Navigation – Plan du site

AccueilNuméros74VariaFunctions and Mechanisms of Code-...


Functions and Mechanisms of Code-Switching in Bulgarian Canadians

Diana Yankova et Irena Vassileva
p. 103-121


Cet article traite d’une analyse de conversations d’immigrants bulgares de première génération, vivant au Canada et de leur recours à des renversements de code ou d’alternance, entre langue première et langue seconde lorsque des difficultés surgissent dans l’appréhension de concepts, d’idées nouvelles, de phénomènes ou de situations qui les déstabilisent. L’enquête montre les résultats de ces interférences et les occurrences de changement ou d’alternance de codes, particulièrement l’occurrence de phrases ou d’expressions en anglais dans des conversations en bulgare.

Haut de page

Texte intégral

Aim and background of the study

1The present study assumes a sociocultural perspective in order to look at “discourse-related switching” (Auer 1998: 8) observed in the discourse of first-generation Bulgarians in Canada. The results are expected to show the types of context where English words, phrases and whole sentences are incorporated into a conversation otherwise held in Bulgarian. An attempt is also made to elucidate the functions of code-switching, i.e. why bilingual speakers code-switch and what factors influence code choice. The investigation also includes analyses of the grammatical units occurring most frequently in the corpus. The article is part of a larger project conducted by a team of researchers, members of the Central European Association for Canadian Studies (CEACS) and funded by the Canadian government.

2 Concerning studies of code-switching in the Slavic languages, according to a recent publication (Lauersdorf 2009) presenting a comprehensive overview of Slavic studies in North America, in over 30 pages of bibliographical references there are about 20 entries dealing with immigrant Slavic languages (prevailingly Yugoslavian, Czech and Slovak) and there is not a single publication on Bulgarian. A handful of the above studies deal directly with code-switching issues and they shall be discussed in the analysis below in order to serve as a basis for comparison.

3Although belonging to the Slavic language family, Bulgarian is structurally different from all the other Slavic languages. From a highly inflected synthetic language historically, it has developed into an analytic language displaying at present a number of features that render it unique among the Slavic family, namely the obliteration of case declensions, the emergence of a post-posited definite article, the absence of a verb infinitive, the evolution of verb forms to indicate unwitnessed, retold and equivocal actions. Therefore, a study of interference between English and Bulgarian is expected to display different findings than other comparative research between English and Slavic languages (Vassileva, Yankova 2013 in print).

Code-switching: theoretical assumptions

4The use of more than one language in one and the same stretch of discourse, where “discourse” is understood in its widest sense as including both spoken and written varieties, as well as extralinguistic forms (Fairclough 1992), is not a new phenomenon. Nevertheless, more in-depth research on the behaviour of bilingual speakers and multilingual speakers began as late as the 1950s, when the first approaches of the phenomenon were prescriptive rather than descriptive and started from the assumption that alternation of languages was primarily due to insufficient knowledge of one or both of the languages concerned or to language interference. For instance, in his fundamental paper on language contact Weinreich (1968:73) maintains that the ideal bilingual is someone who “switches from one language to another according to appropriate changes in the speech situation […], but not in an unchanged speech situation, and certainly not within a single sentence”. The definition presumes that those bilinguals who do not adhere to the rules are “imperfect”.

  • 1 The present study does not deal with cases of natural bilingualism of children growing up in biling (...)

5Without going into details into issues such as language contact, and bilingualism and multilingualism, in what follows we shall discuss some of the most influential lines of research in the study of code-switching (hereinafter CS).1 In the first place, no one has refuted so far Gumperz’s claim (1982:64) that code-switching is predominantly observed in minority groups who speak a native language and the majority language. Moreover, contrary to the above-mentioned essentialist approaches, one cannot but agree with Gumperz (ibid) that “code-switching does not necessarily indicate imperfect knowledge of the grammatical systems in question”. This understanding has more recently been enhanced by a broader definition of bilingualism including “not only the ‘perfect’ bilingual […] or the ‘balanced’ bilingual […] but also various ‘imperfect’ and ‘unstable’ forms of bilingualism, in which one language takes over from the other(s) on at least some occasions and for some instances of language use”. (Dewaele, Wei and Housen 2003: 1). Along these lines, as well as in view of the subjects of the present study, we also start from the assumption that CS is not an indicator of deficiencies in the command of the respective languages but that it is triggered by other, predominantly sociolinguistic and psycholinguistic factors.

6The term has created a number of further controversies concerning its definition depending on the approach involved, and these need to be elucidated here as well. Some authors maintain that CS comprises all instances of language alternation (the term is used by Auer in 1995 as an alternative for CS), others (Kachru 1983, Sridhar & Sridhar 1980) treat only intra-sentential switches as cases of CS, while inter-sentential ones belong to code-mixing. Other linguists (Muysken 2000) use code-mixing to cover both code-switching and borrowing, where the former is intra-sentential and the latter inter-sentential.

Approaches to Code-Switching

7From a more general perspective, three directions from which the phenomenon has been approached so far, can be identified: structural, sociolinguistic and psycholinguistic.

