Code-switching. Grammatical, pragmatic and psycholinguistic aspects

Table of Contents

Introduction

Chapter I. Grammatical/syntactic aspects of code-switching
1. Definitions of code-switching
2. Code-switching vs. borrowing and code-mixing
3. Types of code-switching
4. Grammatical constraints on code-switching
4.1. Types of grammatical constraints: general/universal vs. relativized/probabalistic
4.2. Constraints based on linearity (Poplack)
4.3. Constraints based on dependency (theory of government)
4.4. Constraints considering assymetry of two languages (Matrix Language Frame model by Myers-Scotton)

Chapter II. Pragmatic/discourse aspects of code-switching
1. Functions of code-switching (the functional model by Jakobson)
2. Discourse functions of code-switching (Gumperz)
3. Transactional vs. metaphorical code-switching (Gumperz)
4. Concept of the social arena (Scotton and Ury)
5. Attitudes towards code-switching

Chapter III. Psycholinguistic aspects of code-switching
1. Acquisition of two languages
1.1. Fusion vs. separation of two languages
1.2. Acquisition of the pragmatic functions of code-switching
1.3. Syntactic apects of code-switching in the language acquisition
2. Bilingual competence

Conclusion

References

Introduction

The aim of this paper is to provide a complete overview over the phenomenon of code-switching. In this paper, we will summarize the knowledge currently available on the discourse, linguistic and psycholinguistic aspects of code-switching. Such an overview can be used for different purposes: for seminar reports and papers, and for the preparation for exams in linguistics.

The history of the research of code-switching has undergone various periods that have shown how complex the phenomenon of code-switching is. In the course of research of code-switching it has become clear that code-switching can be investigated from different perspectives. Researchers focussed on code-switching after they had realised that linguistic forms and practices are interrelated. And code-switching, in its turn, embodies not only variation, but the link between linguistic form and language use as social practice. Research from a linguistic and psycholinguistic perspective has focussed on understanding the nature of the systematicity of code-switching, as a way of revealing linguistic and potentially cognitive processes. Research on the psychological and social dimensions of code-switching has largely been devoted to answering the questions of why speakers code-switch and what the social meaning of code-switching is for them. The sociological perspective later goes on to attempt to use the answer to those questions to illuminate how language operates as a social process. Throughout the history of research on code-switching it has been proposed that it is necessary to link all these forms of analysis and that, indeed, it is that possibility that is one of the most compelling reasons for studying code-switching, since such a link would permit the development and verification of hypotheses regarding the relationship among linguistic, cognitive and social processes in a more general way (Heller, Pfaff 1996).

The first chapter of this paper deals with grammatical/syntactic aspects of code-switching. In this chapter, we will give various definitions of code-switching. Since code-switching has some similarities with borrowing and code-mixing, which causes difficulty in precise defining of code-switching, we will provide criteria for keeping these language contact phenomena apart. Secondly, we will define the three types of code-switching: tag switching, intrasentential and intersentential code-switching. In the rest of this chapter, we will take a close look at grammatical constraints on code-switching and show that there exist different approaches to formulating formal, functional principles of code-switching. The best known approach in this concern is represented by Polack’s (1988) equivalence constraint and free morpheme constraint, which are based on the principle of linearity. Further, formal aspects of code-switching based on the notion of government will be illuminated. Since the constraints based on the principle of government don’t account for all cases of code-switching, we will present the notion of L-marking, which is a revised form of the government constraints, as another principle for formulating constraints on code-switching. At the end of this chapter, the Matrix Language Frame model by Myers-Scotton (1992), which is considered to be the most successful approach to defining grammatical constraints on code-switching, will be introduced.

