Dialectal Change in a Northwestern Algerian City The Case of Sidi Bel-Abbes

Nadjouia RAOUD1 University of Algiers 2 [email protected]

ملخص نظرا للتلمس بين اللهجات الناجم عن الزحف الريفي من قرى و مدن مجاورة لمدينة سيدي بلعباس، شمال غرب الجزائر،فإن لهجة مدينة سيدي بلعباس تشهد تغييرا معتبرا.لقد اقترحت دراسات عديدة لتفسير التغيير في اللهجات في البلدان العربية أو الناطقة بالعربية ، إنما الغلبية كانت إما متمثلة في عدد محدود من الصوات أو محصورة على نموذج نظري واحد. يقترح هذا المقال مستويات لسانية أخرى،مثل المستوى المرفلوجي،النحوي، والمعجمي، بالضافة إلى ذالك ،فإن هذا العمل يعرض مزيجا من العوامل المتعلقة بالمتكلمين من حيث شبكاتهم الجتماعية ،مواقفهم، تمثيلتهم وهوياتهم التي تساهم في شرح تغيير اللهجة البلعباسية

Abstract Due to the dialectal contact resulting from a massive migration from different parts of the country, the Arabic variety of Sidi Bel-Abbes, a city in northwestern Algeria, is undergoing change. Countless studies have been suggested to account for dialectal change in Arab and/or Arabic speaking countries; however, most of them have either relied solely on phonological changes such as interdentals to dentals and the status of q/g/;, or suggested a single model to account for dialectal change. In addition to phonological changes, this work examines other linguistic level features, for example, lexical and morphosyntactic markers. Furthermore, it proposes a complex analysis whereby a combination of factors pertaining to the speakers' social relations, attitudes, representations, and identity are considered.

Keywords: Algerian Arabic; dialectal change; morphosyntax; Sidi Bel-Abbes.

1 I would like to express my thanks to the anonymous reviewers for their careful reading of my manuscript and their insightful comments and suggestions, in particular those pertaining to the description of the methodology and procedure as well as to some inconsistencies in the phonetic transcription. Any errors remain my entire responsibility.

© The International Journal of Arabic Linguistics (IJAL) Vol. V Issue 2 (pp. 151-187) 1. Introduction The questions of dialectal diversity, contact, and change in Arabic-speaking communities have attracted the interest of many dialectologists and sociolinguists. There are countless studies on contemporary sociolinguistics in the MENA (Middle-East and North Africa) region. The Maghreb dialects are known through the works of Caubet on Morocco and the Maghreb (e.g., 2002); Aguadé (2003); Aguadé, Cressier and Vincente on Morocco (1998); Messaoudi (2002) on Rabat; Hachimi (2007, 2011) on Fessis in Casablanca; Miller et al. on Arabic urban vernaculars (2007); and Miller and Caubet on Arabic sociolinguistics in the MENA region (2010). Studies in Arabic sociolinguistics and in the Mashreq include Mitchell (1976), El-Hassan (1979) and Sallam (1980) on speakers in Lebanon, Syria, Jordan and Egypt (cf. Owens 2001: 427); Holes 1987 on Bahrain; Owens (2001) and Bassiouney (2009) on Arabic sociolinguistics; Haeri (1997, 2003) on Cairo and the q variant, in particular; Al-Wer (2002, 2007) on Jordan; Germanos (2009) on Beyrouth. Among the linguists having described the linguistic situation in Algeria are Grand' Henry (1976); Boucherit (2004) on Algiers; Morsly (1996, 2012); Taleb-Ibrahimi (1997, 2004); and Dourari (2003, 2011) on the linguistic situation(s) of Algeria. Studies on northwestern Algeria include Siagh (1976) on the northwestern dialects of Ghazaouet, Oran (now the economic and second important city of Algeria), and Tlemcen; Dekkak (1979) on Tlemcen; Benrabah on Algeria, Oran and Ghazaouet (1994, 1999); Bouhadiba (1993, 1998, 2002), Labed (2014) and Guerrero (2015) on Oran; Dendane (1993, 2002, 2007) on Tlemcen; and Madouni (1996) on one Arabic variety of Sidi Bel-Abbes. In Sidi Bel-Abbes, a city in northwestern Algeria, two significantly different dialectal varieties of Arabic coexist: the bedouin dialect of Sidi Bel-Abbes and the old city centre sedentary dialect of Tlemcen. The bedouin/sedentary distinction, epistomized in 14th century Ibn Khaldun’s Muqaddimah, between the so-called sedentary (‘Haɖɑri’) dialects and the bedouin (‘badawi’) dialects, was used by the European dialectologists as early as the late 19th century until the middle of the 20thcentury (for an overview of the developments and approaches in Arabic dialects, see Miller 2004, 2007; Caubet and Miller 2010). The bedouin/sedentary distinction was also used to explain the socio-historical division in Algeria between the so-called bedouin (or nomadic) dialects and the sedentary dialects, due to the two distinct periods of the Arabisation of North Africa (cf. W. Marçais 1961). This classification has survived until now, though, as Miller notes, “[c]ategorizing a dialect X as a bedouin-bedouinized dialect does not mean that the speakers

152 pursue a nomadic bedouin way of life but that they display in their speech a number of features associated with bedouin dialects” (Miller 2007: 5-6). The sedentary dialects share phonetic, morpho-syntactic, and lexical characteristics between themselves; they are distinct from the bedouin dialects, which, in their turn, share many common features with one another. A further division, within the sedentary dialects, was set for the first time between old city dialects (“parlers citadins”) and village dialects (“parlers villageois”) in 1925 by William Marçais and Abderrahmân Guida. This article attempts to answer two major questions: the first one relates to what linguistic features are most involved in the dialectal change in Sidi Bel-Abbes and the direction of this change; the second one is whether change in progress is the result of social variation, koineisation, or other factors. Furthermore, it examines not only phonetic/phonological changes but also morpho-syntactic and lexical differences. All throughout this work, the terms 'bedouin' and 'rural' will be used interchangeably to refer to the old dialect of Sidi Bel-Abbes.

2. Setting the Stage for a Dialectal Study Ech-Charfi (forthcoming) states that “[f]or most of the Arab countries, urban growth started only in the second half of the 20th century, which means that many generations of migrants are still living side by side with those born and raised in the city”. The city of Sidi Bel-Abbes (sometimes referred to here as SBA) has known an important migratory flux not only from the neighbouring rural areas but also from various towns and cities, namely, Oran (now the economic capital and second largest city of Algeria) in the North, Mascara in the East, Saida in the Southeast and Tlemcen in the West. Being formerly a rural place, SBA has come to shelter various populations with their respective life modes: in the outskirts of the city, we find people of different origins, overwhelmingly rural, while people of higher economic standards have lived for generations in the city centre and, in particular, in wealthy neighbourhoods.

153 Figure 1: Sidi Bel-Abbes and adjacent cities [Source: mapnall.com]

The old dialect of Sidi Bel-Abbes belongs to the bedouin (nomadic) group (W. Marçais 1938; Ph. Marçais 1940), itself divided into the A dialects of the Sahara Nomads, the B dialects and the D dialects of the Tell Nomads of Algeria (Cantineau 1940, 1960a, 1960b; for the D dialects, see W. Marçais (1908)2. The bedouin (or rural) dialects are distinguished from the sedentary dialects of the old city centres of Algiers, Constantine and Tlemcen. This study focuses on the impact each of the sedentary Tlemcen and rural Sidi Bel-Abbes has on the new dialectal variety emerging in this city; furthermore, the influence of the capital city Algiers sedentary dialect through the media and the fast growing mobility necessarily makes the description (though partial) of the latter an integral part of this work.

