Issues in Codeswitching

Issues in CS: Competing Theories and Models Issues in Code-Switching: Competing Theories and Models Erman Boztepe1 Teachers College, Columbia University ABSTRACT This paper provides a critical overview of the theoretical, analytical, and practical questions most prevalent in the study of the structural and the sociolinguistic dimensions of code-switching (CS). In doing so, it reviews a range of empirical studies from around the world. The paper first looks at the linguistic research on the structural features of CS focusing in particular on the code-switching versus borrowing distinction, and the syntactic constraints governing its operation. It then critically reviews sociological, anthropological, and linguistic perspectives dominating the sociolinguistic research on CS over the past three decades. Major empirical studies on the discourse functions of CS are discussed, noting the similarities and differences between socially motivated CS and style-shifting. Finally, directions for future research on CS are discussed, giving particular emphasis to the methodological issue of its applicability to the analysis of bilingual classroom interaction. INTRODUCTION Rationale The National Center for Education Statistics (2002) has recently reported that, in the school year 2000-2001, over 3.4 million English language learners (ELL)2 were enrolled in public schools in the United States. In the same school year, these students made up 8.4 percent of over 2,880,000 public school students in New York State and 17.8 percent of New York City’s public school enrollment (New York State Education Department, 2002). Given the pervasiveness of bilingualism and the projections for increasing diversity in this country, continued research on the nature of bilingual speech is of crucial importance and has significant implications for educational research and bilingual classroom discourse alike. Erman Boztepe is a doctoral student in the Applied Linguistics program at Teachers College, Columbia University. His research interests include discourse and identity, language/dialect variation and language attitudes, and bilingual classroom discourse. Correspondence should be sent to Erman Boztepe, 509 West 121st Street, Apt 703, New York, NY 10027. E-mail: erman.boztepe@columbia.edu. 2 English language learners (ELL) are defined as those who, by reason of foreign birth or ancestry, speak a language other than English and (1) either understand and speak little or no English or (2) score at or below the 40th percentile on an English language assessment instrument approved by the Commissioner of Education. The U.S. total presented here is based on the data from 39 states (including the District of Columbia), which reported the number of students who were English language learners and receiving English language instruction during the school year 2000-2001 (National Center for Education Statistics, 2002). 1 1 Issues in CS: Competing Theories and Models As one of the most engaging aspects of bilingual speech, code-switching (hereafter CS) is nevertheless a highly stigmatized form of conversation. It would be reinventing the wheel to argue here for the link between the pejorative attitudes toward CS and the traditions of prescriptivism and semilingualismæwhich still persist today.3 The irony is that such false and unfounded notions are promoted not only in popular culture4 but also by the so-called fathers of modern linguistics, Leonard Bloomfield and Ferdinand de Saussure. For example, the following excerpt is from Bloomfield’s (1927) description of the linguistic profile of a Native American speaker: White Thunder, a man around 40, speaks less English than Menomini, and that is a strong indictment, for his Menomini is atrocious. His vocabulary is small, his inflections are often barbarous, he constructs sentences of a few threadbare models. He may be said to speak no language tolerably. (p. 395) It appears from Bloomfield’s observation that he does not see his informant as a fully competent speaker in either of the languages in his linguistic repertoire. In the same vein, in his now classic work on language contact phenomena, Weinreich (1968) described the ideal bilingual5 as the one who “switches from one language to the other according to appropriate changes in the speech situation (interlocutors, topics, etc.), but not in an unchanged speech situation, and certainly not within a single sentence” (p. 73). Such a characterization assumed by definition the existence of the imperfect bilingual who supposedly has less than ideal competence in either of the languages at his disposal. In turn, CS has become part of the performance of the imperfect bilingual. While it is understandable that these scholars were only reflecting the attitudes of their time, we nevertheless cannot ignore the fact that such notions about the legitimacy of one language or language variety over another have been the major source of inspiration for the deficit hypothesis in the United States and many other countries, and its practical applications in schools. For example, the British sociologist Bernstein’s (1972, 1974) work was mostly taken to imply that the reason why the children of working-class or ethnic minority groups failed in school was that their language was deficient or restricted in some way, which somehow had to be remedied by schools. Fasold’s (1975) description of a teacher-student exchange, which was screened to linguists at the 1973 Linguistic Institute in Ann Arbor, Michigan, illustrates one such corrective program developed by a team of educational psychologists for children alleged to have deficient language abilities: Earnest White teacher, leaning forward, holding a coffee cup: “This-is-not-a-spoon.” Little Black girl, softly: “Dis not no ‘poon.’ White teacher, leaning farther forward, raising her voice: “No, This-is-not-a-spoon.” Black child, softly: “Dis not a ‘poon.” 3 Prescriptivism is known as the view that one variety of language is given an inherently higher value than other varieties and that this ought to be imposed on the whole of the speech community, especially through educational means (Crystal, 1997). Semilingualism, on the other hand, is the popular belief that bilingual speakers who code-switch do so because of their lack of linguistic competence in their repertoire (Edelsky, Hudelson, Flores, Barkin, Altwerger, & Jilbert, 1983). 4 The British Home Secretary David Blunkett has recently suggested that British Asian families should speak English at home. For detailed information, visit http://news.bbc.co.uk/2/hi/uk_news/2261239.stm and http://news.bbc.co.uk/2/ hi/uk_news/politics/2254666.stm. 5 See Li Wei (2000) for a useful set of definitions of the terms commonly associated with bilingualism. 2 Issues in CS: Competing Theories and Models White teacher, frustrated: “This-is-not-a-spoon.” Child, exasperated: “Well, dass a cup!” The reaction of the linguists, after they had finished applauding and cheering for the child, was a mixture of amusement, incredulity and anger. (p. 202-203) In the case of bilingual classrooms, the notion of semilingualism embodies itself in the form of negative teacher attitudes toward students who code-switch in classroom interaction. CS, as with any kind of stigmatized language variety, is seen as a deviation from some norm. In their study on teacher attitudes toward non-standard varieties of American English, Ramirez and Milk (1986) asked teachers of bilingual students to judge four varieties (one standard and three non-standard) in terms of their appropriateness for the classroom, degree of correctness, and students’ academic potential. Teachers consistently judged English/Spanish CS as the least acceptable form in all respects, ranking it even less favorably than ungrammatical English. If this is true, then teachers’ beliefs about their students’ capabilities may strongly influence the level of student achievement. While it may be true that curricular practices of bilingual classroom teachers are shaped by the social conditions beyond the classroom, equally important is the need to consider the social life within the bilingual classroom. In that sense, bilingual classroom discourse has a lot to benefit from research on the nature of bilingual speech and code-switching. Background The study of the alternate use of two or more languages in conversation has developed in two distinct but related directions: Structural and Sociolinguistic. The structural approach to CS is primarily concerned with its grammatical aspects. Its focus is to identify syntactic and morphosyntactic constraints on CS. The sociolinguistic approach, on the other hand, sees CS primarily as a discourse phenomenon focusing its attention on questions such as how social meaning is created in CS and what specific discourse functions it serves. It should be noted at the outset, however, that these approaches are not in contradiction, but complementary to each other. The structural approach tries to identify the structural features of morphosyntactic patterns underlying the grammar of CS, whereas the sociolinguistic approach builds on this in its attempts to explain why bilingual speakers talk the way they do. Given this framework, this paper attempts to provide a critical overview of the competing theories and models most prevalent in the study of the structural and the sociolinguistic dimensions of CS. The paper first sketches the linguistic research on the structural features of CS, focusing in particular on the syntactic constraints governing its operation and the distinction between code-switching and borrowing. It then reviews sociological, anthropological, and linguistic perspectives dominating the sociolinguistic research on CS over the past three decades. Similarities and differences between socially motivated CS and style-shifting are discussed. The paper also reviews empirical studies on the discourse functions of CS. Finally, directions for future research on CS are discussed, giving particular emphasis to the methodological issue of its translation to bilingual classroom interaction. 