In the second chapter, we will deal with pragmatic/discourse aspects of code-switching. In respect to the discourse functions of code-switching one can rely of the following models for identifying the role of code-switching in bilingual speech: the functional model by Jakobson and Halliday (Appel and Muysken, 1987), which contains a referential, directive, expressive and phatic function; the discourse functions as classified by Gumperz (1982), which include quatations, reiteration, message qualification, personalization vs. objectivization, interjections, addressee specification. Firstly, Gumperz’s (1982) classification of the discourse functions of code-switching is the one most frequetly referred to. Secondly, Gumpers suggested distinguishing between transactional and metaphorical code-switching, which is seen by Gumperz to be helpful in identifying the discourse functions of code-switching. On the other hand, by referring to the concept of social arena by Sotton and Ury (Romaine 1989), we will show that the distinction between transactional and metaphorical code-switching is of limited usefulness. The reason for that lies in the great variability of the meanings of code-switched utterances. A simple distinction transactional and metaphorical code-switching does not reflect the complexity of the discourse functions of code-switching. By summarizing the results of different approaches to identifying the discourse functions of code-switching, we will point out the possible pragmatic meanings of code-switching and show the points of difficulty in distinguishing between the discourse functions mentioned.

Extensive research over the last twenty years on language contact phenomena and especially in the field of code-switching has led investigators to agree that bilingual code-switching is governed by certain constraints. However, the development of such formal and functional principles during the course of the linguistic development of bilingual children has not been studied to a larger extent. As far as the acquisition of grammatical constraints is concerned, linguists are facing theoretical problems. The fact that linguists know little about how and when exactly a bilingual child discovers that the use of both languages in its environment is governed by social and pragmatic rules presents one more difficulty. One thing is clear: one should take into account the linguistic development of the child in both languages when examining the acquisition of grammatical constraints on code-switching (Heller Pfaff 1996). We will present some findings of researches on children's acquisition of bilingual speech. Another point of our concern will be the influence of the average bilingual competence on the patterns of code-switching used by bilinguals.

Chapter I. Grammatical/syntactic aspects of code-switching

1. Definitions of code-switching

There have been numerous attempts to provide a precise linguistic characterization of conversational code-switching, especially intra-sentential code-switching. Definition of code-switching has proved to be problematic because it is mostly found in conjunction with borrowing, interference and convergence (Heller, Pfaff 1996). The definition of code-switching in contrast to these other types of language contact phenomena is confounded by the facts that the "codes" themselves are highly variable, are undergoing change, and are often considered "non-standard" particularly when bilinguals are not proficient in the "standard" codes.

In most general terms, code-switching can be defined as the mixing of elements of two linguistic varieties within a single utterance or text. Various authors offer different definitions of code-switching, which accentuate different perspectives of code-switching. The definitions range from based on purely linguistic aspects of code-switching to those based mostly on social aspects of this linguistic phenomenon.

Poplack (1980) defines code-switching as the alternation of two languages within a single discourse, sentence or constituent, which in balanced bilinguals is governed by both extra-linguistic and linguistic factors. Sankoff and Poplack (1981) distinguish code-switching from other possible results of language contact, such as: interference, pidginization, borrowing, calquing, language death, learned use of foreign words, cross-language punning by two criteria:

1. the above mentioned phenomena involve the alteration or replacement of parts of the grammar or lexicon of the language(s) involved, while code-switching does not;
2. and the above mentioned phenomena refer to specialized situations or language functions, while the code-switching is a" widely operative norm of communication in certain types of bilingual communities" (Sankoff and Poplack 1981).

Bokamba (1986) offered a different definition of the terms code-switching and code-mixing: code-switching is the embedding or mixing of words. phrases, and sentences from two codes within the same speech event and across sentence boundaries, while code-mixing is the embedding or mixing of various linguistic units, i.e., affixes, words, phrases, and clauses from two distinct grammatical systems or subsystems within the same sentence and the same speech situation.

Gumperz (1976/1982), for instance, offers a more socially-based definition of code-switching. He defines code-switching as "the juxtaposition within the same speech exchange of passages of speech belonging to two different grammatical systems or subsystems". He also suggests that the minority language is used as the "we code" for in-group relations, while the official language is used as "they code" with out-of-group participants. In addition to it, he defines the concepts of loans and borrowings, which are to be distinguished from code-switching: loans tend to be phonologically, morphologically and syntactically adapted to the rules of the host language. Presumably they have been present in the host language for longer periods of time, while borrowing represents newer innovations to the host language. Gumperz (1976/1982) states: "Bilingual speech is highly receptive to loans. Many items in general use in bilingual communities are known or unacceptable in the monolingual home regions."

