide.So far, however, an integration of these new developments into the Kachruvian model has not taken place, even though they have for many already altered the way we conceive of English generally.Yet in order to integrate English as a lingua franca (ELF) into our model of English and thus our conceptualization of the language, a definition of ELF is required in the first place.In fact, such a definition is still missing despite the recent research focus on English used as a lingua franca.While everybody seems to agree that the object of study is conversations in English between speakers who do not share a mother tongue, we do not yet seem to have a consensus as to the location of such conversations in theoretical, conceptual space.
For instance, we can find quotations such as the following, in which ELF is described as simply a new way of referring to speakers in the Expanding Circle: "since the mid-1990s it has become increasingly common to find EFL speakers referred to [...] as speakers of English as a Lingua Franca (ELF)" (Jenkins 2003: 4).On the other hand, Seidlhofer (2002: 8f.) suggests that ELF transcends the three Kachruvian circles, uniting all speakers of English in cross-cultural communication.While Mauranen (2003: 514) sees ELF as a "variety of English", Juliane House (personal communication) would rather term it a register.Indeed, ELF still needs to be defined more precisely.
If one wishes to define ELF in the terms of the tripartite model, the main question is whether it is a development that the Expanding Circle is making towards its own standard and its own variety, or whether, this not being the case, we should simply state in the model more clearly that the Expanding Circle has acquired ELF as its main function for English.Put differently, the question to answer is whether ELF is a variety -and this question has certainly not yet been answered.As James (2005: 133) suggests, "while the functional essence of the lingua franca (LF) is generally recognized, there is nonetheless a serious striving to adduce empirical evidence for the existence of structural commonalities characterizing the LF in its various manifestations."If such structural commonalities are found, we could certainly argue that the Kachruvian model would have to be re-written at the level of varieties and standards, granting the Expanding Circle the same right to its own standard as the other circles.

Variety status for ELF?
Since the main question in the present paper is whether ELF should count as a variety of English, one first needs to come to the vexed question of what a variety is.While all sociolinguists will subscribe to the view that it is a difficult concept to define, and what is more, that it is even more difficult in practice to decide empirically on the boundaries between different varieties, some basic aspects of the concept are agreed on.'Languages' (another difficult concept) can be divided up into different varieties, i.e. bundles of idiolects that share certain features.These bundles may be determined regionally (also called dialects) or socially (sociolects).To take up Ferguson's (1971: 30) definition: "A variety is any body of human speech patterns which is sufficiently homogeneous to be analyzed by available techniques of synchronic description [...]".Ferguson emphasizes the structural criterion of 'homogeneity' as defining a variety, while of course it remains problematic in practice to decide just how homogeneous a candidate bundle of idiolects will have to be in order to qualify as a variety.Certainly, such thresholds cannot be imposed, but will be judged on a case by case basis by the individual observer.In addition, other approaches to the concept of 'variety' have also incorporated the idea that varieties are frequently as much defined by speakers' perceptions as by inner-linguistic criteria (e.g.Hymes 1977: 123).However, even though speakers themselves may draw boundaries differently from linguists, their boundaries will nevertheless often be based on (salient) linguistic features: 'this is how we say it and not how they say it'.A certain degree of homogeneity thus remains a major factor in determining variety status in the following discussion.
Indeed, much of the debate on the concept of variety in sociolinguistics has focused on what the difference between a language and a variety is (e.g.Haugen's 1966 classic).Yet in the present context of English used as a lingua franca between non-native speakers worldwide, this issue is not so much of interest as the question of whether there are non-native varieties of a language at all and how one can decide if a bundle of non-native idiolects constitute a variety of the English language.In this context, it is the research on New Englishes that proves the most interesting.While Singh (1995: 323) has remarked correctly that a non-native variety of a language is a paradox from the structuralists' point of view, the present article takes the sociolinguistic view that non-native speakers in many post-colonial settings of the world have developed their own varieties (which are different from native varieties and homogeneous in themselves).Once more, as Kachru (1992: 55) states, there are two main characteristics that mark out new varieties: nativisation (the variety has developed its own characteristic structural features) and institutionalisation (the variety has become the standard in the mind for the speakers).The first again refers to linguistic homogeneity, marking the variety off from others, the second to attitudinal aspects.Homogeneity as a major requirement for variety status will be the aspect taken up in the empirical study reported on in this paper.Attitudinal aspects will still need to be addressed, yet it may be difficult for speakers to develop a perception of their own variety unless a certain degree of homogeneity is present.
