Memoire Online - Heritage language maintenance among the berbers of Zrawa (southern Tunisia). An exploratory study

Thesis submitted in partial fulfillment of the requirements of the Master's
Degree in English Linguistics

To my mother and father.
To my brothers and sisters.
To my friends.
To all those who love me and believe in me.

I am grateful to my advisor, Dr. Faiza Derbel, for her direction, support, and valuable comments.

I would like to thank all the participants who asked other individuals to take part in the study.

I would like to thank my parents, brothers, and sisters for their endless love, encouragement, and support.

I would like to thank all my friends for their encouragement, support, and belief in me.

Abstract

The purpose of this study is to investigate the maintenance of Amazigh (Berber) language among the Imazighen (Berbers) of Zrawa, a village in the south-east of Tunisia. The data for the present study were collected during the month of February 2017 using the mixed methods approach. A questionnaire, semi-structured interview, and participant observation were used to collect data. The questionnaire was initially administered to 53 Imazighen from various age groups, various occupations, and both genders. Subsequently, 11 were interviewed after filling out the questionnaire in order to generate in-depth data concerning certain topics included in the questionnaire and to collect data about topics which were not investigated in that questionnaire. The participant observation took place during ten visits to Zrawa, with each visit taking approximately eight hours. The reason behind the use of participant observation was to gather data about the geographic concentration of the Zrawa Amazigh community, including the language used within the community. Results from the study indicate that the factors contributing to AL maintenance in Zrawa are: (a) the geographic concentration of the Amazigh community, (b) the essential role of Amazigh families, (c) Imazighen's positive attitudes towards the Amazigh language, and (d) the perceived close relationship between Amazigh language and identity.

Appendix I. Map of the Amazigh-speech zones in Tunisia based on Pencheon (1968) 94

Appendix J. Map of the Amazigh-speech zones in Tunisia based on Maamouri (1983) 95

Figure 4. 4: Results of statements 4 and 5 on the link between Amazigh language and

identity 63
Figure 4. 5: Results of question 3 about the link between Amazigh language and

AL: Amazigh Language HL: Heritage Language LM: Language Maintenance LS: Language Shift

1.0. Introduction

This chapter is an introduction to the study. It starts by presenting the background, the setting and the rationale of the study. It also highlights the significance of the study. Besides, it describes the design of the study. The chapter ends by providing an overview of the study. 1.1. Background to the Study

Maamouri (1983a, pp. 11-19) states that the linguistic situation of post-colonial Tunisia is complex and argues that six languages are currently used in Tunisia: French, French-Arabic, and four varieties of Arabic: (a) Classical Arabic, which is the «only pure form of the language» (p. 15); (b) Modern Standard Arabic, which is less formal than Classical Arabic;

(c) Tunisian Arabic; and (d) Educated Arabic, which is «a form of `simplified Modern Standard Arabic (arabiya mubassata) and a form of `elevated' Tunisian Arabic (darija muhathaba) or both at the same time» (p.17). French is used in education whereas French- Arabic is used by students in settings other than the classroom and by government officials, members of the professions, and other administrators, in informal situations. As to Classical Arabic, it is used in Qur'an recitations, prayers, religious sermons and talk, and literary creation and criticism. Modern Standard Arabic is used in mass media, political speeches, modern plays, novels, literary magazines and lecture, and primary and secondary schools. As for Educated Arabic, it is mainly spoken by educated Tunisians and used in politics. Tunisian Arabic (TA) is the dominant language in the country. It is less formal than Classical Arabic and Modern Standard Arabic.

As to the Amazigh language (AL), Maamouri (1983a, p. 14) states that it is a marginalized minority language in Tunisia. This implies that it has a number of characteristics. Indeed, Simpson (2001, pp. 579-580) lists 13 characteristics of a minority language. In this context, I will mention only four which, I think, are the most relevant to the plight of AL in Tunisia.

First, a minority language lives in the shadow of a culturally predominant language. Actually, it can be concluded from the use of Arabic in almost all settings that AL lives in the shadow of Arabic. Second, a minority language is not used in formal settings such as administration, education, and mass media; but rather, confined to such domains as the home. As mentioned above, Arabic, but not AL, is used in key domains such as education, politics, mass media, and religion. However, as Maamouri (1983a, p. 14) states, in the Tunisian Amazigh villages, AL is the only language used within Amazigh families. Third, bilingualism is common among the speakers of a minority language. At least, this is the case of the Amazigh-speaking inhabitants of Zrawa, the focus of the actual study (see sections 1.2 and 4.1). Finally, a minority language may have no standardized form. As Ennaji (2005, p. 73) acknowledges, AL is neither standardized nor codified. In a nutshell, AL is neither the majority nor the official language of Tunisia. Indeed, as aforementioned, TA is the dominant language whereas the official language, as Maamouri (1983b, p. 147) acknowledges, is Modern Standard Arabic.

It appears that not being the official or majority language of the country gives AL the aspect of a heritage language. As Cummins (2001) asserts, «in a general sense, the term heritage languages refers to languages other than the official or majority languages of a country» (p.619). Explaining the meaning of the term «heritage languages», Cummins (200,

p. 619) states that it is meant to acknowledge that these languages constitute important aspects of the heritage of individual children and communities and are worthy of financial support and recognition by the wider society. This may be relevant to AL. Indeed, Ennaji (2005, p. 74) states that AL is a basic component of Moroccan and North African culture. Ennaji (2005, p. 76) also states that AL is one of the oldest African languages in the sense that it is the mother tongue of the indigenous inhabitants of North Africa. In a similar vein, Maamouri (1983a, p. 14) acknowledges that AL is the indigenous language of Tunisia.

Taking together the suggestion that AL is a heritage is a heritage, the acknowledgement that AL is the indigenous language of Tunisia, and the fact that the languages spoken by Amerindians, who are the indigenous people of the United States of America, are identified by Fishman (2001, pp. 83-83) as indigenous heritage language leads to the deduction that AL is an indigenous heritage language. Finally, it is worth noting that heritage languages have been referred to using different terms. As Cummins (2001, p. 619) states, the terms «ancestral», «ethnic», «immigrant», «international», «minority», «non-official», «third» (after English and French), «world», «community languages», and «mother tongue teaching» have all been used in different times and in different countries to refer to heritage languages.

Returning to the plight of AL in Tunisia, the regression of this language has been due to a number of factors. As Maamouri (1983a, p.14) states, AL has gradually regressed due to the rapid development of the educational system and the general spread of modern mass media. For Pencheon (1983, pp. 31-32), the regression of AL in Tunisia is attributed to the following factors. First, the geographic dispersal of Amazigh villages and their being surrounded by Arabic-speaking communities have fostered the use of Arabic in interaction with the outside world, including commerce and other transactions. Second, the lack of employment opportunities in the Amazigh villages forces men to migrate to cities where Arabic is the dominant language, leaving behind their families. Living in such urban areas influences the status of AL in the sense that these men end up preferring the use of Arabic in all circumstances which allow its. Third, the liberation of women in the aftermath of the independence has given all Amazigh girls access to education in Arabic. This has resulted in the disappearance of monolingualism among these girls. Finally, there is a lack of solidarity between Amazigh villages such as between Douiret and Chenini and between Tamazret and Taoujout despite their geographical proximity.

The regression of AL in Tunisia is indicated by the decrease in the number of Amazigh villages from 12 in 1968 (see Appendix I) to nine in 1983 (see Appendix J). Indeed, Pencheon (1968) identifies 12 villages where Amazigh is still spoken. Specifically, in Sened, Sakkiet, and Majoura, east of Gafsa, only the elderly people can speak the language. In contrast, Tamazret, Taoujout, and Zrawa, west of Matmata, are still entirely Amazighophone. Moreover, in Djerba only Guellala is still completely Amazighophone whereas one third of Ajim and less than half of Sedouikech are Amazighophone and some 200 or 300 people in El-Mai seem to speak Amazigh. In Tataouine, Douiret and Chenini are still entirely Amazighophone. However, Maamouri (1983a, p. 14) asserts that AL is geographically confined to the villages of Zrawa, Tamazret, Taoujout, Guellala, Ajim, Sedouikech, El-Mai, Chenini, and Douiret. Maamouri (1983a, p. 14) also asserts that traces of AL have completely disappeared from the area of Sened, Majoura, and Sakkiet and reports that AL is occasionally spoken in Tunis and other big cities by the doughnut vendors (ftayriyya), central market porters, and newspaper vendors who had come from different Amazigh villages in search for work. According to a more recent publication (Gabsi, 2003), the Tunisian areas where AL is still spoken today are: Douiret, Chenini, Zrawa, Tamazret, and Taoujout, Guellala, Ajim, Sedouikech, and El-Mai.

As has been just mentioned, Amazigh is still spoken today in many villages such as Zrawa, the focus of this study. The use of AL in Zrawa raises interest in investigating what factors have contributed to AL maintenance. The following section provides details about the village itself and its inhabitants.

1.2. Context of the Study

The current study was conducted in the Amazigh village of Zrawa. This village is officially a part of New Matmata which itself is a district of Gabes, a governorate in the south-east of Tunisia (see Appendix K). Zrawa is located approximately 47 kilometers from

Gabes City and 24 kilometers from New Matmata. It is isolated from the Arabic-speaking neighboring communities. Zrawa is divided into Old Zrawa and New Zrawa. Old Zrawa is a cluster of abandoned old buildings located on top of the mountain with a population reduced to one Amazigh family. New Zrawa, on the other hand, is a small modern village where modern commodities are available, namely running water, electricity, telephones, and the internet. According to Said (a pseudonym), a teacher, New Zrawa was created in the 1978. According to local informants, the road that links Zrawa to New Matmata divides New Zrawa into two territories. Thus, the one to the north-west of the road is part of El-Hamma (Gabes) district and is officially called «Farhat Hachad». The one to the south-east of the road, on the other hand, is part of New Matmata (Gabes) district and is officially designated «Zrawa».

Having given some details about the place where the present research was carried out, it is necessary to provide details about the Amazigh population in Zrawa. Based on the Municipal Vote Register of May 2016 issued by the Ministry of Local Affairs, the population of Zrawa has reached 1328 inhabitants. This number does not include Zrawi (adjective from Zrawa) people who live in «Farhat Hached» and those who have migrated to other Tunisian regions or abroad. There are no statistics concerning the number of Zrawi Amazighophones who have migrated to other regions or abroad. Mohamed (a pseudonym), a member of an Amazigh cultural association, claimed that there are about 5000 Imazighen residing in Zrawa, with its two parts mentioned earlier, and thousands of Zrawi Imazighen families and individuals living in Tunis and abroad.

According to Mohamed and other participants, Imazighen represent the majority of the population in Zrawa and the remainder of the inhabitants consists of Arabic-speaking families and individuals from other regions. Indeed, Mohamed informed me that there were about 50 Arab-speaking inhabitants made up of families and individuals from Dhiba (the governorate of Tataouine) and a six-member family from El-Hamma. In the same vein, Arij (a

pseudonym), a university graduate, told me that the number of Arabic-speaking families settled in Zrawa exceeds that of Arabic-speaking individuals. As a consequence, Amazigh is predominantly spoken in Zrawa. Actually, what is interesting about Zrawa is the impact that the contact between Amazighophones and Arabophones has on Arabic-speaking children and adults. Indeed, Arabophone children learn Amazigh in the neighborhoods when mixing with Amazighophone children. In fact, I witnessed the use of Amazigh by two Arabic-speaking children from El-Hamma when they were conversing with their Amazighophone peers. For adults, as three of my informants told me, can speak Amazigh while others can only understand it. For those who can speak Amazigh they do not use it. Indeed, Mahdi (a pseudonym) told me that his employee can speak Amazigh; however, I noticed that they do converse in TA. For those who can only understand Amazigh they reply in TA whenever addressed in Amazigh.

The Zrawa Amazigh community has some linguistic, economic, social, and religious characteristics.

1.3. Rationale and Objectives of the Study

Factors influencing AL maintenance in Tunisia have not been addressed by linguists. To my knowledge, there is no research about the factors contributing to or hindering AL maintenance among the Tunisian Imazighen. As such, it is important to investigate the factors underlying the current use of AL in Tunisia by a number of speakers, which is estimated by Ennaji (2005) to be around 100.000.

One major aim of the present study is to shed light on the factors contributing to the maintenance of Amazigh in Tunisia. Indeed, this research focuses on four factors that are said to influence LM and supposed to be relevant to AL in Tunisia, and particularly in Zrawa: the geographic concentration of the community, the key role of the families, the positive language attitudes, and the perceived linkage between language and identity.

With emphasis on the factors contributing to AL maintenance in Tunisia, the present study seeks to answer the following research questions:

The study emanates from the researcher's attempt to explore the factors underlying AL maintenance among a sample of Zrawi Imazighen. It attempts to present the factors that most contribute to keeping AL alive. The results of the study are hoped to generate valuable insights which can be of relevance to Amazigh people in Zrawa and, perhaps, to other

Amazigh groups elsewhere. Moreover, the study is significant because it encourages other researchers to investigate AL maintenance in other Amazigh villages and in cities where Amazigh families and individuals live. Furthermore, this study is hoped to draw the attention of sociolinguistics and policy makers to the vulnerable situation of AL in Tunisia, given the fact that it is a dying language, and encourage them to make decisions in order to maintain and revitalize this language which is the indigenous language of the country, and therefore, an important aspect of the national heritage.

1.5. Design of the Study

A mixed methods approach is used in this study in order to provide an in-depth complete understanding of AL maintenance in Zrawa. This approach stipulates the use of quantitative and qualitative data collection methods. In fact, data for the study will be generated by the means of one quantitative data collection technique, namely a questionnaire, and two qualitative data collection methods: participant observation and semi-structured interview. Participant observation will elicit data about the geographic concentration of Zrawa Amazigh community. As to questionnaires and semi-structured interviews will be employed to collect data about the way AL is acquired and used within the Zrawa Amazigh families, the attitudes towards AL and its maintenance, and the link between AL and Amazigh identity. Thus, the study will hopefully uncover the underpinnings of AL maintenance in the Zrawa Amazigh community and provide ground for a discussion of its chances for survival in the future as a minority language.

The study compromises five main chapters. The first chapter is devoted to the Introduction. It describes the background, the setting, the rationale, the design, and the purpose of the study. The second chapter provides a review of the literature relevant to LM. The third chapter deals with the methodology followed by the researcher in order to answer the

research questions. In fact, it provides a detailed description of the participants, the instruments, data collection and data analysis procedures. The fourth chapter presents the main findings of the research and discusses them. Finally, the Conclusion summarizes the major results of the study, lists its major implications and limitations and provides some suggestions for future research.

This chapter begins with highlighting the importance of language maintenance. Then, it presents a review of the literature on language maintenance and the factors contributing to it and those hindering it; hence, facilitating language shift. Reviewing previous research dealing with these factors is thought to be helpful in putting this study into perspective and identifying the related issues.

It seems that definitions of language maintenance (LM) provided by different scholars (Fishman, 1989; Srivastava, 1989; Mesthrie, 2001; Coulmas, 2005) point to the idea of retaining a particular language despite the pressures it faces from another language. This implies that LM occurs in a situation of contact between two groups of people speaking distinct languages. As Fishman (1964) puts it, «the basic datum of the study of language maintenance and language shift is that two linguistically distinguishable populations are in contact and that there are demonstrable consequences of this contact with respect to habitual language use» (p. 33). Not only are the languages spoken by the two groups distinct but also they have different status, that is, the language of one group is more powerful than the language of the other. Thus, the weaker language faces a competition from the stronger one; however, its speakers hold on to it (Mesthrie, 2001). To achieve LM, speakers of the weaker language often seek to transmit it from one generation to another. Assuring inter-generational linguistic continuity is indicative of LM (Fishman, 1989).

It is worth noting that LM is related to language shift (LS) defined by Srivastava (1989) as «a shift from the use of one language to the use of another» (p. 10). As Pauwels (2005, p. 719) points out, studying LM often involves the identification of the domains and situations in which the language is no longer used or is gradually regressing.