8Structuralists have been interested in the “points of code-switching” and the mechanisms that operate and restrict it through the search for universally applicable models (MacSwann 2000, Jake, Myers-Scotton and Gross 2002). The most influential and widely used is Myers-Scotton’s Markedness Model (1993, 1998, 2000) containing the Matrix Language Frame model (MLF) and its recently added sub-models: the 4-M model and the Abstract Level model (Myers-Scotton and Jake 2000a, b, 2001). Since this (or any other) structural model will not be applied in the present study, we shall not discuss it in any detail here.

9From a sociolinguistic perspective, as early as 1929, Sapir accused linguists of being obsessed with their “petty patterns” (1929: 214) without being able to look beyond them where language is intertwined with social, anthropological, cultural and other much more general aspects of human existence. Therefore, a lot of work that followed in linguistic anthropology, sociolinguistics, and other related areas, has delivered interpretive and interactional understandings of code switching in context, or on how code-switching relates to the wider social and cultural context (the sociolinguistic approach). Following the psycholinguistic approach, back in 1954, Vogt focused on the fact that code-switching was an extralinguistic phenomenon: “Code-switching in itself is perhaps not a linguistic phenomenon, but rather a psychological one, and its causes are obviously extralinguistic” (1954: 368), whereas Stroud argues that it is a social phenomenon “that […] cannot really be understood apart from an understanding of social phenomena” (1998: 322) (see also Heller 1988, 1992, Auer 1984, Woolard, 2004).

Code-Switching and Borrowing

10Another unresolved theoretical controversy in the field, which is of crucial importance for the structuralist approach, presents the distinction between CS and borrowing. While Poplack (1978, 1980, 1981) and his followers believe that single words “imported” from another language should be treated differently and do not belong to CS which requires longer stretches of discourse (Poplack, Wheeler, and Westwood 1987, Sankoff, Poplack, and Vanniarajan 1990), a larger group of scholars including Myers-Scotton (1972, 1976, 1983), Bentahila & Davies (1983) assert that code-switching may also consist of insertion of a single word or a phrase.

11Poplack (1980) maintains that if a word is phonologically, morphologically and syntactically integrated in the recipient, base or “matrix language” (to use Myers-Scotton’s term), then it should not be treated as CS. In the present study we will follow Myers-Scotton’s (1993) view according to which (1) CS and borrowings belong to a continuum which is difficult to delineate; (2) borrowings do not necessarily “fill in gaps” in the vocabulary of the recipient language (those she calls “cultural borrowings”) but may also have equivalents there (“core borrowings”); (3) true borrowings may partly be identified as such, taking their frequency of use as a criterion. The latter, however, at least from the point of view of the present research cannot be applied since we deal with Bulgarian as the recipient language for which there are, unfortunately, no frequency dictionaries or databases that could be considered reliable enough. Besides, Bulgarian, like most languages from the former Eastern Bloc, has been in the process of constant and very rapid change since 1990, especially concerning the influx of foreign (primarily English) words and whole discursive structures, so that it is extremely difficult to follow this process. What complicates the problem even more is the fact that the use of English in different social domains and especially age groups varies a lot. Therefore, since the subjects of the investigation have left Bulgaria at various times and have kept different types of contact (if at all) with their home country, it turns out, as will be seen later in the analysis, that some of the English words and phrases they assume they use as inserted, and thus try to explain or translate, have actually, in the meantime, become relatively widely used in Bulgarian.

12We concur with Eastman who maintains that “efforts to distinguish codeswitching, codemixing and borrowing are doomed” (1992:1) and that we have to “free ourselves of the need to categorize any instances of seemingly non-native material in language as a borrowing or a switch” (ibid). In other words, fruitful as they may be for the study of grammatical features of CS, the grammatical approaches fail to account for the elucidation of the reasons why CS occurs.

13Considering the definitions quoted above, we start from the assumption that conversational code-switching is not only a linguistic, but also a social, psychological and pragmatic phenomenon that may be manifested at all levels of language – phonetic, lexical, phrasal, sentential, and the discourse level. We also adopt the view that code-switching may occur both consciously and unconsciously.

The case study

Background of the bilingual communities

14Three waves of Bulgarian immigration to Canada can be distinguished. The first one occurred after the liberation of Bulgaria after 500 years of Ottoman rule in the late nineteenth century, and lasted until the end of the Second World War. Those were mostly economic immigrants and the number of people who moved to Canada at that time is believed to be about 20,000, but for various reasons cannot be precisely established (Yankova, Andreev 2012:40). The second immigration wave was caused by the sociopolitical changes in Bulgaria after the Second World War and the establishment of the communist regime, which led people to flee the country for primarily political reasons. In contrast to the first wave, these were mainly well-educated, skilled professionals. The third immigration wave began after the political changes in 1989 and the economic and political instability that accompanied them. The largest number of these immigrants is formed of well-educated professionals, who had good jobs in Bulgaria but arrived in Canada in order to ensure a better life for themselves and their offspring. The predominant part of the interviewees in this study belongs to this third group of immigrants.

15To date, the statistics accounting for the number of Bulgarian Canadians living in Canada varies between 19 050 (official) and 150,000 (unofficial). The official data is from the Canada 2011 census published by Statistics Canada and is based on population defined by mother tongue (compared to 16 790 in the 2006 census).