2.1 Methodology of research There is no atlas of the regional dialects of Algeria, for example, in the vein of Trudgill and Hughes (1979) English Accents and Dialects, but descriptions of the Bedouin dialect of the “département” (county) of Oran (cf. Cantineau 1940, 1960) and data collected among old Belabbesi speakers show that it shares more features with the speech of Oran than with any other dialect. To trace older features of SBA speech and discover what features have changed, appeared (or disappeared), I appealed to three modes of investigation. The first one is represented by Norms, an acronym for non-mobile, rural, old male speakers (Chambers and Trudgill 1988); the second involves one of the modes of real time observation: to “compare earlier findings with current

2 C dialects are transition zone dialects of a nomadic type now having disappeared (Cantineau 1938). 154 ones: this is the normal procedure of historical linguistics” (Labov 1994: 73-4); and the third is based on apparent time observation (i.e., variation across age categories). The present work is based on fieldwork which I carried out in Sidi Bel-Abbes between December 2007 and September 2013 among over 300 participants. The youngest speaker is 8 and the oldest is 86. For ethical reasons, the names of the consultants have been changed. The data are constituted of 1) notes and diaries based mainly on participant observation, 2) recorded face-to- face and/or telephone conversations with speakers by means of semi-directive and directive interviews, and 3) directive surveys. The first elder speakers were interviewed in the winter of 2007 and investigations on the speech of younger speakers started at about the same time, involving both macro surveys and short interviews of individuals and micro interviews. The total number of the people exceeds 300, but because I was unable to gather sufficient information on some of the SBA dwellers whom I interviewed on the street, for the quantitative analyses, I decided to limit the number to 125 speakers, 52 males and 73 females. Among the 125 people interviewed, 91 of them live in Sidi Lahcen, a suburb about 10 kilometres away from the city centre, and 34 in the city centre. The selection of variables sprang from the linguistic differences found between the speakers under study; this is consistent with the view that search for linguistic change results from the observation that features present before come to be missing. The selection of the linguistic features was also undertaken on the basis of five criteria. The five criteria which I set after a period of participant observation and analysis of the corpus are: absence/change or appearance of new features, high frequency in everyday use, high degree of occurrence in the corpus, inter-personal as well as intra-personal variation, and markers3. Other features were discovered incidentally.

2.2 Some distinctive features of three regional and communal dialects Prior to examining what dialect features have disappeared, changed or been maintained, it is necessary to present the old characteristic features of each of the contributing dialects. Although full acknowledgement is made here of earlier works on the Arabic dialects of the

3 Using Labov's (1972) concepts of indicators -variables that are subject to social class variation- and markers - variables which are subject to both social class and stylistic variation, Trudgill explains that, being “relatively high in a speaker's consciousness” (1986: 10), markers are most likely to be modified by speakers in formal situations.

155 county of Oran by Cantineau (1940, 1960a, 1960b); the bedouin dialects of the North Constantine region (Ph. Marçais 1956) and the Maghrebi dialect (Ph. Marçais 1977); the sedentary dialect of Tlemcen (W. Marçais (1902) and the Arabic dialect of the Ulad Brahim of Saida (W. Marçais 1908); the sedentary dialects of Oran (Cantineau 1940, 1960a, 1960b); the sedentary dialects of Algiers (Millon 1937); the Arabic dialects of Constantine (Cantineau 1938); W. Marçais and A. Guiga (1925); D. Caubet’s questionnaire (2002); and some of the features in de Jong’s (2000) treatment of the bedouin dialects of the Sinai, this work is essentially based on 'time apparent' (cf. Labov 1994: 73-4) dialectal descriptions of corpora collected through fieldwork research that lasted about 6 years. Table 1 below displays some of the linguistic features that my surveys, interviews as well as previous descriptions by dialectologists unveiled of the old varieties of each of the northwestern bedouin dialect of Sidi Bel-Abbes, the northwestern sedentary dialect of Tlemcen, and the capital city Algiers sedentary dialect.

Table 1: Some old features of three regional dialects: SBA, Tlemcen and Algiers

Regions → Features ↓ Old bedouin SBA Old sedentary Tlemcen Old sedentary Algiers

Phonology 1.g, G, q, ʔ: 'Heart' G: Gɑɫᶀ ʔalb Qalb

2. Interdentals/ Interdentals Dentals/stops Dentals/stops dentals léçni!n létni!n létni!n 3. ɶ vs. a/o O ! tomro" ɶ ! t_mr_e a ! tamrae 'You fall ill.masc.'

4. Backing/ fronting: Backing: hɑ:k Fronting: ha:k Fronting: ha:k 'take. imperative.' 5. Affrication: tˢ No: béntè Yes: béntˢè Yes: béntˢè 'My daughter' 6. Bedouin diphthongization x²yr xè!r xè!r 'Better' 'Measured' Mawzu!n muzu!n muzu!n 7.Sedentary yumi!n yumayén yumayén diphthongization Dual:'two days'

156 Table 1 (continued)

8. Bedouin Yes: tʷakʷlè No: taklè Yes: taklè labialisation 'You eat-fem.' 9. Sedentary No: kima Yes: kimʷa Yes: kimʷa labialization: 'as' 10. Bedouin vowel Fi (tlémsa!n) F (tlémsa!n) F (tlémsa!n) maintenance vs. sedentary vowel elision: in Tlemcen 11. Sedentary Vowel bounce “My friend-fem.” s²=ébtè s²=béttè s²=béttè 'Young age' és socr és scor és scor Morpho-syntax 12. a.Verbs in u/aw yébdu yébda!w 'You start' yébda!w b.Verbs in è/ay tétmé<

157 Table 1 (continued)

Regions → Features ↓ 21. Negation in Neutral: Mani!< :a!réf/:a!rfa Mani!< :²!rfa (feminine) Ma-na:raf(ʃɛ))) participles 'I do not know-fem.' nd 22. 2 pers. Ro!=è t

nd 26. 2 pers. nta vs. nti nti:na for both masc. and nta vs. nti masc./fem. indep. fem. pr. distinction: nta/ nti 'you-fem/you. masc.' vs. 'nti:na for both masc./fem.' 27. Marking for some feminine No: ʕzu:ʒè Yes: ʕzu:ʒtè Yes: ʕdjo:ztè nouns: 'My mother- (with affrication) in-law' (and metathesis) Metathesis Metathesis No metathesis 28. 'I owe you' t-sa:l-ni t-sa:l-li t-sa:-li Lexicon 29. 'Something'

4 The q instead of the ; is used in public space by a male Tlemceni speaker, but here the focus is not on the Tlemcen glottal stop but on reflexive verb forms characteristic of Tlemcen Arabic. 158 Table 1 (continued)

Regions → Features ↓ 30. Suffixed forms 'Them ' No: hu:ma Yes: hu:ma Yes: )u!ma 'Today' lyu!m lyu!m lyu!m 31. Reflexive érRo!=è bw²=dè wa=dè pronouns: 'By myself; I alone' 32. Reflexive Ro!=è R²!sè Ro!=è pronouns 'Myself' 33. Possession ta:ʕ dja:l dja:l particle 34. Go: 'he goes' jɔʁda yém<è yém<è 35. Call: 'he calls' Yélca y:iyyé( y:iyyé( 36. Be able to: 'she Tnajjém tandjém Tandjém (with affrication) can' (with affrication) Taqdar ta;dar 37. Do: 'we do' Ndi!ro Né::amlo Ndi!ro 38.Give:'they give' y²:(o iméddo Iméddo 39. 'What'

159 Table 1 (continued)

Regions → Features ↓ 49. 'Yes' wɑ:h jɛ:h ɛ:h 50. 'No' lla la:la la:la 51. 'Corner' q²nt ;ant

2.3 Dialectal Change in Sidi Bel-Abbes To trace older and new features of SBA speech, I relied on my observations, surveys and interviews among speakers living in Sidi Bel-Abbes, cross checking with earlier descriptions of the D dialects (Cantineau 1940, 1960). From among a total of over 300 speakers interviewed, a large number of features were observed, some of which are displayed in table 2 below.

