Two levels can be distinguished in the structure of a language as a system of signs: the level of expression and the level of contents. Every sign of a language will thus be characterized by the unity of these two aspects. We can distinguish therein the signifying (.signans) and the signified (signatum), which correspond to the two levels of the language. Relations between the signifying and the signified in linguistic signs are determined by the relationship between their content and their expression. Relations between signs at the level of contents and at the level of expression are the cause of the complex nature of the linguistic sign.

Translating is often discussed in scholarly circles. Writers talk about it as, obviously, professional translators also do. Even linguists have something to say about this activity, as old as the oldest civilizations. We should like to offer some ideas here on a subject that is so frequently considered. While the ideas are not entirely new, they are results drawn from a lengthy reflection and from the no less lengthy experience of translators. We hope they will indicate some directions that would be fruitful for a continuation of the discussion.

In a number of areas, in particular in the Pacific region, it has been observed that languages have undergone simplification processes of their usually very elaborate grammatical structures, and that such elaborate grammatical features have decayed and in some cases entirely disappeared from some languages, hand in hand with the progressing decay, and falling into disuse, of the traditional cultures of the speakers of such languages. Such phenomena of simplification and decay of grammatical complexities are most readily observable in Papuan languages of the New Guinea area and have also been found in Australian Aboriginal languages. Both these languages have very complex grammatical systems, the Papuan languages even more so than the Australian Aboriginal languages, and the multiplicity of grammatical features in them lend themselves more readily to simplification than may be the case with other languages in the Pacific area, such as Austronesian languages, also called Malayo-Polynesian languages, though such phenomena are by no means absent from them.

Pidgin languages are special reduced interlingual systems of communication created by the need to communicate between speakers of two or more different languages. They originate to fulfil certain communicative requirements, adapt to changes in these requirements, and disappear once they are no longer needed, for pidgin, by definition, are second languages, used by adults and not transmitted (except in the exceptional case of creolization) to a new generation of children. Pidgin languages are found in all parts of the world where trade, warfare, colonialism or tourism has brought members of different speech communities into contact. The lifespan of some is no greater than the length of a summer holiday, others, such as Chinese Pidgin English, remained in existence for more than a century. Whilst for a long time pidgins were put in the category of “marginal languages” and their study correspondingly marginalized, in the most recent past, pidgin studies has become a fast growing and very respectable subfield of modern linguistics. The love-hate relationship which linguists, and indeed educators, historians and others, have had with pidgin languages centres around what I am inclined to regard as their single most important characteristic: their changeability. Pidgins, can be called the chameleons of the world's languages; they grow and contract with communicative requirements, they change their lexical composition to reflect the relative power of communicants, they change their speakers, and their geographical location and to their users look very similar to their own first languages.

Haitian Creole is a unique language, as are all creole languages. Mother tongue of almost six million people—without counting those who have immigrated—it has its own particular ways of expressing in words, of making sentences and forming a discourse. However, for more than two centuries it has been excluded from many circuits of communication. The limits imposed on it have kept it from developing a creativity, at the vocabulary level as well as that of organizing a discourse that would assure a function of plural communication, both oral and written. The present challenge of this long-throttled language has posed an important number of problems in linguistics, teaching, culture and politics because of new communication systems and especially because of education.

In a work that appeared in 1967 but which has lost none of its interest today, P. Alexandre gave a list of the 51 languages of Black Africa (or homogeneous dialectical and linguistic groups) that approached or surpassed a million speakers. For each of them he evaluated available documentation, on a scale of from 1 to 6. He gave the rating 1 to languages for which the documentation was poor (outdated grammars and dictionaries, incomplete or absurd systems of unscientific transcription) and the rating 6 to languages for which works conforming to modern requirements were available (correctly noted sound systems, phonological and phonetic descriptions, annotated texts, dictionaries, stylistic and dialectological studies, audio material). Only three languages merited the rating 6 and only four were rated 5. This means that only seven languages out of about 1200 could be considered as accurately described, that is, offering all the necessary guarantees to the specialist.

On principle, there are no language classifications which are right or wrong, but only classifications which are more or less useful or useless. This statement at the beginning of my paper is intended to indicate the teleological perspective in which I want to view the more general problems involved in the classification of African languages. I shall discuss these within a framework which I derive from the four main components of any language classification, namely:

1. 1. the aims and objectives of language classification,

2. 2. the basic units to be classified,

3. 3. the criteria of classification,

4. 4. the models of language classification.

Given the predominance of multilingualism in many countries there is great need for a new, interdisciplinary approach to the problem. Sociolinguistics—or the sociology of language—provides such an approach and can help to shed new light on a phenomenon which is often discussed in terms that tend to create more problems than they solve. In order to demonstrate how sociolinguistic theory and practice can contribute to the study of multilingualism this paper will attempt to outline briefly some of the basic premises of sociolinguistic theory which appear relevant to the understanding of the phenomenon of multilingualism and also to define multilingualism as a social behaviour in response to a contact situation.