In principle, the ‘habitual, fluent, correct and accent-free use of two languages’ (Paradis, 1986) – or of more than two languages. However on this definition, few individuals qualify as complete bilinguals. It often happens that a bilingual is not equally competent in different aspects of the two languages: they might, for example, have a more restricted vocabulary in one than in the other or might exhibit different abilities in respect of speaking, listening, reading and writing.
Furthermore, many bilinguals use their languages in ways that are domain-specific: one language might be used in the family and one reserved for educational contexts. The imprecision of the term ‘bilingual’ is not helped by a tendency among some psycholinguists to use it when referring to foreign language learners who are relatively advanced but have certainly not achieved a competence that is native-like.
Grosjean (1982) defines bilingualism in terms of language use rather than language proficiency. For him, a bilingual is somebody who needs and uses two or more languages in everyday life. A majority of the world’s speech communities use more than one language; and about half the world’s population is believed to be bilingual in this sense. In addition, ther are many bilinguals who are the offspring of mixedlanguage couples.
An early account of bilingualism (Weinreich, 1968) proposed three types. In compound bilingualism, conditions in infancy are equally favourable for both languages, and words in both are attached to one central set of real-world concepts. Co-ordinate bilingualism occurs when conditions in infancy favour one language over the other; the consequence is that the infant develops two independent lexical systems, though meanings overlap. Subordinate bilingualism occurs when the second language is acquired some time after the first, and so remains dependent upon it. These categories have proved difficult to substantiate. However, the stage at which the two languages are acquired remains an important consideration in recent accounts, which often distinguish simultaneous bilingualism (both languages acquired concurrently), early successive or sequential bilingualism (both languages acquired in childhood but one preceding the other) and late bilingualism (the second language acquired after childhood).
Simultaneous bilingualism arises during ‘primary language development’, which commentators regard variously as occurring during the first three or the first five years of life. Exposed to two languages, infants initially mix vocabulary and syntax from both. In naming objects and actions, they often adopt the first word they encounter, regardless of which language it comes from; though in their morphology they may exhibit a preference for the less complex of their languages.
The unitary language hypothesis concludes that these infants start out with undifferentiated language systems. They begin to distinguish between the two sets of data by restricting each language to particular interlocutors, situations or pragmatic intentions. At the next stage of development, the infant distributes its vocabulary between two separate lexical systems, and becomes capable of translating words from one language to the other. However, the same syntactic rules are usually applied to both systems. In a final stage, the languages become differentiated syntactically, and mixing declines.
An alternative separate development hypothesis maintains that the two languages are distinguished from the start by the infant and that the phenomenon of mixing simply shows two incomplete systems operating in parallel. Simultaneous bilingual acquisition appears to follow a very similar path to monolingual acquisition. There is no evidence that the acquisition process is delayed when more than one language is involved, though early vocabulary levels may be slightly lower in bilingual children. Nor do similarities between the two target languages appear to assist acquisition: an English-French bilingual does not develop language faster than an English-Chinese one.
In successive bilingualism, there is much greater variation between individuals. The time of acquisition of the second language (during the primary period/before puberty/in adulthood) may be a factor; while mastery of the later language may be limited to certain domains. In some cases, the acquisition of the later language is additive, resulting in the use of two systems in parallel. In others, the effect may be subtractive, with the later language replacing the first in some, many or all domains. The acquisition of a second language by an immigrant may even lead to the attrition of the original language if the speaker has to communicate mainly or exclusively with members of the host community.
A distinction is made between adult bilinguals who are balanced and those for whom one language is dominant. A balanced bilingual has been represented (Thiery, 1978) as somebody who is accepted as a native speaker in two linguistic communities at roughly the same social level, has learnt both languages before puberty and has made an active effort to maintain both of them. Fully balanced bilinguals are said to be rare.
Bilinguals may not always be aware of which language is their dominant one, and it has not proved easy to establish dominance. One approach has been to ask individuals which language they are conscious of having spoken first; though many recall acquiring both simultaneously. Another is to ask individuals to express a preference for one of their languages. There may be a relationship between dominance and anxiety, with the dominant language resorted to in times of stress or tiredness.
Experimental methods to determine dominance have included rating bilinguals’ language skills across languages, self-rating questionnaires, fluency tests, tests of flexibility (checking the ability to produce synonyms or draw upon a range of senses for a particular word), and dominance tests where bilinguals read aloud cognates which could be from either of their languages. Even where dominance is established, the situation may not remain constant: the relationship between languages may shift as the individual’s linguistic needs and circumstances change.