Data Collection and interviewee profile

16Data was collected in three manners: recorded interviews, questionnaires, direct observation. Semi-structured interviews of a total of 4 hours were conducted with 16 first-generation Bulgarian immigrants to Canada. The interviewee profile is presented in Table 1 below:

Table 1. Interviewee profile




University degree

Over 40 years of age













17The respondents were legal immigrants who left Bulgaria between 1948–98. The earliest immigrant settled in Canada in 1954, and the last one in 2005. They live in Montreal, Vancouver and Toronto (for more details see Yankova, Andreev 2012 : 42-43).

18All the subjects of the study have Bulgarian as their native language and have retained the language. To the question (in the Interviewee Profile) what language do they speak at home, at work and with friends, almost all speak Bulgarian at home, except for those not married to fellow Bulgarians or those with school-age children, since the children switch languages. At work the subjects speak either English or French, and with friends – a mixture of Bulgarian, English, French and others (e.g. Serbian, Spanish), depending on friends’ nationality. Notably, all subjects are adult bilinguals, i.e. they have emigrated as grown-ups who had already lived long enough in a native-language environment. Part of them went to Canada with some knowledge of English or French, others had to learn the language there.

19The data were primarily collected by one of the authors at the informants’ homes over the course of 1 month. The authors adhered to ethical research principles: the subjects were informed that they were recorded, they had agreed to participate in the study and were guaranteed anonymity.


20According to Sebba there are three types of social factors that define the choice of CS (1) “factors independent of the particular speaker” such as prestige, power relations, economic circumstances; (2) factors that depend on the speakers’ language competence and their social networking, and (3) “factors within the conversations where CS takes place” (2009: 42-43). The first factor represents social circumstances that are practically the same in the case of our subjects; as to the second factor, subjects vary depending on the time spent in Canada, age, certain social factors, and this will need special attention in a separate study. Therefore, the third factor only shall be investigated as a variable in the present paper.

21The methodology combines Gumperz’s conversational functions of CS (1982) with Appel & Muysken’s functions of language (1987), which will be explained below. Probably, the most frequently applied and quoted list of conversational functions of CS is Gumperz’s (1982: 75-79) comprising: (A) Quotations, (B) Addressee Specification, (C) Interjections, (D) Reiteration, (E) Message Qualification, and (F) Personalization versus Objectivization. In the case of quotations the speaker reports someone else’s words either as reported speech or as direct quotation. Through addressee specification the speaker directs his/her utterance to a specific person among the rest of the participants in the conversation. As to interjections, these are so-called fillers, while reiteration, as the term suggests, means repetition of the message in the other language, usually for the sake of clarification and avoidance of misunderstanding or in order to emphasize certain parts of the utterance. Message qualification refers to explanation of the utterance, while personalization versus objectivization indicates the speaker’s involvement. However, the application of Gumperz’s model does not always specify which discourse function the speaker accomplishes by, for instance quotation, interjection or message qualification. Therefore, we will apply the model only in order to explain the functions of CS.

22Appel & Muysken (1987:29) draw on Jakobson’s six functions of language (referential, conative, emotive, phatic, metalingual and poetic) and consider them appropriate for the study of code-switching in combination with the domain approach, which examines patterns of language use across domains in general. This classification shall partly be adopted in this study along with Gumperz’s model.

  1. The referential function – the case where the speakers switch language either because they are not able to find the right word, or the word does not exist. Grosjean (1982: 125) calls this “the phenomenon of the most available word” that saves the speaker time and efforts to find the exact word in the current language spoken.

  2. The directive and integrative function: “By using standard greetings, conventional modes of address, imperatives, exclamations, and questions contacts are made with others and enough of an interactive structure is created to ensure cooperation” (Appel & Muysken 1987:29).

  3. The expressive function: “By making one’s feelings known one can present oneself to others as a unique individual” (ibid).

  4. The phatic function is observed when the speaker tries to keep communication channels open.

  5. The metalinguistic function refers to the case when the speaker makes explicit references to one of the languages involved.

  6. The poetic function denotes cases where speakers change the language when they tell jokes or use word-play in another language, or when they try to avoid taboo words and phrases in the main language used in the particular conversation.

23There have certainly been quite a few attempts at similar classifications (McClure and McClure 1988; Romaine 1989; Nishimura 1997; Zentella 1997, Gardner-Chloros 1991) but they have often been criticized for their lack of clarity and/or confusion of form and function (see e.g., Auer 1995). Bailey (2002 : 77) notes that: “The ease with which such categories can be created – and discrepancies between the code switching taxonomies at which researchers have arrived – hint at the epistemological problems of such taxonomies”. We shall therefore accept Nilep’s (2006 : 10) suggestion “to observe actual interaction rather than [start] from the assumption about the general effects of code switching” and thus make use of some of the classification categories discussed above without, however, sticking to them strictly and we shall try to explain the concrete reasons for code-switching in particular repetitive cases of CS.


24Since the main focus of the interviews was the identity of the first-generation immigrants they were not too conscious or weary of eventual code-switches, in other words they were not aware that their language use might later on be analysed and therefore focused more on the content of their answers. All participants in the interviews, although given a choice between Bulgarian, English or French, chose to speak in Bulgarian.

Functions of code-switching

25In view of the classification adopted for this study, our corpus demonstrated instances predominantly of the referential, the expressive and the phatic function. There were no instances of the directive, the poetic or the metalinguistic functions. The distribution can be seen on Fig. 1.