Table 2: Old and new SBA features Features Old speakers Young Classical/Standard Direction speakers Arabic 1.Interdentals vs. Stops Interdentals Stops Interdentals Sedentary 'Monday' léçni!n létni!n él ;içnayn 2.Diphthongisation vs. Diphthongisation Long vowels Diphthongization: Sedentary Long vowels: 'better' x²yr xè!r x²yr 3.Vowel maintenance Vowel Vowel elision Vowel maintenance Sedentary vs. vowel elision: 'By maintenance f-bél:abba!s fi:-bél:abba!s God's generosity' fi:-bél:abba!s tb²rèk é*- tab²!r²k² é*-*²!) tb²rèk é*-*²) *²) 4. Bounce No bounce Bounce Lexical change Sedentary S²di!qati 'My friend-fem.' s²=ébtè s²=béttè

'Young age' és socr és socr és scor és-sicari/é<

160 Table 2 (continued) Features Old speakers Young Classical/Standard Direction speakers Arabic 5. Feminine marking Unmarked fem. Marked Unmarked feminine Sedentary case: fem.case: case without the a 'Old woman' :zu!j :zu!ja vowel fall and without metathesis: :aju!z(un) 6. Negation in Sedentary imperative verbs la-tji!< ma-tji!< la:-ta;ti 'Don't come' la-tad:è!< ma-tad:è!< la:-tad:i! 'Don't curse (s.o)' 7. Negation in partic m²-qɑrja-:< mé<

10. 'To go' S²ɖɖ;ʃɑwwɑrr; R²!) àa)aba; ca!dara Sedentary ma

'Last night' él ba!ra= él ba!ra= él ba!ri=a Sedentary/ Merger for 'yesterday' and 'last 'Tomorrow' ʁda ʁadwa ʁada!/ʁadan night'8

5 The form =a!ja is a mixed one in the sense that it contains the sedentary lexical item =a!dja without the affrication characteristic of the sedendary dialects of Tlemcen and Algiers. 6 There is no past tense form for 'jocda', 'to go'. 7 See comment below about tables 1 and 2. 8 Cf. footnote 5. 161 If we consider Tables 1 and 2, we notice the emergence of many sedentary features characteristic of Algiers, which, in its turn, has already undergone change towards simplification and merging (e.g., the verb yém<è 'to go' (compare Tables 1 and 2) is now simplified to R²!), the merger (él ba!ra=) for both ja:més 'yesterday' and él ba!ra= 'last night', etc.).

1. Age and Migration Time are Not Sufficiently Revealing In SBA, the use of stopping does not decrease with all old speakers, so age per se cannot explain dialectal variation (see 2.1). I then investigated stopping and origin, which revealed the use of interdentals, dentals and mixed forms among speakers of urban origin (2.2) and rural origin (2.3). Eventually, I examined the correlation between stopping and time of migration, which still proved unsatisfactory (2.3).

3.1 A random sample of stopping and age In order to check for stopping, that is the phonetic change from interdentals to dental stops, quantitative analyses on a sample of 125 speakers were carried out randomly for an initially general picture. The analyses revealed that, although 95,64% of speakers in their twenties (20-29) displayed the use of stopping and 0% of the oldest speakers (76-86) used stopping, much fluctuation was still observed, as 90,34% of speakers aged 70-75 used stopping while younger ones in their thirties, forties, fifties and sixties displayed a lower use of stopping (see figure 2 below)9.

1,5

92,80% 92,19% 95,64% 90,34% 1 75,98% 66,71% 69,64% 59,25% 59,26% Stopping 0,5 0,00% 0 12- 13-14 15-19 20-29 30-39 40-49 50-59 60-69 70-75 76-86

Figure 2: Stopping and age in Sidi Bel-Abbes -a random sample

9 This question will be dealt with in a forthcoming paper. 162 3.2 Stopping and urban origin Perhaps more conclusive are the results of the correlation between stopping and speakers of urban origin in that almost all the speakers of urban origin (95,16%) use stops (fig. 3 below):

Stopping and urban origin 3,32% 1,61%

Dentals Interdentals Mixed forms

95,16%

Figure 3: Stopping and urban origin in Sidi Bel-Abbes

3.3 Stopping and rural origin The high scores of stopping (fig. 3 above) among speakers of urban origin might have led to the conclusion that speakers of rural origin would display a reverse pattern: a quasi totality using interdentals. It is not the case: only 31,74% use interdentals, 36,50% use dental stops, and 31,74% use mixed forms10, as in figure 4 below:

10 Mixed forms involve the use by speakers of both rural interdentals and urban dentals in intra and inter personal communication. 163 Stopping and rural origin

Interdentals 31,74%Mixed forms 31,74%

Dentals 36,50%

Figure 4: Stopping and rural origin in Sidi Bel-Abbes

The variation found among speakers of rural origin and the presence of mixed forms was symptomatic of dialectal change in progress; furthermore, the fact that speakers presented differences in the time of their migration led me to look for a correlation between stopping and time of migration.

3.4 A three generational model of stopping in SBA Trudgill's seminal works on dialects in contact posits that, when speakers of different varieties of a language meet in a new town, a new form emerges; furthermore, he states that this new dialect -a koine- is stabilised after three or more generations of migrants (cf. Trudgill 1986, 1998, 2004). The results of the three generational model of dialectal change in SBA is displayed in figure 5 below:

164 Stopping in SBA A three-generational model

1,2 98,46% 96,68% 1 0,8 g n i 0,6 52,31% p p o t 0,4 22,79% S 0,2 0 Settlers : 1st gen. 2nd gen. 3rd gen. Generations

Figure 5: Stopping and the three generational model in Sidi Bel-Abbes

The koineisation model only partially accounts for dialectal change in Sidi Bel-Abbes, because, despite the conclusive results of focusing after 3 or more generations (fig. 5 above), significant discrepancies are still found among speakers of the same migrant generation (cf. 4.2 and following). Therefore, the intricate situation in which the rapid urbanisation and contact between communal dialects11 is taking place in Sidi Bel-Abbes calls for a complex study, whereby combined factors are suggested, for example, those pertaining to the speakers' social networks (cf. James Milroy 1978; Lesley Milroy 1987, 2002; Lesley Milroy and Matthew Gordon 2003; Bassiouney 2009), communities of practice, and social meaning (cf. Eckert 2000; Hachimi 2007, 2011).

