Language dominance in early and late bilinguals
Una Cunningham, Högskolan Dalarna
Introduction
Bilingual speakers are defined in many ways in the literature, but there is a consensus that bilingual speakers are more or less proficient in more than one language. Most bilinguals are, however, not equally proficient in both of their languages. In addition, the languages are used for different purposes, so
speakers may be have a larger vocabulary and command of the registers in one domain in one language and in another domain in another language. It has also been suggested that the language dominance of an individual is subject to change as the need to use each language changes
Measures of bilingualism
The issue of language dominance in bilingual situations is one which occupies many linguists. This is related to the proficiency of the bilingual in each language. Bilingual speakers can be compared to monolingual speakers of each of their languages (Martin-Jones & Romaine, 1986). The model of the two containers which can both be full, representing native(-like) mastery of both languages, or one full and one half full, representing native-like
competence in one language and less than native-like competence in the other
language. It has, however, been pointed out that this does not present their
skills in a fair or useful way since the proficiency of a bilingual speaker is
essentially different than that of a monolingual (e.g. Grosjean, 1989). This
model also allows for the politically very sensitive issue of what called semi- lingualism. Other researchers, e.g. Skutnabb-Kangas (1984), have claimed that semilingualism is an artefact of the testing system and a reflection of negative attitudes to bilingualism. It is, however, difficult to find any standard other than monolingual speakers to measure by.
One system or two
On the syntactic level, it has been suggested that a child growing up with two languages will move from a single system with lexical material from both languages to having two separate systems. Do children growing up start off with one phonological system and gradually separate the systems of their languages? Do bilingual speakers observe more subtle phonetic differences between their languages? What about the phonotactic systems? At what point can it be said that these systems are separate?
Major (1990) reported a study of American women who had lived for many years in Brazil. He studied the women’s production of initial consonants, in particular the timing of the onset of voicing in such consonants which works differently in English and Portuguese. He found that some of the women approached native-like values in their production of Portuguese voice-onset time (VOT), while others had retained English-like VOT in Portuguese. Those women who had acquired Portuguese-like VOT were found to use it even in their English utterances. This led Major to suggest that it might not be possible for individuals to simultaneously have two separate systems for VOT.
It would seem logical that the shift in VOT displayed by these women
reflects a shift in language dominance. Other timing phenomena might well be
affected in a similar way. This might in any case prove to be a useful way of establishing dominance and charting shifts in dominance, as well as of
monitoring ongoing language attrition on the phonetic level.
Early and late bilinguals
An important distinction can be made between early bilinguals (those who learn two languages simultaneously from birth or sequentially in their early years) and late bilinguals (those who meet their second language after an age which I will not attempt to specify here).
One reason why an individual might be a late bilingual is because of the family’s migration after the early years. Such individuals have been the subject of many second language acquisition studies.
Another quite different reason for this, and the subject of my investigation, is that some students and their parents choose to pursue studies through the medium of a language other than their native language which is also the majority language of the place of residence. This means that even non- language subjects, such as history or physics will be taught in the target language. This is known as Content and Language Integrated teaching and
Learning (CLIL), and is in many ways similar to the kind of immersion classes found e.g. in Canada.
In the Swedish context, CLIL usually involves classes at upper secondary
level which offer one or other of the ordinary Swedish study programmes
(natural sciences or Social Sciences) with varying proportions of the teaching
in English. In recent years such classes have become very popular and are
being offered in many municipalities. Another, even more prestigious variant is the International Baccalaureate programme which is externally examined in Switzerland and where all teaching (except for that in Swedish and foreign languages) is in English. This is offered in a good many schools in Sweden.
Such classes cannot really be compared to immersion classes since the target language has at least as high status as the students’ first language which is in most cases the majority language, Swedish (Hyltenstam, 2004, De Mejía, 2002). These classes have status in the eyes of the students and their parents;
they come about with support from the local authorities and they are generally accepted by both teachers and other students. There is often filtered entry to these classes, in the shape of admission tests which measure students
proficiency in English, maths and sometimes Swedish. Students are therefore motivated and able which should provide good learning conditions.
Concern for L2 (English)
For the student, the point of going in an English-medium class is to improve proficiency in English. The growing importance of the English language in the world, and the perceived advantages native or near-native speakers of English have make the attainment of advanced language skills a worthwhile target.
Schools have tried various ways of ensuring that the students and helped along their way.
If students are to gain proficiency in English it would appear sensible that
they speak English throughout the school day. In some schools there is a rule
forbidding code-switching (Knight, 1990). Since it is generally agreed that
codeswitching is a positive extension of the communicative competence of
bilingual speakers, this is a dubious measure. Other schools accept that students speak Swedish outside the classroom, but note that this is less common in the third year students than in the first years.
Another potential problem is that the teachers on these programmes may not be native speakers of English. Although Phillipson (1992) argues that the native speaker is not the best foreign language teacher, at the very high level of ambition of most of these students, where near-nativeness is the target, a teacher’s non-nativeness is perceived as a problem. There is, of course, not an unlimited supply of native speakers of English in Sweden, at least not who are qualified to teach the required range of subjects.
Even if suitably qualified native speakers can be employed by the schools to teach on these programmes, they will, after a period of residence in
Sweden, usually start to exhibit features of language attrition after a period.
Concern for L1 (Swedish)
When young people enter upper secondary school at the age of 15 or 16 their acquisition of Swedish, their first language, is far from complete.
Normally, a Swedish upper secondary programme will help students to hone
their ability to use academic Swedish, to debate and discuss in Swedish at
various levels, to read academic Swedish of various kinds etc. In the IB
programme, students need only study Swedish in the first, pre-IB year. They
have no Swedish in other subjects and many choose not to continue their study
of Swedish as one of the six IB subjects they choose to study in the second
and third year.
Not only might IB students fail to develop their Swedish in an age- appropriate way, they may potentially even lose some of their attained proficiency in their first language. There may in fact be attrition effects in their Swedish. This attrition can be on various levels: lexical, so the English word is mobilized before the Swedish one; semantic so that there is confusion of false friends such as lås upp – lock up (should be unlock); syntactic, in such expressions as ett glas av vatten (a glass of water – standard Swedish would not have the preposition); phonological as in a loss of the distinction between /y:/ and /i:/ such that bi (bee) and by (village) become homophones (reported in Cunningham-Andersson and Andersson (2004) in connection with simultaneous bilingualism); and phonetic, which is the subject of the present study, “Becoming bilingual”.
In fact, although there is not much work which examines the Swedish of students in English-medium programmes, a study by Washburn (1997) suggests that there is an attrition effect.
Phonetic correlates of dominance
Cunningham Andersson (2003) showed differences in the dominances of
three siblings who were simultaneously bilingual in English and Swedish,
raised according to the one person – one language method (i.e. early
bilinguals). These results are shown in Table 1. The method used was to
examine the relationship between vowel length and postvocalic consonant
length in Swedish and English. Since monolingual native Swedish has
compensatory post-vocalic consonant length adjustment and monolingual
native English does not, the different patterns shown by the children in both
languages allowed them to be analysed as English dominant, Swedish-
dominant or balanced. The calculation used was V:/C – V/C(:) where values over 1 are typical for Swedish speakers and values under 1 more typical for English speakers (McAllister et al 2002).
Table 1. Average V:/C – V/C(:) ratios for the Swedish words vit-vitt and the corresponding English words beat-bit.