Fig. 1. Distribution of functions in the corpus

Fig. 1. Distribution of functions in the corpus

1- The referential function

26Not surprisingly, most of the instances of code-switching demonstrated a dominant referential function – they served to emphasise the message’s denotative purpose in reference to the context or co-text. Three subcategories of the referential code-switches were observed: to refer to culture-specific concepts or reality, to express oneself with a shorter form, and to refer to objects and phenomena the subjects have to deal with primarily in English or French. We will discuss in more detail each of these in turn.

27The first subcategory subsumes code-switches which interviewees resort to, owing to the fact that there is no direct equivalent in the Bulgarian language. For instance:

  • 2 “It is very important to know that although when they select you – we arrived here with approval, i (...)

(1) Mnogo e važno da ze znae, če vǎpreki če kogato te podbirat – došli sme tuk s odobrenie, v smisǎl s viza, s vsičko, legalno sme vlezli v stranata s landed immigrant visa, no v kraina smetka izliza ce….2

28The term “landed immigrant” does not exist in Bulgarian and therefore the interviewee had no choice but to employ the term. Although it has become obsolete in Canadian official documents and replaced by the term “permanent resident” it is still part of the Canadian vocabulary, and is sometimes even present in some government forms. Lack of a Bulgarian equivalent is also the explanation for the CS in example (2):

  • 3 “there’s this element of the glass-ceiling, as it is called, where there’s some kind of discriminat (...)

(2) ….ima edin element na glass-ceiling kakto se nariča, kǎdeto ima njakakav vid discriminacija sprjamo iztočnoevropejci, koeto se prodiktuva ot cjalata tazi kultura na studenata vojna, i horata ot iztočnija blok sa bili, kak da kaža rendered….3

29The idea of a “glass-ceiling” is not part of the Bulgarian conceptual system. For years, there has been no discrimination on the principle of gender in the pay or in the rise to elite positions of females. On the contrary, women were encouraged to perform jobs typically male jobs, especially after World War 2. This equality in job opportunities also explains the virtual absence of any notable or robust feminist movement in modern Bulgaria.

30The following example presents an interesting case: although the word “judge” is present in (3), the code-switch can be explained by the different procedures of obtaining a citizenship. In Bulgaria, after filling in the appropriate documents, followed by an interview with officials at the Ministry of Justice and the respective checks carried out by the Ministry and other governmental bodies if the need arises, the applicants are informed in writing, i.e. there’s no official ritual similar to the Canadian Taking of the Oath at Citizenship ceremony, which is presided by a judge.

  • 4 “Another big event for me was the speech of a….what do you call it, he is a public person, who is a (...)

(3) Drugo goljamo sabitie za men beše rečta na edin, kak se kazva, toj e obštestvena ličnost, kojto e judge, kojto vodi procedurata, kogato stavaš kanadski graždanin, kojto iznese edna reč…. 4

  • 5 “-VAS” is the Bulgarian verb ending for the second person singular Since the grammatical adaptation (...)
  • 6 “You look at a woman and she says: You’re harassing me, you look at a man and he says you’re harass (...)

31Code-switches are sometimes resorted to by bilinguals or multilinguals in order to achieve language economy or express a concept more succinctly or more precisely as in “(4) Pogleždaš edna žena i tja ti kazva ti me harass-vas5, pogleždaš edin mǎž i toj ti kazva ti me harass-vas.6

32“Harass” in the above example is used in the sense of sexual harassment. In order for this notion to be rendered in Bulgarian one would need to use a whole phrase, e.g. “upražnjavam seksualen tormoz” and even then, the content of the concept itself would not be quite clear since this is not considered an offence that is triable under Bulgarian law.

  • 7 “In my time there was a lot of work. There was no welfare then, you know what welfare is.”

33The same holds true for example (5): “Prez mojto vreme imaše mnogo rabota. Togava njamaše welfare, znaete kakvo e welfare7.” The respective L1 term would be “socialna pomošt za bezrabotni” which is much longer than the concise L2 term, “welfare” which communicates the exact semantic message that the addresser strives to convey.

  • 8 “The difficulties in the beginning were immense, because we had to struggle to be recognized as so- (...)
  • 9 “Everybody here wanted me to have the so-called Canadian experience.”

34By far the most frequent use of the referential function of code-switching is to refer to objects and phenomena the interviewees have to deal with primarily in L2 - English or French. In such cases there is an equivalent term or phrase in the Bulgarian language but the code-switch is warranted by the differential use of the language. Usually these are situations, concepts and phenomena that interviewees have to deal with outside of their home and consequently revert to English or French. In other words, most formal contacts are effectuated in English or French and most informal interactions in Bulgarian. This clearly shows in the following instances from the corpus: “(6) Trudnostite v načaloto bjaha dosta golemi, zaštoto trjabvaše da se borim da ostanem za da ni priznajat za refugee-ta, taka narečeni.8” and (7) “Vsički iskaha tuka da imam taka narečenija Canadian experience9.” Or as in:

  • 10 “When I found myself here in 1971 that was mouvement de liberacion des femmes and I entered this wa (...)

(8) Kogato popadnah tuka prez 1971 godina tova beše mouvement de liberacion des femmes i az vljazoh v tazi vǎlna, taka da kažem, i za avortement….i mi se stori, ce ne bila tolkova naprednala tazi strana i če ženite ne sa egales ....kato mǎža10.