2. Identity, Attitudes and Representations Speech Accommodation Theory (SAT), first introduced by Giles 1973’s study, in which speakers were observed to adjust their speech to that of their interlocutors (S’hiri 2009: 320), was extended by Trudgill (1986: 1-4) to describe a situation as “contact between speakers of different regional varieties, and with regionally mobile individuals or minority groups who accommodate,

11 The dialect of Tlemcen is more communal than just regional because Tlemcenis in general consider themselves as a community distinct from others to the point that some women (also heard in the field recordings) speak about non Tlemceni women as “ 'Arbiyaat” (Arab women). 165 in the long run, to a non mobile majority that they have come to live amongst” (ibid., 3). Such accommodation reflects speakers’ attitudes and representations regarding both their own variety and that of their addressees. Representations of and attitudes towards languages or language varieties largely contribute in speakers’ actual linguistic practices and hence dialectal change. As Van Herk states, “society can treat language the same way it treats clothing, the arts, or business, as a thing to be debated and regulated” (Van Herk 2012: 5). In Algeria, the relationship between dialectal variety, prestige, and social identity is also noticeable in the speech of the people in Sidi Bel-Abbes. From the youngest middle school pupils to the oldest speakers, all the speakers interviewed maintain that, in general, old rural SBA speech is considered less prestigious than urban SBA, which, in its turn, is considered less prestigious than Tlemcen Arabic. Despite the positive status of Tlemcen speech, Tlemceni (mostly male) speakers also adjust to the speech of Algiers, considered by all, Tlemcenis and Belabbesis, as having the highest status. This hierarchical ranking is clearly expressed in their judgements not only on their own variety but also on other regional varieties.

4.1 SBA city dwellers vs. SBA rural migrants Rayane, a 22 year old fourth generation migrant of Moroccan origin, says about rural, recently arrived dwellers, that they are easily identifiable from their speech, their inability to get around town, and their dress:

bda!w yéddoxlo lè vil²jwa [the villagers have started to enter (the city)] fél )²dr² [in speech] ma ya::²rfu!< él bl²!yéss )adu!k... [...they don't know how to move around in the city] béss²= yba!nu bél :j²!r w éj jélla!ba…[...but they are spotted through

the voilette and the djellaba]

4.2 SBA dialect as a stigmatised variety When asked about what he thinks about the speech of SBA, Ali, a 65 year old Tlemceni notary public, states:

166 … le parler de Bel Abbès est particulier, c’est guttural, c’est…on a l’impression que c’est paysan… on a l’impression que c’est la/c’est la/ c’est la campagne/des/c’est une campagne non cultivée/qui ont//Y en a certains qui ont beaucoup d’intelligence...[...]...C’est un langage vulgaire/ C’est un langage très vulgaire// […the speech of Bel Abbes is special…it is guttural…we have the impression that it is rural…we have the impression that it is the…it is the…it is the countryside. It is an uncultured countryside. They have… some of them have much intelligence […] It’s a vulgar language. It is a very vulgar language]

Rurality is also stigmatised by its own people, and any attempt to speak another variety (i.e., Algiers) or language (French) is seen as a pushy attempt to disguise one's (rural) origin. For example, Amina, a 14 year old junior school girl of rural origin, expresses vehement contempt for people of rural origin who modify their speech (she was actually aiming at some of her classmates):

ʔanaja naʕraf ;²ᵴl)om/ʕrɔbɛjjɑ/ma jəʃqɑ:ʃ né=ki! lək/ajja hadu:k jʒu: jətkɑbrɔ ʕlijja w jɑhhɑdrɔ lʊʁɑ:t/zaʕma kiman hɑ:kka// rɑ:nɛ ŋgullək//ʃi ʃira:t jʒʊ ʕandɛ/jɔggɔʕdɔʷ/kɛskœ/ ʊ mənna ʊ mənna/ ʊ wa:ʃ rɑ:kɛ/wa:ʃ =abbi!ti / ʊ mənna// [I know their origin…countryside people…I needn’t tell you about them. And then these come to show off at me and they speak languages …like this, you know... Some girls come to me and start “what is”[in French] and so on…”How are you” [with the Algiers accent]…”What do you want?”[with the Algiers accent] and so on]

4.3 Tlemcen as a prestige dialect Speakers of Tlemceni origin who were interviewed consider that the SBA dialect (together with the people who speak it) lacks the refinement and “nobility” of Tlemcen people, illustrated by statements from Khalida and Ali.

167 Khalida, 75 Khalida is a 75 year old educated (A levels) woman who lives with her husband in a relatively posh neighbourhood of SBA. She is the third generation born “child” of a Tlemceni family, as her grandfather came to SBA, where he married her grandmother, who was born in Oran and came to live with her husband in a nearby wealthy neighbourhood. She considers herself as coming from one of the oldest aristocratic families that “made” Sidi Bel-Abbes. She starts introducing herself (in French), stating that nobility is innate and not given to any one:

Quand on est noble, on ne peut pas avoir tout le monde comme nous […]La noblesse, elle ne se vend pas et elle ne s’achète pas. Elle est née en nous memes [When we are noble, we cannot have all the people like us. Nobility cannot be sold and cannot be bought. We are born with it]

When asked how she speaks with SBA people, she admits that she is aware of her speech distinctiveness and the attitudes of Belabbesis to it:

Je ne vous le cache pas/je ne…/ki: n əbda nahdarr/j əbda:w jyéee²=ko ʕlijja/Je ne vous le cache pas ; c'est pas que je suis raciste/mon langage...je suis fière de mes origines/ /[Honestly, when I start speaking, they start laughing at me. Honestly, it’s not that I am a racist, but I am proud of my origins]

Ali, 65 Ali's statement that SBA speech is sometimes used with family for joking reveals the importance, at least inside Tlemceni homes, of Tlemceni speech:

A la maison/le parler bélabbesien/c'est juste pour rigoler/par exemple/pour embêter la maman/ le g au lieu du ;/mais y a un fond qu'on peut pas changer [At home, Belabbesi speech is just for joking...for example, to tease the mum, the g instead of the ;, but we cannot change the substance (of Tlemceni speech)]

168 4.4 The prestige status of French and Algiers Arabic Farida, 22 To the question of whether social class and parents' socioeconomic status is high and shows in a girl's language, Farida replies:

ʃɔ:fɛ/la majoritè lli tʃʊfi:ha tahdarr frɔᵴɔ ɛ ʊ mənna w ʕandha əl lɔɔg²j mli! =/ taʕʕarfɛ bəlli f əd dɑ:r/ ;imma bʷa:ha ;imma mʷha ;imma jku:n ʊsta:d ;imma ʊsta:da/;imma jku:n qɑ:rɛ/malgrè mʷha ma tku:nʃ taxdam wəlla bʷa:ha ma jaxdamʃ// [Look...the majority that you see speaking French and so on... and she has a good language...you know that at home...either her father or her mother is a teacher(mas.) or a teacher (fem.) or educated...even if her mother doesn't work or her father doesn't work.]

Dalila, 14 When asked how people speak when they want to show they are superior, 14 year old Dalila says: x²(r²!t tʁɑjɑr əllahʒa tta!==a bəlʕa:ni w tɔgʕɔd tɑhdɑr hɑ:kkɑ bəl frɔɔᵴɛ kimɑ hɑ:kkɑ/w x²(r²!t tɔgʕɔdd təʒbad ᵊd dzirijja kimɑ hɑ:kka/ʔana: hadʊ lli ma nəbʁɛhɔmʃ/kajni:n zu:ʒ ᵊhna:ja ʃətthɔm// [Sometimes she changes her accent on purpose, and she starts speaking French like this. And sometimes she starts speaking with the Algiers accent like this. I don’t like that sort. There are two of them here that I have seen]

5. Maintenance of Communal Dialects Factors which contribute to the maintenance of communal varieties among the two communities under study are quite complex. They do not rely solely on one single model (i.e., age, origin or time of migration); instead, it is the combination of two or more factors, for example, age together with education and identity, or time of migration and community of practice, which may enhance or inhibit dialectal change.