To come back to the issue of ELF, the idiolects of Expanding Circle speakers engaging in lingua franca conversations would need to share a substantial number of features for ELF to be considered a variety.Even though lingua franca speakers from the Expanding Circle come from a multitude of different language backgrounds and exhibit very divergent levels of English competence, it is nevertheless conceivable that processes of accommodation could result in a new variety, perhaps with unmarked features winning out, as for instance Mauranen (2003: 515) suggests.After all, the New Englishes of the Outer Circle are evidence that independent non-native varieties may develop in exactly this fashion.Lingua franca English in the Expanding Circle thus has the theoretical potential of developing its own variety.However, one further, and essential characteristic of lingua franca communication might prevent such a negotiation.Crucially, ELF situations have the frequent attribute of occurring between ever new conversation partners, so that its speaker community is constantly in flux rather than remaining stable and fixed (cf.James 2005: 135).A user of English as a lingua franca thus has to accommodate to different other speakers from different linguistic and cultural backgrounds with different levels of competence in each speech situation.It is difficult to imagine how a negotiation towards a common standard in all of these ever-changing conversation situations would proceed.Common features in ELF would thus be rather surprising.While speakers in the Outer Circle speech communities have both had the need and the opportunity to use English in many daily situations, for most Expanding Circle speakers, using English as a lingua franca remains a rare adventure with changing partners, not part of their daily lives.
It would thus seem that there is simply not the time and number of conversations available for speakers to negotiate a new, durable common ground, making it rather surprising if common features in ELF were to be found. 3Indeed, research into the nature of ELF has so far resulted more in descriptions of linguistic features that cause successful or unsuccessful communication rather than in descriptions of common features as they are actually used (e.g.Jenkins 2000).Real common features have been identified first and foremost regarding discourse style and pragmatics (for an overview see Seidlhofer 2004: 217f.), suggesting that what unifies lingua franca speech is communication strategies rather than the result of any structural convergence.In any case, research into the structural features of lingua franca English has not matured enough to decide whether ELF does exist or is developing as a variety in its own right.Some new insights, however, are coming forward from a recent research project of mine.

The Euro-English project: Data and methods
While ELF is usually seen as a global development, it is also possible to begin to study it from a regional perspective at first.One good example is English used as a lingua franca in Europe (cf.e.g.James 2000; Jenkins and Seidlhofer 2001;Jenkins et al. 2001).My recently concluded project (Mollin 2006) set out to investigate whether European lingua franca communication has resulted in a new, independent variety of English in Europe, sometimes also called 'Euro-English' (cf.Jenkins et al. 2001).For instance, Jenkins has suggested that a bona fide variety with its own standard is emerging in Europe, likening its development to that of established New Englishes of the Outer Circle: At the start of the twenty-first century, Euro-English is only just emerging as a distinctive variety or group of varieties with its own identity which, like the Asian Englishes, rejects the concept of having to respect British English or American English norms.What has become clear is that English is evolving as a European lingua franca not only in restricted fields such as business and commerce, but also in a wide range of other contexts of communication including its increasing use as a language of socialisation.The progress of the codified Asian Englishes thus indicates the likely future developmental stages of Euro-English.(Jenkins 2003: 38) However, while structural commonalities have been found for the Outer Circle varieties of Asia, European lingua franca English still needed to be researched in this respect.My project accumulated a corpus of English used as a lingua franca between mainland Europeans (citizens of the European Union) of different mother tongues.The Corpus of Euro-English, 400,000 words large, consists of a spoken component (240,000 words) with public discussions as well as public speeches, and a written component (160,000 words) with spontaneous online writing, e.g. from discussion groups and chat rooms.The bulk of the material stems from EU-related contexts, including transcripts of European Commission speeches and press conferences as well as official chats with Commis-sioners. 4The spoken component thus reflects English as it is used daily in a rather closed community, between EU politicians, officials, and journalists, all from various European countries.The online written subcorpus also includes EU politicians to some extent (for instance, chats that European Commissioners held with citizens), but for the most part it represents ordinary citizens communicating with each other on European matters.(For further information on the corpus, cf.Mollin 2006.)The corpus was analyzed in search for common lexicogrammatical and morphosyntactic features across different mother tongues which would separate the European lingua franca usage from a native-speaker standard.Fortunately, a number of hypotheses about possible characteristics of a so-called Euro-English could be drawn upon, extracted from sources such as Alexander (1999), Décsy (1993), James (2000), and Seidlhofer (in Jenkins et al. 2001).In addition, features that are mentioned as frequent errors of European speakers in guidebooks such as Swan and Smith (2001) were also consulted, since these may indicate the direction of convergence to a new common standard as well.A good number of different features, ranging from semantic change for individual words to syntactic variables, could thus be analyzed in the corpus.For instance, the words that had been suggested to have developed new meanings in Euro-English were eventual (used in the sense of 'possible'), actual (in the sense of 'current'), and possibility (in the sense of 'opportunity').Other lexical aspects considered were the use of common verbs such as have, make and do, and fixed phrases.Grammatical variables included non-count nouns used countably, the distribution of the relative pronouns who and which, article use, invariable question tags, the use of prepositions, complementation patterns for nouns, verbs and adjectives, inflectional marking of the third person singular, and finally patterns in aspect use.The findings from the European corpus were then compared against native-speaker patterns with the help of the ICE-GB corpus as well as the BNC in order to discover any European patterns of usage different from the nativespeaker norm.

