Theoretically possible varieties (and their names)

At the turn of the new millennium, a new sociolinguistic project, entitled Language Contact and Focussing: The Linguistics of English in Switzerland, was launched with the support of the Swiss National Science Foundation (hereafter, SNSF). As Rosenberger (Reference Rosenberger2009: 131) informs us, the project was preceded by a pilot study with an overarching research question: what effects does/will/might the use of English in Switzerland have on the relationships between the ethnolinguistic groups within the country? The question as it was posed is indeed ambitious and also highly interesting. It is undoubtedly a socially relevant topic, as Rosenberger (Reference Rosenberger2009: 240) aptly concludes at the end of his book: ‘[English] plays a vital role in the daily life of a large number of Swiss citizens’. Eventually it was decided that the SNSF project should limit itself to exploring ‘the potential formation of an endonormative variety of English in Switzerland’ (Trudgill et al., Reference Trudgill, Watts and Allerton2000). Thus, the PhD theses and publications that resulted from that project centered on the question of whether there existed a national variety of English which had developed as a result of English functioning as an intranational lingua franca, with a research focus on multinational professional environments. The variety whose existence the project was meant to confirm or refute, respectively, was named, somewhat pompously, ‘Pan Swiss English’ (Durham, Reference Durham2003).

In order to illustrate what a potential candidate having the status of a feature of this named-but-not-yet-established variety might look like, Rosenberger (Reference Rosenberger2009: 133) evokes the phenomenon of false friends, well known to all foreign language teachers, arguing that French actuel, German aktuell, and Italian attuale might cause Swiss speakers from all three language groups to use the English actually in the sense of ‘currently’ instead of ‘in fact’. In other words, what is suggested here is that a Swiss variety of English could exist, one that is shared by Swiss Germans, Swiss French and Swiss Italians, in which actual (or actually) is coded as meaning ‘current’ (or ‘currently’). Thus, if a Swiss German says to a Swiss French (or a Swiss Italian): ‘Actually there are 200 people employed in this firm’, not only must s/he mean ‘currently’ by actually, but those s/he interacts with must equally decode it as meaning precisely that. Rosenberger (Reference Rosenberger2009: 133) goes on to explain that:

Rosenberger's proviso at the end of the above quote shows that, for him, Swiss English must remain, for the time being, a ‘theoretically possible variety’ (Reference Rosenberger2009: 95). The ultimate proof of its existence must lie with systematic data analysis. Even though Rosenberger (Reference Rosenberger2009) does not elaborate on it further, it can be assumed that if the development of such a variety should not be empirically confirmed, then the name ‘Swiss English’ must be crossed off the sociolinguist's constantly updated language list. According to this view, which the present article aims to invalidate, the linguistic world is populated with varieties: certain varieties have already been described; others have been named but await confirmation of their existence, while others still need to be discovered.

The notion that a word/name is a surrogate, i.e. that it stands for a thing in the real world, is nothing new. This notion, known as ‘nomenclaturism’, was famously discarded by Ferdinand de Saussure (Reference Saussure1983: 65–6). Instead, Saussure proposed that languages are abstract systems consisting of arbitrary units commonly called words. However, Saussure did not regard the metalanguage of linguistics as relative to the linguistic system to which its vocabulary was, superficially, assignable. Thus, the Saussurean terms la langue and la parole (to denote the community's abstract linguistic code and its execution in communication) are words belonging to the French language only ‘by necessity’, but at a more fundamental level they are ‘languageless’ in the sense that they refer to something outside the linguistic system.Footnote ² On this view, the language of science is believed to operate under different laws and to pose different demands on its users from ordinary language, although both need to be formulated in a language (e.g. scientific French vs. ordinary French).Footnote ³ According to this logic, it is acceptable for a lay speaker to talk about a variety of language as if it existed, based on his/her personal belief in and experience of this variety, whereas the same cannot be said for an academic linguist when making the same statement ex cathedra. It is against this background that Rosenberger's statement must be understood, namely that sociolinguists should not ‘tacitly adopt a hypothesis about a variety […] as a real and acknowledged variety before data is available to describe it sufficiently accurately and actually prove its existence empirically’ (Rosenberger, Reference Rosenberger2009: 99).