169 4.1 Maintenance of rural features among old SBA dwellers Among the maintainers are the 2 oldest representative speakers, one male and one female.

El Hadj, 86 Despite his having lived in the city centre of Sidi Bel-Abbes for over 50 years, 86 year old El Hadj displays the old SBA features: interdentals, diphthongs, vowel maintenance, etc. (cf. table 2) except for old SBA yocda 'to go', which he replaces by new urban SBA yro!=.

Zahra, 86 Zahra was born in Tessala, a rural place near SBA, but she was 2 when, after her mother's death in the 1920s, she settled with her father and his new wife, a Sidi Bel- Abbes dweller. Zahra uses all the rural features of old SBA: interdentals; diphthongisation, and frequently replaces /ɵ/ by [)] in ɵa:nє: )a!nè, ‘also’, as well as the la (ʃ) particle for negative imperatives. She displays variants of 'to go' that are quite similar to those of El Hadj. While she has no occurrences of mʃa, her variation between jɔʁda and rɑ!= are conditioned by tense differences; for example, while rɑ!= is used for past, jɔʁda is used for future reference.

i Rɑ!= : L ʁda rɑ!=ət əl wɑ)rɑn [The following day she went to Oran...] ii jɔʁda: tocdè cadwa [Are you going tomorrow?]

4.2 A closed community of practice: Abdelkader and his coworkers Fixed variables as religion, ethnic or tribal group of a community of practice may not only inhibit linguistic dialectal change but may also preserve the communal dialect, as in the case for Palestinians in Lebanon or Syria (Bassiouney 2009: 124). Fifty year old Abdelkader was born in Ben Badis (Descartes formerly), near Sidi Lahcen, a suburb at about 10 kms away from SBA, where he came at a very early age and has lived ever since. After various jobs and a two month stay in prison for 'trabendo' (illegal commerce), he became a taxi driver then a driver for the Sidi Bel-Abbes Algiers coach line. He is married to two women: one in Sidi Lahcen and the other in Sidi Khaled, a nearby semiurban town. He is holder of a professional typist certificate. Abdelkader belongs to a strong community of practice of SBA male coach drivers originally from rural areas around SBA: they work together, eat together, and sleep in the coach station when 170 they arrive in the middle of the night, to resume work a few hours later. In addition to that, they spend their leisure time together, an additional factor in Abdelkader’s maintenance of his rural speech. During the interview, Abdelkader predominantly uses the rural SBA forms at all linguistic levels: interdentals and diphthongs as well as syntactic and lexical features; however, there is some degree of variation depending on which of his two wives he is speaking about. For example, when speaking about his first (rural) wife, Abdelkader uses old rural SBA features, in particular, the interdentals and the lexical term əʃ ʃa:ʃra ‘boys’:

əl mʊhimm/tzawwaʒt en 77/gʕatt deux ans mʕəl mɑdɑ:m/w rɔħt dərt deux ans ta:ʕ lɑrmɛ//ʕami:n gʕatt ha mʕa l mrɑ w ro=t dərt ʕami:n//ʕandi wəldi lawwal za:jəd f əθ θmənja w səbʕɛ:n/ma:t əɫ-ɫɑh jaħmah/matu:li ga:ʕ əʃ ʃa:ʃra [Anyway, I got married in 77; I stayed two years with the Madame. Then I went to do two years in the Army. I stayed two years in the Army, then I went to do two years...I have a first son, born in seventy eight, who died, God rest his soul. All my boys died]

When he mentions his (Tlemceni) second wife, he uses some urban (and sedentary) phonological: ɖɑrwɑk « now » and lexical variants: [əð ðra:rɛ] 'the children' as well as [bənt] instead of the old rural [

5.3 Maintenance of communal dialects among old speakers of Tlemceni origin In her 2007 and 2011 studies, Hachimi posits that among the Fessis in Casablanca, a new category of speakers has emerged, that of a hybrid Fessi-Casablancan identity, who use Casablancan g instead of Fessi q, while those claiming pure Fessi identity maintain the stereotypical Fessi variant q. This is somehow comparable to Tlemceni male speakers in Sidi Bel- Abbes, with this difference that the glottal stop is substituted to (Algiers) q, and not SBA g. The maintenance of old Tlemcen features is generally observed among old speakers -women, in particular, who maintain the glottal stop, considered too effeminate by male speakers and therefore avoided in public space. Despite the long time spent in Sidi Bel-Abbes, 3rd generation

171 born Khalida has maintained all her communal features. Ali, like Tlemceni men, has levelled out the glottal stop (in public space), but he has maintained many other features, at all linguistic levels.

Khalida (cf. 3.3) Khalida's speech displays much code switching between French and Arabic, and when she uses the latter, it is the Tlemcen variety exclusively: dental stops: tla!ta,'three', )a!du,'these'; the glottal stop: kʷ_n t;è!sék él ;odra; [If fate throws you there (i.e., If you happen to go to Tlemcen)]; vowel elision: fkul xi!)a [Everywhere]; the back ² vowel: -²!)om, not SBA -a!)om 'their father'; the _ vowel: y=_bb r_bbè él kl_!m él=l_w [God likes sweet words], and other Tlemcen features (cf. Table 1).

Ali, 65 Ali is a notary public, who has lived 30 years in SBA, where his father is buried. His speech no doubt displays little variation; however, systematic changes in his use of several variants might also be interpreted as a case of accommodation (or style shifting), perhaps led by sociopragmatic principles. Ali speaks French, mostly Standard French, with rolled 'rs', sometimes switching to the Arabic (mixed Tlemcen/SBA) vernacular. Apart from the glottal stop, which has been completely levelled out, Ali's speech displays very few SBA phonological features, many of which coexist with the Algiers q, in particular in ga:l 'to say', Ali has maintained many of the Tlemcen phonological features:

ɶ : mɶcrè!b 'Morocco'; long vowels in verbs: ma

172 of the interview, said with both SBA phonological g and morphosyntactic ah, Ali maintains Tlemcen ɔ/u, instead of SBA a)):

verbs: ro!)om/rè!)om 'they are' j rabbi:w ‘they comfit/preserve’ regular (i.e.Tlemcen) plural adjectives: l ʕabasiyyi!n 'Belabbesi people'

At the lexical level, and apart from a few SBA markers, namely, kin terms and stereotypical Belabbesi interjections of the sort: xl²d²!r é -²!k 'God ruin your father's house', almost all the Tlemceni lexical features are maintained:

n aʕʕamlɔ 'we put' dya!l/ék 'your/s' (but ta!: to denote provenance: és s²n:² ta!: tlémsa!n) wa!< 'what'; wasmék 'what's your name?' na::amlo 'we do'; ném<è 'I go'; m²rr² 'once'.