The Euro-English project: Results
The overall result of the study was that, foremost, speakers stick to native-speaker standard usage and make individual 'errors', if one wishes to name them so, depending on mother tongue and English competence generally.There were hardly any common features that united lingua franca speakers, even in a context such as the EU, where speakers use English frequently with each other and do have the opportunity to negotiate a common standard.For most variables, it is surprising how few deviations from the native-speaker standard there are at all.One is tempted to conclude that the common ground that lingua franca speakers resort to is not a new negotiated common variety but the standard variety that they have all been taught at school.There is no indication that English used as a lingua franca between speakers of different European mother tongues in the context sampled by the corpus has produced a new, coherent variety.
A good example of a low rate of deviations and usage true to the native-speaker model is the -s inflection marking the third person singular in the present tense, which Seidlhofer (in Jenkins et al. 2001: 16) has predicted will become lost in European lingua franca English.In the corpus, concordances of the third person singular pronouns he, she and it were searched for cases in which the verb requires a final -s (in the simple present, present progressive, and present perfect).In over 2,700 instances of these, the -s marker is only missing in a whole of 16 cases, which computes as a rate of deviation from the British English standard of no more than 0.58%.
Similarly, while Seidlhofer (ibid.) has suggested that European lingua franca English may be marked by an interchangeability of the relative pronouns who and which, the analysis of the Corpus of Euro-English reveals no such tendency.The concordances for who and which were checked for unusual combinations (i.e.inanimate referents with who and animate referents with which).For who, out of the 436 cases in which the relative pronoun referred to a simple noun phrase, only eight violated a native British pattern (as checked against the BNC), an error rate of 1.83%.For which, the rate lies even lower at 0.91% (12 out of 1324 occurrences).Again, the European speakers sampled in the corpus have not developed any new common standard in this respect.
The use of the present perfect could serve as a third example for findings from the Euro-English project which do not support the view of lingua franca English as a new variety.Both Décsy (1993: 15) and James (2000: 35) predict, based on the knowledge that this is a difficult area for learners of English to master, a loss of the present perfect in Euro-English.A first quantitative analysis of the corpus, in contrast, suggests that the present perfect is used even more frequently by the European speakers than by British speakers: while spoken Euro-English features 60.44 present perfect constructions per 10,000 words, spoken British English (as in the spoken component of ICE-GB) has only 37.55 per 10,000 words.However, qualitative analyses for selected high-frequency verbs (go, make and take) show that the present perfect in the European corpus is used in exactly the same contexts as in the British corpus.It is the sampling of the corpus itself that is responsible for the high frequency of perfective constructions.These occur mostly in briefings and press conferences, in which officials naturally do less of narrating the past than of explaining the present and recent events with a bearing on today.
Interestingly, looking at specific constructions which require the perfect in native-speaker English and which are known to be problematic for some groups of learners (such as constructions with for + time period and since + point in time), we find that the rate of using a non-perfective lies between 15 and 36%.However, these deviations from the nativespeaker standard are distributed over a range of alternative constructions, so that again we cannot speak of a new common standard in European lingua franca English.