The importance of LM has been emphasized by many scholars. For example, Dorian (1987, p. 64) states that although it is difficult to maintain a language, there are reasons for undertaking maintenance efforts. Among these reasons, she mentions three: (a) LM can lead to a reversal of negative attitudes towards a particular language; (b) LM can contribute to the transmission of traditions and customs over generations; and (c) LM can be economically beneficial in the sense that it they provide not only employment for teachers and translators of these languages but also greater opportunities for business and international trade. In the same vein, Tamis (1990, p. 499) points out that learning the mother tongue helps children enhance their relations with their family members especially their parents. Likewise, Garner (1988, pp. 42-47) states that enabling children to talk with their elders, most usually their grandparents, is an important reason for language maintenance for both Russian and Swedish parents.

Language maintenance and language shift as a separate field of study dates back to the last part of the 19^th century. According to Fishman (1964, p. 32), this field has its origins in the literatures which continued from the latter part of the 19th century into the Second World War days, the 1953 Conference of Anthropologists and Linguists, and the work of Uriel Weinreich and Einar Haugen. The study of LM and LS, as Fishman (1964, pp. 33) points out, is divided into three sub-fields: (a) habitual language use at more than one point in time or space under conditions of intergroup contact; (b) antecedent, concurrent or consequent psychological, social and cultural processes and their relationship to stability or change in habitual language use; and (c) behavior toward language in the contact setting, including directed maintenance or shift efforts.

The literature on LM and LS points to a multitude of factors affecting LM and LS. Some scholars did outline typologies of LM and LS factors. For example, Giles, Bourhis and

Rosenthal (1977) proposed the ethnolinguistic vitality framework which includes 14 factors affecting minority language vitality. These factors are divided into three types: status, demography, and institutional support factors. Giles et al. (1977) introduce status factors as «a configuration of prestige variables of the linguistic group in the intergroup context» (p. 309). This means that the more status an ethnolinguistic group has, compared to other groups, the more likely it is to maintain its language. Status factors are broken down into economic, social, socio-historical, and language status factors. The demographic factors are connected to the sheer numbers and geographic distribution of group members. These factors are broken down into «group distribution factors» (national territory, group concentration, and group proportion) and «group number factors» (absolute numbers, birth rate, mixed-marriages, immigration, and emigration). The institutional support factors refer to «the degree of formal and informal support a language receives in the various institutions of a nation, region, or community» (p. 309). That is, the more support a minority language receives from both governmental and nongovernmental institutions, the more vitality it has. Giles et al. (1977) describe the aforementioned factors as being «objective» and acknowledge that «a group's assessment of its vitality may be as important as the objective reality» (p. 318). As a consequence, Bourhis, Giles and Rosenthal (1981) proposed the subjective vitality questionnaire as a way to assess in-group members' perception of in-group and out-group vitality on each of the factor included in the «objective» vitality model.

Similarly, Edwards (1992, pp. 49-50) suggests that a minority language situation can be influenced by 33 factors. He classifies these factors into 11 categories, namely demography, sociology, linguistics, psychology, history, politics, geography, education, religion, economics, and the media, with three factors attributed to each category. For instance, demography factors include numbers and concentration of speakers, extent of the language, and rural-urban nature of setting. Likewise, sociology factors have to do with socioeconomic

status of speakers, degree and type of language transmission, and nature of maintenance or revival efforts. Similarly, psychology factors consist of language attitudes of speakers, aspects of the language-identity relationship, and attitudes of majority group towards minority group. Edwards acknowledges that these are not the only factors.

In the Tunisian context, little research has been carried out on the factors affecting AL maintenance. Maamouri (1983, p. 14) states that the regression of AL has been due to the development of the educational system and the spread of mass media. In the same vein, Pencheon (1983) argues that geographic dispersal, migration to big cities, access to education, and lack of solidarity between proximate Amazigh villages have all contributed to the regression of AL. Battenburg (1999) goes as far as to state that AL in Tunisia is facing gradual death. In his study of AL in Douiret, Gabsi (2003, p. 291) points out that the sociolinguistic factors leading to AL regression are: a) the low prestige of AL as a mother tongue; b) the constant migration of Imazighen from Douiret to other major Tunisian cities; and c) the modernization of Imazighen's way of life. Moreover, Hamza (2007), in his study of AL in Douiret, Chenini, Guellala, and Tamazret, states that positive attitudes towards Arabic have promoted the shift from the use of AL to the use of Arabic. Such attitudes are indicated by the fact that most of the 100 participants in the study favored the use of Arabic as an in-group language of communication. Also, the parents among these participants viewed Arabic as a tool to gain access to education and a vehicle for social mobility. In addition, Gabsi (2011) indicates that LS is underway in the villages of Douiret and Chenini. Based on the situation of AL in these two villages, Gabsi (2011) suggests that in Tunisia there is no generation of Amazigh monolinguals and that multilingualism is the norm among the Tunisian AL speakers.

It seems that all these studies have highlighted the decline of the AL in Tunisia. However, there is no investigation, at least to my knowledge, on the factors which explain AL

maintenance in one particular village such as Zrawa, Gabes, given the fact that this language has disappeared from the villages of Sened, Majoura, and Sakkiet in the govenorate of Gafsa (Maamouri, 1983). The following two sections review literature on the most important factors of LM and LS.

This section deals with some of the factors of LM discussed in the literature on language maintenance and shift.

The geographic concentration of the speakers of a minority language affects the maintenance of that language. According to Holmes (2013, pp. 64-65), living in the same neighborhoods and meeting on a daily basis help families belonging to a minority group maintain their language. Holmes (2013) gives the examples of four immigrant communities: the Chinese community in the United States of America, the Greek community in New Zealand, and the Indian and Pakistani communities in Britain. To start with, Chinese people who live in the Chinatown areas of big cities are much more likely to maintain a Chinese dialect as their mother tongue through to the third generation than those who move outside the Chinatown area. As to the members of the Greek community in Wellington, New Zealand, they have established shops where imported foodstuffs from Greece are sold and where they speak Greek to each other. Likewise, Pakistani and Indian communities living in British cities have established their own shops where Pakistanis use Panjabi, and Indians use Gujerati, with each other.

As argued by Giles et al. (1977, p. 313), the concentration of a minority ethno-linguistic group in a given area (territory, region, or country) contributes to the maintenance of its language in the sense that its members have the opportunity to use that language as a means of daily communication. They mentioned the example of Canada where French Canadian

families living in isolated areas outside Québec and being in contact with English Canadians have lost their French language within few generations.

Many studies have shown the role of geographic concentration in LM. Laleko (2013, p.95), who investigated the vitality of Russian as a heritage language in the United states, found that it was somewhat easy for Russian immigrants living in large cities and metropolitan areas to maintain ties with their heritage language and culture due to the existence of the relevant infrastructure, including small businesses that cater to the Russian- speaking population, most remarkably food stores and restaurants, bookstores, art galleries, hair and beauty salons, medical offices, various real estate and insurance agencies, and religious organizations. In his study of LS among Chinese-Americans, Li (1982, pp. 118-

119) indicates that residing in Chinatown assists in resisting LS; however, Chinese- Americans, especially school children, residing outside Chinatown will usually adopt English as their mother tongue. In the same vein, Al-Khatib and Al-Ali (2010) reviewed previous research on the maintenance of minorities languages among the five minority groups residing Jordan (Kurds, Armenians, Gypsies, Chenchens, and Circassians). They observe that the Gypsies, Chechens, and Circassians have maintained their ethnic languages and cultures thanks to their residence in co-ethnic neighbors within close-knit communities, with limited contact with the majority language and culture; however, the Armenians and the Kurds have shifted towards the dominant language and culture because of being dispersed across the large urban centers. 2.3.2. Family

The family can play a key role in LM. As Fishman (1991) points out, the family «has a natural boundary that serves as a bulwark against outside pressures, customs and influences. Its association with intimacy and privacy gives it both a psychological and a sociological strength that makes it peculiarly resistant to outside competition and substitution» (p. 94). In

the same vein, Brown (2008) states that the home or family is «the last bastion in terms of language maintenance» (p.1). Results from Renz's (1987) survey of respondents of Portuguese origin in California indicate that family was ranked as the most important factor in linguistic and cultural maintenance.

Parents, in particular, play a key role in LM. Lieberson and Curry (1971, p. 146) indicate that what language parents choose to speak to their children affects, to a large extent, LM or LS. Parents play a crucial role in transmitting the mother tongue either by teaching it to their children (Okamura-Richard, 1985, p. 63; Sridhar, 1988, p. 84) or by speaking it at home (Sridhar, 1988, p. 84).

Many studies confirm the role of parents in LM. Nesturik (2010), in her investigation into LM among 50 Eastern European immigrant parents living in the USA, found that the majority of the participants in the study communicated with their children in their native language at home in order to maintain it. Al-Sahafi (2015) also investigated the role of ten Arab parents in the maintenance of Arabic in New Zealand. The findings reveal that the participants were determined to maintain and transmit the Arabic language to their children. In a similar vein, Gomaa (2011) examined the efforts of five Egyptian families living in Durham, United Kingdom, to transmit Egyptian Arabic to their children. The results show that, as a strategy to maintain Egyptian Arabic, the parents insisted that their children speak it at home. For example, when their children spoke to them in English, the parents replied in Egyptian Arabic and told them to speak Egyptian Arabic as long as they were at home. In her study of a Turkish family living in Western Pennsylvania, United States of America, Tatar (2015) found that the parents had tried their best to teach Turkish to their children, including speaking it with them most of the time. Likewise, Becker (2013) studied three Korean immigrant families living in West Michigan, USA. The findings reveal that parents in the three families encouraged their children to use Korean at home and wanted them to achieve

high proficiency level in Korean. Moreover, Abdelhadi (2017) in his study of LM among the Arabic-speaking community in Toowoomba, Australia, found that parents implemented many teaching strategies, such as translation, teaching writing, and reading stories, in order to maintain the Arabic language and transmit it to their children.

Language attitudes are defined by Myers-Scotton (2006) as «subjective evaluations of both language varieties and their speakers, whether the attitudes are held by individuals or by groups» (p. 120). Reviewing the literature on language attitudes, Giles, Hewstone, and Ball (1983) indicate that the term includes:

Language evaluation (how favorably a variety is viewed), language preference (e.g., which of two languages or varieties is preferred for certain purposes in certain situations), desirability and reasons for learning a particular language, evaluation of social groups who use a particular variety, self-reports concerning language use, desirability of bilingualism and bilingual education, and opinions concerning shifting or maintaining language policies. (p.83)

Language attitudes can be studied in different ways. Indeed, O'Rourke (2011, pp. 24-26) mentions three approaches to the study of language attitudes: the societal treatment of language, the indirect assessment within the speaker evaluation model, and the direct measurement with interviews or questionnaires. The social treatment of language includes techniques such as content analysis, observational analysis, participant observation, and ethnographic study of language choices. These techniques do not involve directly asking respondents about their attitudes. As to the indirect assessment of language attitudes, it involves the use of the matched-guise technique whereby listeners are asked to rate tape- recorded voices spoken by the same speaker but in different languages on a number of personal traits such as sociability, leadership, ambition, and sense of humor. This technique

aims at eliciting the listeners' attitudes towards the different voices on the basis of their attitudes towards the different languages. Finally, the direct assessment of language attitudes is done through questionnaires and interviews addressing particular aspects of language.

Particularly speaking, positive language attitudes towards a minority language can contribute to its maintenance. These attitudes, according to Holmes (2013), include associating the minority language with high value and regarding it with pride. Holmes (2013) believes that such attitudes assist with the efforts to use the language in a diversity of domains so that people can restrain the pressure from the majority group to adopt their language. In the same vein, Bradley (2002) states that «perhaps the crucial factor in LM is the attitudes of the speech community concerning their language» (p. 1).

Many researchers have addressed attitudes towards heritage languages. For example, the findings from Becker's (2013) study suggest that the participants believed that maintaining Korean was critical for their children's Korean identity formation and for family communication. In Tatar's (2015) study, the parents of the immigrant Turkish family living in Western Pennsylvania show positive attitudes towards learning and speaking Turkish as their HL at home and express a strong desire for their children to learn Turkish. Indeed, the father believed that it was essential for his children to learn Turkish because it is their heritage language and because they need it in their communication with the members of the family in Turkey and of the Turkish immigrant community in the US. As for the mother, she stated that learning Turkish as an additional language would have a positive impact on children's brains. As to the children, they believe that Turkish is part of their identity and a means of communication with Turkish relatives in Turkey. Furthermore, the results from Gomaa's (2011) study (see section 2.3.2) indicate that parents believed that it was very important for them to transmit Egyptian Arabic to their children for two reasons. First, they considered Egyptian Arabic to be a means to passing on their Egyptian culture and traditions.

Hence maintain the children's Egyptian identity. Second, they believed that speaking Egyptian Arabic would allow their children to communicate with their grandparents and extended family living in Egypt.

It seems that there is a link between language and ethnic identity which Phinney (2003) defines as «a dynamic, multidimensional construct that refers to one's identity or sense of self as a member of an ethnic group» (p. 63). Indeed, ethnic-group membership is often based on which language one speaks (Trudgill, 2000; Padilla & Borsato, 2010). For example, Trudgill (2000, p. 54) gives two cases in which ethnic-group membership is defined by language. The first case is found in one suburb in Accra in Ghana where there are more than eighty languages the native speakers of which often identify themselves as being members of a tribe or group on the basis of which these languages is their mother tongue despite the fact that most of them are bi- or tri-lingual. The second case is found in Canada where the two ethnic groups, namely English Canadians and French Canadians, are defined mainly by whether their mother tongue is English or French. In addition, using the language of one's ethnic group can serve as an indication of the involvement in the social life and cultural practices of that group (Phinney, 1990, p. 505). Moreover, language can be viewed as an essential component of ethnic identity. Actually, socialization, the process by which a child acquires appropriate behavioral alternatives, takes place through the language used within the family and the community, which makes language an essential constituent of ethnic identity (Padilla & Borsato, 2010, pp. 11-12). Furthermore, language can be a marker of one's ethnic group (Mesthrie & Tabouret-Keller 2001; Kotzé, 2001; Ennaji, 2012). Viewing language as a marker of ethnic identity can contribute to its maintenance. As Holmes (2013, p. 64) argues, a minority language is likely to be maintained longer in areas where it is considered as an important symbol of ethnic identity. This is the case of Polish and Greek people immigrants

in many countries regarded their languages as a primary condition for preserving their identity. Consequently, they have maintained their languages for three to four generations. The same is true for Greek immigrants in USA, Australia, and New Zealand. In the same vein, O'Rourke (2011) states that «the strength of a minority language can be predicted by the degree to which speakers value their language as a symbol of group or ethnic identity» (p. 19).

The close relationship between language and ethnic identity is confirmed by results from several studies. Hatoss (2005) studied LM among 50 families in the Hungarian community of Brisbane, Australia, using questionnaires and telephone interviews. The sample for the study was made up of 50 families. The findings suggest that most of the participants (93%) demonstrated a strong attachment to Hungarian and regarded it as an important tool for maintaining their identity. Additionally, Chiung (2001) investigated the relationship between mother tongue and ethnic identity among 244 Taiwanese students in Taiwan. These students were speakers of different native Taiwanese languages. Since Mandarin Chinese was the official language of Taiwan for more than 50 years, there was a decline in the use of Taiwanese native languages. The Results indicate that the maintenance of one's ethnic language is a contributing factor to the maintenance of one's ethnic identity while the erosion of one's ethnic language does not necessarily result in the erosion of one's ethnic identity.

However, findings from other studies indicate that there is no tight linkage between language and ethnic identity. For example, Ahn (2008) studied ethnic identity and language use among 20 United States-born Koreans living in Hawaii. The results show that none of the participants identified themselves as Americans even though English was their first language. In addition, Kang (2004) investigated the relationship between heritage language maintenance and ethnic identity among 18 Korean and Chinese immigrant college students (eight Chinese and ten Korean) in New Jersey, the United States. The findings from the study

reveal that the Chinese and Korean participants did not consider their heritage languages to be markers of their ethnic identities. Moreover, Bentahila and Davies (1992) in their study of LS in Morocco conclude that there is an unnecessary relationship between language and ethnic identity.