3 Issues in CS: Competing Theories and Models Definitional Issues As with any aspect of language contact phenomena, research on CS is plagued by the thorny issue of terminological confusion. Not all researchers use the same terms in the same way, nor do they agree on the territory covered by terms such as code-switching, code-mixing, borrowing, or code-alternation. In particular, at issue here is the perceived distinction between the terms code-switching and borrowing (Gysels, 1992; Myers-Scotton, 1992; Poplack, 1980, 1981) on the one hand, and code-switching and code-mixing (Kachru, 1978; 1983; Sridhar & Sridhar, 1980) on the other. Several criteria have been proposed to distinguish between these two pairs of concepts. But before looking at them more closely in the next section, a consideration of the definitions of some basic terms in language contact, together with a working definition of CS that I adopt in the remainder of this paper, may be useful. The term code is a relatively neutral conceptualization of a linguistic varietyæbe it a language or a dialect.6 However, not many researchers really explicate the term in their definitions. Milroy and Muysken (1995), for example, define CS as “the alternative use by bilinguals of two or more languages in the same conversation” (p. 7). They use code-switching as a cover term under which different forms of bilingual behavior are subsumed. The term intra-sentential is used to refer to switching within the sentence, in contrast with the term inter-sentential used for switches between sentences as the relevant unit for analysis. Myers-Scotton (1993b) also uses code-switching as a cover term and defines it as “alternations of linguistic varieties within the same conversation” [italics added] (p. 1). Other researchers (e.g., Gardner-Chloros, 1991) also emphasize that switching can occur not only between languages but also dialects of the same language. In the same vein, Gumperz (1982) refers to the term as “the juxtaposition within the same speech exchange of passages of speech belonging to two different grammatical systems or subsystems” (p. 59). I will be using the term code-switching as an umbrella term in the remainder of this paper to cover the phenomena of alternating between two languages or dialects of the same language within the same conversation. At times I will also refer to code-alternation in a similar sense, but this should not be confused with the technical definition of the term (see below). Some researchers (e.g., Auer, 1995) use the term code-alternation as a hyponym to replace CS, but it is marginally used in that sense. The term alternation is, in fact, used in the literature to refer to instances of one language being replaced by the other halfway through the sentence, and it is mostly, but not always, associated with longer stretches of CS. The term insertion, in contrast, mostly correlates with occurrences of single lexical items from one language into a structure from the other language. In this sense, the terms represent two distinct but generally accepted processes at work in CS utterances (Muysken, 1995, 2000). Others (Kachru, 1983; Singh, 1985; Sridhar & Sridhar, 1980), however, reserve the term code-switching for inter-sentential switches only, and instead prefer to use code-mixing for intra-sentential switches. The reason is that only code-mixing (i.e., intra-sentential CS) requires the integration of the rules of the two languages involved in the discourse. But as far as the structural constraints are concerned, the intra- vs. inter-sentential distinction can equally well 6 Romaine (1995) mentions ‘monolingual code-switching’ extending the meaning of code-switching to include what others (Labov, 1966; Trudgill, 1974) would call style-shifting in monolingual speech: “I will use the term ‘code’ here in a general sense to refer not only to different languages, but also to varieties of the same language as well as styles within a language” (p. 121). 4 Issues in CS: Competing Theories and Models distinguish the two types of switches. So it largely remains as a matter of individual preference, but at the same time it creates unnecessary confusion. Still others (e.g., Muysken, 2000) avoid using the term code-switching as a cover term because they believe that switching suggests alternation only, as in the case of switching between turns or utterances, but not necessarily insertion. Instead, they prefer to use code-mixing as a hyponym to cover both code-switching (intra-sentential only) and borrowing (e.g., Pfaff, 1979). More importantly, however, Pfaff (1979) raises, along with Poplack (1980), the question of the need to distinguish between code-switching and borrowing.7 This is a much more complicated issue than the perceived distinction between code-switching and code-mixing, and it will be discussed as part of the structural approach to CS in the next section. STRUCTURAL DIMENSIONS OF CODE-SWITCHING In addressing the question of whether CS is rule-governed or random linguistic behavior, Labov (1971) once concluded that “no one has been able to show that such rapid alternation is governed by any systematic rules or constraints and we must therefore describe it as the irregular mixture [italics added] of two distinct systems” (p. 457). But CS research has come a long way over the past three decades. We know today that CS is not a purely idiosyncratic behavior, but rather it occurs at specific switch points. A basic question researchers have now been asking relates instead to the formulation of the syntactic constraints on where switching can occur within the sentence.8 Code-Switching vs. Borrowing Before the structural approach to CS could address its supposedly central question of how free the alternation is between the two languages from a structural point of view, it faced another issue to resolve: If lexical borrowings are to be excluded from the analysis of CS utterances, the boundaries between CS and lexical borrowing have to be clear before the analysis actually starts. Of the foreign words in code-switched utterances, what constitutes CS and what constitutes lexical borrowing? This problem can in fact be traced back to what Weinreich, Labov, and Herzog (1968) called the transition problem: Because language change is a diachronic process, we cannot really determine at what point in time a particular lexical item gained the status of a loanword in the recipient language. Also, the fact that bilingual communities in urban contexts where language change is supposedly rapid tend to be diffuse with no clear norms makes it even more difficult to study variation synchronically. There are two contradictory approaches as to whether and how to distinguish between the two terms. One group of researchers associated with Poplack (1978, 1980, 1981)9, have argued that lone other-language items are fundamentally different from longer stretches of switches. They 7 The concept of borrowing will be discussed as it relates to the study of code-switching only. Also, I will be using a borrowed item in the same sense as a loan word. 8 Note that only intra-sentential code-switching is relevant to the question of syntactic constraints. 9 i.e., Poplack, Wheeler, and Westwood (1987), Sankoff and Poplack (1981), and Sankoff, Poplack, and Vanniarajan (1990). For abbreviation purposes, these authors will be referred to as “Poplack and her associates.” 5 Issues in CS: Competing Theories and Models proposed morphosyntactic and phonological integration of foreign words into the recipient language as criteria for establishing the status of such single words. Most researchers (Bentahila & Davies, 1983; Myers-Scotton, 1993a), on the other hand, have chosen to deal with the problem by claiming that the perceived distinction between the two processes is not really critical to analyses of bilingual speech. Moreover, unlike the first group of researchers, they acknowledged single-word (i.e., insertions) and multiple-word (i.e., alternations) occurrences as two forms of CS, rather than as distinct processes to be distinguished from each other. According to Poplack and her associates, borrowing and CS are in fact based on different mechanisms. Using participant observation performance data of CS from the bilingual Puerto Rican community in New York City, she proposed three types of criteria to determine the status of non-native material in bilingual utterances. These include whether or not single lexical items from a donor language in code-switched utterances were (1) phonologically, (2) morphologically, and (3) syntactically integrated into what she called the base language.10 She identified four possible combinations of integration as shown in Table 1. According to this approach, in cases where a lexical item shows (a) only syntactic integration (Type 2), or (b) only phonological integration (Type 3), or (c) no integration at all (Type 4), it is considered to be an instance of CS. In contrast, cases where a lexical item shows all three types of integration (Type 1) constitute borrowing. While it did capture some generalizations and received confirmation from empirical studies in other bilingual communities, the criterion of phonological integration was later discarded due to its highly variable nature. The intermediary category has since been identified as nonce borrowings. TABLE 1 Poplack’s (1980) Identification of Code-Switching Based on the Type of Integration into the Base Language Type Levels of Integration Into Base Language Code-Switching? Phonological Morphological Syntactic 1 ÷ ÷ ÷ No 2 ¥ ¥ ÷ Yes 3 ÷ ¥ ¥ Yes 4 ¥ ¥ ¥ Yes Nonce borrowings are single lexical items or bound morphemes which are syntactically and morphologically integrated into the base language, but which may or may not show phonological integration. They differ from established borrowings in that they do not meet the criteria of frequency of use or degree of acceptance. In this approach, lexical borrowing is seen as a continuum ranging from established loanwords to nonce borrowings as shown in Figure 1. Its advantage is that it allows for lone other-language items to achieve the status of loanwords in time through an increase in their frequency and their adoption by monolinguals. But notice that neither CS is considered to be part of such a continuum nor are nonce borrowings seen as instances of CS (Poplack, Wheeler, & Westwood, 1987). 10 Base Language is the main language in a code-switched utterance to which a majority of phonological and morphological features of discourse can be attributed. See also matrix language vs. embedded language in Myers-Scotton (1993a). In this paper, the terms base language and recipient language will be used interchangeably. 