While some other features are not yet clearly decided pending research based on larger corpora (such as noun complementation or the semantic change of eventual to mean 'possible'), one further feature that was analyzed may speak for the view that commonalities in ELF, if found, are not structurally meaningful, but rather a function of the communication strategies that lingua franca speakers pursue.This is the "over-use" of common verbs, as proposed by Seidlhofer (2004: 220) to be characteristic of lingua franca English generally.A count of the highfrequency verbs suggested by Seidlhofer (have, do, make, take, put), naturally only used as full verbs,5 in the spoken components of the Corpus of Euro-English and ICE-GB shows a much stronger preference for have, a slightly stronger preference for make and take as well as a slightly weaker preference for do and put on the part of the European speakers.A closer look at instances of have shows that while in many contexts the use of have is entirely in line with native-speaker usage, there are also many instances in which have is certainly not wrong, but may sound a little clumsy to native speakers, who may have chosen different verbs (see examples 1-3).
(1) I tried to have a dialogue on fair play with all of you with all of you, who follow this matter... (esdf-123) (2) [...] I think it is in all fairness probably best not to have a definitive judgement at this point [...] (esdf-112) (3) [...] on the European level we had the strongest consolidation for the for the budget only Finland has a stronger consolidation of its public finances [...] (esms-054) While each example on its own would not be out of place in a nativespeaker corpus, their mass taken together accounts for the over-use of have in the European corpus.Similarly, make seems to be used as an allpurpose verb.While its strongest collocates in verb+object constructions are to make a decision, to make progress, to make a proposal etc., all of which are native-like, we also find examples like sentences 4-6, in which make does not sound entirely appropriate to native-speaker ears.
(4) My first son is an economist, the second makes research in biology.(ewdc-001) (5) But it was secular turkey that made the genocide of Armenians and Kurds... (ewmg-328) (6) EU makes many wrong things but this problem does not come from EU... (ewmg-307) So even though not all of the common verbs suggested are actually over-used by the European lingua franca speakers, at least have and make point to the conclusion that in some cases the LF speakers prefer a certain all-purpose verb to other alternatives.Two reasons may account for this.Firstly, the speaker may not know a more appropriate verb, having a restricted vocabulary.Secondly, the speaker may choose a word which her interlocutor is certain to know over one that might cause difficulties in understanding.In either case, the common feature of using have and make more frequently than native speakers appears not to be a sign that ELF is a variety in its own right with its own stable characteristics, but a direct result of the communication purpose: getting meaning across from one non-native speaker to the other.Certainly, this communication purpose may (and perhaps will in the future) lead to a common structural variety, yet it seems that ELF has not arrived there yet.The results of the study presented above suggest that the function of getting meaning across has not led to common structural features.
Therefore, if it is the case that the characteristics of ELF depend on the situation and communication purposes of the speakers, we should indeed classify ELF not as a variety, but indeed as a register (as House has done) in the Hallidayan sense6 .Halliday et al. (1964: 77) distinguish "varieties according to users" (what has been defined as variety above) from "varieties according to use", which they term register.A register, language used for a specific function rather than by a specific group, has more potential for heterogeneity than a variety (what Halliday et al. would call dialect), and may thus better describe ELF.Characteristics of ELF, as the result of lingua franca communication strategies, could be shorter utterances, a smaller range of vocabulary generally, or the avoidance of what Seidlhofer (2004: 220) has called "unilateral idiomaticity", i.e. the use of fixed expressions and idioms which the interlocutor is unlikely to be familiar with.The perspective that ELF is a functional rather than a structural phenomenon also explains why much research on English used as a lingua franca has so far been concentrated in the field of pragmatics.
To sum up at this point, what the findings from the Euro-English project suggest is that ELF is no structurally coherent variety as Old and New Englishes are.Rather, it is a phenomenon at the level of language function, which means that in order to integrate it into Kachru's three circle-model, we do not need to disturb the original tripartition and would not group the Expanding Circle communities within the normdeveloping Outer Circle.It should suffice for us to be aware that the Expanding Circle uses English predominantly for lingua franca purposes.

Implications for teaching
The debate about the nature of ELF has from the very beginning also been a debate about teaching standards in the Expanding Circle.Many voices have put forward the argument that if Expanding Circle speakers use English mainly for lingua franca communication, English teaching should prepare them more for this than for communication with native speakers.Yet while it is certainly true that teachers and producers of teaching materials should do their best to make English teaching as relevant to the students as possible, the perspective that ELF is no stable variety as such would make it a bad teaching standard, as indeed most protagonists agree.Furthermore, there is a broad consensus that the needs of learners of English should not be presupposed by us linguists.Rather, we should allow them their own say in the question of which English they ought to be taught.