Notably, Rosenberger (Reference Rosenberger2009) did not choose to postulate, say, a ‘Swinglish hypothesis’: his book bears the title The Swiss English Hypothesis. Why is this so? Presumably, both designations existed prior to the start of the SNSF project. One reason, it might be conjectured, is that ‘Swinglish’ sounds the same as ‘Swenglish’ (the English spoken in Sweden). Another, ideologically motivated, explanation why ‘Swinglish’ cannot form the starting-point for any serious ontological investigation in sociolinguistics, is that analogous lay language terms formed by the process of blending, and with a relatively high frequency of usage (e.g. ‘Chinglish’, ‘Singlish’, ‘Fringlish’, ‘Spanglish’), tend to be associated with incorrect and mixed language use, often including a heavy L1 accent in the target language. However, this is precisely what the SNSF researchers ruled out categorically even before embarking on their projects, namely that Swiss English might be ‘Swiss’ (i.e. ‘Pan Swiss’) also on the phonological level. Instead, the search launched was one for a variety which would consist exclusively of morpho-syntactic and lexico-semantic features.Footnote ⁴ As Rosenberger (Reference Rosenberger2009: 130) explains,

From a lay speaker's perspective, this restriction must appear question-begging, as, it could be argued, either one speaks Swiss English – mind, body and soul – or one doesn't. Of course, one can speak English and be aware, say, that one's accent sounds ‘Swiss German’. Still, what one speaks is English. The truth is that the whole project of ‘Pan Swiss English’ comes straight out of the linguistic laboratory, or, which amounts to the same thing, out of the linguists' heads, i.e. it never was grounded in the researchers' personal linguistic experience, let alone the project leaders'. Their direct personal experience preceding the actual formulation of a ‘Swiss English hypothesis’ was, presumably, that Swiss people from the different linguistic areas tend to use certain morphological forms and syntactic structures in nonstandard ways when speaking/writing English. ‘Swiss English’ is certainly not a variety that Swiss Germans, Swiss French or Swiss Italians would think of themselves as being ‘speakers of’; if at all, it could be argued, they would more likely think of themselves as speakers of ‘Swiss German English’, ‘Swiss French English’, and ‘Swiss Italian English’. It is quite telling that the question of lay recognition was not considered by the project leaders or their researchers as worth probing into. If ‘Pan Swiss English’ did turn out to exist, therefore, it would be a variety that the Swiss don't know (yet) that they are users of in certain contexts, i.e. not until the SNSF linguistic team informs them of it: ‘We can assure you that what you speak at work is English, though whether it is “Standard British English”, “Standard American English” or another (attested) variety, is up to the phonologist to decide. However, based on the grammatical forms you produce and the way you use certain words, we can confirm that the variety of English you use in professional settings is called “Pan Swiss English”.’Footnote ⁵ It is far from clear whether the hypothetical users of Swiss English would be happy about the discovery of a variety they can call ‘their own’ as well as a ‘Pan Swiss linguistic identity’. But what is more worrisome is the idea that this newly gained ‘self-knowledge’ might lead to a reexamination of previous notions of ‘English’ on the part of Swiss lay speakers, possibly fostering a linguistic insecurity comparable to that encountered by Roy Harris on arriving in Hong Kong in the late 1980s, as his example of a Hong Kong Chinese, asking the readers of the South China Morning Post column whether what he spoke was ‘English’ or ‘Chinglish’, makes clear (Harris, Reference Harris1989: 41). Harris goes on to remark that ‘[these] are not signs of a linguistic community in good health. They are signs of a linguistic community in desperate need of therapy.’Footnote ⁶ Lay speakers already worry enough (and often quite unwarrantedly) about whether, for instance, their English is ‘good enough’ for the communicational tasks at hand. Scientific linguistics is not meant to add to these linguistic concerns any new ones, such as Swiss people starting to wonder whether what they speak/write can still be called ‘English’.

It is the code that makes me say it this way – or is it me?