Government policies in favor of a minority language can contribute to its maintenance. The following are examples of maintenance efforts undertaken by governments in three different countries: Morocco, Australia, and New Zealand. To start with, Sadiqi (2010, p. 35) emphasizes the support that AL receives from the Moroccan government. In fact, she states that the establishment of the Royal Institute for Amazigh Culure (Institue Royale pour la Culture Amazighe en Maroc) in 2001 in Rabat marked the beginning of the process of the institutionalization of AL. She also states that AL has been included in Morocco's public school curriculum as a result of an agreement signed between the Institute and the Ministry of Education in 2003. In the Australian context, Wurm (1974, p. 217) mentions that the Australian government established the Australian Institute of Aboriginal Studies in Canberra in 1961 for the purpose of promoting research on aspects of aboriginal culture in the widest sense. He adds that the main interest of the institute has been in linguistics, which has resulted in the rediscovery of many languages and in government's positive attitudes towards and greater interest in aboriginal languages. As to the New Zealand context, Caldas (2013, p. 361) emphasizes the efforts of the New Zealand government to maintain the indigenous Maori language. Such endeavors include funding the establishment of the Maori Language Commission (1987), Maori radio stations, and the Maori Television Service (2003), the compilation of a Maori dictionary, the development of Maori school immersion and bilingual programs, and the tuition to attend university courses offered in Maori. All these efforts aimed at promoting community and, especially, family use of Maori. Indeed,

many Maori reported learning the language from taking classes, watching TV, or listening to radio.

The influence of education on LM is clearly evidenced in the literature on the topic. Li (1982, p. 11) points out that there is often a connection between parents' educational level and LM in the sense that educated parents would be more aware of maintaining the mother tongue. Similarly, Clyne (1988, p. 31) indicates that parents who are educated may contribute to LM by devoting time and efforts to the maintenance of their minority group's heritage and culture. According to Dorian (1987, p. 65), Irish Gaelic has been maintained because it is accepted as a legitimate language of study in Ireland, whereas in Scotland the status of Gaelic has been eroded because it is generally not available as a school subject option. He states that:

Irish school children ... are most unlikely to be denied the opportunity to study Irish if

they wish it; over most of Highland Scotland, school children and their parents are still

told either that there are no teachers available to teach Gaelic or that there is no room in

A related point, in her study of Greek in Australia, Tamis (1990) points out that including Greek in the educational system leads to a corresponding increase in its social value and/or acceptability. She states that:

The introduction of MG (mother tongue Greek) as an examinable subject for tertiary entrance requirements in 1973 and its teaching in certain Australian Universities and Colleges of Education ... further increased not only the functional value of the mother tongue, but also stipulated the acceptability of MG in the Greek community. (p.498)

Religion may be an important factor affecting LM. Actually, church services and activities are often, or can be, conducted in the mother tongue of a minority group. For example, in his study of Greek-Australians, Smolicz (1985, p. 26) indicates that Greek is used in the Greek Orthodox Church in Australia. In the same vein, Sridhar (1988, p. 80) points out the importance of religion for LM through ritual and prayer. Putz (1991), in his study of a German Australian speech community, indicates that "membership of a religious denomination ... seems to promote language maintenance which, in turn, underlines the importance of a combination of domains, i.e. religion, ethnicity and language" (p. 487).

It appears that Smolicz (1985), Sridhar (1988), and Putz (1991) share the view that religion contributes to LM. Such view is different from that of Fishman (1972). Indeed, Fishman (1972) suggests that religion substitutes for and preserves ethnicity but it is not sufficient for LM. He also suggests that the function of language in the church is primarily to safeguard the faith of the people in an urban situation, and therefore, the use of the vernacular in the mass is not sufficient for LM.

In the case of AL, religion has caused a negative impact on the maintenance of this language. In fact, the close relationship between Arabic and Islam has facilitated the dominance of Arabic over AL (Ennaji, 2005, p. 10). Besides, the spread of Arabic throughout North Africa has been due the necessity of using Arabic in praying (Sadiqi, 1997, p. 9). Furthermore, one should be literate in Arabic in order to understand the Quran (Ennaji, 20025, p. 10). In a nutshell, Arabic is viewed as the language of Islam.

The media can be a key domain in LM. Putz (1991, p. 484) points out that German television in Australia promotes interest in contemporary German culture and language, resulting in an increase in the use of the mother tongue, with specific reference to the younger

generation. Resnick (1988, p. 93) states that it is common for immigrant groups to broadcast or publish in their language. Government support for the media could also assist in language and culture maintenance. An example of such support is highlighted in Cartwright's (1987) study of the Anglophone minority in Quebec, where book publishing, theatrical, radio and television production, and regional newspapers and magazines have received government support.

Socio-cultural organizations, including clubs and associations, may contribute to LM. Clyne (1988) and Putz (1991), in their studies of LM among German-Australian communities, conclude that not only do social clubs increase LM, but also they are, in the words of Putz (1991), "one of the most important domains promoting language maintenance in ethnic communities" (p.487). Fishman (1972, p. 49) stated that the role of cultural organizations in LM is more important than the press or broadcasting. Indeed, McCarty (2012, p. 561) indicates that such American organizations as the Advocates for Indigenous California Language Survival, the American Indian Language Development Institute, the Indigenous Language Institute, the National Alliance to Save Native Languages, and the Stabilizing Indigenous Languages Symposium played a major role in the development and passage of the 1990/1992 Native American Languages Act (NALA) which stipulates that it is the federal government's responsibility to promote the rights of indigenous Americans, including the use of their languages is schools. In 2006, NALA was reinforced by the Esther Martinez Native American Languages Preservation Act which provides financial support to language nest preschools, native language survival schools, teacher preparation, and materials development.

The degree of LM can be higher in rurally isolated communities in contrast to urban communities where people are in contact with other languages. Thompson (1974, p. 10) points out that urbanization leads to language shift. Indeed, in his study of the 1970 census in the United States of America, Thompson (1974, p. 14) identifies two important trends regarding the childhood residence of Mexicans: The predominant use of Spanish was evident among the first generation, who came from a rural background (or childhood) while the use of Spanish decreased in the second and third generations, as the number of people from a rural background or childhood decreased. Note here that the shift from Spanish to English was related to the shift from rural areas to urban areas. Likewise, in her study of language maintenance and contraction among two groups of Tashelhit (a variety of AL) speakers in southern Morocco, Hoffman (2006, p. 150) reports that the Native Tashelhit-speaking men who have migrated with their wives and children to the towns in the 1970s have become bilingual in Arabic, and their children grow up as Arabic-dominant speakers; many of them even go so far as to reject their ethnic identity. In the same vein, Srivistava (1989) points out that an agrarian and rural society would be more supportive of the minority language than an industrialized and urban society where the co-existence of a number of speech communities results «in a state of constant competition and conflict in learning each other's language» (p. 15). Similarly, Sadiqi (1997, p. 15) in her study of AL in Morocco indicates that AL regression in rural areas is slower than in the urban areas because no language competes with it in those areas. In a similar vein, Holmes (2013, p. 66) believes that «resistance to language shift tends to be last longer in rural than in urban areas» (p. 66). She cites the example of Ukrainians in Canada. Indeed, Ukranian has been maintained better among those living on farms than those living in cities. Also, in New Zealand, Maori is still used as a language of daily communication in rural areas where most of the inhabitants are Maori.

The aforementioned view is different from that held by Fishman (1964). Indeed, he argues that LM efforts are greatest in the city. He states that:

Language revival movements, language loyalty movements, and organized language maintenance efforts have commonly originated and had their greatest impact in the cities. Intelligentsia and middle class elements, both of which are almost exclusively urban, have frequently been the prime movers of language maintenance in those societies which possess both rural and urban populations. (p. 53)

I have so far dealt with some of the factors contributing to minority LM. In the next section, I will focus on factors promoting LS.

A number of factors contributing to LS are revealed in the literature: family, social factors, economic factors, government policy, and media. These factors will be discussed below.

We have seen in section 2.3.2 that the family is a factor which strongly promotes LM. Nevertheless, it could cause LS. Parents in particular may facilitate LS. In fact, in her study of Welsh, Lewis (1975) points out that "the relation of the child's language to that of the family is naturally determined to a considerable extent by expressed or revealed competence of the parents" (p. 107). She also indicates that, in certain instances, where both parents are bilingual, there is a high degree of English monolingualism among the children. Moreover, the results from Bentahila and Davies' (1992) study reveal that some Amazigh Moroccan parents encouraged their children to use Arabic rather than Amazigh because the latter would not help them earn their «daily bread». Similarly, in her study of Spanish maintenance and linguistic shift towards English among Chicano adolescents in two speech communities in Austin, Texas, USA, Galindo (1991, p. 113) found that a crucial factor in the shift from Spanich to English was parents' decision not to teach Spanish to their children. This had to

do with parents' negative school experiences. Indeed, parents, including Galindo's, still remembered when they were punished for speaking Spanish school and made to feel inferior. Thus, these parents did want their children to be discriminated against because of their language and culture.

In addition to parents, older siblings may also influence the degree of LS among the young. In other words, when the older siblings use their mother tongue, which has been influenced by the majority language, to communicate with the younger siblings, they facilitate LS. Indeed, Sridhar (1988) states that «the presence of older siblings whose language is increasingly affected by the mainstream language may make the younger child's control over the ethnic tongue less secure» (p. 81).

Another way in which the family can promote LS is through intermarriages. Demos (1988, p. 170) points out that intermarriage causes a decline in the use of the mother tongue. Furthermore, Paulston (1987, p. 35) indicates that in cases of intermarriage, there is a shift of one partner to the language of the more socio-economically favorable group. In the same vein, Giles et al. (1977, p. 314) suggest that the increase in the rate of mixed-marriages may lead to LS in the sense that the more prestigious variety is more likely to be used at home, including the interaction with children. Moreover, Holmes (2013, p.62) argues that intermarriages can accelerate LS in contexts where monolingualism is the norm. A typical example of this is the German community in Australia. Indeed, in intermarriage between a German-speaking man and an English-speaking woman, English is often used at home and to children. Also, in the Cherokee community in Oklahoma, USA, when a Cherokee-speaking person marries outside the community, the children grow up to be monolingual in English.

The shift towards another language may be due to the prestige or social value associated with that language. That is, people choose a language due to the social value associated with

it. Mougeon, Beniak, and Valois (1985) state that «the language being shifted to enjoys wider currency and greater prestige» (p. 455). In her study of language attitudes in Barcelona, Woolard (1984) indicates that the shift from Castilian to Catalan is due to the fact that the latter is more prestigious than the former. Tamis (1990) points out that «the acquisition and learning of MG (mother tongue Greek) is further inhibited or promoted by the attitudes of the overall society and the prestige that it carries amongst teachers and educators" (p. 498).

Language shift can be the result of the length of residence in an area where a language other than the mother tongue is spoken. In his study of Japanese in the USA, Okamura- Bichard (1985, pp. 82-83) indicates that the length of residence can lead to language shift in the sense that regression in the mother tongue is related to the period of time spent outside Japan. Huls and van de Mond (1992), in their study of two Turkish families living in the Netherlands, conclude that the more the length of residence is, the more is the shift from Turkish to Dutch. 2.4.4. Access

Access can be a factor of LS. As Cartwright (1987, p. 204) states, the main reason behind learning the language of the dominant society is that it provides access to certain areas, important in everyday life such as residence, employment, and daily shopping. In a similar vein, Srivastava (1989, p. 22) points out that immigrants, in the USA, switch to English because it is English which determines access to goods and services.

Employment can play a major role in LS in the sense that people should be able to speak the language of the working environment in order to be employed, especially in the more skilled areas of employment. This factor has been widely-studied in the literature. A general pattern amongst linguistic minorities is the use of the mother tongue at home and the use of

the dominant language in the working environment. The results from Gynan's (1985) study of Spanish in the USA, for example, reveal that Spanish (the mother tongue) is spoken at home while English (the language of the host society) is spoken in the working environment. In the same way, Paulston (1987, p. 53) points out that immigrants may have to learn a language for purposes of employment and this has proven to be a major cause of LS.

The importance of speaking the language of the working environment in order to be employed, and the subsequent language shift which occurs, is discussed in detail by Lieberson and Curry (1971, pp. 131-133). This study emphasizes the fact that immigrants, in the USA, who are not able to speak English are "handicapped" (p. 132) and could even be discriminated against if they are unable to speak English. This research suggests that the language shift due to employment could also «influence the possibility of mother tongue shift between the generations» (p.133). In addition, Cartwright (1987), in his study of language usage in Quebec, points out that in the working environment the younger Francophones use English more often (as opposed to French) in comparison to the older respondents. It can, therefore, be seen that occupation promotes the use of language of the host community.

Furthermore, Schlieben-Lange (1977) and Eckert (1980) in their studies of Occitan in France indicate that most of the younger members of the Occitan community are monolingual French speakers. Eckert, for example, states that «the adult population of the community is consciously transitional - they have encouraged their children to leave the region to find work, and in preparation for this they have raised them as monolingual French speakers» (p. 1059).

Putz (1991, p. 486) summarizes the facts mentioned above in his study of German in Australia. He indicates that the fact that work environment is a meeting place of people from various backgrounds necessitates the need for a lingua franca, which facilitates LS. He said that:

The work domain seems to be a rather negative factor in the promotion of language maintenance; the multi-ethnic situation at the work-place and the prevalence of people from different language backgrounds creates a need for a lingua franca, i.e. English, in order to facilitate communication among workers. (p.486)

The pattern of LS can be accelerated by moving either to another area of the same country or to another country in search of better living conditions. Giles et al. (1977) state that «migrants who move in an area where linguistic groups are in overt or covert competition appear to be willing (for obvious economic reasons) to adopt the language and culture of the dominant rather than that of the subordinate linguistic group» (p. 314). According to Giles et al (1977, p. 315), this has been the case of Welsh speaking families who had moved at the turn of the nineteenth century to England or to the English-dominated and industrially- developed areas of south Wales in order to find jobs. These families had lost the Welsh language within less than two generations. Pencheon (1983, p. 31) highlights the role of migration from villages to cities in the regression of AL in Tunisia. Similarly, Ennaji (1997; 2005) points out that the regression of AL in Morocco was due in part to the migration of Moroccan Amazighophones from rural villages to urban areas. Likewise, Hoffman (2006, p.

150) reports that Moroccan native speakers of Amazigh who have migrated permanently to cities have shifted towards Moroccan Arabic.

Government policy towards minority groups and their languages can promote LM (see section 2.3.5). But, it can also facilitate LS. Miyawaki (1992, pp. 358-363), in his study of Ainu, Okinawan, and Korean linguistic minorities in Japan, indicates the unfavorable official attitudes towards these communities. To begin with, the Ainu population numbers about 25.000, but only ten people are able to speak the Ainu language fluently and these are over 70

years old. The decline of the Ainu language was due to the assimilative language policy of the Japanese government. An example of such policy is the prohibition to teach Ainu language at school to Ainu children. Like the Ainu language, the Okinawan dialect has declined as a result of the government's assimilation policies such as the penalization of the Okinawan children who speak their language at school. Such decline was manifested in the fact that only Okinawan people over 50 years old can speak their dialect. The situation of the Korean language is not so different from that of the Ainu and Okinawan languages. The Japanese government assimilated 80% of the 642 Korean schools, established during the Korean War (1950-1953) to the regular Japanese schools and caused many others to collapse, which had a destructive effect on the maintenance of the Korean language.

We have seen in section 2.3.8 that the media can greatly promote LM. Nevertheless, it can contribute to the process of LS. Maamouri (1983, p. 14) points out that AL regression in Tunisia is due in part to the general spread of mass media. Young (1988, p. 325) indicates that the media can contribute to the process of LS especially when there is government support for the use of a particular language in the media such as television. In their study of language shift in Morocco, Bentahila and Davies (1992, p. 198) state that the mass media especially television, is one of the factors contributing to the regression of the AL in Morocco.