6 Issues in CS: Competing Theories and Models FIGURE 1 The Continuum for Levels of Borrowing in Code-Switching Utterances (Poplack, Wheeler, & Westwood, 1987) ESTABLISHED LOANWORD NONCE BORROWING Morphologically/Syntactically/Phonologically Integrated Recurrent (individual) Widespread (community) Accepted Restricted Lexicon Morphologically/Syntactically (+/- Phonologically) Entire Lexicon (Content Words) At the other end of the continuum are those who claim that assimilation may not always be the defining criterion to distinguish borrowing from CS. Myers-Scotton (1992, 1993a) rejects morphosyntactic integration as a basis for distinguishing between CS and borrowing because she sees them as universally related processes such that both concepts are part of a single continuum. She therefore argues that a categorical distinction between CS and borrowing need not be made, yet she proposes frequency as the single best criterion to link borrowed forms more closely with the recipient language mental lexicon. She also disagrees with those researchers (e.g., Bentahila & Davies, 1983; Sridhar & Sridhar, 1980) who argued that one of the major characteristics of borrowed items is to fill lexical gaps in the recipient language. Instead, she argues that not all established borrowings actually occur due to the perceived absence of an equivalent term in the recipient language culture. Inspired by Haugen’s (1953) comment that “borrowing always goes beyond the actual ‘needs’ of language” (p. 373), she then draws a distinction between what she calls cultural borrowings and core borrowings. Cultural borrowings are those lexical items that are new to the recipient language culture. Core borrowings, on the other hand, refer to those lexical forms that have “viable” equivalents in the recipient language, and hence, do not really meet any lexical need in the base language (1993a, p. 169). It is only this type of borrowing which Myers-Scotton (1993a) considers to be part of a continuum involving lone other-language items in code-switching. Moreover, in cases where the language of the core borrowed item has a higher symbolic value (cf. Bourdieu, 1991) than that of the recipient language, the social prestige associated with the donor language motivates the non-integration (e.g., phonological) of any type of borrowed item. She then goes on to suggest that educated bilingual speakers may practice elite closure by consciously pronouncing borrowed items as closely to the originals as possible. The important point in Myers-Scotton’s argument is that, unlike Poplack and her associates, she does not see CS and borrowing as two distinct processes, nor does she see such a distinction to be critical. Gysels (1992) takes this idea even one step further on the basis of her French data in urban Lubumbashi Swahili by claiming that whether a lone other-language item is a switch or borrowing in fact cannot be determined because the same form may be interpreted as either a borrowed item or a code-switch depending on the overall discourse structure. Similarly, on the basis of his work among Turkish/Dutch bilinguals in the Netherlands, Backus (1996) also rejects morphosyntactic integration as a criterion for distinguishing switches from 7 Issues in CS: Competing Theories and Models borrowings, claiming that it lies, at least partially, within the individual speaker’s motivations to ascribe status to single-word foreign items in the recipient language. Indeed, there seems to be very little reason to distinguish borrowing from code-switching for purposes of formulating grammatical constraints on the surface syntactic level. After all, there are more similarities than differences between the two concepts. This does not of course mean that morphological and syntactic integration are not reliable criteria to distinguish the two processes. But given Weinreich, Labov, and Herzog’s (1968) transition problem, it may be impossible to systematically categorize instances of foreign elements as either CS or borrowing. Therefore, I have to agree with Eastman (1992) who states that “efforts to distinguish codeswitching, codemixing and borrowing are doomed” (p. 1), and that it is crucial that we “free ourselves of the need to categorize any instance of seemingly non-native material in language as a borrowing or a switch” (p. 1) if we want to understand the social and cultural processes involved in CS. Grammatical Constraints An early but influential contribution to the linguistic aspects of CS is Poplack’s (1980, 1981) study where she argued for the word-order equivalence between the languages involved. She proposed what she calls the free morpheme constraint and the equivalence constraint as being operative at the point of the switch in Spanish/English bilingual utterances. Other researchers such as Lipski (1978) and Pfaff (1979) also suggested similar constraints from the perspective of linear equivalence, but Poplack’s (1980) study was the first attempt to suggest explanatory principles. These constraints bear mentioning here because they form much of the later discussions: (1) The Free Morpheme Constraint Codes may be switched after any constituent in discourse provided that constituent is not a bound morpheme. (2) The Equivalence Constraint Code-switches will tend to occur at points in discourse where juxtaposition of L1 and L2 elements does not violate a syntactic rule of either language (Poplack, 1980, p. 586). The first constraint predicts that a switch is disallowed between a lexeme and a bound morpheme unless the item is phonologically integrated into the base language. It limits the potential switch sites to word boundaries only. For example, in the case of Spanish/English CS, *EAT-iendo is not permissible unless the verb stem is phonologically adapted into Spanish. The second constraint, on the other hand, predicts that CS will occur at points where the surface structures of the two languages map onto each other. Again, in Spanish/English CS, switches may not occur between nouns and adjectives in the noun phrase because attributive adjectives in English typically precede the head noun, whereas in Spanish they follow it. Poplack (1980) suggested universal validity for both constraints, but several researchers provided counter-evidence from different languages, notably Bentahila and Davies (1983) from their Moroccan Arabic/French corpus, Berk-Seligson (1986) in Spanish/Hebrew, and Belazi, 8 Issues in CS: Competing Theories and Models Rubin and Toribio (1994) from Italian/English, to name just a few. The counter-examples for the free morpheme constraint came especially from agglutinative languages such as Turkish (Hankamer, 1989), partially because, in such languages, each component of meaning is productively expressed by its own morpheme, which are then affixed to the stem. I will give here an example from my own corpus of data in Turkish/English CS: (1) Sen-inle bu konu-da CONFLICT-imiz var. Æ you-PREP this issue-PREP conflict-POSS PRONOUN (1ST PLURAL) exist Æ We (You and I) have a conflict (disagreement) over this issue. The constraint predicts that the Turkish bound morpheme -imiz (our) would not be affixed to the phonologically unintegrated English root conflict, yet it does. Clyne (1987) has also explicitly questioned the universality of a set of constraints using performance data from German/English and Dutch/English bilingual speakers in Australia. He also reports some examples that violate the free morpheme constraint, but more importantly, he further develops the notion of triggering that he introduced in earlier work (Clyne, 1967). In triggering, an item of ambiguous affiliation (i.e., one belonging to the speaker’s two languages) triggers off the switch from one language to the other. Free morpheme and equivalence constraints are also criticized for not recognizing the notion of asymmetry, which is arguably central to many language contact situations. Joshi (1985), for example, suggested that CS is characterized by an asymmetry with respect to the degree of participation of the languages involved. He wrote “speakers and hearers generally agree on which language the mixed sentence is ‘coming from.’ We can call this language the matrix language and the other language the embedded language” [italics in original] (p. 190-191). This brings us to the discussion of the asymmetrical models of insertion. Myers-Scotton’s (1993b) technical definition of CS captures the very aspect of asymmetry: “CS is the selection by bilinguals or multilinguals of forms from an embedded language (or languages) in utterances of a matrix language during the same conversation” (p. 4). Inspired by Joshi’s (1985) work, this definition departs from the notion of insertion and it forms the basis of what Myers-Scotton (1992, 1993a) calls the Matrix Language Frame model. In fact, we have already discussed one aspect of this modelæthat it makes no distinction between borrowing and switching as far as morphosyntactic integration is concerned. This approach differs from what Poplack and her associates have continuously claimedæthat lexical borrowing and CS are distinct phenomena and thus, lexical borrowings should be excluded from analyses of CS. But another central assumption behind Myers-Scotton’s model is that code-switched utterances have an identifiable matrix language, and that there is always an asymmetrical relationship between the matrix language (ML) and the embedded language (EL) such that the matrix language dominates a mixed clause according to the following three principles: (1) The Morpheme Order Principle In ML + EL constituents consisting of singly occurring EL lexemes and any number of ML morphemes, surface morpheme order (reflecting surface syntactic relations) will be that of the ML. 9 Issues in CS: Competing Theories and Models (2) The System Morpheme Principle In ML + EL constituents, all system morphemes which have grammatical relations external to their head constituent (i.e., which participate in the sentence’s thematic role grid) will come from the ML. (3) The Blocking Hypothesis In ML + EL constituents, a blocking filter blocks any EL content morpheme which is not congruent with the ML with respect to three levels of abstraction regarding subcategorization. (Myers-Scotton, 1993a, p. 83-120) According to the first principle, the matrix language determines the order of the elements in ML + EL constituents. Given the distinction that the model makes between content morphemes (e.g., nouns/verbs) and system morphemes (e.g., inflections/articles), and that there is a fundamental difference in their distribution, the second principle requires that function morphemes can only be drawn from the matrix language. Finally, the Blocking Hypothesis restricts the role of the embedded language even more by allowing only certain embedded language content morphemes to occur in mixed constituents. But how do we identify the matrix language in the first place? Myers-Scotton (1993a) identifies it as the language of more morphemes in any code-switched utterance, suggesting a discourse-oriented approach: The language of the conversation is the matrix language, which, by the way, is not unproblematic as the matrix language may change even within the course of a single conversation. Other ways of determining the matrix language were also proposed from a structural point of view, notably that of Joshi (1985) and Treffers-Daller (1994). On the whole, however, as Nishimura (1986) notes, researchers who work on typologically similar language pairs such as Spanish and English tend to assume symmetrical models of switching which depart from alternation (e.g., Poplack, 1980, 1981), whereas those working on typologically different languages such as Turkish and Dutch develop asymmetrical models of insertion which emphasize the contrast in terms of the degree of participation of the languages involved (e.g., Backus, 1996; Myers-Scotton, 1992, 1993a). More recently, some researchers (e.g., MacSwan, 1999) have also attempted to apply CS data to Chomsky’s (1995) minimalist framework. In this approach, the mechanisms for monolingual grammar are seen as necessary and sufficient for explaining bilingual CS. Researchers in the studies that we have looked at so far focused on (1) the extent to which similar constituents match in syntax across the language pairs involved in CS, and (2) the potential role of categorical equivalence, rather than word order equivalence, in the structural operation of CS. Still others analyzed CS within the theory of government and binding (Chomsky, 1981). Their basic unit of analysis was dependency, rather than equivalence, such that a switch cannot occur between two constituents if they are lexically dependent on each other (DiSciullo, Muysken, & Singh, 1986). There are also different patterns of switching in different language contact situations