Interestingly, Timmis' (2002) survey of 400 English students in 14 different countries revealed that the learners are highly oriented towards a native-speaker standard and would like to approximate this standard as closely as possible.Native-like English appears to be a clear status symbol, while English with a recognizable accent as well as using English "my own way" and sometimes "saying things which native speakers think are grammar mistakes" (ibid.: 244) are undesirable to the students.Similarly, Dalton-Puffer et al. (1997) found a clearly negative attitude among Austrian advanced learners of English towards their own non-native accent, valuing native accents more highly.Finally, Murray's (2003) survey of Swiss English teachers reveals that the non-native teachers (more so than native speakers teaching in Switzerland) are rather conservative as far as teaching standards are concerned and would not accept a teaching model based on European lingua franca English.Murray's very plausible explanation is that the non-native teachers have invested heavily in their near-native English competence and do not want to see this achievement devalued (ibid.: 160).Generally, decades of foreign language teaching based on a native-speaker model seem to have made a good deal of persuasion necessary for a move away from the native-speaker teaching model, both for teachers and students.
Findings from the Euro-English project (Mollin 2006) support this view.Over 400 academics from several disciplines and from all over Europe were surveyed in an e-mail questionnaire study regarding their attitudes towards Euro-English and their judgements of what good English is or is not.For example, the participants were asked to indicate their agreement with certain statements.Two may be of special interest in the present context.The first was: "I am not bothered about mistakes that other learners of English make as long as I understand what they want to say."A majority of respondents (59.3%) agreed with this statement, while 25.88% obviously do mind other people's mistakes in lingua franca conversations.Interestingly, one respondent added: "I agree, so long as they are not my students", confirming the view that compliance with the native-speaker model is seen as a must in a pedagogical setting, even if perhaps not in the real life of lingua franca communication.This ties in nicely with attitudes towards the statement "Schools should teach English not as the native speakers speak it, but for efficient international communication", which had more opponents (43.43%) than supporters (33.10%), with 23.47% of neutral respondents.Whereas respondents think communicating at all is more important than a perfect conformity with native-speaker standards in LF conversations, they nevertheless tend towards keeping the native norms as teaching models.
To conclude, I would agree with Gnutzmann (2005: 116f.), who argues that L1 norms should, also in keeping with learners' own desires and aspirations, remain the teaching model in ELT.However, our understanding should be that "[a] model is an idealisation, from which one can diverge.In this sense, the primary function of a model is to offer orientation for the learners and not to act as a frame of reference to signal errors."(ibid.: 117) Furthermore, in line with the previous findings, students should be exposed to the structures of native-speaker English, but also be made aware of the strategies of lingua franca communication, of how to effectively accommodate to somebody whose mother tongue is also not English.This would involve pragmatic strategies (House 2002) and core phonology (Jenkins 2000) as well as the avoidance of difficult structures and infrequent words (see above).Incidentally, this is a training that could and should also be given to native speakers (cf.also Seidlhofer 2004: 226ff.for suggestions for lingua franca competence teaching).

Conclusion
The three-circle model of English use world-wide does have its limitations, yet it is well-established and should thus not easily be discarded.While a re-writing of the model ought most of all to take care of the fusion of several different levels of analysis (nations, speakers, functions, varieties), it does not seem that the increasing use of English as a lingua franca would call for a major alteration of the model.Rather, ELF should be included in the model as a specific function of English, particularly in the Expanding Circle.
Why is ELF not included as the variety of a norm-developing Expanding Circle?Corpus evidence from the Euro-English project has been cited to support the view that ELF does not count as a variety as such, since it appears to lack coherent features marking it off from other varieties.Instead, we ought to consider (and conduct more research to this end), whether ELF can be conceptualized as a functional phenomenon concerning English world-wide, emerging from the strategies of lingua franca communication.
Finally, in discussing implications for teaching, I have argued that since we have no ELF variety, and since learners are oriented towards the native-speaker standard, the native norm should remain the teaching model in the Expanding Circle.However, this does not imply that training for successful lingua franca communication should not take place.