The approach taken by those interested in finding a ‘Swiss’ variety of English is underpinned by a code-based view of language, which is why individuals themselves become mere representatives, whose linguistic behavior connects with other representatives of the same group because they have internalized a system based on a certain practice (here, the practice of communicating with other Swiss co-nationals via English for professional reasons). Thus, individual speakers are seen as users of systems constraining their freedom of choice. In other words, they speak according to the system. It is not particularly surprising to find that the typical investigation done within the ‘Varieties of English’ paradigm adopts such a code-based view. After all, sociolinguistics was born out of traditional dialectology, which in turn was an offspring of historical linguistics. For the latter discipline, treating units of analysis (lexicon, morphology, syntax, and phonetics) as independent of each other was a handy way of describing linguistic varieties for which there were no real, but only hypothetical, speakers around. In other words, because the historical linguist usually thinks of a variety under scrutiny as ‘dead’, there is nothing remarkable about focusing his/her research exclusively on, e.g., verb forms in Early Modern English (e.g. Abbott, Reference Abbott1953). Traditional dialectology was equally characterized by a tendency towards essentialism. Thus, the so-called ‘folk speech’ spoken by elderly rural speakers was supposed to represent the ‘natural’ output of a system still uncorrupted by urban speech ways. This or that word or grammatical form thus became a feature of this or that dialect (the same features were usually also present in other rural dialects). Analogously, the researchers of the SNSF project consider the possibility that certain features might be genuine features of Swiss English (e.g. the extended use of the infinitive where Standard English prescribes the gerund), while at the same time these very features are also characteristic of other (identified) ELF varieties.Footnote ⁷

It is precisely by operating within this framework that mistakes come to be treated as ‘code-given’. Thus, in one system a certain feature may count as a mistake, while in another system it does not. Dröschel's distinction (Reference Dröschel2011: 72) between mistake and deviation must be considered in this light. Following Kachru (Reference Kachru1983), she notes that:

Based on my experience as a teacher of English in German-speaking Switzerland, I find it difficult to uphold this distinction. For instance, I recollect a number of my Swiss students at the grammar school systematically writing in their grammar tests: I want that you…; I look forward to see you; I live here since 2 years. They did so collectively (some of them consistently) and during several years, up to the very day of their final exams. I would always mark these structures as mistakes. One could argue that these were learners' mistakes, made by speakers who had not yet mastered the target structure, in spite of all the tutoring classes they had taken. What we are invited to believe is that once the learners leave the school environment and eventually find a job in a multinational company, their mistakes become deviations, on the grounds that the Swiss employees are no longer learners of the language, but use English to communicate (via email or in business meetings) with other native and non-native speakers of English, including Swiss from linguistic areas other than their own. The deviations, according to this reasoning, are no longer mistakes triggered by a speaker's underlying mother-tongue system, but the result of productive processes, ‘transfer’ and ‘simplification’ (Dröschel, Reference Dröschel2011) being two of them, exclusive to Swiss (professional) contexts in which English is used as a lingua franca.

The interesting question at this point is whether the structures used by the Swiss speakers in lingua franca situations are the result of a conscious choice. Suppose I mastered the structure look forward to + gerund as a student of English at grammar school. Would I choose to use the structure look forward to + basic form in lingua franca situations when working in a multilingual company in Switzerland? Or does the use of this structure occur below the level of conscious awareness, i.e. the code-based explanation? Perhaps those Swiss who (variably) use look forward to + basic form in the lingua franca situations do so because that is what they did already as learners of English at grammar school? The phenomenon of accommodation is certainly a pertinent one on that score, even though why an individual accommodates in a particular situation, i.e. his/her motivation, is very complex and hence hardly amenable to scientific inquiry. Possible reasons for accommodating in a situation where English is used as a lingua franca are manifold, e.g. because one does not want to sound too ‘posh’ or to give the impression that one thinks of oneself as the ‘better speaker’. For instance, I am aware of shifting towards ‘Hong Kong English’ (whatever that might be) when interacting with local Hong Kongers, but I am not sure if I do so for the same reasons all the time, and how much of it is voluntary at all. Conversely, there is no reason why a proficient non-native speaker of English should necessarily accommodate at all when interacting with other, less proficient non-native speakers. As an alternative model to the code-based view of language, Harris (e.g. Reference Harris1998a: 29; Reference Harris, Harris and Wolf1998b: 235) has proposed that the communicational possibilities of interactants, including what they say or write, are always determined by three factors (or parameters), namely, what human beings in general are ‘physiologically and mentally equipped’ to do (biomechanical factors), what specific individuals are ‘collectively conditioned’ to do (macrosocial factors), and what they are ‘individually aiming to do in given circumstances’ (circumstantial factors). Considering individual Swiss speakers' linguistic behavior from the latter perspective would be a very different enterprise from trying to ascertain which variety (endonormative vs. exonormative) these speakers are speakers of.