which are mostly related to the typological characteristics of the language pairs involved in CS. In other words, much of the disagreement in the structural approach to CS stems from the different types of models and syntactic constraints proposed to explain the general properties of CS for which universal validity was suggested. Despite these disagreements over the properties of patterning, however, there is also widespread agreement on the question of bilingual speakers’ degree of competence: “It was found that code-switching is a quite normal and widespread form of bilingual interaction, requiring a great deal of bilingual 10 Issues in CS: Competing Theories and Models competence” (Muysken, 1995, p. 177). In fact, little if any evidence has been found in research to date that CS leads to semilingualism or that CS is random linguistic behavior. On the contrary, it was suggested that CS itself constitutes the norm in many stable bilingual communities, and that “satisfaction of this norm requires considerably more linguistic competence in two languages” (Poplack, 1980, p. 588). So, if balanced bilinguals are linguistically competent in both languages in their repertoire, then the question becomes why two distinct codes exist at all within a single conversation or even an utterance? What specific discourse functions does CS serve in conversation? What factors influence a speaker’s code choice? We now turn to look at the answers suggested in the literature to these central questions of the sociolinguistic aspects of CS research. SOCIOLINGUISTIC DIMENSIONS OF CODE-SWITCHING The Meaning of Social Meaning Alternation between codes is the norm rather than the exception in many communities around the world today. In their agenda-setting article on switching between the standard and the non-standard dialects11 in a small town called Hemnesberget in Norway, Blom and Gumperz (1972) found that alternating codes among the local people was both patterned and predictable. Using an integrated ethnographic and linguistic approach, they identified two different types of code choice: situational switching and metaphorical switching. Situational switching occurs when participants redefine each other’s rights and obligations. For example, teachers deliver formal lectures in the standard dialect (i.e., Bokmål), but if they want to encourage open discussion, then they will shift to the local dialect (i.e., Ranamål). It assumes a direct relationship between the social situation and code choice. Metaphorical switching, on the other hand, is triggered by changes in topic rather than the social situation. For example, in clerk-resident exchanges at the community administration office, Blom and Gumperz (1972) observed that while greetings and inquiries about family affairs took place in Ranamål, conversations about the business transaction occurred in the standard dialect. Blom and Gumperz also introduced three types of social constraints which presumably affect the code choices of speakers: (1) setting, (2) social situation, and (3) social event (cf. Hymes’ (1967, 1972) social units of speech situation and speech event). Setting refers to the physical environment in which the social life of speakers operates. Social situation is defined as “particular constellations of [speakers], gathered in particular settings during a particular span of time” for a certain activity (p. 423). Finally, social event is a particular definition of the same social situation at a particular point in time. The switching of codes illustrated in the clerk-resident interaction for metaphorical switching echoes Erving Goffman’s notions of front stage and back stage. While the standard dialect is associated with front stage behavior, the local dialect symbolizes in-group solidarity, and creates islands of back stage within the office. In fact, Gumperz (1982, 1992) himself talks about conversational CS in his later work as contextualization cues where he sees the code, the dialect, and even style switching processes, as well as prosodic features of speech and formulaic 11 Ironically, one of the most influential articles on code-switching comes from alternation between dialects, rather than languages. 11 Issues in CS: Competing Theories and Models expressions, as implicit ways of conveying meaning as part of the interaction between speakers. In urban institutional contexts (e.g., workplace, school), although speakers may share a common lingua franca at a surface level, those from different ethnic or social class backgrounds often lack in their conversational exchanges a common set of contextualization cues, as a result of which misunderstandings may occur. As Gal (1989) notes, such “misunderstandings are heard by those in control of the institutions not as linguistic differences but as indications of personal qualities, and thus as objective grounds for rejection and devaluation of those attempting access [to material resources]” (p. 352). In the case of teacher-student interaction in bilingual classrooms, this occurs when teachers negatively evaluate bilingual students’ use of CS as a discourse strategy. Other researchers, notably Auer (1984, 1988, 1995), Alfonzetti (1998), and Sebba (1993) further developed Gumperz’s interactional perspective by employing conversation analysis (CA) techniques in their research in order to analyze performance data on CS. Specifically, Auer’s sequential approach to code-switching is made manifest in his following statement: “Any theory of conversational code-alternation [his term] is bound to fail if it does not take into account that the meaning of code-alternation depends in essential ways on its ‘sequential environment’” (p. 116). That is, the meaning of code-switching needs to be interpreted in relation to the preceding and following utterances. For Auer, the sequential embeddedness of meaning in bilingual conversation is “relatively independent” of its social meaning for the community (p. 132). The significance of Blom and Gumperz’s (1972) study therefore lies in their attempt to define social meaning largely as a product of individual interactions to the extent that it is created and negotiated locally, echoing, in a sense, LePage and Tabouret-Keller’s (1985) notion of acts of identity: [T]he individual creates for himself the patterns of his linguistic behavior so as to resemble those of the group or groups with which from time to time he wishes to be identified, or so as to be unlike those from whom he wishes to be distinguished. (p. 181) Blom and Gumperz placed much of the responsibility within the individual such that they saw stable patterns as generated from individual code choices, but not vice versa. Their approach allows the individual speaker the kind of flexibility LePage and Tabouret-Keller (1985) talk about in their description of the acts of identity quoted above: “[T]he same individual need not be absolutely consistent in all his actions. He may wish to appear as a member of the local team on some occasions, while identifying with middle-class values on others” (Blom & Gumperz, 1972, p. 421). This differs considerably from Fishman’s (1965, 1972) macro-level approach to language choice where he focuses on the correlations between code choice and types of activity. Inspired by Ferguson’s (1959) seminal article on diglossia, he is primarily concerned with stable norms of choice and habitual use of language in which there is an almost one-to-one relationship between codes and activities: “‘Proper’ usage dictates that only one of the theoretically co-available languages or varieties will be chosen by particular classes of interlocutors on particular kinds of occasions to discuss particular kinds of topics [italics in original] (Fishman, 1972, p. 437). In other words, social meaning lies not within the act of switching itself, but in the perceived association between speech activities on the one hand, and norms of language choice on the other. It is the stable patterns of use that give meaning to individual choice. This is made manifest in Fishman’s (1965, 1972) key concept of domain, 12 Issues in CS: Competing Theories and Models which he develops in relation to some corresponding typical role relationships as shown in Table 2.12 TABLE 2 The Scheme of Relationships in Fishman’s (1972) Domain Analysis Domain Interlocutor Place Topic Family Parent Home How to be a good son or daughter Friendship Friend Beach How to play a certain game Religion Priest Church How to be a good Christian Education Teacher School How to solve an algebra problem Employment Employer Workplace How to do your job more effectively Fishman’s (1971) typical example is English/Spanish CS between a boss and his secretary, both Puerto Ricans. The boss makes exclusive use of English as he dictates a letter to his secretary, but then switches to Spanish for an informal conversation with her about the addressee. Note that social meaning lies not within the act of switching here, but in the correlation between the type of activity and code choice (e.g., Spanish for informal conversation vs. English for business). The tension between macro- and micro-sociolinguistic dimensions of CS has shaped much of the later discussions in the study of the social aspects of code choice. For example, on the basis of his work in Kru/English CS in Liberia, Breitborde (1983) vehemently claims that the social meaning of CS cannot afford to ignore the societal regularities and the macro-level organization of social relationships, which arguably give meaning to individual choices. In the case of the clerk-resident interaction at the community administration office in the Hemnesberget study, Breitborde (1983) would therefore argue that the teacher’s behavior of switching back to the local dialect is in fact indicative of the underlying social regularities rather than an individual strategy to redefine the social situation. I would argue that Fishman’s (1965, 1972) model of domain analysis is too deterministic to explain CS in urban contexts. It tells us very little about what the speaker accomplishes as a result of alternating between available codes in his linguistic repertoire. Societal factors do form the basis, at least partially, of the contextual interpretation of code choice, but certainly not at the expense of determining language choice in all cases per se. On the other hand, I find the current practice of the conversation-analytic approach in CS too isolated from the macro-level factors which, if not determine, at least provide a general framework for its interpretation. There is evidence in research, especially from African data (e.g., Blommaert, 1992), which suggests that the social meaning of CS cannot be accounted for by local factors only. As Gal (1983) accurately pinpointed, “neither the more macro approaches nor those giving primacy to micro variables constitute a conceptually unified group” (p. 64). She sees norms associating codes with general spheres of activity as “not rules to be obeyed, but requisite knowledge to build on in conveying one’s communicative intents” (p. 69). Both approaches, in their current form, fail to capture this link between macro- and micro-level factors in speakers’ interpretation of CS utterances. 