In terms of surface structure, code-switching may involve linguistic units of different sizes: whole sentences, clauses, phrases, single words, tags, interjections and formulaic expressions. The frequency of each of these types of structure varies widely across speakers and situations.

2. Code-switching vs. borrowing and code-mixing

The distinction between borrowing, code-switching and code-mixing can be made from two different viewpoints: from the point of view of language competence and from the perspective of grammaticality.

Thus, from the point of view of language competence, it has been argued that these types of phenomena can be linked with different types of competence and degree of fluency in the two languages. Borrowing can occur in the speech of those with only monolingual competence. Whereas code-switching and code-mixing imply some degree of competence in the two languages, although the precise relation between competence and types of code-switching is disputed.

The most precise distinction between borrowing and lexical code-switches from the perspective of grammar is provided by Myers-Scotton (1992). According to Myers-Scotton, borrowings and code-switches can be distinguished with help of the following criteria:

1. structural integration, which can be divided into phonological and syntactic integration. (phonologically and syntactically integrated items are treated as borrowings);
2. frequency of occurrence (frequently used lexemes are more likely to be borrowings than those that are used rarely);
3. membership in matrix lexicon.
The distinction between code-switching and code-mixing seems to be more problematic than the distinction between code-switching and borrowing. The greatest difficulty in this respect is to separate code-switching from code-mixing in the cases of intrasentential code-switching, i.e. when the codes are switched within one clause.

Researches noticed that code-switching cannot occur at any point of the sentence. Interesting about this is the fact that, even where the languages differ syntactically, the intrasentential code-switching tends to be regular and gives the impression of fluency of bilingual speech. Such observations led to the supposition about the existence of grammatical constraints on code-switching. Bilingual speech is fluent in case when a bilingual speaker follows these constraints and it is not fluent if a bilingual does not switch grammatically. That is why researchers agree on the following way of keeping code-switching apart from code-mixing: the notion of “code-switching” is used for clean grammatical switching, which does not violate the rule of either languages, whereas the notion “code-mixing” is used for sequences which violate the grammar rules of one or both languages. In other words, the bilingual speaker, when using both languages in the same utterance or conversation, violates syntactic or pragmatic constraints on code-switching established by the language use of the bilingual community.

3. Types of code-switching

There are three major types of code-switching:

1. tag-switching, in which tags, exclamations and certain set phrases in one language are inserted into utterance in the other language. The tags etc. serve as an emblem of the bilingual character of an otherwise monolingual sentence. That is why Poplack (1980) has named this type of switching emblematic switching.
2. inter-sentential code-switching, in which a change of language occurs at a clause or sentence boundary, where each clause or sentence is in one language or the other. In other words it is a switching at a sentence level. Inter-sentential code-switching may serve to emphasize a point made in the other language, signal a switch in the conversation participants, indicate to whom the statement is addressed, or to provide a direct quote from, or reference to, another conversation.
3. intra-sentential code-switching, in which switches occur within a clause boundary. In other words, intra-sentential code-switching represents switching at the clause, phrase level, or at word level if no morphological adaptation occurs and the mentioned above criteria for the distinguishing code-switches and borrowings are observed.

4. Grammatical constraints on code-switching

4.1. Types of grammatical constraints: general/universal vs. relativized/probabalistic

All of the grammatical constraints on code-switching can be classified in two groups: general or universan and relativized or probabilistic contraints.

General/universal constraints are supposed to hold for the majority of cases. These constraints on code-switching describe which switches are theoretically grammatical, in other words, which switches are allowed.

Relativized/probabilistic contraints do not explore which switches are allowed, but rather attempt to establish which switches are more frequent ones. In other words, the probabilistic approach does not aim at describing all of the possible points of switching. It rather explores only the real, objectively observed cases of code-switching and makes conclusions on constraints in accordance with the frequency of switches.