What is more, mistakes do not fall out of a clear blue sky. Therefore, the question should always be: from whose point of view is something alleged to be ‘mistaken’? What linguists term a deviation may well be termed a mistake by non-linguists. The important point here is that there is no objective court of appeal to turn to. Thus, Swiss speakers of English who have been refining their language skills for years may protest that what is described as ‘Swiss English’ by linguists is just ‘ungrammatical English’ as far as they are concerned. In fact, for a number of Swiss using the basic form in conjunction with look forward to is simply a mistake, with no exceptions. For others, this is what they say or write, and thinking of this structure as a ‘mistake’ is simply not an issue (until someone happens to direct their attention to it). Judging something to be a mistake is always context-sensitive: especially for language teachers, this insight may go against common sense, as mistakes are commonly thought of as having an objectively valid explanation, existing independently of the three Harrisian parameters outlined above. However, when teachers of English proclaim that look forward to is ‘always’ followed by a verb in the gerund form, they do not refer to ‘descriptive rules’ of English, but rather invoke norms of usage. As Harris has pointed out (Harris, Reference Harris2006: 16), rules are not realia, and ‘the expression descriptive rule is a nonsense’. Harris (Reference Harris2006: 18) also notes that teaching your students that learning English is a matter of learning the rules of English is wrong because ‘it misrepresents the role of language (not just the English language) in human communication’. It is important, in this context, to understand that marking something as a ‘mistake’ in a grammar test does not imply that the one marking, by doing so, implicitly embraces the view that speaking (or writing) English is a ‘rule-governed’ activity.

‘Swiss English’ made practical?

The studies made under the aegis of finding ‘Swiss English’ are conspicuously silent on any practical implications that would be involved should the quest turn out to be successful. Only Rosenberger discusses briefly, at the end of his book, how Swiss English relates to questions of English language teaching: ‘…the results of the analysis make it absolutely clear that no specifically Swiss variety of English is available as a suitable alternative teaching model’ (Rosenberger, Reference Rosenberger2009: 235). The immediate question arising here is how ‘Swiss’ the variety studied would have needed to be in order to qualify as a ‘suitable alternative teaching model’. Rosenberger (Reference Rosenberger2009, 235) goes on to say:

What is interesting in this respect is that the ‘deviating’ linguistic features of the ELF variety used by the Swiss professionals are more or less all features, it seems to me, that grammar school teachers constantly encounter in the spoken and written English of their Swiss students (in all linguistic parts). If this is true, then the question arises as to whether these features could be incorporated at all into the various teaching materials used in Swiss schools? Would an imaginary unit on ‘English in Swiss multinational companies’ feature dialogues between Swiss employees of different linguistic backgrounds, and if yes, would they speak with a Standard, a ‘Swiss English’, or a mixed grammar? Another question is whether it would be desirable, as part of such a unit, to show Swiss people using English towards other Swiss (i.e. in interactions not involving any non-Swiss). The point is that devising an English learner's book for official use in Swiss schools which features modern (professional) Swiss of a German, French and Italian linguistic background interacting in English is likely to raise concerns from different directions within the political spectrum, and might indeed be construed as sending out the wrong (Anglophile) signal. In fact, teachers of French in German-speaking Switzerland and teachers of German in the French-speaking part already face considerable difficulties rousing their pupils' interest in the other national language. These are issues of real-world relevance, and any (SNSF funded) research that gives priority to the description of systems and their linguistic features (i.e. their frequency and distribution), while leaving the above issues to be dealt with in a successor project, if at all, is, from an integrational point of view, like putting the cart before the horse. What needs to be understood in all of this is that the current political and public debates concerning English as a ‘national’ language of Switzerland are too important to allow for, and perhaps be informed by, a parallel scientific inquiry as to whether ‘Pan Swiss English’ exists or not.