As we have seen in section 2.3.6, education can contribute to LM. However, access to education can lead to LS. Pencheon (198, p. 31) suggests that the access of Amazigh girls to education in the aftermath of the independence had contributed to the regression of AL in Tunisia in the sense that it had led to the disappearance of monolingualism among Amazigh females. In the same vein, Maamouri (1983a) stated that the rapid development of the

educational system in Tunisia had contributed to the regression of AL. Clyne (1988, p. 31) points out that the lack of education or lower education leads to isolation from the dominant host community and hence language maintenance. This implies that education facilitates LS. Ennaji (1997, p. 29) indicates that the access of Amazigh-speaking children to free education since Moroccan independence had fostered the regression of AL in Morocco.

This chapter started with showing the importance of LM. Besides, it provided a description of two models of LM and LS factors, which places the study within the research on LM and LS in general, and cited five publications which had shed some light on the situation of AL in Tunisia. Then, ten factors of LM were referred to, namely geographic concentration of speakers, family, language attitudes of speakers, aspects of the language- identity relationship, government policy, education, religion, the media, socio-cultural organizations, and urban-rural nature of setting. Some of the factors contributing to LS were also dealt with. These consisted of family, prestige, length of residence, access, employment, migration, government policy, the media, and education. It is should be noted that factors such as family, education, government policy, and the media are ambivalent in the sense that they can facilitate either LM or LS, which denotes the close relationship between these two terms. It is predicted that among the ten factors of LM mentioned earlier, only four can contribute to the maintenance of AL in Zrawa, namely geographic concentration, family, language attitudes of speakers, and aspects of language-identity relationship.

It would be better at the end of this chapter to recall the four research questions:

The following chapter provides a detailed description of the methodology employed in the present study as an attempt to find answers to these questions.

This chapter presents the methodology used in the study. It begins with explaining the methodological approach of the study. Next, it describes the participants and the data collection instruments employed in the study. It also provides a description of the procedures that the researcher followed in order to collect and analyze the data.

Before providing a description of the methodological approach used in the study, it is important to talk a little bit about the methodological approaches employed in previous studies on LM and LS. Indeed, researchers on LM and LS used a variety of methods: quantitative, qualitative or a combination of both. Quantitative methods examining LM consist mainly of census data analysis (e.g., Lieberson & Curry, 1971; Thompson, 1974) and survey questionnaires (e.g., Allard & Landry, 1992; Putz, 1991; Young, 1988; Okamura- Bichard, 1985).

Qualitative methods such as participant observation (e.g., Tandefelt, 1992), case studies (e.g., Prince, 1987; Paulston, 1987; Miyawaki, 1992), and interviews (e.g., Boeschoten, 1992; Tandefelt, 1992; Zhang, 2008) have been used to investigate language maintenance and shift.

Language maintenance and shift has also been explored using a mixed method approach, that is, a combination of both quantitative and qualitative methods. For example, Bentahila and Davies (1992) used questionnaires and interviews. Clyne (1988) employed taped interviews, questionnaires, study of institutions, and analysis of census statistics.

As to this study, it is a mixed methods research defined by Creswell (2008) as a «research in which the inquirer or investigator collects and analyzes data, integrates the findings, and draws inferences using both qualitative and quantitative approaches or methods in a single study or a program of study» (p. 526). In fact, I used a combination of quantitative and

qualitative methods, which resulted in having quantitative and qualitative data. Thus, quantitative data were elicited by survey questionnaires. However, qualitative data were generated by participant observation and semi-structured interviews.

The reason for using such research is that the use of both qualitative and quantitative approaches will provide a more complete understanding of the four factors contributing to AL maintenance in Zrawa. Indeed, the survey questionnaires gathered a great deal of unobservable data which had to do with the Amazigh participants' language attitudes, self- reported language use, and perceptions of the relationship between language and ethnic identity. The semi-structured interviews generated in-depth data concerning language attitudes, perceptions of the link between language and ethnic identity, and the role of the family in LM. The reason behind the use of participant observation was to collect data about the Zrawa Amazigh community in terms of actual language use and geographic concentration.

As stated by Creswell (2007), using a mixed methods approach involves collecting both quantitative and qualitative data at the same time and merging the data to form one interpretation. As a matter of fact, the quantitative and qualitative data for the study were collected concurrently during my ten visits to the village of Zrawa in the month of February of the year 2017, with each visit taking approximately eight hours. These data will be merged in order to answer the research questions.

This section provides an explanation of the sampling methods used in choosing the study participants. It also identifies the characteristics of these participants and reveals the access problems the researcher faced during their selection. It should be mentioned here that in accordance with the code of ethics in social sciences, the participants were informed about

the aim of the study and their real names were replaced by pseudonyms (Christians, 2005). Also, the names of the places were changed.

Sampling techniques were used to select not only the persons who filled out the questionnaire but also those who responded to the interview questions. Thus, the selection of the respondents to the questionnaire was done using two kinds of sampling: convenient sampling, which, according to Saumure and Given (2008), means that «individuals who are the most ready, willing, and able to participate in the study are the ones who are selected to participate» (p.124), and snowball sampling, which, as Morgan (2008) states, «uses a small pool of initial informants to nominate other participants who meet the eligibility criteria for a study» (p.815). Convenient sampling resulted in having 14 participants. Snowball sampling had to do with asking six participants (four males and two females) to recruit other people to take part in the study. As a consequence, the number of participants reached a total of 53 Amazigh persons. As for the respondents to the interview questions, they were drawn from the ones who filled out the questionnaire on the basis of convenient sampling. Indeed, 11 individuals, namely three single young adults (two males and one female), five male parents, and three grandparents, accepted to be interviewed.

Due to the difficulty in having access to Amazigh women, it was necessary to ask participants to administer the questionnaire to women from their social networks. Actually, two participants, namely a man and a young woman, hand- distributed the questionnaire to 19 females. I had direct access only to five female participants. The uneasiness to have access to women explains the small number of females with whom semi-structured interviews were conducted. Also, 16 Amazigh persons, including two females, refused to take part in the study.

The participants represented various age groups, various occupations, and both genders (see Appendices C and D). Concerning participants' gender, there were 29 males and 24 females. As to age, the participants belonged to different age groups. Indeed, the age of male participants was between 15 and 82 years whereas the age of the female participants ranged from 13 to 90 years. As to occupation, the participants were from different occupations.

All the participants reported that their first language was AL. Besides, 51 among them reported that they were AL-TA bilinguals. A 90-year-old female participant reported that she knew some TA. However, a 66-year-old female reported that she was monolingual in AL. 3.3. Description of Data Collection Instruments

This section presents the techniques employed in the data collection process, namely survey, interviewing, and observation.

Participant observation was carried out by the researcher. Basically, it was used to discover the language used within the Zrawa Amazigh community and to know about the geographic concentration of this community. Less importantly, it was employed to generate background data like the social, economic, and religious characteristics of Zrawi Imazighen.

The questionnaire (see Appendix A) was made up of 15 items designed to collect information about language background, language use and attitudes, and perception of the link between language and ethnic identity. It was divided into three sections:

1. The first section included 6 questions meant to generate data about personal and language background. As to the source of the questions, the first three questions were the researcher's own creation, the fourth and fifth questions were adapted from Mah (2005), and the last question was taken from Næssan, Monaghan, and Mühlhäusler (2010).

Ten interview questions were designed (see Appendix E). These were of one type and were related to various topics.

All the interview questions were open-ended (e.g. «Whose responsibility is it to keep the Amazigh language alive?). Two questions among these, namely «Do you speak only Amazigh to your children?» and «Do you ask your children to speak the Amazigh language at home or around Amazigh people?» were subject to variation depending on the interviewee's age and marital status.

The interviewees were asked questions related to various topics. Actually, questions were about the role of the family (e.g. «Do you speak only Amazigh to your parents?»), language attitudes (e.g. «How important is the Amazigh language for you?), and the linkage between Amazigh language and identity («How do you see the link between Amazigh language and identity?»).

This section specifies the way the three data collection methods, mentioned in the previous section, were employed.

The participant observation involved the direct observation of linguistic behavior, the informal interviewing of some of the participants with whom I had direct contact, and connecting with as many Amazigh persons as possible. Making Amazigh contacts helped me have access to data about the unobservable characteristics of Imazighen and to other Imazighen. At the end of each visit, data were recorded in the form of field notes written up in a research log. These data were used in the reporting of the findings from participant observation.

The questionnaire was administered to 53 persons. Twenty-seven questionnaires were filled in my presence but a male teacher and a female graduate, who were part of my Amazigh contacts in Zrawa, hand-distributed 11 and 15 questionnaires separately to people from their social networks.

The questionnaire was initially written in English then translated into Standard Arabic before being administered. It took about 7 minutes to be filled out. The researcher, the male teacher, and the female graduate informed the questionnaire respondents about the purpose of the study. The participants received no financial incentives.

Eleven among the participants were interviewed after filling out the questionnaire. The 11 semi-structured interviews were conducted during six days at different points of time and in various stress-free settings depending on the interviewees' availability (see Appendix D). The interview questions (see Appendix E) were initially written in English then translated

into Standard Arabic given. Nevertheless, the interviews were conducted in TA in order to make them sound like casual conversations and to make the interviewees feel at ease.

The interviews were audio-recorded using a smart phone then transcribed verbatim. Indeed, pauses, grunts, such as «mmm», unfinished words, inaudible material, nonverbal communication, like laughter, were included in the form of comments (see Appendix H). Then, the most important excerpts from the interviewees' statements, which had to do with several themes addressed in the interview guide, were translated into English in order to be used as quotes when reporting the results from the interviews.

This section represents the techniques employed in the analysis of qualitative data obtained from participant observation and the 11 interviews and quantitative data obtained from the 53 questionnaires.

To analyze the data from participant observation and semi-structured interviews, I employed an analytic method known as the qualitative or latent content analysis (Julien, 2008). This technique was used to derive meaning from the field notes and interview transcripts.

Using qualitative content analysis, field notes and interview transcripts were analyzed for content related to themes which had been set beforehand. As for data from interview transcripts, it was coded by identifying participants' expressions and statements related to the following themes: language acquisition, the importance of AL, the maintenance of AL, the linkage between Amazigh language and identity, and language attitudes. Relevant participants' expressions and statements were used as quotes to support ideas in the final report. As to field notes, they were analyzed for content related to four themes: linguistic behavior, social life, economic activities, and religious affiliation.

Analysis of quantitative data, that is questionnaire data, involved data screening, data coding, and data description. Data screening means making certain that (a) responses are legible and understandable, (b) responses are complete, and (c) all of the necessary information, such as participant's gender, age, occupation, and first language, had been included. As for data coding, responses were transformed into numerical codes then written in the form a codebook in order to be calculated. As to data description, it was done using descriptive statistics which involved the calculation of percentages for responses to the statements and questions making up the second and third sections of the questionnaire, respectively.

This chapter introduced the research approach adopted in the study. Besides, it described the participants and the data collection instruments. Next, it accounted for the procedures of data collection and analysis. The following chapter will include an analysis and discussion of the findings from participant observation, questionnaire, and interview data.

This chapter merges the results of the analysis of data generated by the questionnaire, the semi-structured interviews, and the participant observation in order to answer the research questions. Thus, it consists of four sections, each dealing with one of the four factors which are suggested to promote AL maintenance in Zrawa, namely family, geographic concentration, language attitudes, and perceived language-identity link. It also provides a discussion of these results in the light of the literature in the field.

Results reveal that the Zrawa Amazigh community is geographically concentrated. The first thing which indicates is the use of AL in the streets by most people. Indeed, when I visited Zrawa for the first time, I felt as if I were in a foreign country because I heard people speaking a language I am not accustomed with. As Figure 4.1 below shows, Zrawa buildings are geographically concentrated. Imazighen, according to informants, constitute most of

Zrawa inhabitants. This implies that Amazigh families are geographically concentrated. It is worth noting that these informants told me that while the total number of Zrawa Amazigh population ranges from 3000 to 5000 people, the number of «Arabs», who have moved from Al-Hamma and Dhiba to reside or work in Zrawa, does not exceed 56 persons. This denotes that Imazighen are the indigenous inhabitants of Zrawa. All this makes of AL the majority/dominant and the indigenous language of Zrawa. This favorable condition led Alaa (a parent) to state that:

Migration to cities and marrying outside the Amazigh community can lead to the loss of the Amazigh language. But as long as we live in this village it won't. This village is the secret behind the survival of this language. The existence of the village of Zrawa has

In this extract Alaa emphasizes the role of migration to cities, where TA is the dominant language, and inter-marriages in AL loss. This suggests that living in Zrawa, which is a village where AL is the majority language, and marrying within the Amazigh community helps maintain AL. There is evidence from the literature in the field (of LM and LS) that these two factors affect LM. Emphasizing the effect of inter-marriages on language maintenance, Demos (1988) points out that intermarriage causes the regression of the mother tongue. In the same vein, Paulston (1987) and Giles et al. (1977) indicate that inter-marriages can lead to language shift towards the more socio-economically prestigious language. The role of migration from rural areas to cities in the shift from AL towards Arabic was highlighted by Pencheon (1983), Hoffman (2006), and Ennaji (1997) (see section 2.4.6 for more details).

Being geographically concentrated, the Zrawi Imazighen have the chance to meet on a daily basis and, as I observed, this is what they actually do. From participant observation I learnt that Amazigh men are likely to meet each other in the streets, in shops, at the post

office, at the mosque, at the local hospital, at the workplace (e.g., construction fields), at souks, including the souk of Zrawa, and in the cafés. As for the women, they may meet in the streets, at shops, at the post office, at the nursery and primary schools (when accompanying their children), at the local hospital, and at the souk of Zrawa, as well as other souks. As to children (from three to eighteen years), the males among them may meet each other in the street, at the youth club (for teenagers) and at schools (nursery, primary, preparatory or secondary) while the females are likely to meet in the street and at schools. It should be noted here that there is no mixing between males and females due to the conservative aspect of the village.

During all these in-group encounters between Zrawi Imazighen, the only language used is AL. For example, when I first went to a local cultural association I found Mohamed, a member of the association, conversing with a young woman in Amazigh. Another example, Amazigh was used in interactions between Mansour, a post office employee and the public. In fact, when I walked into the post office for the first time, I found the post office employee speaking to an old woman in Amazigh. A further example, when strolling down the village streets I usually encountered people, either men or women, conversing with each other in Amazigh. A fourth example, whenever I passed by the Zrawa primary school, I heard children speaking Amazigh. A fifth example, observing language use in a shop and a library, I noticed that Imazighen used Amazigh to each other. A sixth example, Amazigh secondary school students usually met at the bus station and on the bus and used AL as a language of communication. Indeed, I took that bus ten times and the language I heard students using is Amazigh. Intersestingly, it happened that I witnessed a conversation between students: one was using Amazigh and the other was using TA. A seventh example, many times did I encounter children speaking to each other in Amazigh. An eighth example, I accompanied three Amazigh young men to their workplace (a construction field). They used Amazigh to

each. A tenth example, AL is used in cafés. In fact, I had coffee one time in Café Ennakhil, one time in Café Eljanoub, and seven times in Café Ennasr (note that these names are unreal). In all these times, I usually heard people conversing in Amazigh. An eleventh example, I witnessed a converstion between Mohamed and a four-year-old child and another one between him and a two-year-old girl. Both conversations were conducted in Amazigh. A twelfth example, as I observed, Amazigh was used in interactions between teenagers at the youth club. A final example, as I usually went from New Matmata to Zrawa by taxi, sometimes I heard some passengers in the car chatting to each other in Amazigh.

What can be concluded from all these examples is that AL is the language of communication within the Zrawa Amazigh community. Such conclusion is supported by the fact that all the respondents to the questionnaire reported that they use Amazigh with anyone who speaks it other than the family members. It is also supported by data from the interview with Alaa (a parent). Indeed, Alaa asserted that «within ... the Amazigh community, like the village of Zrawa, the language of communication is the Amazigh language». He even went as far as to claim that «the language of communication between Imazighen in Tunisia is exclusively the Amazigh language.»

