Code-switching

Definition
The term codeswitching (or code-switching) refers to the alternation between two or more languages, dialects, or language registers in the course of discourse between people who have more than one language in common. Typically one of the two languages is dominant; the major language is often called the matrix language, while the minor language is the embedded language.


 * (Code-switching) "occurs when a bilingual introduces a completely unassimilated word from another language into his speech." (Haugen 1956:40)
 * "Codeswitching ... is the selection by bilinguals or multilinguals of forms from an embedded variety (or varieties) in utterances of a matrix variety during the same conversation" (Myers-Scotton 1993:3).

Subtypes

 * intersentential codeswitching
 * intrasentential codeswitching

Examples

 * German-Spanish
 * Aber du #sabes que eres, o sea, tú sabes que# du bist gut, zum Beispiel besser in Mathe als in Deutsch. #Eso sabe.
 * Aber du weißt, dass du bist, also, du weißt, dass du  bist gut,  zum Beispiel  besser  in Mathe  als in Deutsch. Das weiß er.
 * English-German
 * Nein er kam doch erst um neun, #you know?
 * Nein er kam doch erst um neun, weißt du?

The first example (from Edel 2007) shows, in its first sentence, a case of intrasentential codeswitching, where the switch occurs between words or phrases (it may also occur within the same word). Both sentences together provide an example of intersentential codeswitching, where the switch occurs between sentence boundaries. The second example illustrated a third (and rarely noted) form of codeswitching, the so called 'tag switches', where a tag phrase from the second language is embedded into the matrix language.

Code-switching vs. borrowing
The phenomenon is to be differentiated from borrowing. In general, the easiest way to find out whether a word or phrase is a borrowing in a given language is to determine whether the word is used by monolinguals of that language or not: Code-switches are only used by bilingual speakers. Borrowings, by contrast, are regularly used by monolinguals of a given language and that they have, to an extent at least, been adapted to the phonological system of the recipient language. However, as some linguists assume that the process of integration happens gradually, some forms of borrowings cannot be easily distinguished from code-switches. A more or less reliable test mentioned in the relevant literature relies on frequency (borrowings occur much more frequent than code switches).

Code-switching and code-mixing
The terms code-switching and 'code mixing' are oftentimes used synonymously, though 'code-mixing' is often used for intrasentential codeswitching only.

Matrix language and embedded language
In situations of code-switching a distinction is made between the matrix language and the embedded language. The matrix language is the language that is dominantly used during a conversation and whose grammar is - in most cases - applied to the overall sentence structure. The embedded language is the one from which switches may originate. Some models of code-switching models supply tests allowing one to identify the matrix language. So far, no reliable test has however been found. Models like the one of Muysken (2000) even consider the differentiation between matrix and embedded language unnecessary.

History of code switching
Code-switching has been known since the early twentieth century, when the first recognizable observations concerning bilingual research were recorded (Ronjat 1913, and later Leopold 1939-49). However, the phenomenon was not investigated for a long time. In the entire first half of the twentieth century and in large parts of the second half, code-switching was considered something that occurred randomly, without a logical pattern behind it, as a result of imperfect second language learning.

The perspective on code switching began to change in the nineteen seventies when Blom and Grumperz (1972) published an article in which they presented a survey of their studies of a Norwegian village. Blom and Grumperz discovered that members of the village spoke two dialects of Norwegian and used them according to specific situations. In the following years, more scholars conducted research on the systematic character of code-switching. From the late seventies on, there has been a lively debate going on, producing various models predicting (constraints on) code-switching.

Types of code switching and reasons for code switching
One of the first categorisations of code-switching was provided by Appel and Muysken (1987). They distinguish between five reasons for why speakers code-switch:


 * Referential function
 * Directive function
 * Expressive function
 * Phatic function
 * Metalinguistic function

Speakers will use the referential function of code-switching to compensate for shortcomings in the matrix language. This may either make up for lexical gaps in the matrix language, or help the speaker to maintaining a smooth speech flow. The directive function refers to a situation in which a speaker either wants to associate with, or dissociate themselves from other interlocutors. The phatic function signals a change in 'tone'. The metalinguistic function occurs when speakers comment on a specific feature of a language by using the other language.

Although this model covers a number of functions, it cannot really answer the question of why speakers use code switching. In their study on Ranamål and Bokmål -- the two variants spoken in the mentioned example in northern Norway -- Blom and Gumperz argue that there are mainly two functions of code switching: Situational and metaphorical code switching.

The question of why speakers switch codes was addressed by psychologists like Howard Giles (1973) who took an audience centred approach. Giles stated that speakers code-switch in order to either a.) Associate with the interlocutor or b.) Dissociate from the interlocutor.

The underlying assumption is that an audience will evaluate a speaker more beneficially who tries to put effort into showing his ties to an audience, while they will disapprove of those who distance themselves from the audience.

Finally, there are also pragmatic approaches to code-switching.

Structural models of code switching
Four mayor approaches to the structural study of code-switching canb be distinguished:


 * Descriptive accounts (Timm 1975, Pfaff 1979)
 * Accounts involving surface constraints and a 'third grammar' for code-switching (Poplack 1980; Sankoff and Poplack 1981'')
 * Principle-based accounts with code-switching specific mechanisms and constraints (Belazi et al. 1994; Bentahila and Davies 1983; Di Sciullo et al. 1986; Joshi 1985; Woolford 1983'')
 * Principle-based accounts without code-switching specific mechanisms and constraints (Mahootian 1993, 1996a, 1996b; Myers-Scotton 1993)

Descriptive accounts can be regarded as reactions to the assumption that code-switching occurs randomly and unpredictably. Timm and Pfaff (1979) were able to show that there are specific rules restricting code-switching, e.g. the constraint disallowing switches that would violate the surface order of either the matrix language or the embedded language. The 'three grammar approach' states that bilingual speakers must not only inherit one grammar for each language, but build a third, code-switching grammar as well. There is a wide range of principle-based approaches, for example Di Scullio et al.'s (1986) 'government constraint', which states that elements in a government relation have to be from the same language. Finally, there are principle-based accounts that assume no special mechanisms and that rely on general principles of sentence structure, rather than code-switching specific rules, as for example the 'head-compliment principle model'.

As mentioned earlier, each of these models has been subject to criticism. For example, some models have been claimed not to be able to predict code-switches; other models may allow for some such predictions, but have deficiencies on other parts. The only model that has not been severely criticized is that of Muysken (2000), which tries to incorporate insights gained in all previous models, with a stronger focus on sociolinguistic than on structural aspects.

Origin
The term "code-switching" first appeared in Hans Vogt's (1954) review of Uriel Weinreich's Languages in Contact (1953). Weinreich had used the phrase "switching codes," apparently borrowed from information theory (e.g. Fano 1950). Roman Jakobson (e.g. Jakobson 1953) and Einar Haugen (e.g. Haugen 1956) were among the earliest linguists to develop the notion.

Other languages
German Sprachwechsel, Codewechsel, Codeswitching (de)