5.4 A closed network of Tlemceni maintainers Now living in a relatively posh neighbourhood, Farid is a 35 year old pastry cook (and owner of a pastry shop), who came from Tlemcen a few years ago. His speech, which he intends to be refined, is careful, his voice and manners are soft. A factor which may have made him maintain his communal language (i.e. dentals and affrication, but not the glottal stop, which he replaces by g) may be that he has remained in a strong social closed network of Tlemcenis: he is married to a Tlemceni woman, with whom he has 2 children; he has Tlemceni friends; he is also keeping with the (Tlemceni) tradition, as he specializes in Turkish/Tlemceni pastry.

6. Variation and Accommodation Speakers of distinct dialectal varieties finding themselves in everyday talk exchanges may resort to short term accommodation, the latter very often resulting in long term accommodation, which, in its turn, brings about significant changes, namely the adoption of new features or features of one of the contributing dialects and the disappearance of features belonging to either dialect. It is, in other words, 'how language can reveal social relationships, such as how each of us, as social beings, adapts our language to suit the situation and the audience' (Van Herk 2012: 3). In Morocco, Hachimi found out that linguistic levelling among Fessis in Casablanca correlates

173 more with the social meaning of features than with the time of migration and that questions of contextand identity are more relevant in accommodation processes ( Hachimi 2007). As for Alger ia, the question of the construction of a new urban vernacular remains open, “due mainly to lack of comparative studies between Oran and Algiers” (Miller 2007: 22).

6.1 Phonological accommodation and identity projection The 'identity projection model' involves accommodation that does not necessarily correspond to actual addressees but to “images, or stereotypes, of the group the interlocutor belongs to, or of a socially attractive group not actually represented in the immediate context” (Auer 1998, in Kerswill 2002: 21). Some other Tlemceni male speakers are found to accommodate not to their interlocutors but to their representations of the speech of the people they are talking about. This holds particularly for phonological variation, illustrated by extracts from conversations containing g and q in the verb 'to say' by 2 male speakers: Ali and Hamid.

Ali, 65 From recordings of long conversations with notary public Ali, it may be said that his variation (or accommodation) is not solely attributable to the mixing characteristic of first or second generation migrant speakers. Ali’s mixed use of g and q cannot be attributed to mere coincidence because it has been observed to follow a systematic pattern: in the verb 'to say', the predominance of Ali's use of q when reporting what other people (men and women) say to him (i.e., qalli and qatli) and total absence of it when he reports what he says may indicate his avoidance of the variant q, while his use of q in his addressees' mouth is used to express more refinement than its harsher SBA counterpart g (the features involved are in bold):

gɔtlah tləmsa:n/qalli ɛ:h/ouais/qalli Sidi laxdar /gɔtlɔ wasmék ənta:ja/qalli ʔana:.../gɔtlɔ nta tləmsa:nɛ qalli lla/ʔa:na m əlmdija/[…] jé lwi è èksplikè kil ètè dé tlèmsèn/qalli l marra lləwwla n ətla:ga b … [I said to him “ You're Tlemceni ?” He said to me “ Yes ”...He said to me “Sidi Lakhdar”. I said to him “What's your name ?”He said to me “ I am...”I said to him “ You're a Tlemceni ?” He said to me “No, I'm from Medea” [...]I explained to him that he was from Tlemcen. He said to me “It's the first time I meet...”]

174 During the two hour face to face interview, he does not use a single Tlemcen glottal stop ʔ, and, in addition to his hypercorrected n-ətla:ga ‘I meet’, said with g instead of q (one of the exceptions in SBA), the occurrences of q and g when he reports conversations with different people are distributed as follows:

i With a man from SBA, Ali is shown to have adopted the SBA g in galli all throughout his conversation (he displays exclusive use of the g variant in the 24 instances), however keeping the (Tlemcen/Algiers) morphosyntactic o/u form: gut lʊ 'I said to him'.

ii Right in the middle of the interview, he alternates between g and q, and this time he is reporting a conversation with a man he met in Tlemcen :11 gutlu, 8 qalli, and (only)2 galli, which means that (almost) every time Ali reports his own speech, he uses gutlu, a mix of SBA phonological g and Tlemcen morphosyntactic u, but when he reports the other man’s speech, he uses qalli.

Hamid, 32 Hamid is a conductor on the Tlemcen Algiers coach line. Another example of identity projection is found in his sudden shift from SBA speech to Algiers when he mentions Algiers, where each of the two occurrences of 'I said to him'(in bold) are said in the two dialects:

gʊt lah l post rɑ:h ʒa:j ʕla ʔəsmɛ// rɑ:h ʒa:j m éd dza:jər// wa:ʃ qɔtlɔ// [I said to him 'The post is intended for me’. It’s coming from Algiers. What did I say to him?]

6.2 Spatial linguistic segregation among Tlemcenis Sid Ahmed, 32 Although Sid Ahmed has almost focused his urban SBA speech, he says that his use of the Tlemcen dialect is restricted to family and in group interlocutors while the SBA variety is used in public space:

məʃʃi y a des mots qui glissent/xɑ:(ər on a gardé da:jmənn f əd dɑ:r/ on parle él la)ja/n-ahɑdrɔ b él la)ja ət ta:ʕna// [It's not that words slip...because we have always kept at home...the dialect : we speak in our dialect].

175 In the extract below (where he codeswitches between Arabic and French), Sid Ahmed says he practices short-term accommodation even in public space. He exhibits such SBA features as g in jgu:l 'he says' and q in yétla!qa 'to meet'; the (yé)n passive form:

[...]par exemple, des fois, nku:n avec des amis… j'ai des amis belabbésiens nku:n mʕa:hɔm…dès que nətla:qa un tlemcenien, je change systématiquement de parler…jənxalʕɔ…jgʊl lək j’ai rien compris… majaffahmɔ wa:lʊ// […for example, sometimes, I am with…I have Belabbesi friends…I am with them…as soon as I meet a Tlemceni, I systematically change my speech…they are surprised…they say 'I have not understood anything'. They don’t understand anything.]

6.3 From short term accommodation to change among speakers of rural origin Many speakers of rural Belabbesi origin are also found to practice linguistic variation, in that, similarly to Tlemcenis (e.g., Sid Ahmed), they accommodate their speech according to their interlocutors, the place or setting in which they find themselves.

Samir, 20 Samir is a university student, who was born in rural Telagh, south of SBA (cf. map in Figure 1). He says that he uses his community’s (rural) variant lʁa 'to call' at home and the urban variant :²yy²( outside and with his peers. The direct interview I had with Samir displays variation in many features; for example, he sometimes uses both variants to mean 'to call' and the urban variant to mean 'to call someone on the telephone', though the latter meaning is also at times expressed by means of the rural variant lʁa. Samir says he uses the rural features with his mother, then he alternates between the rural and urban variants when he addresses his sister, while he uses the urban variant outside (cf. the Appendix). This variation among young speakers indicates dialectal change in progress, as Samir’s short term accommodation followed by long term accommodation will most probably lead to dialectal change, whence the urban features will completely supplant the rural ones irrespective of the communication context in which he will engage.