12 Note that in Labov’s (1966) now classic study of the social stratification of English in New York City, the patterns of variation could emerge only in relation to the a priori identification of macro-level social categories. 13 Issues in CS: Competing Theories and Models Tabouret-Keller (1983) makes the point that the higher the predictability of code choice, the more likely the act of switching is an instance of conforming to societal patterns. For him, conforming to norms implies no choice at all on the part of the speaker other than the choice to conform, and therefore, “a distinction is necessary between a predictable switch and an unexpected one” (p. 143). This is where the theoretical concept of markedness comes into play in CS. Myers-Scotton’s (1993b, 1993c) Markedness Model is arguably an attempt to incorporate the micro- and the macro- perspectives into CS research. But the basic assumption of the model is Fishman’s (1972) normative framework: “Habitual language choice in multilingual speech communities is far from being a random matter of momentary inclination” (p. 437). Myers-Scotton (1987) argues that any code choice is indexical of norms of society at large, yet norms only determine the relative markedness of choices, rather than the choices themselves. She sees code choices as a function of negotiations of position between the speakers, rather than as situated behavior. Speakers use the codes in their repertoire to index the rights and obligations holding between the participants. Following Grice’s (1975) cooperative principle in its structure only, she formulates a negotiation principle as underlying all code choices in bilingual speech, for which she claims universality and predictive validity: “Choose the form of your conversation contribution such that it indexes the set of rights and obligations which you wish to be in force between speaker and addressee for the current exchange” (Myers-Scotton, 1993b, p. 113). She proposed several related maxims to account for such switching phenomena. But she is also at pains to make clear her ambitious goals, as she has considerably revised them over the years as new data came in (see Myers-Scotton, 1980, 1983, 1993b). Note that the model is principally developed on the basis of the researcher’s work on Swahili/English CS in Kenya only. She identifies three maxims operative in bilingual conversation: The Unmarked Choice Maxim requires the speaker to switch from one unmarked (i.e., expected) code to another on the basis of situational changes during interaction such that the unmarked code changes. This first maxim is reminiscent of Fishman’s (1971) example of boss-secretary interaction. The Marked Choice Maxim applies when the speaker chooses to negotiate the rights and obligations balance for such purposes as increasing social distance or creating an aesthetic effect. Finally, The Exploratory Choice Maxim occurs when an unmarked choice in accordance with community norms is not obvious from situational factors. It applies in cases where, for example, there is a clash of norms and role relationships as in the case of a conversation between a brother and a sister at the brother’s place of business in the presence of other customers, as opposed to home, their usual place of meeting. The sister uses Lwidakho, their shared mother tongue, which signifies solidarity. The brother, on the other hand, speaks in Swahili, the national lingua franca, to let his sister know that she is being treated as a customer (Myers-Scotton, 1993b, p. 144-145). Proponents of the conversation analysis approach sharply criticized the Markedness Model for its adoption of Fishman’s (1965, 1972) approach. In particular, Meeuwis and Blommaert (1994) argue that the model is a static and mistaken view of indexicality and social behavior where speakers are described as simply followingæor, not followingærules for already existing norms. They also accuse the model for leaving no room for the constitutive nature of talk of the social structure as well as its ignorance of diachronic language change in the history of the community. In more recent work, however, Myers-Scotton (Myers-Scotton, 1998, 1999; Myers-Scotton & Bolonyai, 2001) has reconsidered the model within Elster’s (1986, 1989) 14 Issues in CS: Competing Theories and Models rational action theory in an attempt to develop an “extended version” of it (1999, p. 1260). In this modified approach, she argues that CS is best explained by the optimal use of speakers’ resources in their linguistic repertoires. In other words, speakers engage in CS because, through conscious calculation of costs and benefits, they discover that the rewards of CS will be greater than those of maintaining a monolingual discourse pattern (cf. Sperber & Wilson, 1986). But whether human action can be the outcome of conscious calculation is the subject of much debate in sociology. For reasons which will be clear below, Bourdieu (1990), for example, argues that by virtue of what he calls the habitus (see below), individual actions should be seen as an encounter between already possessed predispositions to act in certain ways and an immediate contextual situation (Thompson, 1991, p. 17). Bourdieu (1977b, 1991) is to be singled out, among other things, for his original contribution to language contact phenomena. The explanatory value of his theory comes from its ability to integrate micro-level linguistic variation with macro-level societal factors. But we cannot proceed into the details of his linguistic theory unless we understand what he means by the following two key conceptsæthat of habitus, and symbolic capital. Habitus refers to one’s complex system of dispositions to act and behave in certain ways. It is responsible for one’s attitudes, perceptions and practices, and is grounded in one’s early childhood experiences. These dispositions are both structured and durable. They are structured in the sense that they inevitably reflect the social environment within which they were acquired. They are also durable because they are deeply rooted in one’s life history. Symbolic capital, in sociological terms, is the end-product of one’s linguistic competence. It refers to the honor and prestige accumulated through one’s linguistic practices on a par with his or her position in the social structure. There are also other forms of capital, namely economic (e.g., cash and assets), cultural (e.g., knowledge, skills, formal education), and social (e.g., networks of interpersonal relations). In his theory of practice, Bourdieu (1977a) assumes a fundamental link between one’s actions and interests, which one may consciously or unconsciously pursue. In the same vein, he sees a strong correlation between one’s linguistic utterances and the particular contexts, or, to use his term, linguistic markets in which those utterances are produced (cf. Blom & Gumperz’s (1972) characterization of setting, social situation, and social event). Given the fact that the properties of linguistic markets endow linguistic expressions with a certain value, part of one’s language socialization involves knowing when and how to produce utterances that are highly valued in those markets (i.e., contexts). To better illustrate the point let us consider Bourdieu’s (1977b) following example of the use of code-switching and style-shifting by an old woman from a village in Béarn, a province in southwestern France: [The old lady] at one moment used “provincialised French” to address a shopkeeper’s wife, a young woman originating from another large market town in Béarn; […] the next moment, she spoke in Béarnais [the local dialect] to a woman who lived in the town but who was originally from [the villages] and more or less of her own age; then she used a French that if not “correct” was at least strongly “corrected” to address a minor official in the town; and finally she spoke in Béarnais to a [roadworker] in the town, […] aged about fifty. (p. 657) According to Bourdieu, it is the speaker’s assessment of the contextual cues and the anticipation of the likely reception of his/her linguistic utterances that serve as internal constraints on his/her code choices. To put it another way, all utterances are in a sense 15 Issues in CS: Competing Theories and Models euphemized: “What is said is a compromise between what would like to be said and what can be said” (p. 663). But the important point here is that the chances for accessing the linguistic variety with higher prestige is differentially distributed in societies, including the so-called monolingual nation-states (Edwards, 1985). Ironically, however, even those who have little or no access to the standard, or the legitimated variety accept its legitimacy and its correctness.13 This is what Bourdieu (1991, p. 51) means by symbolic domination and misrecognition of language: An “impalpably inculcated” process of acquisition of the arbitrary social construction by those who benefit least from the hierarchical relations of power.14 But as Gal (1989) notes, Bourdieu agrees at the same time that linguistic form does not have power per se, but rather it is the speakers who attribute value and power to languages or language varieties. Therefore, the future of the legitimate language and the reproduction of linguistic capital depend on the school system, a fact constantly and deliberately distorted by the state, the media, and the educational system itself in capitalist societies: [T]he educational system succeeds in legitimating a particular linguistic variant exactly by denying this, by creating a misrecognition of the variant’s imposition, presenting it instead as an inherently better form. The linguistic insecurity of working-class and minority speakers and their hypercorrection are the results of symbolic domination [italics in original]. (Gal, 1989, p. 353-354) Heller’s (1992, 1995a, 1995b) research focuses on this very aspect of symbolic domination within the larger sociopolitical context in Canada. In her analysis of classroom discourse of French-language minority education in Ontario, she makes the point that CS is one of the most powerful and potentially effective strategies at the disposal of French/English bilingual students to collaborate with or resist the monolingualizing and standardizing efforts of the school. Bourdieu’s theory sees the creation of social meaning both as a reflection of the social system and simultaneously as constitutive of it. Other researchers (e.g., Genesee & Bourhis, 1982, 1988; Gibbons, 1987) also recognized the need for a comprehensive model which takes into account not only macro-level societal factors but also micro-level situational and attitudinal ones. In his research on Cantonese/English CS among Hong Kong university students, for example, Gibbons (1987) emphasizes that code choices are made against the background of social factors as well as those related to the immediate situation. Discourse Functions of CS Bourdieu’s theory has been criticized for emphasizing the dimension of power at the expense of downplaying the role of solidarity in vernacular varieties (e.g., Mesthrie, Swann, Deumert, & Leap, 2000). But we will focus on the arguably neglected aspect in Bourdieu’s work hereæ what functions CS serves in bilingual discourse, and what factors influence code choice. 