The sociolinguistic situation in Algeria:

The sociolinguistic situation in Algeria is diverse and complex because of the existence of a melting pot of languages: Standard Arabic (hereafter SA) which is according to Algerian Constitutions (1963,1976, 1989, 1996, 2002) the official and first national language; Dialectal Arabic or Algerian Arabic including its regional varieties (hereafter AA) which is the native language of the majority; the middle languages2 which are intermediate forms between CA and AA (Modern Standard Arabic MSA, Educated Spoken Arabic ESA, and Literary Arabic); Berber (with its different varieties) which is also the native language of a considerable minority and was recently recognized as a national language (Algerian constitution 2002), and French. As a result of the co-existence of these languages, a wide range of sociolinguistic phenomena are observed in Algeria swinging between a bilingual situation and a diglossic one.

The fact that most countries in the world today are multilingual but the policies are in most cases monolingual is a reality that is reflected in the Algerian context. Officially, Algeria is an Arabic Muslim state whose sole official and the first national language is CA. Arabic is supposedly used by all members of the speech community. However socio-linguistically speaking, Algeria is a multilingual country where two or more varieties coexist resulting in complex sociolinguistic phenomena. These are diglossia, bilingualism, code switching and borrowing. Diglossia was first introduced as a term by the French linguist William Marçais (1930) as ‘La Diglossie Arabe’3. Then it has been introduced by Ferguson (1959) as a linguistic theory based on his study of four languages, Arabic, Greek, Haitain Creole and Swiss German. He defines diglossia as: “A relatively stable language situation in which, in addition to the primary dialects of the language (which may include a standard or regional standards), there is a very divergent, highly codified (often grammatically more complex) superposed variety, the vehicle of a large and respected body of written literature, either of an earlier period or in another speech community, which is learned largely by formal education and is used for most written and formal spoken purposes but is not used by any sector of the community for ordinary conversation.” (Ferguson, 1959b: 336) According to Ferguson (1959),

Diglossia is the side-by-side existence of two structurally and historically related language varieties, a High variety (hereafter H) and a low variety (hereafter L), throughout a community. The most important differences between the two varieties according to Ferguson are the linguistic features of H and L (lexicon4, grammar5, phonology6) their respective uses (i.e. they are functionally differentiated)7 and their sociolinguistic differences in terms of prestige8, literary heritage9, the way they were acquired10 and standardization11. Within the Algerian speech community, the diglossic features are present to some extent covering only part of Algerian linguistic situation. CA and AA are genetically related however they differ from each other in terms of grammar, vocabulary and pronunciation as well as in terms of prestige, acquisition, literary heritage, function and standardization. CA is the ‘H’ variety, it is the language of the Koran and Muslim identity and it is used for formal and public settings. The ‘L’ variety (AA) is reserved for use in informal and intimate contexts. Fishman (1972) developed the notion of diglossia to cover not only genetically related varieties but also genetically unrelated ones as far as they are functionally distinguished. Applying the term to monolingual as well as bilingual societies Fishman (1972b:92) states: “Diglossia exists not only in multilingual societies which recognize several languages and not only in societies that utilize vernacular and classical varieties, but also in societies which employ several dialects, registers, or functionally differentiated varieties of whatever kind”(Fishman, 1972:92) The extended definition of diglossia can cover another part of the situation in Algeria. The existence of French as another H variety which is genetically unrelated to CA and AA, and Berber as another L variety that is genetically different from AA and CA; display the following diglossic relations within the Algerian context: French/AA, French/Berber, CA/Berber. Algeria is also a bilingual community. Bilingualism has been defined in different ways by different linguists. Following Bouamrane’s definition (1986: 15) which is a combination of many linguists’ definitions, bilingualism is defined as:

The conditions of obtaining the corpus:

The data gathered for this study is based on more than fourteen hours of audiorecorded conversations. The speakers were not aware that their speech was being recorded. The fourteen hours of audio-recorded conversations are not all what we have recorded. Many recordings have been deleted either because the speakers talk exclusively in AA, or because the quality of the sound was not good. Some conversations were conducted in noisy environments where other speakers talk as in a university class, or in a secondary school yard. Other conversations were carried on in places where there was music or traffic noise as in a cafeteria, in a street, or in a bus. Thus we have only retained recorded conversations that include clear passages of CS. Yet even some of those conversations included unclear passages where the speech was too low or too fast to understand it. There are also passages where turn takings were not respected and speakers spoke at the same time producing interrupted utterances. Recording spontaneous conversation has not been an easy and a straightforward task especially in the beginning. It was time consuming and sometimes disappointing. We have already cited some conditions that constitute a challenge in obtaining good recordings (i.e., the quality of the sound, overlap of the speech, interruption, noisy environment…etc), the other challenge concerns the informants who recorded the conversations.