As for inter-group encounters between Imazighen and, in the terms of informants, «Arabs», I noticed that Imazighen often converged with the interlocutor in the sense that they switched to TA. Indeed, all the Zrawi Amazigh persons I communicated with, including those who took part in the study, used TA when speaking to me. Another example, during my sixth visit to the village of Zrawa, I met Alaa and went with him to Café Ennasr. He joined a group of about six people sitting around a table. I expected him to use Amazigh but he used TA. When I asked him about the reason behind using TA instead of AL, he told me that there had been two «Arabs» among the six men he had joined. Moreover, I noticed that when a group of Imazighen is joined by a person whom they do not know or they know that he is not an AL

If someone talks to me in Tunisian Arabic, I should answer him in Tunisian Arabic

speaker, they automatically switch to TA. For example, when I first went to Café Ennasr, I told the waiter about the research I was carrying out. Consequently, he invited me to have a cup of coffee with him and five other Imazighen. The five men had been conversing in Amazigh, but as soon as I joined them they were courteous enough to switch to TA. A further example, during one of these visits to Zrawa, I met five teenagers at the square where the souk is held every Sunday in order to ask them to take part in the study. They had been chatting to each other in Amazigh. However, they switched to TA as soon as I approached them. Furthermore, as Figure 4.2 below indicates, 94 % of the questionnaire respondents that they switch to TA whenever they are addressed in TA. This confirms Bradley's (2013) statement that «one of the basic principles of human interaction is that there is accommodation to the speech of the interlocutor, that is, one adjusts to the speech repertoire, abilities, and preferences of the people one is speaking to» (p. 5). Salah (a grandparent) summarized the rule governing language use by stating that: «We use Amazigh with those who speak Amazigh and Arabic with those who speak Arabic.» It should be noted here that rarely do the Imazighen switch to TA when being in Zrawa because they have limited contact with Arabs, given the fact that they form the great majority of Zrawa inhabitants.

However, it happened once that Alaa, a participant who became a close friend, did spoke to me in Amazigh in order to make fun of me. Additionally, some Imazighen did use Amazigh in my presence. For instance, during my seventh visit to Zrawa, I met Ali (a baker), Alaa, Mansour (a post office employee), and another Amazigh man near the post office. All of them spoke to me in TA, but when addressing each other they often switched to Amazigh. Another example, when I visited Zrawa for the sixth time, I went with Alaa to a youth club. We met a teenager. Alaa addressed him in Amazigh in my presence. Using Amazigh in my presence raises the following question: Did they consciously exclude me, or did they do so in service of solidarity and maintenance? An explanation for this is provided by Mahdi (a parent) who stated that: «When being at a mixed encounter between Imazighen and Arabs, Imazighen normally use Tunisian Arabic but unconsciously they do speak Amazigh to each other because it is the language they are accustomed to using in their interactions.»

In a nutshell, the geographic concentration of Zrawa Amazigh community facilitates the maintenance of AL as it allows the community members to meet on a daily basis and to use Amazigh as the language of communication in such frequent encounters. As Giles et al. (1977) indicate, the concentration of a minority ethnic group in a given geographic area, whether it is a region, a territory, or a country, fosters the maintenance of the group's language by virtue of the fact that the group members have the opportunity to use their language as a means of daily communication. In the same vein, Holmes (2013) argues that the geographic concentration of the speakers of a minority language contributes to the maintenance of that language through its daily use. Findings from Li's (1982), Laleko's (2013), and Al-Khatib and Al-Ali's (2010) studies (see section 2.3.1) confirm the role of geographic concentration in the maintenance of minority languages.

We have just seen that the geographic concentration of the Zrawa Amazigh community is a crucial factor AL maintenance in the sense that it promotes its use as a means of daily

communication within the Amazigh community. Given the fact that the Amazigh families represent the cornerstone for the Zrawa Amazigh community, it is necessary to deal with the efforts of these families, if ever, to maintain AL.

Data from questionnaires, semi-structured interviews, and participant observation indicate that the Amazigh families play an essential role in AL maintenance in Zrawa. To begin with, in response to the questionnaire item «Whom do you use Amazigh with?» (see Appendix A), all participants reported that they use Amazigh with their family members, namely grandparents, parents, children, aunts, uncles, siblings, cousins, nephews, and nieces. This implies that AL is the language of communication within Zrawa Amazigh families. Actually, two participants, Alaa and Mahdi (parents), did speak Amazigh to members of their families in my presence. These conversations were conducted in Amazigh. To be precise, Alaa used it to his wife and children and Mahdi spoke it to his little daughter and his sister-in-law in Amazigh. Furthermore, when strolling down the village streets, I often encountered women speaking to their children in Amazigh.

The idea that AL is used within the Zrawi Amazigh families is further supported by details taken from the semi-structured interviews with single young adults, parents, and grandparents (see Appendix F). In answer to the question «Do you speak only Amazigh with your parents?» the three single young adults (Khalifa, Arij, and Hadi) had the following responses. Khalifa and Arij stated that they use only AL when conversing with their parents. Hadi, on the other hand, reported that he uses only AL when speaking to his mother and uses mostly AL when conversing with his father. Interestingly, he reported that he uses TA when speaking to his father about politics. This is a case of metaphorical code-switching which means that the change of the conversation topic requires the change of the language used (Wardhaugh, 2010).

As to parents (Alaa, Ahmed, Abderrahman, Mahdi, and Mohamed), they had different responses to the question «Do you speak only Amazigh to your children?» While some parents reported that they use Amazigh most of the time when speaking to their children, other parents stated that they converse with their children only in Amazigh. Indeed, Alaa and Ahmed said that their conversations with their children are conducted exclusively in Amazigh. As Alaa , for instance, asserted, that «when the child grows up, he/she understands that this [Amazigh] is the language of communication with his father and mother and with the other family members». He also stated that «within the family ... the language of communication is the Amazigh language.» In contrast, Abderrahman, Mahdi, and Mohamed stated that they usually speak to their children in Amazigh and use TA only when having guests who do not understand Amazigh. This is a case of situational code-swiching. That is, the situation of the conversation decides the language used (Wardhaugh, 2010). It seems that using Amazigh in parents' interactions with their children contributes to its maintenance. As pointed out by Sridhar (1988), parents can facilitate the maintenance of the mother tongue by using it at home. The results from studies such as Becker (2013) and Gomaa (2011), mentioned in section 2.3.2, emphasize the important role that parents play in language maintenance through using their minority languages at home.

When asked the question «Do you speak only Amazigh to your children and grandchildren?» the grandparents (Hammouda, Salah, and Salwa) stated that they usually speak to their children and grandchildren only in Amazigh. For instance, Salah reported: «As I told you, it [Amazigh]'s our language. At home, it's the language we use, we don't have another language». Salwa insists on using Amazigh with their grandchildren who understand it but also with those who do not. Indeed, she said:

One of my daughters is married with an Arab man from Skhira [a region in the governorate of Sfax]. When she and her two children come from Skhira to visit me, I

usually address my two granddaughters in Amazigh. The one who understands Amazigh does reply and the one who does not keeps staring at me. I speak in Amazigh whether she understands it or not.

Salwa's persistence to use AL with her grandchild, who does not understand it, is an implicit message that she should learn the language. A further point, the extract indicates a case of language loss and suggests its cause. That is, the fact that the grandchildren cannot understand AL is likely to be due to her being the offspring of a mixed marriage between an Arabophone man and an Amazighophone woman and to her contact with the majority language, the language of her father's family. It should be mentioned here that only one female out of the 22 participants, 11 males and 11 females, who reported that they are married, cited that her spouse is not Amazigh. This suggests that endogamy is the norm within the Zrawa Amazigh community. Endogamy can foster AL maintenance. As, Giles et al. (1977) indicate, that minority groups are likely to maintain their language when the rate of inter-marriages is low.

As for language policy within Amazigh families, most of the respondents to the interview questions asserted that at home Amazigh is spoken naturally. In response to the question «Do your parents ask you to speak the Amazigh language at home or around Amazigh people?» the single adults reported that their parents do not ask them to use AL when speaking to them to do; rather, they do it naturally and intuitively. For example, Arij said: «They don't ask or oblige me to speak Amazigh. Speaking Amazigh runs in the blood». As reported here, here, the parents of these respondents do not ask them to use AL.

Answering the question «Do you ask your children to speak the Amazigh language at home or around Amazigh people?» parents, except for Abderrahman, stated that they do not ask their children do speak Amazigh at home or around the other Imazighen naturally. As Alaa said:

There is no need to ask [children to speak Amazigh at home and within the community]. Naturally, when the child grows up, he/she understands that this [Amazigh] is the language of communication with his father and mother and with the other family members. I do not ask my child to speak Amazigh outside home because he may have friends and teachers who do not understand Amazigh, so he should talk to them in the language that all people know, which is Tunisian Arabic. However, within the family and the Amazigh community, like the village of Zrawa, the language of communication is the Amazigh language.

In this excerpt, Alaa maintains that Amazigh is the language of communication within the family and the community; however, the language of communication with people who do not understand AL is TA which he acknowledges as being the dominant language of the country (that is, Tunisia).

As to the grandparents, they were asked «Do you ask your children and grand children to speak the Amazigh language at home or around Amazigh people?» In response to this question, they reported that their children and grandchildren use Amazigh with them and with the other Imazighen without being asked to do so. Salah summarized the situation by saying:

We use Amazigh with those who speak Amazigh and Arabic with those who speak Arabic, and our children do the same. The child notices what his father does; he speaks to an Amazigh person in Amazigh and to an Arab person in Arabic, and he imitates him.

This extract indicates that Zrawi Amazigh parents transmit not only AL to their children but also the rules governing the use of that language.

The idea that, within the families of ten of the 11 respondents, there is no explicit policy for which language to use within the family supports Caldas' (2012) observation that for most families, the family language use is unconscious and has basically been dedicated by history

and circumstances over which the family has no control and that the majority of parents do not deliberately plot and plan what language to use within the family.

Zrawa Amazigh families represent not only a domain of AL use but also a setting where AL is acquired. Indeed, in response to the questionnaire item «What is your first language?» all the participants reported that their first language is AL. This suggests thay they had acquired it at home. In addition, in answer to the interview question «How and when did you acquire the Amazigh language?» all the respondents reported that they had acquired AL at an early age within their families. According to them, when they had been born, they had found their parents and other family members speaking Amazigh. Moreover, their parents, together with the other relatives, had always addressed them only in Amazigh. For instance, Alaa explained:

I acquired it [Amazigh] at home. When the Amazigh child is born, the first speech he hears is in the Amazigh language. He hears it from his father, his uncle [father's brother], his grandfathers, his grandmothers, his uncles [mother's brothers] and from all the family. This means that the language of communication between Imazighen in Tunisia is exclusively the Amazigh language.

What Alaa says in this extract supports the idea that Zrawa amazigh families serve as a domain of AL use and acquisition. In the same vein, Hadi (a single young adult) stated the following: «Concerning Amazigh, it's my first language. I acquired it spontaneously and intuitively (...) I got it in the milk». It should be mentioned that the expression «got in the milk» means transmitted naturally from parents to children.

When asked «How do you facilitate your children's acquisition of the Amazigh language?» the parents with children asserted that this can be achieved through the daily use of Amazigh in conversations with children. The strategies they use consist of speaking AL to children on a daily basis in order to make it easy for children to acquire the language. Two

among the five respondents, Mohamed and Ahmed, reported that they give their children feedback when they insert lexical items from Arabic or French into utterances produced in Amazigh. For example, Mohamed stated that when his children use Arabic or French words during a conversation conducted in AL, he teach them the equivalent of those words in Amazigh.

As to the acquisition of TA, all of the interviewed participants asserted that they had acquired TA as a result of daily contact with its speakers; however, the way such contact occurred differs from one participant to another.

To start with, the three grandparents reported that they had acquired Tunisian Arabic as a result of the contact with the speakers of TA. Salwa said that she had migrated to Tunis with her family at the age of 4. There, she acquired TA in the Arabic-speaking neighborhoods when playing with her Arab peers. Hammouda, on the other hand, stated that he had acquired TA as a result of the contact with the Arabic-speaking teacher of Quran (the «middib»). As to Salah, he asserted that he had acquired TA from the conversations he had witnessed between Zrawi Imazighen and Arabic-speaking persons, who had come from other regions, such as

El-Hamma, to work in Zrawa. Those conversations, as Mohamed stated, had been conducted in Tunisian Arabic due to the fact that the newcomers had been unable to understand AL.

Like the grandparents, parents stated that the contact with the speakers of TA had resulted in their acquisition of this language. In fact, Mohamed reported that he had acquired TA at school through the contact with Arabic-speaking teachers. However, the other parents asserted that they had acquired it as a result of the contact with both Arabic-speaking teachers at school and Arabic-speaking people in settings other than schools. For example, Ahmed pointed to a situation in which Imazighen and Arabs may come into contact saying: «As we don't have jobs in our region [Zrawa], we migrate to other regions to work there and, of

course, when you have contact with those people [Arabic-speaking people] you acquire Tunisian Arabic».

In the same way, the single young adults reported that their acquisition of TA was a consequence of the contact with Arabic-speaking people. Khalifa said that mixing with speakers of Tunisian Arabic either at school or in the street and the contact with Tunisian-Arabic-speaking TV programs had been the ways he had acquired TA. Arij stated that her acquisition of TA had started at the age of six at school. As for Hadi, he said that he had acquired TA at school, but he added that, in addition to school, he had acquired it as a result of watching TV programs broadcasting in TA and of making Arabic-speaking contacts.

These details suggest that TA, which is the dominant language in Tunisia (Maamouri, 1983a), is not acquired within the Amazigh families but rather through the contact either with the speakers of TA outside home in the street or the TA-speaking media, especially TV. This supports the idea that Amazigh is the language of communication within these families.

A related point, informants told me that Amazigh children under the age of six are, in sociolinguistic terms, monolingual in Amazigh and that these children's contact with TA, and indeed its speakers, starts when they attend school. For example, Alaa told me that his daughter who has not attended school yet cannot speak AL. It appears that the situation of Amazigh children in Zrawa is similar to the situation of children belonging to ethnic minorities in the United States of America, which Campbell and Schnell (1987) emphasize by stating that:

In nearly every major centre in the United States we find large concentrations of homes in

which English is not the dominant language of communication. In these homes, children

up to the age of five or six regularly acquire and use the home language for all their

sociocultural and basic physical needs. As a consequence, through natural language

acquisition process found in all human societies, these children, in Chomskyan terms,

acquire nearly all of the phonological, grammatical and semantic rules that identify them as native speakers of that language. (p. 178)

This is inconsistent with Gabsi's (2011) observation that, in the villages of Douiret and Chenini, Amazigh children are being raised as bilinguals in Amazigh and TA as a result of using both languages at home. Gabsi (2003) adds that these children are more competent in TA than in Amazigh.

Finally, in answer to the interview question «Whose responsibility is it to keep the Amazigh language alive?» the 11 respondents showed agreement on the idea that Amazigh families are responsible for keeping AL alive. Nevertheless, Salwa and Arij said that schools are also responsible for AL maintenance. Besides, Arij and Mahdi shared the belief that keeping AL alive is also the responsibility of the government.

In brief, the results mentioned above indicate that the Amazigh family is a domain of AL acquisition and use, which contributes to AL maintenance. As Sridhar (1988) indicates, parents can maintain their mother tongue by speaking it at home. Additionally, the effect of parents' language choice on LM or LS is emphasized by Lieberson and Curry (1971). Likewise, Clyne and Kipp (1997) point out that «intergenerational transmission is heavily dependent on home language use» and that «the home language question does enable predictions of future use of the languages concerned» (p. 451). In the same vein, Okamura- Bichard (1985) indicates that mother tongue transmission is unlikely to happen if parents do not teach it to their children their mother tongue. Fishman (1991) summarizes all this by stating that the family is «the most common and inescapable basis of mother tongue transmission, bonding, use and stabilization» (p. 94). The role of the family in LM, emphasized in the current study, is confirmed by findings from research carried out by other researchers such as Renz (1987), Nesturik (2010), Al-Sahafi (2015), Gomaa (2011), Tatar

In contrast, the family can be a factor of LS. This happens when parents from minority ethnic group do not transmit their languages to their children by not teaching and using these languages at home. The results from Bentahila and Davies' (1992) and Galindo's (1991) studies, aforementioned in section 2.4.1, support this idea.