13 See Queen (in press) for a description of Turkish/German bilinguals in Germany within Bourdieu’s framework. Bourdieu (1977b) notes that “what are called linguistic conflicts arise when the possessors of the dominated competence refuse to recognize the dominant language æ and with it the monopoly of linguistic legitimacy which its possessors arrogate to themselves æ and claim for their own language the material and symbolic profits that are reserved for the dominant language” (p. 664). 14 16 Issues in CS: Competing Theories and Models We know today from several studies that CS functions primarily as a symbol of group identity and solidarity among members of the speech community (Beebe, 1981; Gal, 1978, 1979; Milroy, 1987). In fact, Gumperz (1982) referred to the two codes in switching as the we-code and the they-code, categorizing them in terms of their primary function—i.e., solidarity. While the former is associated with in-group relations and informal activities, and is aesthetically undervalued, the latter refers to the majority language that often serves as the communication tool for out-group relations with the mainstream community. Grosjean (1982) provides a concise but comprehensive outline of the factors that potentially explain speakers’ choice of we-code or they-code (Table 3). TABLE 3 Grosjean’s (1982, p. 136) List on Factors Influencing Language Choice FACTORS INFLUENCING LANGUAGE CHOICE _____________________________________________________________________________ Participants Situation •Language proficiency •Language preference •Socioeconomic status •Age •Sex •Occupation •Education •Ethnic Background •History of speakers’ linguistic interaction •Kinship relation •Intimacy •Power relation •Attitude toward languages •Outside pressure •Location/Setting •Presence of monolinguals •Degree of formality •Degree of intimacy Content of Discourse •Topic •Type of vocabulary Function of Interaction •To raise status •To create social distance •To exclude someone •To request or command In her ethnographic study of the language shift process in a border town called Oberwart in eastern Austria, Gal (1978, 1979) looked at the language choice patterns of bilingual speakers of Hungarian/German in a variety of social contexts. Oberwart is traditionally an agricultural community, but has undergone rapid social change due to economic developments in the area, which gave the natives ample opportunities to work in waged jobs as opposed to doing peasant work. As a result, an opposition is created between peasant and worker values, which are represented in the two languages of the community. While Hungarian symbolizes the traditional peasant culture, German is associated with access to material resources and modernity. Using implicational scales, Gal demonstrated that there is a strong correlation between the individual’s language choice patterns and his or her age such that while older speakers preferred Hungarian, younger speakers chose German even in cases where their interlocutors addressed them in Hungarian. This, in turn, led her to conclude that there is a language shift in progress in the 17 Issues in CS: Competing Theories and Models community such that activities which were previously associated with Hungarian are now associated with German. A related finding of Gal’s (1978, 1979) study is that the interlocutor is the most critical factor influencing a speaker’s code choice.15 Given the social values symbolized by each language, she looked at the role of speakers’ contacts in the community, i.e., social network, on their language choice. She discovered that there was a high correlation between speakers’ patterns of language choice and their social network. Milroy (1987) took the idea of the relationship between social network and code choice even one step further. In her study of the vernacular working class speech of three inner city communities in Belfast, she found that the dense and multiplex nature of a working class individual’s social network gives rise to its imposing the vernacular form, which symbolizes in-group solidarity, on his or her code choice.16 A similar relationship between social network and language choice has also been found among second generation Chinese/English bilinguals in Newcastle upon Tyne in England (see Li Wei, 1995; Milroy & Li Wei, 1995). Several researchers put forward the idea that code-switching accomplishes for the bilingual what style-shifting does for the monolingual (e.g., Romaine, 1995). In her Oberwart study, Gal (1979) also argued that CS and style-shifting occur in “complementary distribution” depending on the linguistic means available to one’s interlocutor: “Style-shifting occurs only where conversational language switching does not” (p. 118). While both CS and style-shifting may serve the same kind of functions in conversation, I would argue that it would be too naïve to assume such mutual exclusivity in terms of the distribution of their occurrences across one’s interlocutors. For example, it is not uncommon to encounter situations where the bilingual speaker uses CS in interacting with a monolingual speaker simply to create an aesthetic effect, or to claim expertise in an area, or even to impose authority on a social inferior. In fact, this kind of occurrence is not absent from Gal’s corpus of data either. Until very recently, no serious attempt has been made to understand how style-shifting and CS co-exist in bilingual speakers’ speech. In their study of the Panjabi/English bilingual community in London, Gardner-Chloros, Charles, and Cheshire (2000) found that speakers used CS “as a further dimension to the monolingual means which are available” (p. 1305). They concluded that CS is an additional tool at the disposal of bilinguals, the effect of which “was almost always over and above what could be achieved monolingually” (p. 1335). Some researchers also considered CS within the speech accommodation theory (SAT) (Giles & Powesland, 1975; Giles & Smith, 1979). The theory posits that speakers adjust their speech style as a way of expressing their attitudes or intentions toward their interlocutors. The two key concepts introduced by SAT are convergence and divergence. While the former refers to accommodating toward the speech style of one’s interlocutor, the latter signals a shift away from it. The notion of convergence is considered to convey a sense of solidarity. In contrast, divergence is a means to create social distance from one’s interlocutor through which social disapproval is communicated. Bourhis, Giles, Leyens, and Tajfel’s (1979) study gives support to this latter aspect of accommodation. In their study of intergroup behavior (Tajfel, 1974) among 15 Beebe (1977) also demonstrated that the ethnic background of the interlocutor (i.e., listener) influences the code choice of Chinese/Thai bilingual teachers. 16 A high-density network, as opposed to a low-density network, is the one where one’s contacts also know each other. In a multiplex network, on the other hand, individuals interact with each other in more than one capacity (e.g., neighbor and co-worker at the same time). 18 Issues in CS: Competing Theories and Models Flemish university students in Belgium, they found that Flemish-speaking students frequently used CS as a way of dealing with a perceived ethnic threat coming from an out-group Francophone speaker by helping them create social distance. Bell (1984, 1991) also sees the interlocutor, or the audience, as the main motivation behind variation in speech style. In his approach, accommodating toward an audience is not limited to monolingual style-shifting only, but applies to all codes and levels of one’s linguistic repertoire, including switching between languages. On the basis of three language contact situations around the world,17 Gumperz (1982) identifies six basic discourse functions that CS serves in conversation to illustrate its most common uses. These are (1) Quotations, (2) Addressee Specification, (3) Interjections, (4) Reiteration, (5) Message Qualification, and (6) Personalization versus Objectivization. Quotations are occurrences of switching where someone else’s utterance is reported either as direct quotations or as reported speech. Gal (1979) argued that “all one needs to know to predict the language in which most quotes will be spoken is the language in which the original utterance was spoken” (p. 109). This, nevertheless, may not always be the case (e.g., Auer, 1995). In addressee specification, the switch serves to direct the message to one particular person among several addressees present in the immediate environment. Interjections, on the other hand, simply serve to mark sentence fillers as in the insertion of the English filler you know in an otherwise completely Spanish utterance. Reiteration occurs when one repeats a message in the other code to clarify what is said or even to increase the perlocutionary effect of the utterance. For example, a Spanish/English bilingual mother may call her children who are playing on the street first in Spanish, but if they do not listen, then in English. Gumperz (1982) defines message qualification as an elaboration of the preceding utterance in the other code in Mann & Thompson’s (1986) sense. Finally, personalization versus objectivization signals the degree of speaker involvement in a message as in the case of, for example, giving one’s statement more authority in a dispute through CS. Gumperz’s (1982) categorization of conversational functions of CS is not unproblematic. In at least three of the cases above, the items do not really tell us what the speaker accomplishes in conversation through switching codes. In quotations, for example, we still do not know what is achieved other than the fact that speakers generally tend to report utterances in the language in which they were originally spoken. A similar problem arises with interjections, and message qualification as well. The question of what specific discourse functions are fulfilled by inserting, for example, an English sentence filler in an otherwise Spanish utterance still remains largely unanswered. Similar typologies proposed by other researchers are not less problematic, often confusing functions with forms of CS (Gardner-Chloros, 1991; Saville-Troike, 1982). For example, the three most common functions in Gardner-Chloros’ (1991) study on French-Alsatian CS in Strasbourg are what Gumperz (1982) would call quotations, addressee specification and reiteration. Notice that most of the functions identified by Gumperz and others can be marked either through lexical means, or by prosodic features or even gestures in monolingual conversation. This, in turn, supports the idea that CS is in fact an additional strategic device, and only one of the contextualization cues at the disposal of bilingual speakers. 17 Situation 1 is Slovenian/German code-switching near the border of Austria. Situation 2 involves Hindi/English code-switching in New Delhi. Situation 3 is Chicano Spanish/English code-switching in the U.S. (Gumperz, 1982, p. 73). 