176 7. Levelling Out of Communal Features Applied to Arab (and Arabic-speaking) situations, social networks, resulting from economic, political and demographic factors -such as urbanisation- encourage social mixing, thus leading to linguistic change, precisely because of the “flexibility” of social variables such as social class and education, which enhance linguistic change from above; for example, a doctor of working-class origin may climb up the social ladder and acquire higher-class values and habits, including language (Bassiouney 2009: 124). When faced with an interlocutor with a dialectal variety different from their own, Belabbesis of rural SBA or Tlemcen origin have been observed to level out (or drop) their communal dialect features for those of another variety, may it be a prestige, new or local one, and this to varying degrees, at different linguistic levels, and for different reasons and purposes. The choice from among the g, q, and ; in Arabic and/or Arabic speaking countries, for example, has received much attention. In the Mashreq, the glottal stop has until recently been a prestige national variant in Egypt even among educated speakers in formal situations though q is reappearing on the urban Egyptian scene (Haeri 1997: 154-6). Owens (2001) notes that studies by Sallam (1980: 93), Abd-el Jawad (1981), Bakir (1986), Haeri (1991, 1998) and Daher (1998) revealed ; as a prestige variant among women while q was preferred by men. While ; is a national variant in many Middle-Eastern countries, because of its prestige position as a variant of the capital cities -for example, Cairo for Egypt and Beirut for Lebanon- it has remained a regional variant in the Maghreb cities of Tlemcen and Fes, not gaining prestige outside their communal sphere.

7.1 Levelling out of phonological features among speakers of rural origin Education and the opportunities of contact with urban speakers which it offers is not always sufficient for the acquisition of urban speech because neither education nor social networks alone enhance dialectal change, and even when this occurs, the change does not involve all the linguistic levels. For example, the acquisition of urban stops by 22 year old Houria, a young settler/first generation speaker does not render her speech urban. She exhibits a mixed style at the morphosyntactic level (e.g., the presence of the preposition ʕla after the transitive verb 'to call') and at the lexical level (rural lʁa and urban :²yy²(( for 'to call'), and even more so in

177 another conversation when she is asked to describe a situation of extreme anxiety, as when she recalls a day when her father fell seriously ill:

Houria: lʁi:t ʕla ʒɑ:rna/ki: ʒa hʊwwa ʊ martah/da:rlah əlməfta:h fi jəddah ʊ wxɑrajla) lsa:nah/wəlla ma ni:ʃ ʕa:rəf/ʒa:mɛ wa:ħəd mrɔee gʊdda:mna kimɑ nhɑ:k/mɑ:mɑ r²!=ét/ mʕa:) ʊ ʒɑ:rna tani:k [I called our neighbour. When he came with his wife, he put a key in his hand and took his tongue out or I do not know what...never had a person been ill before us like that. My mum went with him, and so did our neighbour] N.: ʃku:n əd da:h ləʂ ʂbè(²!r [Who took him to the hospital?] Houria: ʒɑ:rnɑ lli lʁi:tlah/gɔtlah rwɑ:=/əlli jəskənn ħda:na/ʊ ʕɑjjɑ(na ta:ni l ʕammɛ/ jassəknʊ f éz zɔɳqa t ta:ʕna/bəʂʂɑ=/ajja lʁɛna:lhɔm/béss²= di:k əl li:la/mani:ʃ ʕa:rfa hada:k ən nhɑ:r...// [Our neighbour, whom I called/ I asked him to come...the one who lives nextdoor and we also called my uncle/ they live in our street, but.. Then we called them, but that night...I don't know...that day!]

7.2 Levelling out: A three generational model and social networks among Tlemcenis Unlike his brothers and sisters, who have all become doctors and pharmacists, to respect the familly's tradition, Adel, Khalida’s 32 year old son, gave up school when he was a teenager. Being a 4th generation born child of Tlemceni origin, Adel displays none of the Tlemcen communal features; adding to this, the social networks he has woven with many friends from Oran and the community of practice (fishing) in which he has engaged have obviously contributed in the focusing of his SBA variety:

;anè ga!:éd n

178 campfire with my family. Now I am looking at H'miyyaan12, with those tables, tables for a day rent]

7.3 A community of practice and an open Tlemceni network Bilal is a 17 year old college young man of Tlemceni parents, who had come to settle in the 1990s in suburban Sidi Lahcen, where he was born. His father died and his mother is very sick. Bilal is usually found with his football playmates. I have not had the opportunity to observe Bilal's speech at home, but with his peers outside, his speech is characteristic of mixed (old and new) SBA speech, with no Tlemcen features. When he addressed me in the presence of his (male) schoolmates, Bilal used mainly the variety he uses with them: the rural features, including the interdentals, which he realises as dental stops when asked to repeat, as in the two extracts below; for example, the rural lexical ù² is also immediately converted into m²m² ‘my mother’. Whether he intended this correction as an attempt to adjust to urban SBA Arabic, this equally means that he considers the rural variety as less prestigious (and perhaps less intelligible or tougher) than the new urban variety. More than that, the fact that he uses rural SBA, and not urban SBA, may be further evidence that his playmates, with whom he constitutes a community of practice, also use rural SBA Arabic. Below are some examples of interdentals/stops (extracts 1 and 2) and ù²/m²m² (extract 3):

Extract 1 I.: winta taqro l anglais [When are you studying English?] Bilal: n)ar léçni!n [lit. Monday('s day)] I.: winta [When?] Bilal: n)²r létni!n [Monday('s day)] Extract 2: I.: cadwa qaysa!< tédoxlo [Tomorrow, what time do you start school?] Bilal: éç çménya [Eight o'clock] I.: qaysa!< [What time?] Bilal: ét tménya [Eight o'clock]

12 Adel uses the term H'miyyaan in a derogatory sense to refer to wandering (nomadic) people, comparable to the Roms in France. 179 Extract 3 This extract might confirm the idea that he uses rural SBA with his peers Bial: nro= m:a ù² [I'll go with my mother] N.: kifa) [What?] Bilal: m:a m²m² [With my mother]

Conclusion The old rural Arabic dialect of Sidi Bel-Abbes is changing, and this change is at first towards an urban Belabbesi variety. Because they are aware of the stigmatisation that characterises their communal variety, young speakers of rural origin are the least maintainers. As to Tlemcenis, and despite their pride in their Tlemceni culture and language, the situation is not that straightforward. On the one hand, young Tlemceni male speakers accommodate to SBA Arabic in Belabbesi public space at all linguistic levels, including the SBA g feature, particularly in male dominated spheres. On the other hand, older women are the strongest maintainers of their traditional dialects while their male (old) counterparts accommodate, at least at the phonological level, to the speech of Algiers, not because of the prestige capital city status that it enjoys but because its feature q enables them to avoid the effeminate glottal stop ; and, by the same token, the too “vulgar”, “guttural”, “uncultured”, “rural” SBA g. Furthermore, the question of the linguistic level changes is important when considering change from above13, as change in SBA involves the phonetic /phonological features in the first place (e.g., variation in G, q, ;), then come the lexical features (e.g., əð ðra:rɛ ‘the children’; y :²yy²( ‘to call’). As to many morphosyntactic features, they are, if they change at all, the last to be modified because either they are below the level of the speakers’ consciousness in the Tlemcen dialect (e.g., verbs such as y rabbi:w ‘they comfit/preserve’) or, in the case of the Sidi Bel-Abbes dialect, their too high salience makes them weak candidates for change from above (e.g., the ah suffix for object particle masculine pronoun). Speakers of the same migrant generation are not all equal when it comes to embracing urban SBA features, and thus the three generational model alone cannot guarantee dialectal change. Neither can modern, urban identity aspirations be considered alone, as the acquisition of

13 'Change from above' is conscious change towards dialectal features characteristic of a higher class (or prestige or (locally) dominant) language or language variety by speakers of less prestige languages or varieties (cf. Labov 2007). 180 new urban speech requires time. Therefore, a multimodal approach is more capable of explaining the dialectal change occurring in Sidi Bel-Abbes, for example, a combination of time of migration and social networks together with identity aspirations: for even the combination of identity and time of migration will not necessarily lead to urban speech if the speaker does not engage in the social networks of the target variety and values. Thus, if they identify with urban SBA speech and values, Belabbesi dwellers of Tlemceni origin whose social networks offer them opportunities of contact with the urban SBA dialect will acquire it after three or more generations. Likewise, speakers of rural origin who aspire to a prestige variety and whose social network is open to urban speech are more likely to acquire urban features than those whose network is limited to their own closed community provided the former have spent enough time in the city of Sidi Bel-Abbes to acquire them. The question of dialectal change however is not that straightforward and calls for more thorough and systematic studies on Algerian dialects and their relation with one another.