Using AL within Amazigh families as well as within the Amazigh community, implies that it is highly valued among its speakers. As Ennaji (2005) puts it, «the languages chosen for communication are generally those that people favor. The more a language is liked or appreciated, the more it is used in different domains. Similarly, a language that is disliked by speakers will be used less frequently» (p.157). The following section focuses on the attitudes of Zrawi Imazighen towards their language, AL.

Quantitative (questionnaire) and qualitative (interview) data suggest that the participants have positive attitudes towards their language, namely AL. As Figure 4.3 below shows, 98 % of the participants reported that they are proud of being speakers of AL. This suggests that Amazigh is highly valued among its speakers and is a source of pride for them. Being a source of pride denotes the historicity and the prestige value of AL for its speakers. The historicity of AL has to do with the fact that it is the indigenous language of Tunisia (Maamouri, 1983a).

120%
100% 96%
		91%

80%

		64% 62%
60%
40%
20%

0%

Imazighen have the right to speak Amazigh in the presence of those who do not understand it My ties with Tunisian Arabic are weaker than my ties with Amazigh

Figure 4.3 also reveals that 99 % of the participants believed that maintaining AL is necessary. This denotes the high value that AL has among these participants. It is worth noting that Abderrahman (a parent) is the only participant who thought that AL maintenance is unnecessary. Nevertheless, he reported that he is proud of being an AL speaker. The semi-structured interviews reveal further information about participants' thoughts concerning whether or not maintaining AL is necessary. Indeed, the question «Do you think that maintaining the Amazigh language is necessary? Why? and How can this be achieved?» was asked to the 11 respondent. To begin with, the three single adults, in response to this question, stated that it is necessary to maintain AL because it is a language of heritage and identity. For instance, Hadi said that «It's normal that one maintains his language because if a human being loses his language, he would lose a main part of his civilization.» As for the ways to maintain AL, the three respondents believed that keeping the language alive is primarily the responsibility of Amazigh families. In addition, associations, as Arij asserted, can play a significant role in AL maintenance through teaching it to non-Amazigh speaking

children. As for Hadi and Khalifa, they thought that teaching AL at school to all Tunisian children contributes to its maintenance.

The same question («Do you think that maintaining the Amazigh language is necessary? Why? and How this can be achieved?») was also directed to the parents. All of them stated that AL maintenance is necessary. For example, Mahdi said: «It [Amazigh]'s the identity of Imazighen and of Tunisians in general. That's why one should maintain it». Another example, Ahmed stated that it is necessary for an Amazigh person to maintain AL because it represents not only identity but also ancestors' heritage. A further example, Alaa thought that AL must be maintained because it a component of the Tunisian culture, .As to the way to maintain AL, Mohamed thought that it can be maintained through the contact between Imazighen from different regions. Indeed, he asserted: «We are in contact with Imazighen in Morocco, Algeria, Siwa, Taoujout, Tamazret, and Douiret through the means of associations». For Mahdi and Mahdi, acquiring AL at home contributes to its maintenance. Alaa pointed to the contribution of associations to maintaining Amazigh saying that:

This [AL maintenance] is now in the hands of the civil society since the government has not reacted yet in favor of this issue. Also, there are no demands from Imazighen for the governments' reaction or for teaching this language [Amazigh] at schools. Personally, I think that cultural associations contribute to promoting, maintaining, and enriching this language.

In this excerpt Alaa highlights the role of cultural associations in AL maintenance in the absence of government's institutional support for the language.

Like parents and single young adults, grandparents did respond to the question «Do you think that maintaining the Amazigh language is necessary? Why? and How this can be achieved?». All of them stated that maintaining AL is necessary. When asked about the reasons behind the importance of AL maintenance, Hammouda said: «It [Amazigh]'s my

language» while Salah reported: «Because it [Amazigh]'s the language of our parents and ancestors; here, in Zrawa, it's our language» whereas Salwa answered: «I love it [Amazigh]; I admire it (...) For me, it's beautiful; it's something very beautiful». The best way to maintain AL, as the grandparents asserted, is to transmit it from parents to children.

Returning to Figure 4.3, it is also demonstrated that 79 % of the participants thought that Imazighen have the right to speak Amazigh in the presence of those who do not understand it. This implies that these participants are attached to their language and see it as a source of prestige. As learnt from participant observation, many of the participants I communicated with did use AL with each other in my presence. These ideas indicate that these participants have positive views about the public use of AL in the presence of monolingual Arabic speakers. As Bradley (2013, p. 1) points out, the way the public use of a minority language in the presence of monolingual majority speakers is viewed, affects attitudes towards that language. In contrast, 13 % of the participants believed that Imazighen should not speak Amazigh in the presence of people who do not understand it. Indeed, the ones who filled the questionnaire in my presence among these participants told me that speaking Amazigh in the presence of those who do not understand is illogical. The remaining eight percent of the participants were neutral.

Another language attitude that Figure 4.3 above shows is that 81 % of the participants reported that their ties with TA are weaker than their ties with AL. This denotes that the shift from AL towards TA is unlikely to occur. Indeed, Ennaji (2005) states that «the more a language is liked or appreciated, the more it is used in different domains» (p. 157).

Additionally, in response to the questionnaire item «Are you against the abandonment of the Amazigh language by its speakers?» 100 % of the participants reported that they are against the abandonment of AL by its speakers. This reveals the participants' attachment to their language. Such attachment is further emphasized in the responses to the interview

questions «How important is the Amazigh language for you?» In fact, in answer to this question respondents from the three groups asserted that AL is important. In fact, three parents (Mohamed, Ahmed, and Mahdi) and a single young adult (Arij) stated AL equals identity. For example, Mohamed said: «For me as an Amazigh man, the Amazigh language is my identity. I won't abandon it». Abderrahman (a parent) and Khalifa (a single young adult) reported that AL is important because it is the language of ancestors. For the grandparents, the fact that AL is their mother tongue accounts for importance. As to Alaa (a parent), he asserted that the importance of AL lies in its being both the language of communication within the family and the indigenous language of North Africa. Actually, he said:

The Amazigh language, for us, is the language of communication that we have been accustomed with. We got it from our fathers and mothers as our mother tongue. It represents the indigenous language of the peoples of North Africa (...) and it remains the only connection between the Imazighen of North Africa.

What Alaa wants to say in this excerpt is that AL is trans-national in the sense that it unites

Imazighen from the different countries of North Africa. This explanation emphasizes one of

the facets of nationalism, which Sharara (cited in Paulson, 1987) highlights by stating that: The nation ... is a wider conception than the state, greater than the people, and meaningful than the fatherland. It is not necessary for a nation to have one state or one fatherland, or to be composed of one people, but it must have its own language, its own history, its own ideals, its own shared aspirations, its own shared memories, and its own natural links which bind its members in two respects, the moral and the economic. (p. 39)

As for Hadi (a single young adult), AL is important because it is not only a language of communication with his family, his friends, and the other members of the Amazigh community but also a component of his culture and identity.

It is deduced from what has been mentioned above that participants' positive language attitudes facilitates LM. This confirms Holmes' (2013) observation that positive attitudes towards a minority language contribute to its maintenance. It is also consistent with Becker's (2013), Tatar's (2015), and Gomaa's (2011) studies (see section 2.3.3 for more details) which reveal that positive attitudes towards one's language play an important role in its maintenance. Conversely, speakers' negative attitudes towards their minority language lead to language shift. In fact, Bentahila and Davies (1992) found that some Amazigh parents encouraged their children to learn Arabic because it was more practical in everyday life than AL, which facilitated the shift from Amazigh to Arabic. In the same vein, Galindo (1991) found that one of the factors causing the regression of Spanish among the Chicano adolescents in Austin, USA, was the parents' unwillingness to teach Spanish to their children. Interestingly, in some families, parents' efforts to maintain their language, which are motivated by their positive attitudes towards it, are often undermined by their children's negative attitudes towards the language, manifested in their unwillingness to learn it. In his study of language maintenance among 18 Chinese families residing in Philadelphia (USA) Zhang (2008) found a contradiction between parents' and children's attitudes towards Chinese language. The interviews which Zhang conducted with parents indicated that some of them viewed Chinese as being beneficial for children's cognitive development; others believed that it is closely linked to ethnic identity; and still others saw it as a way to reinforce family bonds. In contrast, the interviews with children revealed their lack of motivation, or even resistance, to learning Chinese.

Positive attitudes towards AL do foster its maintenance in the sense that such attitudes promote its use as a language of communication at home, that is Amazigh families, and within the Amazigh community. Another factor which contributes to AL maintenance among

the Imazighen of Zrawa is the perceived tight link between Amazigh language and identity. This is the focus of the next section.

Questionnaire and interview data reveal that most of the participants believe that there is a close relationship between Amazigh language and identity. As Figure 4.4 below demonstrates, 96% of the participants believed that AL is the most salient marker of Amazigh identiy. This confirms Mesthrie and Tabouret-Keller's (2001) and Kotzé's (2001) view that

The Amazigh language is the most salient marker of the Amazigh identity Speaking the Amazigh language is a prerequisite for being Amazigh

Figure 4.4. Results of statements 4 and 5 on the link between Amazigh language and identity.

From participant observation, I learnt that there are other markers of Amazigh ethnic identity such as biological heritage, location of origin (that is being the indigenous inhabitants of Zrawa), the tattoos on the faces of elderly women, celebration of the Amazigh New Year on 14 January of every year, female traditional clothing, and female names, such as «Yizza», «Bookha», «Masyoogha», and «Sasiyya». What is interesting about names is that although some Zrawi Imazighen have official Arabic names, their co-ethnics call them by the Amazigh counterparts of these names. For example, people whose names are Khadija (a female name),

Belgacem, Mohamed, and Abdullah are often called «Jaja», «Kissi», «Hammou», and «Abali», respectively. What I noticed is that these identity markers are not as salient as the Amazigh language.

In order to get in-depth details about the perceived link between language and ethnic identity, the participants in the semi-structured interviews were asked the question: «How do you see the link between Amazigh language and identity?» Their responses showed that there is disagreement between them on whether or not AL is a marker of the Amazigh identity. The majority, namely seven respondents, thought that the Amazigh identity is indicated by AL use. For instance, Alaa asserted: «You don't find a Zrawi Amazigh person who doesn't speak the Amazigh language.» Hammouda went even further when he stated that anyone who does not speak the AL is not considered Amazigh. However, four respondents, namely Ahmed, Abderrahman, Arij, and Khalifa, believed that AL is not a marker of Amazigh identity in the sense that not speaking the language does not deprive an Amazigh person from his ethnicity. For example, Arij and Khalifa stated that there are many Amazigh people who do not speak AL. The view hold by Ahmed, Abderrahman, Arij, and Khalifa denotes that language use does not necessarily indicate ethnic identity. This view is supported by details from participant observation, which reveal that speaking AL does not mean being Amazigh. Indeed, it happened that when I was walking in the street with Mohamed we encountered four children speaking to each other in Amazigh. He informed me that two amo,g those children were Arab. Besides, Mahdi informed me that his «Arab» employee, who has been working for him for 15 years, can speak Amazigh. Furthermore, Alaa told me about other Arab inhabitants of Zrawa, who can speak AL. It is worth mentioning that Imazighen do not consider such people to be Amazigh regardless of the fact that they can speak AL. Interestingly, when I asked Mahdi about the way to differentiate between a native and a non- native AL speaker, he informed me that there is a particular Amazigh sound which is

mispronounced by non-native AL speakers and only those who acquired AL at home as their mother tongue, that is Imazighen, can pronounce it correctly. Additionally, I was told about some Arabs, living in Zrawa, who can understand AL but cannot speak it. This is the case of an Arab woman who has been working as a teacher a Zrawa nursery school.

Figure 4.4 also shows that 86 % of the participants thought that speaking AL is a prior condition for being Amazigh. This implies that AL is an essential component of the Amazigh identity and a criterion for group membership. Such result supports Trudgill's (2000), Padilla and Borsato's (2010), and Ennaji's (2012) observation that language is a criterion for ethnic group membership (see section 2.3.4).

Furthermore, Figure 4.5 below indicates that 87 % of the participants thought that AL loss would result in the loss of Amazigh identity, which suggests that AL maintenance is an important factor in the maintenance of the Amazigh identity. This finding is similar to that in some previous research. As a matter of fact, Hatoss (2005), who investigated LM among the Hungarian community in Brisbane, Australia, found that the maintenance of the Hungarian language is seen as a vehicle for maintaining the Hungarian identity. Similarly, Chiung (2001), who used a questionnaire and chi-square test in his study of the relationship between mother tongue and ethnic identity among 244 Taiwanese students in Taiwan, found that the maintenance of one's ethnic language is a contributing factor to the maintenance of one's ethnic identity.

Figure 4.5. Results of question 3 about the link between Amazigh language and identity.

More details about the tight link between Amazigh language and identity come from responses to the interview question «What do you lose if you were to lose the Amazigh language?» Indeed, nine out of the 11 respondents, namely two parents (Mahdi and Ahmed) and two single young adults (Arij and Khalifa), thought that AL loss would lead to the loss of the Amazigh identity. Besides, three respondents, to be precise one parent (Mohamed) and two grandparents (Hammouda and Salwa) went as far as to claim that AL loss means the end of life. For example, Salwa stated: «What it is left for me in this life (...) if I lose Amazigh? Life is over». In contrast, Abderrahman said that he would not lose anything in case Amazigh is lost. Both Salah (a grandparent) and Alaa (a parent) rejected the likeliness of AL loss. Salah, for example, argued that AL would not be lost as long as it used at Amazigh homes and within the Amazigh community.

The disagreement among the respondents on the existence of a language-identity relationship, mentioned earlier, reflects the contrast between the results from Hatoss's (2005) and Chiung's (2001) studies and those from Ahn's (2008), Kang's (2004), and Bentahila and Davies's (1992) studies (see section 2.3.4). Indeed, studies carried out by Hatoss (2005) and

Chiung (2001) indicate that language is tightly linked to ethnic identity. However, research findings from Ahn (2008), Kang (2004), and Bentahila and Davies (1992) suggest that there is no relationship between language and ethnic identity. Bentahila and Davies (1992), for instance, point out that the use of AL does not necessarily means the identification with Imazighen by stating that:

The ability to speak Berber [AL] is thus evidently not felt to be a necessary condition for self identification as a Berber [Amazigh]; and, indeed, this view was upheld by 83% of the informants, who replied negatively to the question `Is it necessary to speak Berber to be a Berber?' Moreover, the fact that a majority of those who did speak Berber fluently and habitually chose not to describe themselves as Berbers [Imazighen] at all suggests a striking lack of correspondence between use of the language and identification with the group. (p. 202)

As a concluding remark to this section, the majority of the participants believed that Amazigh language and identity are closely related in the sense that AL is both a marker of ethnic identity and a criterion of ethnic group membership and that its loss would lead to the loss of the Amazigh identity. Such relationship promotes AL maintenance. As stated by Holmes (2013), a minority language is likely to be maintained longer in areas where it is considered to be an important symbol of ethnic identity.

To sum up, the examination and analysis of the data revealed the following key findings. First, the geographic concentration of Zrawa Amazigh community gives the community members the opportunity to use AL as a language of communication, which fosters the maintenance of the language. Second, the Amazigh families contribute to AL maintenance by virtue of their being a domain where the language is acquired and used. Third, the participants, in general, reported that they have positive attitudes towards AL, which

promotes its use at home and within the community and, hence, its maintenance. Finally, most of the participants believed that there is a close relationship between Amazigh language and identity. Carrying such belief does contribute to AL maintenance.