35In examples (6) – (8), the concepts of refugees, Canadian experience, mouvement de liberacion des femmes, avortement, egalite, are all connected with situations the interviewees had to deal with outside of their home, in a more or less formal English or French-speaking environment. When they have to retrieve the term for the concept, it naturally emerges in the language of its predominant (and perhaps only) use.

2 - The expressive function

36The second in number of incidence is the group of code-switches whose dominant function is expressive. These switches are oriented to the addresser of the message and reflect his/her response to a situation. They do not modify the denotation of what is being said, but provide additional information concerning the internal state of the addresser – his emotions or attitudes. As for instance in:

  • 11 “I continue to be always Bulgarian, le bulgare, le bulgare, le bulgare, even the Hispanophones I ha (...)
  • 12 “I realized that my city was Montreal without having seen it, because it was 3 o’clock in the morni (...)

(9) Prodǎlzavam da sǎm si vinagi Bulgarian, le bulgare, le bulgare, le bulgare, daže ispanofonite, s koito sam rabotil mi vikat il bulgaro11“ or : (10) Az razbrah, če mojat grad e Monreal bez da sǎm go vidjal, zaštoto beše 3 časa sutrinta, tǎmno, ništo ne se vižda, edin highway, avtobus….12

37Some researchers (Romaine 1995) have discerned an analogy between style-shifting in monolinguals and code-switching in bilinguals. Bailey (2009: 358) following Zentella (1997) holds the view that if languages are not ascribed specific functional domains, “the search for a function of a particular switch may be akin to trying to explain why a monolingual speaker selects one synonym or phrasing or another”. In examples (9) and (10) there is an attempt to create an aesthetic, emotional effect and to underscore the difference between “us” and “them”, or in Gumperz’s terms this exemplifies a switch between the “we-code” and the “they-code” (Gumperz 1982), categorizing the message in terms of solidarity. In example (9), the nationality “Bulgarian” is rendered in three other languages – English, French and Spanish to reinforce more eloquently the idea of the Bulgarian identity that the speaker feels he has retained. Resort to the English word “highway” in example (10) can be explained not because this word is missing in the vocabulary of the Bulgarian language, but again as a proof of this old-new, before-now, us-them dichotomy characteristic of immigrants. It is meant to express the differences with home, to embrace of a new way of life, where even most immediate surrounding objects are different.

3 - The phatic function

  • 13 “…starting from the food and going to the way you express yourself. That’s it.”
  • 14 “I think I have a Bulgarian passport, bon, Canadian passport.”

38There were few instances of phatic code-switches found in the corpus for understandable reasons. The phatic function, which is meant to check the working of the channel of communication, to initiate, or discontinue communication, to attract or keep the addressee’s attention, was not dominant in a pre-agreed, structured interview, where the interviewer and the interviewee had each other’s full attention. Therefore, there was practically no need for this function and this is displayed in the corpus. The few examples of code-switches serving a phatic function are given below starting with: “(11) …kato počneš ot hranata i stigneš do načina, po kojto se izrazjavaš. That’s it.13” or (12) “Az mislja, če imam bǎlgarski, bon, kanadski pasport.14

39The choice of code-switches in examples (1) to (12) was based on their being illustrative of one more or less dominant function. The corpus was replete with instances where several functions intertwined in one stretch of discourse, as is certainly characteristic of most kinds of communication. Some of these examples will be adduced further on in the article.

40As mentioned above, three functions of CS were not observed in the corpus. This fact could tentatively be explained by the following facts: (1) The directive and integrative function is expected to ensure cooperation proved redundant in the case of these interviews since the interviewer knew the subjects personally quite well; (2) The absence of the metalinguistic function could also be explained by the already described format of the interviews – they were primarily directed toward the content rather than the ways and means of expression and the subjects code-switched unconsciously, aiming mainly at making themselves understood correctly; (3) The poetic function, according to its definition, is not really clearly delineated from the expressive function, so that, from a practical point of view, all such instances were classified under the ‘expressive function’, while from a theoretical point of view we believe that the two functions do not need to be distinguished.

Compensatory mechanisms

41What we would like to focus on now are some of the mechanisms employed by the interviewees when they switch codes and what is achieved by the particular CS. The corpus presented instances of reiteration, message qualification, quotations, interjections, as well as resort to calques. We will adduce some examples which present a mixture of functions and mechanisms, demonstrating the complex factors that accompany the process of switching between two or more languages.

  • 15 “an impression is created in people’s minds that they are something different, that they are a grou (...)

(13) ….sǎzdava se vpečatlenie v sǎznanieto na horata, če te sa nešto različno, če sa njakakva grupa, kojato se tretira kato nešto različno – different, kato the other, v kavički složeno i tova prodǎlžava, kakto kazvat da se reinforce by mass media and culture črez filmi15.

42In the above example we are witnessing a reiteration – the word “different” is code-switched with the aim of highlighting, of amplifying the meaning, then comes a reformulation, or message specification by the code-switched ‘the other’ with the idea of further emphasizing the distinctness, the difference.