19 Issues in CS: Competing Theories and Models DIRECTIONS FOR FUTURE RESEARCH From an applied perspective, research on CS may help us understand classroom interaction and how differences in interactional norms in bilingual and multilingual classrooms influence not only the learning environment but also the level of student achievement. Even this brief review of the issues in the structural and the sociolinguistic dimensions of CS shows that achieving such ambitious goals is not an easy task, and may be possible only from a multi-disciplinary perspective. A successful analysis of syntactic constraints alone, for example, cannot of course guarantee attitude change, because understanding may not necessarily or automatically lead to altered views. Lipski (1978) once wrote: since the role of individual idiosyncratic factors seems to be an important aspect of code-switching, in that among groups of approximately equal bilingual abilities, some code-switch more than others, a complete determination of the sufficient conditions for code-switching probably lies beyond the reach of behavioral sciences. With regard to the linguistic constraints, however, the path toward an eventual model seems more clearly indicated. (p. 261) Probably no one believes this today. I do agree that grammatical models of CS with high levels of explanatory value are both an essential and a useful first step in increasing self-awareness, but they may not necessarily be sufficient to explain why stigmatized varieties including, but not limited to, CS, still persist today given the value of standard forms for social advancement (Bouchard-Ryan, 1979). Another concern regarding the nature of the relationship between the two approaches is also raised by Muysken (1995), who argues that “the sociolinguistic study of CS cannot proceed without a solid, theoretically based ‘structural analysis’” (p. 178). This statement echoes my earlier comments regarding the structural and the sociolinguistic approaches to study CS as being complementary to each other, with each producing additional information to build an overall theory of bilingual speech. An equally important dimension that should definitely be one component of such an overall theory is the psycholinguistics of mixed language processing in bilingual speakers (see Grosjean, 1997, for a recent review of the relevant research). In order to meet the challenge set in the beginning of this paper, it is my belief that more ethnographic studies of bilingual classroom interaction are needed. Such studies should ultimately aim to link micro-level analyses of classroom interaction with broader questions of social reality. Only then may it be possible to get insights into the role of CS as a discourse strategy. Finally, CS studies to date have mostly focused on synchronic variation. Few studies have indeed looked at diachronic change, though there are some exceptions (e.g., Gal, 1979; Pfaff, 1991). But we still know little about whether and how CS in a bilingual community can be a sign of attrition for a minority language. In the United States, it is the continuous flow of immigration that keeps Spanish vibrant. It is this vibrancy that makes the high level of first language attrition among second- and third-generation bilinguals more difficult to recognize (Hernández-Chavez, personal communication). But we know little about the role of CS in this process. Much work therefore remains to be done in this particular area of research. 20 Issues in CS: Competing Theories and Models CONCLUSION This study explored the issues in the structural and the sociolinguistic dimensions of CS. In the structural framework, we have observed that researchers have suggested competing models derived from current syntactic theories to explain the switching phenomena. Symmetrical models of alternation such as equivalence-point switching try to explain CS in terms of word order equivalence. Asymmetrical models of insertion, on the other hand, are based on the premise that there is a difference in the degree of participation of the languages involved in CS. We have also observed that the differences in models are due in large part to the differences in typological characteristics of the language pairs in question. In the sociolinguistic framework, much of the debate is characterized by a tension between macro- versus micro-sociolinguistic approaches as a point of reference to define social meaning. The basic premise of the macroperspective is that there are societal norms of code choice, which, in turn, are associated with certain types of activity. From the micro- perspective, code-switching is seen as a strategic tool at the disposal of speakers through which social reality is created, and conversational functions ranging from signaling dual membership in the two communities to simply emphasizing a message are conveyed. After all, it is not always what you play, but how you play the game. REFERENCES Alfonzetti, G. (1998). The conversational dimension in code-switching between Italian and dialect in Sicily. In P. Auer (Ed.), Code-switching in conversation: Language, interaction and identity (pp. 180-211). London: Routledge. Auer, P. (1984). Bilingual conversation. Amsterdam: John Benjamins. Auer, P. (1988). A conversation analytic approach to codeswitching and transfer. In M. Heller (Ed.), Codeswitching: Anthropological and sociolinguistic perspectives (pp. 187-213). Berlin: Mouton de Gruyter. Auer, P. (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 (pp. 115-135). New York: Cambridge University Press. Backus, A. M. (1996). Two in one: Bilingual speech of Turkish immigrants in The Netherlands. Tilburg, The Netherlands: Tilburg University Press. Beebe, L. M. (1977). The influence of the listener on code-switching. Language Learning, 27, 331-339. Beebe, L. M. (1981). Social and situational factors affecting the communicative strategy of dialect code-switching. International Journal of the Sociology of Language, 32, 139-149. Belazi, H. M., Rubin, E. J., & Toribio, A. J. (1994). Code switching and X-bar theory: The functional head constraint. Linguistic Inquiry, 25, 221-237. Bell, A. (1984). Language style as audience design. Language in Society, 13, 145-204. Bell, A. (1991). Audience accommodation in the mass media. In H. Giles, J. Coupland & N. Coupland (Eds.), Contexts of accommodation: Developments in applied sociolinguistics (pp. 69-102). Cambridge: Cambridge University Press. 21 Issues in CS: Competing Theories and Models Bentahila, A., & Davies, E. D. (1983). The syntax of Arabic-French code-switching. Lingua, 59, 301-330. Berk-Seligson, S. (1986). Linguistic constraints on intrasentential code-switching: A study of Spanish/Hebrew bilingualism. Language in Society, 15, 313-348. Bernstein, B. (1972). A sociolinguistic approach to socialization with some reference to educability. In J. J. Gumperz & D. Hymes (Eds.), Directions in sociolinguistics: The ethnography of communication (pp. 465-497). New York: Holt, Rinehart, and Winston. Bernstein, B. (1974). Class, codes and control, vol. 1. London: Routledge. Blom, J. P., & Gumperz, J. J. (1972). Social meaning in linguistic structures: Code-switching in Norway. In J. J. Gumperz & D. Hymes (Eds.), Directions in sociolinguistics: The ethnography of communication (pp. 407-434). New York: Holt, Rinehart, and Winston. Blommaert, J. (1992). Codeswitching and the exclusivity of social identities: Some data from Campus Kiswahili. Journal of Multilingual and Multicultural Development, 13, 57-70. Bloomfield, L. (1927). Literate and illiterate speech. American Speech, 2, 432-439. Bouchard-Ryan, E. (1979). Why do low prestige varieties persist? In H. Giles & J. St. Clair (Eds.), Language and social psychology (pp. 145-157). Cambridge: Cambridge University Press. Bourdieu, P. (1977a). Outline of a theory of practice. Cambridge: Cambridge University Press. Bourdieu, P. (1977b). The economics of linguistic exchanges. Social Science Information, 16, 645-668. Bourdieu, P. (1990). The logic of practice. Cambridge: Polity Press. Bourdieu, P. (1991). Language and symbolic power. Cambridge, MA: Harvard University Press. Bourhis, R. G., Giles, H., Leyens, J. P., & Tajfel, H. (1979). Psycholinguistic distinctiveness: Language divergence in Belgium. In H. Giles & R. St. Clair (Eds.), Language and social psychology (pp. 158-185). Oxford: Blackwell. Breitborde, L. B. (1983). Levels of analysis in sociolinguistic explanation: Bilingual code switching, social relations, and domain theory. International Journal of the Sociology of Language, 39, 5-43. Chomsky, N. (1981). Lectures on government and binding. Dordrecht: Foris. Chomsky, N. (1995). The minimalist program. Cambridge, MA: MIT Press. Clyne, M. (1967). Transference and triggering. The Hague: Nijhoff. Clyne, M. (1987). Constraints on code-switching: How universal are they? Linguistics, 25, 739-764. Crystal, D. (1997). The Cambridge encyclopedia of language. Cambridge: Cambridge University Press. DiSciullo, A.-M., Muysken, P., & Singh, R. (1986). Code-mixing and government. Journal of Linguistics, 22, 1-24. Eastman, C. M. (1992). Codeswitching as an urban language contact phenomenon. Journal of Multilingual and Multicultural Development, 13, 1-17. Edelsky, C., Hudelson, S., Flores, B., Barkin, F., Altwerger, B., & Jilbert, K. (1983). Semilingualism and language deficit. Applied Linguistics, 4, 1-22. Edwards, J. (1985). Language, society and identity. Oxford: Basil Blackwell Ltd. Elster, J. (1986). Introduction. In J. Elster (Ed.), Rational choice (pp. 1-33). Oxford: Blackwell. Elster, J. (1989). The cement of society. Cambridge: Cambridge University Press. Fasold, R. (1975). Review of J. R. Dillard’s ‘Black English.’ Language in Society, 4, 198-221. 22 Issues in CS: Competing Theories and Models Ferguson, C. A. (1959). Diglossia. Word, 15, 325-340. Fishman, J. A. (1965). Who speaks what language to whom and when? La Linguistique, 2, 67-88. Fishman, J. (1971). Sociolinguistics. Rowley, MA: Newbury House. Fishman, J. A. (1972). Domains and the relationships between micro- and macro-sociolinguistics. In J. J. Gumperz & D. Hymes (Eds.), Directions in sociolinguistics: The ethnography of communication (pp. 435-453). New York: Holt, Rinehart, and Winston. Gal, S. (1978). Variation and change in patterns of speaking: Language shift in Austria. In D. Sankoff (Ed.), Linguistics variation: Models and methods (pp. 227-238). New York: Academic Press. Gal, S. (1979). Language shift: Social determinants of linguistic change in bilingual Austria. New York: Academic Press. Gal, S. (1983). Comment. International Journal of the Sociology of Language, 39, 63-72. Gal, S. (1989). Language and political economy. Annual Review of Anthropology, 18, 345-367. Gardner-Chloros, P. (1991). Language selection and switching in Strasbourg. Oxford: Oxford University Press. Gardner-Chloros, P., Charles, R., & Cheshire, J. (2000). Parallel patterns? A comparison of monolingual speech and bilingual codeswitching discourse. Journal of Pragmatics, 32, 1305-1341. Genesee, F., & Bourhis, R. Y. (1982). The social psychological significance of code-switching in cross-cultural communication. Journal of Language and Social Psychology, 1, 1-27. Genesee, F., & Bourhis, R. Y. (1988). Evaluative reactions of language choice strategies: Francophones and Anglophones in Quebec City. Language and Communication, 8, 229-250. Gibbons, J. (1987). Code-mixing and code choice: A Hong Kong case study. Clevedon: Multilingual Matters. Giles, H., & Powesland, P. F. (1975). Speech style and social evaluation. London: Academic Press. Giles, H., & Smith, P. M. (1979). Accommodation theory: Optimal levels of convergence. In H. Giles & R. N. St. Clair (Eds.), Language and Social Psychology (pp. 45-65). Baltimore, MD: University Park. Grice, H. P. (1975). Logic and conversation. In P. Cole & J. L. Morgan (Eds.), Syntax and semantics, vol. 3: Speech acts (pp. 41-58). New York City: Academic Press. Grosjean, F. (1982). Life with two languages: An introduction to bilingualism. Cambridge, MA: Harvard University Pres. Grosjean, F. (1997). Processing mixed language: Issues, findings, and models. In A. M. de Groot & J. F. Kroll (Eds.), Tutorials in bilingualism (pp. 225-254). Mahwah, NJ: Lawrence Erlbaum. Gumperz, J. J. (1982). Discourse strategies. Cambridge: Cambridge University Press. Gumperz, J. J. (1992). Contextualization and understanding. In A. Duranti & C. Goodwin (Eds.), Rethinking context (pp. 229-252). New York: Cambridge University Press. Gysels, M. (1992). French in urban Lubumbashi Swahili: Codeswitching, borrowing, or both. Journal of Multilingual and Multicultural Development, 13, 41-55. Hankamer, J. (1989). Morphological parsing and the lexicon. In W. D. Marlsen-Wilson (Ed.), Lexical representation and process (pp. 392-408). Cambridge, MA: MIT Press. 