The next chapter will conclude the present study by summarizing the main findings reached in this research, listing the implications, acknowledging the limitations, and making recommendations for future research.

This is the last chapter in the study. The first section summarizes the main results of the present study concerning the five contributing factors to the maintenance of the AL among the Zrawa Amazigh inhabitants. The second section lists and explains the implications drawn from this research. The third section highlights the contributions of the study. The forth section acknowledges the limitations of this study. The chapter ends by making some recommendations for future research.

Study findings from questionnaire, observation, and interviewing have revealed that the Zrawa Amazigh community is geographically concentrated. This allows its members to meet on a daily basis and to use AL as the language of the street, that is, the language of daily communication. The frequent use of AL within the community means provides favorable conditions for its maintenance.

There is confirmation that Zrawa Amazigh families play a crucial role in the maintenance of AL. The family is believed to be a naturalistic setting where AL is acquired, a domain where AL is used as a means of communication between family members, and a major agent in the process of keeping AL alive.

There is evidence that participants have positive attitudes towards AL. Indeed, all of them thought that Imazighen should not abandon their language. Besides, the great majority of these participants reported their pride in being speakers of AL and their belief that maintaining AL is necessary. Moreover, approximately four fifths of the participants reported

their attachment to AL and their belief in Imazighen's right to use AL in the presence of those who do not understand it.

The study results have demonstrated that most of the 53 participants believed that there is a close relationship between Amazigh language and identity. In fact, most of them considered AL to be the most salient marker of Amazigh identity. In addition, the majority among them thought that speaking AL is a prerequisite for being Amazigh and believed that the AL loss would result in the loss of the Amazigh identity. As revealed in previous research on LM, if a language is seen as a symbol of ethnic identity it is likely to be maintained; and this is the case for AL in Zrawa.

The results of this study have three implications. The first of these is that, as unprecedented, the results of this study have provided insights into the underpinnings of AL maintenance. They have revealed that four factors contribute to the maintenance of AL among the Imazighen of Zrawa. These factors, as the study findings have indicated, are: the Zrawa Imazighen's geographic concentration, the key role of the Zrawa Amazigh families, the Imazighen's positive attitudes towards AL, and the perceived close relationship between Amazigh language and identity.

As to the second implication, this research indicates that the chances of maintaining AL in Zrawa in the future are high given the fact that teenagers, as well as single young adults, have positive attitudes towards AL and see it as closely related to the Amazigh identity. Once married, they are likely to transmit it to their children.

As for the third implication, this study suggests that AL would be maintained in Zrawa as long as it is the majority language and as long as it is used as the home and community language of communication.

As mentioned earlier (in the first section of the literature review), the maintenance of AL in the Tunisian context is under-researched. Therefore, this study contributes to the research on LM in Tunisia and, more generally, in the Maghrib through its investigation into the factors underpinning the maintenance of AL which is Tunisia's indigenous language (Maamouri, 1983) in Zrawa, one of the nine Tunisian villages where AL is still spoken today (Maamouri, 1983; Gabsi, 2003).

Besides, the study has provided evidence that Zrawa Imazighen are attached to AL. This may encourage the Tunisian government to pass acts aiming at promoting and preserving Imazighen's linguistic rights. Such linguistic rights concern not only individuals but also communities. Individuals' linguistic rights, according to Hamel (1997), refer to the rights «to learn their mother tongue, to enjoy education through the medium of that language, to use it in socially significant official contexts, and to learn at least one of the official languages in one's country of residence» (p. 1). As to communities' linguistic rights, Phillipson (1992) states that these include right to «establish and maintain schools and other educational institutions, with control of curricula and teaching in their own languages ..., [and also] autonomy to administer matters internal to the groups, at least in the fields of culture, education, religion, information, and social affairs, with the financial means ... to fulfill these functions» (p.2).

Furthermore, the researcher's relatively easy access to the Zrawa Amazigh community may encourage other researchers to carry out further research on AL maintenance and shift in Zrawa and in other the other villages where AL is still used today.

This study is not without limitations. Indeed, it has three limitations. The first of these emanates from the short time that the field work took, namely one month.

The second limitation has to do with the small number of participants. This is due to three facts: First, many Imazighen who promised to participate in the study did not fulfill their promise. Second, many Imazighen refused the idea of participating in the study altogether. Third, a number of Imazighen took some questionnaire forms in order to fill them out but they did not do that.

As to the third limitation, it is related to the absence of interviews with female parents due to the fact that these are staying at their homes, which makes them out of reach, and this made it impossible to explore the in-depth opinions of Amazigh females concerning the different issues related to AL maintenance.

Based on the aforementioned limitations of the present study, future research on the maintenance of AL is recommended to involve larger samples from the Amazigh population in Zrawa and in other regions and to take longer time. Besides, further research studies should consider the attitudes and beliefs of the secluded Amazigh women, which have to do with AL maintenance.

Abdelhadi, M. (2017). The role of education in the maintenance of Arabic language among the Arabic-speaking community in the regional city of Toowomba, Australia. Fourth Asia Pacific Conference on Advanced Research. Asia Pacific Institute of Advanced Research: Melbourne. Retrieved August 13, 2017, from https://goo.gl/WAHSzQ

Ahn, J. K. (2008). Ethnic identity and language use of the heritage Koreans in the US. The Linguistic Association of Korea Journal, 16 (3), 61-79. Retrieved June 5, 2017, from https://goo.gl/HwMG71

Al-Khatib, M., & Al-Ali, M. N. (2010). Language and cultural shift among the Kurds of Gordan. SKY Journal of Linguistics, 23, 7-36. Retrieved August 13, 2017, from

Allard, R. & Landry, R. (1992). Ethnolinguistic vitality beliefs and language maintenance and loss. In W. Fase, K. Jaspaert & S. Kroon (Eds.), Maintenance and loss of minority languages (pp. 171-196). Amsterdam: Benjamins.

Al-Sahafi, M. (2015). The role of Arab fathers in heritage language maintenance in New Zealand. International Journal of English Linguistics, 5 (1). Retrieved December 3, 2016, from https://goo.gl/WsoE2F

Becker, D. J. (2013). Parents' attitudes toward their children's heritage language maintenance: The case of Korean immigrant parents in West Michigan (Master thesis, Grand Valley State University, 2013). Master Theses, 59. Retrieved November 28, 2016 from https://goo.gl/nQNLhs

Baker, C. (2001). Foundations of bilingual education and bilingualism. Clevedon: Multilingual matters Ltd.

Battenburg, J. (1999). The gradual death of the Berber language in Tunisia. International Journal of the Sociology of Language, 137, 147-161.

Bentahila, A. and Davies, E. E. (1992). Convergence and divergence: Two cases of language shift in Morocco". In W. Fase, K. Jaspaert & S. Kroon (Eds.), Maintenance and loss of minority languages (pp. 197-210). Amsterdam: Benjamins.

Boeschoten, H. (1992). On misunderstandings in a non-stabilized bilingual stuation. In W. Fase, K. Jaspaert & S. Kroon (Eds.), Maintenance and loss of minority languages (pp. 83-98). Amsterdam: Benjamins.

Bourhis, R., Giles, H. & Rosenthal, D. (1981). Notes on the construction of a 'subjective vitality questionnaire' for ethnolinguistic groups. Journal of Multilingual and Multicultural Development 2, 145-155.

Bradley, D. (2002). Language attitudes: The key factor in language maintenance. In D. Bradley & M. Bradley (Eds.), Language endangerment and language maintenance (pp. 1-9). Available from https://goo.gl/LF8rZR

Brown, N.A. (2008). Language shift or maintenance? An examination of language usage across four generations as self-reported by university age students in Belarus. Journal of Multilingual and Multicultural Development, 29 (1), 1-15.

Caldas, S. J. (2012). Language policy in the family. In B. Spolsky (Ed.), The Cambridge handbook of language policy (pp. 351-373). Cambridge: Cambridge University Press.

Campbell, R. N. & Schnell, S. (1987). Language conservation. The Annals of the American Academy of Political and Social Science, 490, 177-185.

Cartwright, D. (1987). Accommodation among the Anglophone minority in Quebec to official language policy: A shift in traditional patterns of language contact. Journal of Multilingual and Multicultural Development, 8 (1&2), 187-212.

Chiung, W. T. (2001). Language and ethnic identity in Taiwan. Paper Presented at the 7^th North American Taiwan Studies Conference. University of Washington: Seattle. Retrieved June 7, 2017 from https://goo.gl/MFWQhA

Christians, C. G. (2005). Ethics and politics in qualitative research. In N. K. Denzin & W. S. Lincoln (Eds.) The SAGE handbook of qualitative research (3^rd ed., pp. 139-164). Thousand Oaks: Sage Publications.

Clyne, M. G. (1988). The German-Australian speech community: Ethnic core values and language maintenance. International Journal of the Sociology of Language, 1988 (72), 67-83.

Clyne, M. & Kipp, S. (1997) Trends and changes in home language use and shift in Australia, 1986-1996. Journal of Multilingual and Multicultural Development, 18 (6), 451-473.

Coulmas, F. (2013). Sociolinguistics: The study of speakers' choices. Cambridge: Cambridge University Press.

Creswell, J. W. (2008). Mixed methods research. In L. M. Given (Ed.), The SAGE encyclopedia of qualitative research methods (pp. 526-529). Los Angeles, CA: Sage Publications.

Cummins, J. (2001). Heritage languages. In R. Mesthrie (Ed.), Concise encyclopedia of sociolinguistics (pp. 619-620). New York: Elsevier Science.

Demos, V. (1988). Ethnic mother tongue maintenance among Greek orthodox Americans. International Journal of the Sociology of Language, 69, 59-71.

Dorian, N. C. (1987). The value of language-maintenance efforts which are unlikely to succeed. International Journal of the Sociology of Language, 68, 57-67.

Dresing, T., Pehl, T., & Schmieder, C. (2015).Transcription conventions: Software guides and practical hints for qualitative researchers (3^rd ed.). Marburg. Available Online: http://www.audiotranskription.de/english/ transcription-practicalguide.htm

Eckert, P. (1980). Diaglossia: Separate and unequal. Linguistics, 18, 1053-1064.

Edwards, J. (1992). Sociopolitical aspects of language maintenance and loss: Towards a typology of minority language situations. In W. Fase, K. Jaspaert & S. Kroon (Eds.), Maintenance and loss of minority languages (pp. 197-210). Amsterdam: Benjamins.

Ennaji, M. (1997). The sociology of Berber: Change and continuity. International Journal of the Sociology of Language, 123, 23-40.

Ennaji, M. (2005). Multilingualism, cultural identity, and education in Morocco. New York: Springer.

Ennaji, M. (2012). The Arab world: Maghreb and the Near East. In J. Fishman (Ed.), The handbook of language and ethnic identity (2^nd ed., Vol.2, pp. 407-422). Oxford: Oxford University Press.

Fishman, J. A. (1964). Language maintenance and language shift as a field of inquiry. Linguistics, 9, 32-70.

Fishman, J. A. (1972). Language in sociocultural change. California: Stanford University Press.

Fishman, J. A. (1989). Language and ethnicity in minority sociolinguistic perspective. Clevedon: Multilingual matters Ltd.

Fishman, J. A. (1991). Reversing language shift: Theoretical and empirical foundations of assistance to threatened Languages. Clevedon: Multilingual matters Ltd.

Fishman, J. A. (2001). 300-plus years of heritage language education in the United States. In J. K. Peyton, D.A. Ranard & S. McGinnis (Eds.), Heritage Languages in America Preserving a National Resource. Retrieved from https://goo.gl/9EFb9k

Gabsi, Z. (2003). An outline of the Shilha (Berber) vernacular of Douiret (Southern Tunisia) (Doctoral dissertation, University of Western Sydney, 2003). Retrieved January 1, 2017 from https://goo.gl/JzTuhw

Gabsi, Z. (2011). Attrition and maintenance of the Berber language in Tunisia. International Journal of the Sociology of Language, 211, 135-164. Retrieved January 31, 2017, from https://goo.gl/6NsXaS

Galindo, L. (1991). A sociolinguistic study of Spanish language maintenance and linguistic shift towards English among Chicanos. Lenguas Modernas, 18, 107-116. Retrieved August 13, 2017, from https://goo.gl/PgX7os

Garner, M. (1988). Ethnic languages in two small communities: Swedish and Russian. International Journal of the Sociology of Language, 72, 37-50.

Giles, H., Bourhis, R. Y. & Taylor, D. M. (1977). Towards a theory of language in ethnic group relations. In H. Giles (Ed.), Language, ethnicity and intergroup relations (pp. 307348). London: Academic Press.

Giles, H., Hewstone, M. & Ball, P. (1983). Language attitudes in multilingual settings: Prologue with priorities. Journal of Multilingual and Multicultural Development, 4, 81-100.

Giles, H. & Johnson, P. (1987). Ethnolinguistic identity theory: A social psychological approach to language maintenance. International Journal of the Sociology of Language, 68, 69-99.

Gomaa, Y. A. (2011). Language maintenance and transmission: The case of Egyptian Arabic in Durham, UK. International Journal of English Linguistics, 1 (1). Retrieved December 19, 2016, from https://goo.gl/j5HhTD

Google Earth. (n.d). Zraoua Nouvelle [New Zrawa]. Retrieved August 23, 2017 from https://goo.gl/cGR7Me

Gumperz, J. J. (1982a). Language and social identity. Cambridge: Cambridge University Press.

Gynan, S. N. (1985). The influence of language background on attitudes toward native and nonnative Spanish. Bilingual Review, 12 (1&2), 33-42.

Hamel, R. E. (1997). Introduction: Linguistic human rights in a sociolinguistic perspective. International Journal of the Sociology of Language, 127, 1-24.

Hamers, J. F. & Blanc, M. H. A. (2000). Bilinguality and bilingualism (2^nd ed.). Cambridge: Cambridge University Press.

Hamza, B. (2007). Berber ethnicity and language shift in Tunisia (Doctoral dissertation, University of Sussex, 2007). Retrieved March 3, 2017, from https://ethos.bl.uk/Logon. do;jsessionid=B3554FF34A4B8353662878BD67146CC3

Hatoss, A. (2005). Do multicultural policies work? Language maintenance and acculturation in two vintages of the Hungarian diaspora in Queensland, Australia. In J. Cohen, K. T. McAlister, K. Rolstad & J. MacSwan (Eds.), Proceedings of the 4th International Symposium on Bilingualism, pp. 1001-1009. Retrieved June 10, 2017 from http://www.lingref.com/isb/4/076ISB4.PDF

Hoffman, K. (2006). Berber language ideologies, maintenance, and contraction: Gendered variation in the indigenous margins of Morocco. Language and Communication, 26, 144167.

Holmes, J. (2013). An introduction to sociolinguistics. Malden, MA: Blackwell Publishing.

Huls, E. & Van de Mond, A. (1992). Some aspects of language attrition in Turkish families in the Netherlands. In W. Fase, K. Jaspaert & S. Kroon (Eds.). Maintenance and loss of minority languages (pp. 99-116). Amsterdam: Benjamins.

Julien, H. (2008). Content analysis. In L. M. Given (Ed.), The SAGE encyclopedia of qualitative research methods (pp. 120-122). Los Angeles, CA: Sage Publications.

Kang, B. A. H. (2004). Heritage language maintenance, acculturation, and identity: Chinese and Korean 1.5 generation immigrants in New Jersey (Master Thesis, Ohio State University, 2004). Retrieved December 3, 2016, from https://goo.gl/8Jb9ew

Kotzé, E. F. (2001). Ethnicity and language. In R. Mesthrie (Ed.), Concise encyclopedia of sociolinguistics (pp. 324-329). New York: Elsevier Science.

Laleko, O. (2013). Assessing heritage language vitality: Russian in the United States. Heritage Language Journal, 10(3), 89-102. Retrieved December 3, 2016, from https://goo.gl/GYhQsG

Lewis, E. G. (1975). Attitude to language among bilingual children and adults in Wales. International Journal of the Sociology of Language, 4, 103-125.