  • 16 “I hadn’t had peanut butter before, peanut butter”

43More often than not, there is a code-switched word in English or French, followed by a repeated word or phrase in Bulgarian, trying to gloss it. The speaker cannot immediately think of the respective word or phrase in L1, but then when it comes to him or her, it is reiterated in Bulgarian as in: “(14) Az ne bjah opitval peanut butter, făstăčeno maslo.16” Interestingly enough there were instances of the opposite - a word or phrase initially uttered in Bulgarian, and then clarified further in French or English:

  • 17 “My father is from Suhindol and I feel close to the countryside, to the villagers – les paysans – a (...)

(15) Bašta mi e ot Suhindol i az se čuvstvam blizo do seloto, seljanite – les paysans – i mislja, ce bǎlgarinǎt ima uvaženie kǎm zemjata i tazi čerta bǎlgarska ja namiram v indiancite17.

44The speaker in the above example obviously wants to make sure that she has used the right Bulgarian word and just in case is repeating it in French in order to elucidate the intended meaning. Reiteration and repetition can also have the purpose of creating coherence within the speech of the bilingual, without sounding impolite or arrogant. There are three instances of CS in the example below.

  • 18 “I’m in contact with people, such as for example the musician A.L., for example, who has concerts a (...)

(16) Kontaktuvam si s hora, kato naprimer muzikanti kato A. L., primerno, kojato pravi koncerti i se opitva da, kak se kazva – to promote…..anyway (2 interjection), da….gledai kak ne mi idva bǎlgarskata duma, da utvǎrzdava ili prezentira bǎlgarskata kultura črez muzika18.

45Using the English verb “promote” can be classified under the third subgroup of referential meaning above. Then there’s the interjection “anyway” which seems to convey the feeling that the bilingual speaker has at that moment assumed the role of an English-language speaker and the previous code-switched word has acted as a trigger for maintaining the change of codes. The last instance of a CS is the word “prezentira” which is a calque of the verb “present”. The interviewee is not even aware that she is code-switching, she is trying to look for the Bulgarian word for ‘promote’ and thinks she has found it by means of the calque.

  • 19 “They say in Quebec: You are all Quebecers.”
  • 20 “And she said “I’ll cheer for her – how do you say that – I’ll clap my hands, I’ll support her.”

46The interviewees did not resort quite often to quotations – there were only three instances in the corpus – one of the switches was to French, the other to English and the third to Serbian as in: (17) “V Quebec kazvat: Vous êtes tous Québécois.19” or (18) “I tja kaza: I’ll cheer for her – kak se kazvaše tova – az šte i rǎkopljaskam, šte ja podkrepjam.20

47Quotations provide the speaker with another voice with which to encode a message, it allows for a shift of point of view. They are rendered in the language they were spoken. This is in keeping with the rule postulated by Gumperz (1982:82) that “a message is quoted in the code in which it was said”.

  • 21 “That was the moment I realized what country I had been living in.”
  • 22 “They did not come to Canada because of family matters.”

48Another characteristic of the speech of the interviewees was the extensive use of language transfer on all language levels. Starting from phonetic interference from L2 and transfer into L1 (Quebec pronounced as [kebek], c.f. the standard Bulgarian pronunciation [kvebek]) to lexical: (19) “Tova beše momentǎt, v kojto realizirah v kakva strana sǎm živjala.21”to phraseological: (20) “Te ne doidoha v Kanada poradi vǎprosi na familija22

49Calques can be considered as part of the covert interaction (Marian 2009: 163) in the language production of bilinguals since they are used in L2 in a way that is semantically or syntactically inappropriate for L2. The linguistic means by which code-switching is performed is exceptionally interesting from a research point of view. Issues connected to the phonology, lexis, syntax and discourse characteristics of code-switching such as words from which grammatical class are most often switched, the morphophonemic integration of code-switched items, the creation and characteristics of an inter-language, etc., merit a more detailed discussion and remain outside the scope of the present article.


50Undoubtedly, the complex phenomenon of code-switching allows for a multitude of diverse interpretations depending on the type of perspective – linguistic, sociolinguistic, psycholinguistic and on the type of tasks researchers set themselves. On the basis of our recorded interviews we have tried to shed some light on the process of CS in Bulgarian Canadians. Many reasons for CS have been highlighted by previous research, ranging from difference in the discourse strategy, aspiration to solidarity with a group, prestige, lack of formal knowledge in the languages, among others.

51The results of our study give sufficient grounds to conclude that the most frequent function of CS is to refer to concepts, ideas, phenomena, situations, interactions that speakers have to deal with in L2, or culture-specific concepts and realia that cannot be referred to in L1, therefore leading to expressing the concept with a word or phrase from L2.

52The study suggests that several reasons for CS can be singled out. The most common one is the uneven distribution in the use of L1 and L2, with L1 mostly used for personal, informal, family communication, while reserving the formal functions – work, communicating with official bodies, for the L2. Hence the harder retrieval of words or phrases which come from the less frequently used domain in that language. Another reason for CS that arose in the process of this study is that speakers resort to switching in order to express a message more succinctly, or more clearly in one language than in another. Some concepts and ideas are much more easily rendered in English than Bulgarian.

53The above claims are based both on the presented results and on the overall discussions of these problems with the interviewees, which cannot be included here since the texts would be too long.