Two problems arise in this respect. Firstly, it is difficult to ask people to record their everyday conversations even if they know you, trust you and know that it is for a research purpose and their names will not be revealed. Yet people generally like privacy16. This is what makes the informants reluctant and selective in recording the conversations. This has been reflected in their recordings which were very few and disappointing. They for instance record parts of conversations and choose the one in which they don’t participate. Sometimes you gave them the recording equipment for weeks and then they gave it back to you with small amount of recordings simply saying that they forgot to use it. The second problem is that some informants17 were selective in choosing the interlocutors. They tend to record conversation with interlocutors who use French more than AA or whose speech contains whole utterances in French thinking that this is what is meant by alternation between AA and French. Because this was the first time that we were dealing with such technique of gathering a corpus, we did not expect such obstacles. Yet these kinds of problems opened our eyes to the fact that we need to discuss these things with the informants in order to avoid them, and to explain to them what we need from these recordings instead of just giving them the equipment and telling them to record conversations. Moreover, the informants needed time in order to adapt themselves to the recording equipment and to the idea of recording their conversations. The other thing is that we have asked them to record whole conversations whatever time they take and to record even the conversations that are mostly spoken in AA, because they may contain interesting instances of CS and because it cannot be expected when people will use French words or expressions. Analysing the data was another challenge because we were dealing with AA which is not a codified dialect and has no written norms or grammatical books. So in some cases we had to decide subjectively on how to transcribe some particles; for instance the particle ‘’18 sometimes is transcribed as a verb (to have) and sometimes as a preposition (to, at, for).

The approach advocated in the study:

This work has a micro-sociolinguistic perspective which relies basically on qualitative methods. We resort to grammatical qualitative description of the data rather than a quantitative analysis for different reasons or considerations. Many researchers in code switching research and especially those working on the morpho-syntactic aspects of CS structure within the sentence rely on spontaneous conversations of recorded data gathered from micro sociolinguistic environments. These methods of research give them more freedom in analyzing and interpreting sociolinguistic phenomena and it has been efficient in deriving some general constraints. In addition the field of contact linguistics in general and code switching research in particular has tremendously advanced in the last decades due to the findings of these qualitative studies that have provided a large body of CS literature from typologically divers languages (Myers-Scotton, 1993, Muysken, 1987, Boumans, 1998, Backus, 1996, Romaine, 1989, Gardner-Chloros, 2009). The qualitative examination of different data sets, show that even some CS strategies (i.e. double morphology, bare forms, bilingual compound verbs and flagged CS) are indeed universal strategies that are shared by many bilingual speakers.

In fact many authors have drawn on a variety of recent studies and data sets that have a qualitative perspective in approaching CS from a theoritical perspective including Myers-Scotton (2002) who has revised and extended her model, and Boumans (1998) who has propose his Monolingual Structure Approach. Muysken (2000) has also relied on these qualitative findings to propose his bilingual speech taxonomy21. On the other hand some scholars have referred to the difficulty of quantitative analysis. Boumans (1998: 154) states that: “quantitative information becomes useful only if it exists for all categories that are in complimentary distribution”. Boumans (ibid: 154) further explains that “in addition to counting CS instances, a word count for sufficiently large samples of the monolingual would be required”. For instance in order to count English lexical insertions that incorrectly lack Nigerian Arabic (NA) definite article22 (bare forms), Owens (2005) compare the English inserted bare nouns into NA matrices with those English inserted nouns that are marked for the definite article and with those that are correctly unmaked23. Then these English insertions in CS corpus are globally contrasted with the monolingual corpus. Furthermore researchers such as (Poplack, 1981, 1990, Polack and Meechan, 1998, Owens, 2005) who advocate for a quantitative approach, use statistical analysis program such as the CLAN software and the Combo statistical software in the interpretation of bilingual data. These statistical softwares are however not available to us. their