The development of bilingual proficiency--a sequential analysis.

Abstract

Codeswitching is in the following analyzed as an indication of bilingual proficiency level. By applying Auer's sequential approach to codeswitching data from a longitudinal study of a Danish-Turkish bilingual setting I establish an (implicational) hierarchy of codeswitching types. The hierarchical relations between the different types are correlated with the monolingual aspect of the informants' integrated bilingual competencies. The correlation between the development of codeswitching and the development of L2 Danish paves the way for establishing a needed method of determining relative bilingual competence. I show that the way in which the codeswitching is utilized by the informants is finely attuned to the individual's linguistic proficiency.

Key words

codeswitching implicational scale

development of codeswitching

Turkish-Danish

1 Auer in action

Bilingual proficiency entails the ability to maintain, and to switch between two available codes. The full understanding of this intriguing skill still keeps researchers busy, and without doubt codeswitching contributes to keeping bilingualism interesting.

But bilinguals do not always utilize this means of communication, which is often considered a breach of appropriate communicative behavior. Furthermore, codeswitching is an acquired skill, the use of which may diminish with adulthood. There are certain indications that at least some aspects of codeswitching are strongly connected with teenage in a multilingual environment. Thus codeswitching may signal a phase which ends as the teenager steps in to adulthood.

It is by no means generally accepted in society at large that codeswitching is positively related to overall linguistic competence. In this paper I will therefore study how codeswitching develops among Turkish-Danish grade school children from the age of seven through the age of 15, and how this ability is related to their development of Danish as a second language.

2 Developing codeswitching--when and how

In the following I study the different aspects of bilingual verbal behavior, that is, codeswitching, and their relations to overall linguistic proficiency, represented by the acquisition of Danish. As an outcome I propose a method of measuring bilingual proficiency based on truly bilingual premises.

In order to study general bilingual competence I have analyzed data from all nine years of the Koge-Project (see introduction to this volume), to see how codeswitching develops in parallel to the L2 Danish aspect, over the students' nine years of grade school.

My analysis is based on Auer's sequential approach (Auer, 1984a, 1995, 1998, 1999) and the patterns of codeswitching which he proposes, in order to get a clearer picture of how codeswitching develops.

The development of codeswitching behavior has been studied by Holmen & Jorgensen (1997, 2000), who put forward a sketchy developmental path for the bilinguals' use of codeswitching for communicative (pragmatic) purposes, compared with their overall linguistic competence. But I am able to detail this developmental path even further and distinguish between at least three stages of codeswitching behavior.

My informants were 10 Turkish-Danish bilinguals whose group-conversations were tape-recorded each year during the nine-year period. The informants were equipped with individual microphones taping each individual's contribution to the conversations.

3 In search of bilingual proficiency

My working hypothesis is that there is a connection between speakers' overall linguistic ability (competence) and their different uses of codeswitching. Codeswitching enters the linguistic repertoire at some point at which it seems to be frequent, and at a later stage it seems to become less frequent. Here we shall look at what happens in between.

It is clear from the outset that the way bilinguals utilize their two available languages changes over time, and that time in the sociolinguistic environment in Koge equals progression in overall linguistic ability.

From a linguistic point of view no really useful way of determining and describing levels of bilingual proficiency exists, and the methods proposed all have the same flaw: they are based on a monolingual concept of bilingualism. Researchers have for the past 10-15 years reminded us that bilinguals "are not two monolinguals in one person" (Grosjean, 1995, p. 259), and this has led to a theoretical dogma, which shows up in many disguises. Li Wei puts it like this:

 
   It is important to emphasize that bilingual speakers have a unique 
   and psychological profile; their two languages are constantly in 
   different states of activation: they are able to call upon their 
   linguistic knowledge and resources according to the context and 
   adapt their behavior to the task in hand. 
 
   Li Wei (2000, p. 17) 

The traditional way of measuring or determining a relative bilingual proficiency is through a monolingual test in each of the two available languages. In my opinion this approach misses the point. It is true that the bilingual can choose to act monolingually by choosing to speak in only one of the two available tongues. This does to some degree justify traditional thinking, but still leaves aside a vital and characteristic aspect of bilingual verbal activity--namely codeswitching and codemixing. And working with codeswitching data one soon realizes that most of the traditional theoretical linguistic machinery prevents progress in this respect, and nontraditionally based ideas need to be tested.