Li, W.L. (1982). The language shift of Chinese-Americans. International Journal of the Sociology of Language, 1982 (38), 109-124.

Lieberson, S. & Curry, T.J. (1971). Language shift in the United States: some demographic clues. International Migration Review, 5, 125-137.

Maamouri, M. (1983a). The linguistic situation in independent Tunisia. In P.M. Payne (Ed.), Language in Tunisia (pp. 11-21). Tunis: Institut Bourguiba des Langues Vivantes.

Maamouri, M. (1983b). Illiteracy in Tunisia: An evaluation. In P.M. Payne (Ed.), Language in Tunisia (pp. 139-157). Tunis: Institut Bourguiba des Langues Vivantes.

Mah, B. (2005). Ethnic identity and heritage language ability in second generation canadians in Toronto (Master Thesis, Ryerson University, 2005). Theses and Dissertations, 74. Retrieved December 3, 2016 from https://goo.gl/T3Qj2z

McCarty, T. L. (2012). Indigenous language planning and policy in the Americas. In B. Spolsky (Ed.), The Cambridge handbook of language policy (pp. 544-569). Cambridge: Cambridge University Press.

Mesthrie, R. & Tabouret-Keller, A. (2001). Identity and language. In R. Mesthrie (Ed.), Concise encyclopedia of sociolinguistics (pp. 165-169). New York: Elsevier Science.

Mesthrie, R. (2001). Language maintenance, shift, and death. In R. Mesthrie (Ed.), Concise encyclopedia of sociolinguistics. New York: Elsevier Science.

Myers-Scotton, C. (2006). Multiple voices: An introduction to bilingualism. Malden, MA: Blackwell Publishing.

Miyawaki, H. (1992). Some problems of linguistic minorities in Japan. In W. Fase, K. Jaspaert & S. Kroon (Eds.), Maintenance and loss of minority languages (pp. 357-368). Amsterdam: Benjamins.

Morgan D. L. (2008). Snowball sampling. In L. M. Given (Ed.), The SAGE encyclopedia of qualitative research methods (pp. 815-816). Los Angeles, Calif: Sage Publications.

Mougeon, R., Beniak, E. & Valois, D. (1985). A sociolinguistic study of language contact, shift and change. Linguistics, 23, 455-487.

Næssan, P.,Monaghan,P., & Mühlhäusler,P. (2010). Family language policies for Indigenous language maintenance and revival. Retrieved January 24, 2017, from https://goo.gl/Tnrjva

Nerteruk, O. (2010). Heritage language maintenance and loss among the children of Eastern European immigrants in the USA. Journal of Multilingual and Multicultural Development, 31 (3), 271-286.

Okamura-Bichard, F. (1985). Mother tongue maintenance and second language learning: A case of Japanese children. Language Learning, 35 (1), 63-89.

O'Rourke, B. (2011). Galician and Irish in the European context: Attitudes towards weak and strong minority languages. Hampshire & New York: Palgrave Macmillan.

Padilla, A. M. & Borsato, G. N, (2012). Psychology. In J. Fishman (Ed.), The handbook of

language and ethnic identity (2^nd ed., Vol.2, pp. 5-17). Oxford: Oxford University Press.

Paulston, C. B. (1987). Catalan and Occitan: Comparative test cases for a theory of language maintenance and shift. International Journal of the Sociology of Language, 31-62.

Pauwels, A. (2005). Language maintenance. In A. Davies & C. Elder (Eds.), The handbook of applied linguistics (pp.719-737). Malden, MA: Blackwell Publishing.

Prince, E. F. (1987). Sarah Gorby, Yiddish folksinger: a case study of dialect shift." International Journal of the Sociology of Language, 67, 83-116.

Putz, M. (1991). Language maintenance and language shift in the speech behavior of

German-Australian immigrants in Canberra. Journal of Multilingual and Multicultural Development, 12 (6), 477-491.

Pencheon, T. (1968). La Langue berbère en Tunisie et la scolarisation des enfants berbèrophones. Revue Tunisienne des Science Sociales, 13, 173-186.

Pencheon,T. G. (1983). La langue berbère en Tunisie et la scolarisation des enfants berbérophones [The Berber language in Tunisia and the scolarization of Berberophone children]. Tunisia. In P. M. Payne (Ed.) Language in Tunisia (pp. 23-34). Tunis: Institut Bourguiba des Langues Vivantes.

Phillipson, R. (1992). Linguistic imperialism. Oxford: Oxford University Press.

Phinney, J. S. (1990). Ethnic identity in adolescents and adults: Review of research. Psychological Bulletin, 108 (3), 499-514. Retrieved January 31, 2017, from https://goo.gl/DefcW1

Phinney, J. S. (2003). Ethnic identity and acculturation. In K. M. Chun, P. B. Organista & G. Marin (Eds.), Acculturation: Advances in theory, measurement, and applied research (pp. 63-82). Washington, D.C.: American Psychological Association.

Renz, B. B. (1987). Portuguese broadcasting in linguistic and cultural maintenance in northern Califomia. International Journal of the Sociology of Language, 68, 23-40.

Trudgill, P. (2000). Sociolinguistics: An introduction to language and society. London: Penguin.

Resnick, M. (1988). Beyond the ethnic community: Spanish-language roles and maintenance in Miami. International Journal of the Sociology of Language, 69, 89-104.

Sadiqi, F. (1997). The place of Berber in Morocco. International Journal of the Sociology of Language, 123, 7-21.

Sadiqi, F. (2010). The teaching of Amazigh (Berber) in Morocco. In J. A. Fishman(Ed.), Handbook of language and ethnicity (2nd ed., Vol. 2, pp. 33-44 ). Oxford : Oxford University Press.

Saumure, K. & Given, L. M. D. (2008). Convenience sample. In L. M. Given (Ed.), The SAGE encyclopedia of qualitative research methods (pp. 124-125). Los Angeles, CA: Sage Publications.

Simpson, J. M. Y. (2001). Minority languages. In R. Mesthrie (Ed.), Concise encyclopedia of sociolinguistics (pp. 579-580). New York: Elsevier Science.

Schlieben-Lange, B. (1971). The language situation in Southern France. Linguistics, 191, 101108.

Smolicz, J. J. (1985). Greek-Australians: a question of survival in multicultural Australia. Journal of Multilingual and Multicultural Development, 6 (1), 17-29.

Sridhar, K. K. (1988). Language maintenance and language shift among Asian-Indians: Kannadigas in the New York area. International Journal of the Sociology of Language, 69, 73-87.

Srivastava, R. N. (1989). Perspectives on language shift in multilingual settings. International Journal of the Sociology of Language, 75, 9-26.

Tamis, A. M. (1990). Language change, language maintenance and ethnic identity: The case of Greek in Australia. Journal of Multilingual and Multicultural Development, 11 (6), 481-500.

Tandefelt, M. (1992). Some linguistic consequences of the shift from Swedish to Finnish in Finland. In W. Fase, K. Jaspaert & S. Kroon (Eds.), Maintenance and loss of minority languages (pp. 171-196). Amsterdam: Benjamins.

Tatar, R. (2015). Parents' role in their children's development and maintenance of the heritage language: A case study of a Turkish-American immigrant family (Master thesis, Indiana University of Pennsylvania, 2015). Theses and Dissertations, 1273. Retrieved December 3, 2016, from https://goo.gl/jm785B

The Tunisian Ministry of Local Affairs. (2016). Al-tandhim al-baladi [The Municipal Vote Register]. Retrieved from https://goo.gl/bQrQEJ

Thompson, R. M. (1974). Mexican American language loyalty and the validity of the 1970 census. International Journal of the Sociology of Language, 2, 6-18.

Wang, Y. (2013). Language, culture, and identity among minority students in China: The case of the Hui. London: Routledge.

Wardhaugh, R. (2010). An introduction to sociolinguistics. Malden, MA: Wiley-Blackwell.

Woolard, K. A. (1984). A formal measure of language attitudes in Barcelona: a note from work in progress. International Journal of the Sociology of Language, 47, 63-71.

Wurm, S. A. (1974). Language policy, language engineering and literacy in New Guinea and Australia. In J. Fishman (Ed.), Advances in language planning (pp. 205-220). Retrieved from https://goo.gl/sZdk7c

Young, R. L. (1988). Language maintenance and language shift in Taiwan. Journal of Multilingual and Multicultural Development, 9 (4), 323-338.

Zhang, D. (2008). Between two generations: Language maintenance and acculturation among Chinese immigrant families. El Paso: LFB Scholarly Publishing.

Section I: Please answer the following questions.

Section II: For each of the following statements, please circle whether you Strongly Agree, Agree, Neutral, Disagree, or Strongly Disagree.

3. Imazighen have the right to speak Amazigh in the presence of those who do not understand it.

7. If someone talks to me in Tunisian Arabic, I should answer him in Tunisian Arabic.

Section III: Please circle the appropriate answer.

ÉÏ

ÔÈ

ÖÑÇ

ÉíÛíÒÇãáÇÇÈ äíËÏÍÊÊ / ËÏÍÊÊ äã Úã .1

ÉÈÓÇ

äãáÇ ÉÈ

ÇÌáÅÇ

áæÍÉÑÆ

ÇÏ

ÚÖæ

ÁÇ

ÌÑáÇ

:ËáÇËáÇ

ãÓÞáÇ

áÇæÎáÇ ÁÇ

äÈ

ÊÇãÚáÇ ÁÇ

äÈ - ã

ãÚáÇ ÁÇ

äÈ

-Ê

ÇÞíÞÔáÇ -

ÁÇÞÔáÇ - ÊáÇÇ

ÎáÇ -

áÇæÎáÇ

-ÊÇãÚá

Ç -

ÇãÚáÇ

-ÁÇÈáÂÇ

ÏÇ

ÏÌáÇ

-ÊÇæÎáÇ ÁÇ

äÈ

ÉæÎáÅ

Ç ÁÇäÈ -ÊáÇÇ

Îá

Ç ÁÇ

äÈ -

ãåÊÛá äÚ Ûí

ÒÇãáÇ

íáÎÊ

ÖÑÇÚÊ áå.2

ãÚä .1

áÇ .2

Table 1 Distribution of male participants by age group
Age group (years old)		Number of Participants
13-20		12
21-30		3
31-40		6
41-50		3
51-60		2
61 and more		3
Table 2 Distribution of male participants by marital status
Marital status	Number of participants
Single	18
Married	11
Table 3 Distribution of male participants by occupation
Occupation			Number of participants
Student			8
Construction field worker			4
Retired			3
Company worker			2
Baker			2
Carpenter			2
Unemployed			2
Waiter			1
Pastry maker			1
Farmer			1
Secondary school teacher			1
Nurse			1
Post office employee			1

Table 1 Distribution of female participants by age group
Age group (years old)		Number of Participants
13-20		5
21-30		10
31-40		3
41-50		3
61 and more		3
Table 2 Distribution of female participants by marital status
Marital status	Number of participants
Single	13
Married	11
Table 3 Distribution of female participants according to occupation
Occupation			Number of participants
Housewife			11
Unemployed			8
Student			4
Shopkeeper			1

	Interviewee's pseudonym	Age	Place of interview	Date of interview
Grandparents	Salah	63	An association	February 12, 2017
	Hammouda	74	An association	February 12, 2017
	Salwa	78	An association	February 14, 2017
Parents	Mohamed	55	An association	February 9, 2017
	Abderrahman	53	An association	February 12, 2017
	Ahmed	40	A café	February 13, 2017
	Alaa	41	A club	February 15, 2017
	Mahdi	47	A workshop	February 16, 2017
Single young adults	Khalifa	33	An association	February 12, 2017
	Arij	24	An association	February 14, 2017
	Hadi	24	A café	February 14, 2017

Appendix G. Transcription Symbols (Dresing, Pehl & Schmieder, 2015; Gumperz, 1982)

Material	Symbol
Pause	(...)
Overlapping	Speech overlaps are marked by //. At the start of an interjection, // follows. The simultaneous speech is within // and the person's interjection is in a separate line, also marked by //.
Incomprehensible	( )
Laughter	(laughter)
Discontinuations	/
Interviewer	I
Participant	P

P: The Amazigh language, for us, is the language of communication that we have been accustomed with. We got it from our fathers and mothers us our mother tongue. It represents the indigenous language of the peoples of North Africa (...) and it remains the only connection between the Imazighen of North Africa.

P: I acquired it [Amazigh] at home. When the Amazigh child is born, the first speech he hears is in the Amazigh language. He hears it from his father, his uncle [father's brother], his grandfathers, his grandmothers, his uncles [mother's brothers] and from all the family. This means that the language of communication between Imazighen in Tunisia is exclusively the Amazigh language.

I: What are the means that have helped your children acquire the Amazigh language?

P: It is (...) the language we use to communication with each other, that is, I communicate with

I: Do you ask your children and grand children to speak the Amazigh language at home or around Amazigh people?

P: There is no need to ask [children to speak Amazigh at home and within the community]. Naturally, when the child grows up, he/she understands that this [Amazigh] is the language of communication with his father and mother and with the other family members. I do not ask my child to speak Amazigh outside home because he may have friends and teachers who do not understand Amazigh, so he should talk to them in the language that all people know,

which is Tunisian Arabic. However, within the family and the Amazigh community, like the village of Zrawa, the language of communication is the Amazigh language.

P: On the national level, it [the maintenance of Amazigh]'s necessary because Amazigh is one of the sources of the Amazigh culture which, in turn, represents the Tunisian culture. Culturally speaking, the Tunisian culture dates back to 3000 years, and it should have a number of components. Among these we find a main constituent which is the Amazigh language. This language still exists today in this country.

P: This [maintenance of Amazigh] is now in the hands of the civil society since the government has not reacted yet in favor of this issue. Also, there are no demands from Imazighen for the governments' reaction or for teaching this language [Amazigh] at schools. Personally, I think that cultural associations contribute to promoting, maintaining, and enriching this language which, as I think, is a national asset.

P: It [Amazigh] isn't something concrete to be lost. In case Tunisia loses Amazigh, I can tell you what might happen, but me as an individual I won't (...) won't lose it.

P: It is something that is unlikely to be lost. The next generations may lose it and this will be a national cultural loss. Migration to cities and marrying outside the Amazigh community can lead to the loss of the Amazigh language. But as long as we live in this village it won't. This village is the secret behind the survival of this language. The existence of the village of Zrawa has helped us maintain the language.»

P: We use Amazigh and some Arabic words because, we like it or not, the Amazigh language in Tunisia mu/ some concepts, words, and notions have become unknown due to (...) maybe because there is no revival of the language, there is no studies centers to protect the language and introduce the true language to the young generations. We have started losing some, if not many, concepts. For those words related to modern sciences and technologies, we borrow them from Arabic or French or English, as other people [speakers of other languages] do. If Arabic suffers from this, one can't blame the Amazigh language.

I: The phenomenon that a language borrows words from another language is found in (...) in all languages. If you examine any language, you will find that it includes words borrowed from another language. This isn't our concern here. It seems your knowledge about this phenomenon has led you to say that the Amazigh language includes words from Arabic or from other languages. The question is whether or not you use only what other Amazigh people consider as Amazigh language, when communicating with your children.

P: To my children I use what I believe to be an Amazigh language with some Arabic words which have been used in place of words lost from the Amazigh language.

P: Always with some words lost from the Amazigh language and no longer exist, words which we don't know and we didn't take it from our parents.

I: Do you mean that speaking the Amazigh language implies that the speaker is Amazigh, don't you?

P: It's not necessarily, but 95%, true. You don't find an Amazigh person who doesn't speak the

Aamzigh language. However the opposite is true: a person from Arab origins may speak the

P: The point I got is that a person who doesn't speak the Amazigh language and lives in Zrawa,

where most of the inhabitants speak the language, acquire it as a matter of necessity.

P: At school (...) then in the street, in Tunis where I was born and lived my childhood.

Appendix I. Map of the Amazigh-speech Zones in Tunisia Based on Pencheon (1968)

Appendix J. Map of the Amazigh-speech Zones in Tunisia Based on Maamouri (1983)