54The present study is also in line with some more recent tendencies to look at migrant communities’ language performance “locally and ethnographically” (Bucholz and Skapoulli 2009: 2), without opposing it to a dominant nation-state language. Categories like “transmigration” (Stein 2008) and “super-diversity” (Blommaert 2010, among others), are attempts at explanations going beyond typically European monolingual nation-state ideologies and understanding of code-switching as a natural phenomenon (not as an exception in need of research!) reflecting the increased mobility of people, languages, cultures and artefacts. Bailey (2009) points to the negative effects of essentialist ideologies in Europe, reflected strongly in the educational systems as well. Whether this is the case in officially bilingual Canada is a political dispute we would not enter into here, but it seems an interesting point of departure for future research.

55From a purely linguistic point of view it seems that modern sociolinguistics is dominated by the Danish school of scholars who have introduced new terms such as “languaging”, and “poly-languaging” under the conditions of superdiversity and start from the assumption that the traditional view of “a language” no longer reflects and serves the needs of present-day multicultural societies (Jorgensen 2010, Moller 2009). This approach is even more applicable to the well-recognized multicultural Canadian society and is definitely worth employing in further investigations.

Haut de page


Appel, R. and P. Muysken (1987). Language Contact and Bilingualism. London: Edward Arnold.

Auer, Peter (1984). Bilingual Conversation. Amsterdam: John Benjamins.

_____ (1995). “The pragmatics of code-switching: A sequential approach”. In L. Milroy & P. Muysken (eds.), One speaker Two Languages: Cross-disciplinary Perspectives on Code-switching. New York: Cambridge University Press. 115-135.

_____ (1998). Code-Switching in Conversation: Language, Interaction and Identity. London: Routledge.

Bailey, Benjamin (2001). “The Language of Multiple Identities among Dominican Americans”, Journal of Linguistic Anthropology 10(2). 190-220.

_____ (2009). “Multilingual forms of talk and identity at work”, in Peter Auer, Li Wei (eds.), Handbook of Multilingualism and Multilingual Communication. Berlin: Walter de Gruyter. 355-369.

Blommaert, Jan (2010). The Sociolinguistics of Globalization. Cambridge: Cambridge University Press.

Buchholz, Mary & Elena Skapoulli (2009). “Introduction: Youth language at the intersection: From migration to globalization”. Pragmatics 19(1). 1-16.

Dewaele, Jean-Mark, Alex Housen and Li Wei (2003). Bilingualism: Beyond Basic Principles. Clevedon: Multilingual Matters.

Eastman, Carol (1992). “Codeswitching as an Urban Language-Contact Phenomenon” in Carol Eastman (ed.), Codeswitching. Clevedon, Philadelphia, Adelaide: Multilingual Matters 1-18.

Fairclough, Norman (1993). Discourse and Social Change. Cambridge: Polity Press.

Gardner-Chloros, P. (1991). “Ni tu ni vous: principes et paradoxes dans l’emploi des pronoms d’allocution en francais contemporain”. In Journal of French Language Studies 1 (1). 139-155.

Gumperz, John (1982). Discourse Strategies. Cambridge: Cambridge University Press.

Heller, Monica (1988). Codeswitching: Anthropological and Sociolinguistic Perspectives. Berlin: Mouton de Gruyter.

_____ (1992). “The Politics of Codeswitching and Language Choice”, in Carol Eastman (ed.), Codeswitching, Clevedon: Multilingual Matters. 123-142.

Jake, Janice, Carol Myers-Scotton and Steven Gross (2002). “Making a Minimalist Approach to Codeswitching Work: Adding the Matrix Language”, Bilingualism: Language and Cognition 5(1). 69-91.

Jorgensen, J. Normann (2010). “Languaging. Nine Years of Poly-Lingual Development of Young Turkish-Danish Grade School Students”,. Copenhagen Studies in Bilingualism, the Koge Series, Volume K15-K16. University of Copenhagen. volume I-II

Lauersdorf, Mark Richard (2009) “Slavic Sociolinguistics in North America: Lineage and Leading Edge”. Linguistics Faculty Publications. Paper 32.

MacSwan, Jeff (2000), “The architecture of the bilingual language faculty: Evidence from code-switching”, n Bilingualism: Language and Cognition, 3(1), 37-54.

McClure, Erica and Malcolm (1988). “Macro- and Micro-sociolinguistic Dimensions of Code-switching in Vingard (Romania)”, in Monica Heller (ed.) Codeswitching: Anthropological and Sociolinguistic Perspectives. Berlin: Walter de Gruyter. 25-51.

Mühlhäusler, Peter (1981), “Structural expansion and the process of creolization”, in Theoretical Issues in Pidgin and Creole Studies, Valdman and Highfield (eds.), 19-56. New York.

Muysken, Pieter (2000), Bilingual Speech: A Typology of Code-mixing. Cambridge: Cambridge University Press.

Myers-Stanton Carol (1998). Codes and Consequences: Choosing Linguistic Varieties. New York: Oxford University Press.

_____ and Jake, Janice (2000a), “Explaining aspects of codeswitching and their implications”, in Janet Nichol (Ed.). One mind, Two Languages: Bilingual Language Processing. Oxford: Blackwell. 91-125 .

_____ and Jake, Janice (2000b), “Four types of morphemes: Evidence from aphasia, codeswitching and second language acquisition”, In Linguistics 38,6. 1053-1100.

Nilep, Chad (2006), “Code Switching” in Sociocultural Linguistics. In Colorado Research in Linguistics. June 2006. Vol. 19. Boulder: University of Colorado. 1-22.