Muysken (1995) argues that absolute constraints, that can be invalidated by as few as one counterexample, are less appropriate for performance data, particularly data which arise from quite complex factors, not all of which are always under control. Just making a general statement about which type of switch is not likely to occur misses the point that some types of code-switching are less frequent than others, within a given corpus.

Muysken (1995) points out that it is as important to consider the non-occurring switches as the ones that do occur. Muysken argues as well that, since we do not know how the grammar and the lexicon interact with other psychological faculties to produce actual speech, we clearly cannot ignore phenomena such as frequency of ocurrence and regularity. Two complications arise, however. Firstly, frequency of swithing may result from the conventionalisation of a certain type of switch, rather than from a cinical grammatical factor. Secondly, we do not yet know enough about the relation between frequency distributions of specific grammatical patterns in monolingual speech data and properties of the grammar to handle frequency in bilingual data with any assuarance.

The universal constraints proposed in the literature cluster around two fundamental grammatical and psycholinguistic concepts: linearity and dependency. In the following, we will present the both approaches to formulating grammatical constraints on code-switching.

4.2. Constraints based on linearity (Poplack)

The focus of interest in theoretical linguistic circles that study the phenomenon of code-switching has been the establishment of universal linguistic constraints that govern this type of behaviour in bilingual communities.

In the recent studies of code-switching, researchers agreed on the fact that, even where languages differ syntactically, the intra-sentential switching tends to be regular and to give impression of fluency rather than disfluency of bilingual speech. Many instances of code-switching seem to be grammatical in both languages, either because they take place at sentence or clause boundaries where there are no morphosyntactic dependencies or because they involve single words or fixed phrases in structures which are not problematic (the later are excluded as borrowings by some researchers). That is why all of the proposed constraints deal with intra-sentential code-switching, since intra-sentential code-switching is the only type that involves the interaction of the grammars of the two languages (Heller, Pfaff 1996). The first explicit statement of the principle of linearity can be found in Poplack’s observation (1980): “Code-switching will tend to occur at points in discourse where juxtaposition of first language and second language elements does not violate a syntactic rule of either language, i.e. at points around which the surface structurea of the two langauges map onto each other”.

Poplack (1988) has established two constraints on intra-sentential code-switching which seem to be part of the mental grammar of the bilingual speakers in her study. These constraints were derived from Poplack's and other's observations of code-switching behaviour, and are, therefore descriptive, not prescriptive. They serve to predict points where intra-sentential code-switching might occur, rather than to prohibit bilingual speakers from switching languages at a particular part of the sentence. And, particularly in the case of the free-morpheme constraint, they do not predict accurately all of the time; they may be violated occasionally even by the most proficient bilingual speakers. Poplack (1988) proposed the following constraints:

The equivalence constraint"which requires that the surface word order of the two languages be homologous in the vicinity of the switch point". According to this constraint, word order immediately before and after a switch point must be grammatically possible in both languages.

The free morpheme constraint" which prohibits mixing morphologies within the confines of the word". According to this constraint, no switches are allowed between stem and affix, and few within idiomatic expressions and set phrases. Poplack (1988) observed that, while idiomatic expressions are often considered to behave like bound morphemes, and therefore tend not to be switched, a small number of switches within idiomatic expressions occurred in the speech of her respondents.

This approach has been found to be less productive than originally hoped. Such apparent "rules" governing switch-points as had been observed in some communities were found not to hold in others (Bentahila and Davies 1983; Berk-Seligson 1986; Romaine 1986). Romaine has concluded from a study of Panjabi-English switching that the principal constraint on occurrence of code-switching, Poplack's "equivalence constraint will fail in any case where two typologically different languages are mixed". The studies on Spanish/Hebrew code-switching by Berk-Seligson (1986), and French/Arabic code-switching by Bentahlia and Davies (1983) have also proved that in pairs of genetically unrelated languages neither of the two constraints is valid. Berk-Seligson (1986) reported different types of violation of the equivalence constraint: the omission of the Spanish determiners, copulas, and prepositions and prepositional phrases are just a few examples of them.

[...]