I therefore suggest codeswitching as our point of reference for determining the totality of the different aspects of bilingual proficiency. Bilingualism means controlling two languages and using all of the linguistic know-how you have acquired (are in possession of) as an integrated whole and to use it appropriately for the when and why, including the conversational partners' linguistic competence (see Jorgensen, 1998a, p. 142). Based on this assumption it is possible to connect the different aspects of bilingual linguistic competence. In this view, bilingualism is an integrated skill, not two sets of monolingual skills who happen to cohabit the same speaker.

My goal is to render it probable that specific levels of codeswitching behavior are correlated with specific levels of monolingual proficiency. For this purpose, it is necessary to compare the way in which my informants develop their codeswitching, with the way their monolingual L1 Turkish and their monolingual L2 Danish verbal behaviors develop.

What the integrated concept of bilingualism entails, is that its development can be described along an implicational hierarchy of codeswitching types or functions, and overall proficiency. The main inspiration comes from Comrie (1996):

 
   ... The intuition that underlies the discussion of the present 
   section is a very simple one: the hierarchy subject > direct 
   object > nondirect object > possessor defines the ease of 
   accessibility of relative clause formation, that is, it is in 
   some intuitive sense, easier to relativize subjects than it is 
   to possessors etc.... 
 
   If a language can form relative clauses on a given position on 
   the hierarchy, then it can also form relative clauses on all 
   positions higher (to the left) on the hierarchy... 
 
   Comrie (1996, p. 148) 

I propose the same type of hierarchy for bilingual proficiency. With its different developmental stages of codeswitching, the hierarchy provides a short-cut to determine the overall bilingual proficiency, but also to proficiency in both monolingual aspects. The different stages refer to certain codeswitching types or configurations.

Type 1 > > Type 2 > ... > Type X

Each codeswitching type (type 1, type 2, etc.) reflects a level of proficiency. Movement up the hierarchy happens in accordance with added proficiency. The stages are implicationally connected and really a part of bilingual proficiency. Once you are able to make use of a certain aspect of the hierarchy it is implied that you are also able to work at all lower positions as well.

4 The sequential approach--an outline

Auer's sequential approach has a CA framework as its methodological instrument, whereby verbal interaction is prepared for microscopic study, particularly of the dialogic construction of meaning. Auer (1984a) came up with the sequential approach to codeswitching as a reaction against the semantic approach put forward by Gumperz (1982). Gumperz' distinction between situational and metaphorical codeswitching was at the time seen as a breakthrough in codeswitching research. But contrary to Gumperz' static notion of context Auer defines context as a dynamic and negotiable part of a sociolinguistic situation. According to Auer, language choices taking place in bilingual conversation "are not determined by situational parameters." The choice of one language over the other "is part of a complicated business of defining the situation" (Auer, 1984a, p. 4). Auer's redefinition of context means that context is no longer available for categorizing codeswitching--context is no longer a global phenomenon--but a local product.

Prior to the actual sequential patterning, Auer chooses an interactional approach to the overall analysis of a given conversation, in order to understand codeswitching as a resource for linguistic communication. Only by taking this stand as a starting point, is it possible to reach a plausible analysis, according to Auer. The understanding of the social processes taking place in and during (a) conversation (monolingual or bilingual) is a prerequisite for ascribing meaningful sense to specific instances of codeswitching. Only a sequential approach can achieve this:

 
   ... it will hopefully become clear that any theory of 
   conversational codeswitching will fail if it does not 
   take into account that the meaning of code-alternation 
   depends in essential ways on its "sequential environment." 
 
   (Auer, 1995, p. 116) 

The kind of social processes Auer talks about in relation to the occurrence of codeswitching, is found in the work of several researchers. Jergensen (1998b) found that codeswitching is (and could be) used for "linguistic power-wielding" in conversations between bilingual youngsters in a Danish-Turkish environment. Cromdal (2000) on the other hand, has shown how codeswitching works as the focal point in organizing a group-task. The group studied followed a more or less unconscious division of labor between their two available languages in relation to the …