Poplack, Shana (1978). “Dialect acquisition among Puerto Rican bilinguals.” In Language in Society 7. 89-103.

_____ (1981). “Syntactic structure and social function of code-switching.” In by Duran, R (ed.), Latino Language and Communicative Behavior, 169-184. New Jersey: Ablex Publishing Corp.

_____ and Meechan, Marjory 1995. “Patterns of Language Mixture: Nominal structure in Wolof-French and Fongobe-French bilingual discourse”. In L. Milroy and P. Muysken (eds.), One Speaker, Two Languages, 199-233. Cambridge: CUP.

Romaine, Suzanne. (1989), Bilingualism. Oxford: Basil Blackwell.

Stein, Mark (2008). Transmigration. Münster: Universität Münster manuscript.

Stroud, Christopher (1998), “Perspectives on Cultural Variability of Discourse and some Implications for Code-switching”, in Peter Auer (ed.), Code-Switching in Conversation: Language, Interaction and Identity, 321-348. London: Routledge.

Vassileva, Irena; and Yankova, Diana (in print). “Observations on the Language of First Generation Bulgarian Immigrants to Canada”. International Journal of Applied Linguistics.

Woolard, Kathryn (2004). “Codeswitching”. In Alessandro Duranti (Ed.) A Companion to Linguistic Anthropology. Blackwell. 73-94.

Yankova, Diana and Andreev Andrei (2012), “Cross-cultural and transnational identity: a case study of Bulgarian immigrants in Canada”, British Journal of Canadian Studies, 25/1: 37-56.

Haut de page


1 The present study does not deal with cases of natural bilingualism of children growing up in bilingual families.

2 “It is very important to know that although when they select you – we arrived here with approval, in the sense of a visa, and everything, we entered the country legally with a landed immigrant visa, but in the end it turns out that….”

3 “there’s this element of the glass-ceiling, as it is called, where there’s some kind of discrimination against Eastern Europeans, which is dictated by this whole culture of the cold war and the people from the Eastern Bloc have been, how should I say, rendered….”

4 “Another big event for me was the speech of a….what do you call it, he is a public person, who is a judge, who leads the procedure when you become a Canadian citizen, who held a speech.”

5 “-VAS” is the Bulgarian verb ending for the second person singular Since the grammatical adaptation of English words into the Bulgarian discourse will be the subject of a follow-up study, we are not going to discuss it in this article.

6 “You look at a woman and she says: You’re harassing me, you look at a man and he says you’re harassing me.”

7 “In my time there was a lot of work. There was no welfare then, you know what welfare is.”

8 “The difficulties in the beginning were immense, because we had to struggle to be recognized as so-called refugees.”

9 “Everybody here wanted me to have the so-called Canadian experience.”

10 “When I found myself here in 1971 that was mouvement de liberacion des femmes and I entered this wave, so to say, and also for avortement and it seemed to me that this country was not that advanced and that women aren’t egales to men.”

11 “I continue to be always Bulgarian, le bulgare, le bulgare, le bulgare, even the Hispanophones I have worked with call me il bulgaro.”

12 “I realized that my city was Montreal without having seen it, because it was 3 o’clock in the morning, dark, you can’t see a thing, a highway, a bus…”

13 “…starting from the food and going to the way you express yourself. That’s it.”

14 “I think I have a Bulgarian passport, bon, Canadian passport.”

15 “an impression is created in people’s minds that they are something different, that they are a group that is treated as something different – different, like the other, in quotation marks, and this continues as they say to be reinforced by mass media and culture in films.”

16 “I hadn’t had peanut butter before, peanut butter”

17 “My father is from Suhindol and I feel close to the countryside, to the villagers – les paysans – and think that Bulgarians have respect for the land and this Bulgarian trait I find in the Indians.”

18 “I’m in contact with people, such as for example the musician A.L., for example, who has concerts and is trying to, how you say it – to promote….anyway, yes, see how the Bulgarian world does not occur to me, to strengthen or present Bulgarian culture through music.”

19 “They say in Quebec: You are all Quebecers.”

20 “And she said “I’ll cheer for her – how do you say that – I’ll clap my hands, I’ll support her.”

21 “That was the moment I realized what country I had been living in.”

22 “They did not come to Canada because of family matters.”

Haut de page

Table des illustrations

Titre Fig. 1. Distribution of functions in the corpus
Fichier image/png, 13k
Haut de page

Pour citer cet article

Référence papier

Diana Yankova et Irena Vassileva, « Functions and Mechanisms of Code-Switching in Bulgarian Canadians »Études canadiennes / Canadian Studies, 74 | 2013, 103-121.

Référence électronique

Diana Yankova et Irena Vassileva, « Functions and Mechanisms of Code-Switching in Bulgarian Canadians »Études canadiennes / Canadian Studies [En ligne], 74 | 2013, mis en ligne le 01 juin 2015, consulté le 30 novembre 2021. URL : ; DOI :

Haut de page


Diana Yankova

New Bulgarian University

Irena Vassileva

European Polytechnic University

Haut de page

Droits d’auteur


Haut de page
  • Logo AFEC
  • OpenEdition Journals
Rechercher dans OpenEdition Search

Vous allez être redirigé vers OpenEdition Search