In Lutry they speak…

The SNSF researchers all concur that English is currently not developing specifically ‘Swiss’ characteristics (Durham, Reference Durham2007: 239; Rosenberger, Reference Rosenberger2009: 240; Dröschel, Reference Dröschel2011: 330). Instead, Rosenberger (Reference Rosenberger2009: 236) concludes at the end of his study that ‘the use of English as a lingua franca […] is a functional variety as opposed to a linguistic variety defined by formal properties’. My main critique in this article has been directed towards the idea of a scientific linguistics that describes ‘for the sake of science’, regardless of the consequences that finding (and publicizing) Swiss English might entail for politicians, teachers, pupils, parents, etc. However, those who believe in a descriptive linguistics would disagree with me and argue that the Swiss English project was meant, from the very beginning, to yield new insights into the current debates within pidgin/creole studies regarding, among other things, theories of language genesis and linguistic universals. On this view, the Swiss English project could thus be seen as having a greater purpose than merely a description of a yet-to-be-established variety, namely that of informing sociolinguistic theory more broadly, which in turn allows the discipline of general linguistics to advance.

The integrational linguist, in turn, is not concerned with linguistic typologies and historical explanations of abstract systems, as there are no languages and varieties to be counted in the first place. Languages are discursive constructs; they are talked about (by lay people as well as linguists) as if they existed. For this reason, there are no language names whose objects of reference need to be ontologically verified with a greater urgency than others. Thus, trying to prove the existence of ‘English’, i.e. a high-frequency, publicly shared sign with an attested history of (written) usage, is no more justified than trying to prove the existence of an ad hoc (private) language name, for which the sociolinguist has no attested history. The question ‘Does X-an English exist?’ cannot be asked in a communicational vacuum. Questions and answers, in fact, always presuppose someone who asks the question and someone who answers it. This is not to say that such a question never makes sense: it certainly does, but as a contextualized question. A personal experience will serve as a case in point here. After moving with my family from Lausanne (in French-speaking Switzerland) to Hong Kong, my five-year-old daughter, who was starting to come to terms with the new ‘local’ language, i.e. English, asked us how one would say this or that ‘in Lutry-language’. Lutry is a small village, 2 miles from the city of Lausanne, where we had been living for 3 years before moving. My daughter's coinage is likely to puzzle the linguist committed to the ‘ontologically tough’ position on languages: for the latter, the name ‘Lutry-language’ must obviously be a ‘false’ name in the sense that a village does not have its own ‘language’; in Lutry, a ‘language’ is indeed spoken, namely French (or is it Swiss regional French?), and this is, as far as I remember, how I interpreted ‘Lutry-language’. The linguist will add that, possibly, Lutry may have its own (local) dialect, in which case ‘Lutry-language’ actually turns out to be ‘Lutry-dialect’. However, child language cannot be assessed in this way, as children do not know the difference between ‘languages’ and ‘dialects’ (Do adults necessarily know? And what is the correct answer anyway?), which is why speculating whether my daughter meant ‘dialect’ when she said ‘language’, or even whether she meant anything at all, is pointless. What we need to realize once and for all is that ‘Swiss English’ and ‘Lutry-language’ are both meaningful names in context, suitable for the communicational purposes that they had to serve. Signs, the integrational linguist will argue, are equally real for children and adults: surely, my daughter must have known what she meant by ‘Lutry-language’, but asking oneself, qua scientist, what she ‘really’ meant by ‘Lutry-language’ (as if she lacked the correct terminology) is futile. Names do not possess context-independent reference, nor do metalinguistic terms like ‘language’ and ‘dialect’ have determinate meanings.