23 Issues in CS: Competing Theories and Models Haugen, E. (1953). The Norwegian language in America: A study of bilingual behavior. Philadelphia: University of Pennsylvania Press. Heller, M. (1992). The politics of codeswitching and language choice. Journal of Multilingual and Multicultural Development, 13, 123-142. Heller, M. (1995a). Language choice, social institutions, and symbolic domination. Language in Society, 24, 373-405. Heller, M. (1995b). Code-switching and the politics of language. In L. Milroy & P. Muysken (Eds.), One speaker two languages: Cross-disciplinary perspectives on code-switching (pp. 158-174). New York: Cambridge University Press. Hymes, D. (1967). Models of the interaction of language and social settings. Journal of Social Issues, 23(2), 8-28. Hymes, D. (1972). Models of the interaction of language and social life. In J. J. Gumperz & D. Hymes (Eds.), Directions in sociolinguistics: The ethnography of communication (pp. 35-71). New York: Holt, Rinehart & Winston. Joshi, A. (1985). Processing of sentences with intrasentential code switching. In D. R. Dowty, L. Kattunen, & A. Zwicky (Eds.), Natural language parsing: Psychological, computational and theoretical perspectives (pp. 190-205). Cambridge: Cambridge University Press. Kachru, B. B. (1978). Code-mixing as a communicative strategy in India. In J. E. Alatis (Ed.), Georgetown University Round Table on Languages and Linguistics (pp. 107-124). Washington, DC: Georgetown University Press. Kachru, B. B. (1983). On mixing. In B. Kachru (Ed.), The Indianization of English: The English language in India (pp. 193-207). New Delhi: Oxford University Press. Labov, W. (1966). Social stratification of English in New York City. Washington, DC: Center for Applied Linguistics. Labov, W. (1971). The notion of ‘system’ in Creole languages. In D. Hymes (Ed.), Pidginization and creolization of languages (pp. 447-472). Cambridge: Cambridge University Press. Le Page, R. B., & Tabouret-Keller, A. (1985). Acts of identity: Creole-based approaches to ethnicity and language. Cambridge: Cambridge University Press. Li Wei. (1995). Variations in patterns of language choice and code-switching by three groups of Chinese/English speakers in Newcastle upon Tyne. Multilingua, 14, 297-323. Li Wei. (2000). Dimensions of bilingualism. In Li Wei (Ed.), The bilingualism reader (pp. 3-25). New York: Routledge. Lipski, J. M. (1978). Code-switching and the problem of bilingual competence. In M. Paradis (Ed.), Aspects of bilingualism (pp. 250-264). Columbia, SC: Hornbeam Press. MacSwan, J. (1999). A minimalist approach to intrasentential code switching. New York: Garland. Mann, W. C., & Thompson, S. A. (1986). Relational propositions in discourse. Discourse Processes, 9, 57-90. Meeuwis, M., & Blommaert, J. (1994). The ‘markedness model’ and the absence of society: Remarks on codeswitching. Multilingua, 13, 387-423. Mesthrie, R., Swann, J., Deumert, A., & Leap, W. L. (2000). Introducing sociolinguistics. Philadelphia: John Benjamins Publishing Company. Milroy, L. (1987). Language and social networks. Oxford: Basil Blackwell. Milroy, L., & Wei, L. (1995). A social network approach to code-switching: The example of a bilingual community in Britain. In L. Milroy & P. Muysken (Eds.), One speaker two 24 Issues in CS: Competing Theories and Models languages: Cross-disciplinary perspectives on code-switching (pp. 136-157). New York: Cambridge Univ. Press. Milroy, L., & Muysken, P. (1995). Introduction: Code-switching and bilingualism research. In L. Milroy & P. Muysken (Eds.), One speaker two languages: Cross-disciplinary perspectives on code-switching (pp. 1-14). New York: Cambridge University Press. Muysken, P. (1995). Code-switching and grammatical theory. In L. Milroy & P. Muysken (Eds.), One speaker two languages: Cross-disciplinary perspectives on code-switching (pp. 177-198). New York: Cambridge University Press. Muysken, P. (2000). Bilingual speech: A typology of code-mixing. Cambridge: Cambridge University Press. Myers-Scotton, C. (1980). Explaining linguistic choices as identity negotiations. In H. Giles, W. P. Robinson, & P. M. Smith (Eds.), Language: Social psychological perspectives (pp. 359-366). New York: Pergamon Press. Myers-Scotton, C. (1983). The negotiation of identities in conversation: A theory of markedness and code choice. International Journal of the Sociology of Language, 44, 115-136. Myers-Scotton, C. (1987). Code-switching and types of multilingual communities. In P. Lowenberg (Ed.), Georgetown University Roundtable on Languages and Linguistics (pp. 61-82). Washington, D.C.: Georgetown University Press. Myers-Scotton, C. (1992). Comparing codeswitching and borrowing. Journal of Multilingual and Multicultural Development, 13, 19-39. Myers-Scotton, C. (1993a). Duelling languages: Grammatical structure in codeswitching. New York: Oxford University Press. Myers-Scotton, C. (1993b). Social motivations for codeswitching: Evidence from Africa. Oxford: Oxford University Press. Myers-Scotton, C. (1993c). Common and uncommon ground: Social and structural factors in codeswitching. Language in Society, 22, 475-503. Myers-Scotton, C. (1998). A theoretical introduction to the markedness model. In C. Myers-Scotton (Ed.) Codes and consequences: Choosing linguistic varieties (pp. 18-38). New York: Oxford University Press. Myers-Scotton, C. (1999). Explaining the role of norms and rationality in codeswitching. Journal of Pragmatics, 32, 1259-1271. Myers-Scotton, C., & Bolonyai, A. (2001). Calculating speakers: Codeswitching in a rational choice model. Language in Society, 30, 1-28. National Center for Education Statistics. (2002). Overview of public elementary and secondary schools and districts: School year 2000-01. Retrieved 17 November 2002 from http://nces.ed.gov/pubsearch/pub sinfo.asp?pubid=2002356 New York State Education Department. (2002). New York, The state of learning: Statewide profile of the educational system. Retrieved 17 November 2002 from http://www.emsc.nysed.gov/irts/ch655_200 2/home.html Nishimura, M. (1986). Intra-sentential code-switching: The case of language assignment. In J. Vaid (Ed.), Language processing in bilinguals: Psycholinguistic and neuropsychological perspectives (pp. 123-143). Hillsdale, NJ: Lawrence Erlbaum. Pfaff, C. (1979). Constraints on language-mixing: Intrasentential code-switching and borrowing in Spanish/English. Language, 55, 291-318. 25 Issues in CS: Competing Theories and Models Pfaff, C. (1991). Turkish in contact with German: language maintenance and loss among immigrant children in Berlin. International Journal of the Sociology of Language, 90, 97-129. Poplack, S. (1978). Dialect acquisition among Puerto Rican bilinguals. Language in Society, 7, 89-103. Poplack, S. (1980). Sometimes I’ll start a sentence in English y termino en español: Toward a typology of code-switching. Linguistics, 18, 581-618. Poplack, S. (1981). Syntactic structure and social function of codeswitching. In R. P. Durán (Ed.) Latino language and communicative behavior (pp. 169-184). Norwood, NJ: Ablex. Poplack, S. (1993). Variation theory and language contact. In D. Preston (Ed.), American dialect research: An anthology celebrating the 100th anniversary of the American Dialect Society (pp. 251-286). Amsterdam: Benjamins. Poplack, S., Wheeler, S., & Westwood, A. (1987). Distinguishing language contact phenomena: Evidence from Finnish-English bilingualism. In P. Lilius & M. Saari (Eds.), The Nordic languages and modern linguistics 6 (pp. 22-56). Helsinki: University of Helsinki Press. Queen, R. (2003). Language ideology and political economy among Turkish-German bilinguals in Germany. In B. Joseph, J. DeStefano, N. Jacobs, & I. Lehiste (Eds.), When languages collide: Perspectives on language conflict, language competition, and language coexistence (pp. 201-221). Columbus, OH: Ohio State University Press. Ramirez, A. G., & Milk, R. D. (1986). Notions of grammaticality among teachers of bilingual pupils. TESOL Quarterly, 20, 495-513. Romaine, S. (1995). Bilingualism. Oxford: Blackwell. Sankoff, D., & Poplack, S. (1981). A formal grammar for code-switching. Papers in Linguistics, 14, 3-45. Sankoff, D., Poplack, S., & Vanniarajan, S. (1990). The case of the nonce loan in Tamil. Language Variation and Change, 2, 71-101. Saville-Troike, M. (1982). The ethnography of communication: An introduction. Oxford: Blackwell. Sebba, M. (1993). London Jamaican. Harlow: Longman. Singh, R. (1985). Grammatical constraints on code-switching: Evidence from Hindi-English. Canadian Journal of Linguistics, 30, 33-45. Sperber, D., & Wilson, D. (1986). Relevance: Communication and cognition. Cambridge, MA: Harvard University Press. Sridhar, S. N., & Sridhar, K. (1980). The syntax and psycholinguistics of bilingual code-mixing. Canadian Journal of Psychology, 34, 407-416. Tabouret-Keller, A. (1983). Comment. International Journal of the Sociology of Language, 39, 139-148. Tajfel, H. (1974). Social identity and intergroup behavior. Social Science Information, 13, 65-93. Thompson, J. B. (1991). Editor’s introduction. In J. B. Thompson (Ed.), Language and symbolic power (pp. 1-31). Cambridge, MA: Harvard University Press. Treffers-Daller, J. (1994). Mixing two languages: French-Dutch contact in a comparative perspective. Berlin: Mouton de Gruyter. Trudgill, P. (1974). The social differentiation of English in Norwich. Cambridge: Cambridge University Press. 26 Issues in CS: Competing Theories and Models Weinreich, U. (1968). Languages in contact. The Hague: Mouton. Weinreich, U., Labov, W., & Herzog, M. (1968). Empirical foundations for a theory of language change. In W. Lehmann (Ed.), Directions in historical linguistics (pp. 97-195). Austin: University of Texas Press. 27