181 APPENDIX

What follows is an extract about the use of ‘to call’ from an interview I had with Samir, a 22 year old student and first generation speaker of rural origin.

I.: w é<

Translation of the extract above [I.: And the old woman (Samir's mother), you say to her lcè l amina [rural 'call Amina'] or :²yy(è l amina[call Amina]? Samir: No, when I'm at home, I say call [lcè l] Amina' I.: You say lcè Samir: Yes, at home, I say lcè. I.: So at home you speak like that and when you go out, there's another way of speaking. Samir: No, on the telephone, when I want to speak on the telephone I say [n:²yy²(la)] 'I'll call that gentleman'. I.: OK. To your friend, when you go out...you go to his place... You say to them [:²yy²(] 'call Mister X or [lco] 'call Mister X'? Samir: :²yy(o ('call') Mister X I.: To Amina? Samir: No, sometimes, I say :²yy(o ('call') and sometimes I say lco 'call').…]

182 References

Aguadé, J. (2003). Notes on the Arabic Dialect of Casablanca (Morocco). AIDA 5 Proceedings, Cadiz, pp. 1-8. https://www.academia.edu/2466275. Aguadé, J., Cressier P. and Vicente, A. (eds) (1998). Peuplement et Arabisation au Maghreb Occidental. Dialectologie et histoire. Revue des mondes musulmans et de la méditerranée. Madrid-Zaragoza, Casa de Velazquez, Universidad de Zaragoza, 175p. Al-Wer, E. (2002). Education as a Speaker Variable. In A. Rouchdy (ed.) Language Contact and Language Conflict Phenomena in Arabic. New York: Routledge-Curzon, 41-53. Al-Wer, E. (2007). The Formation of the Dialect of Amman: From Chaos to Order. In Miller, C., Al-Wer, E., Caubet, D. and Watson, J. (eds) Arabic in the City: Issues in Dialect Contact and Language Variation. London: Routledge. Bassiouney, R. (2009). Arabic Sociolinguistics. Edinburgh: Edinburgh University Press. Benrabah, M. (1994). Attitudinal Reactions to Language Change in an Urban Setting. In Sulaiman (ed.) Issues and Perspectives in Arabic Sociolinguistics. London: Curzon Press Ltd., 213-25. Benrabah M. (1999). Les filles contre les mères. Revue de Linguistique et didactique des langues (Lidil), Les parlers urbains (Grenoble) 19, 11-28. Boucherit, A. (2004). L’arabe parlé à Alger. Aspects sociolinguistiques et énonciatifs. Alger: ANEP. Bouhadiba, F. (1993). In Loci for Norm and the Arabic Language Continuum: in Defence of MSA. Cahiers de Dialectologie et de Linguistique Contrastive, vol. IV, Cahier 1. Université d'Oran. Bouhadiba, F. (1998). Continuum linguistique ou alternance de codes? Essai d'analyse dynamique des faits: Etat des lieux. Cahiers de Linguistique et Didactique (GRLDLD) N° 1 : 1-11. I.L.E. Oran: Algérie. Cantineau, J. (1938). Les parlers arabes du département de Constantine. Actes du IVe Congrès II, 849-863. Cantineau, J. (1940). Les parlers arabes du Département d’Oran. Revue Africaine 84: 220-231. Cantineau, J. (1941). Les parlers arabes des Territoires du Sud. Revue Africaine 85, 72-77. Cantineau, J. (1960a). Cours de phonétique arabe. Paris: Librairie C. Klincksieck. Cantineau, J. (1960b). Études de linguistique arabe. Paris: Librairie C. Klincksieck. 183 Caubet, D. (2002). Questionnaire de dialectologie du Maghreb (d’après les travaux de William Marçais, Marcel Cohen, G.S. Colin, Jean Cantineau, David Cohen, Philippe Marçais, S.levy). EDNA Esudios de Dialectologia Norteafricana e Andalusi 5, 73-92. Caubet, D. and Miller, C. (2010). Arabic Sociolinguistics in the Middle East and North Africa (MENA). In Ball M. (ed.) Sociolinguistics around the World. Londres/New York: Routledge, 238-256. Dekkak, M. (1979). Sex Dialect in Tlemcen: an Algerian Urban Community. PhD thesis, London: SOAS. Dendane, Z. (1993). Sociolinguistic Variation in an Urban Context: the Speech Community of Tlemcen. Unpublished thesis. University of Oran. Dendane, Z. (2002). Aspects of women's speech in Tlemcen. Cahiers de Linguistique Didactique, Nouvelle Série ½. Oran, 38-46. Dendane, Z., (2007). Sociolinguistic Variation and Attitudes towards Language Behaviour in an Algerian Context: The Case of Tlemcen Arabic. Unpublished dissertation. University of Oran. Dourari, A. (2003). Les malaises de la société algérienne, crise de langue et crise d'identité. Algérie: Casbah. Dourari, A. (2011). Politique linguistique en Algérie : Entre le monolinguisme d’Etat et le plurilinguisme de la société”. Politiques linguistiques en domaine francophone, Vienne. http://www.cnplet.net/file.php/1/cnplet_mina/navi-horiz/doc_recherche/4.pdf. Ech-Charfi, A. (forthcoming). The Expression of Rural and Urban Identities in Arabic. to appear in Reem Bassiouney & Keith Walters (eds.). The Routledge Handbook of Arabic and Identity. Routledge. Eckert, P. (2000). Linguistic variation as social practice: the linguistic construction of identity in Belten High. Malden: Blackwell Publishers. Grand' Henry, J. (1976). Les parlers arabes de la région du Mzab (Sahara algérien). Brill: Leiden. Germanos, M-A. (2009). Identification et emploi de quelques stéréotypes, traits saillants et autres variables sociolinguistiques à Beyrouth (Liban), thèse de Doctorat. Université de la Sorbonne Nouvelle (Paris 3).

184 Guerrero, J. (2015). Preliminary Notes on the Current Arabic Dialect of Oran (Western Algeria). Romano-Arabica XV. University of Bucharest Center for Arabic Studies. Hachimi A. (2007). Becoming Casablancan: Fessi in Casablanca as a case study. In Miller, C. et al. (eds) Arabic in the City: Issues in Dialect Contact and Language Variation. London-New York: Routledge-Taylor, 97-122. Hachimi, A. (2011). Réinterprétation sociale d’un vieux parler citadin maghrébin à Casablanca. Langage et Société, n°138. Paris. Haeri N. (1997). The Sociolinguistic Market of Cairo. Gender, Class and Education. London/New York: Kegan Paul International. Haeri, N. (2003). Sacred Language, Ordinary People: Dilemmas of Culture and Politics in Egypt. Palgrave Macmillan: New York. Holes, C. (1987). Language Variation and Change in a Modernising Arab State: The Case of Bahrain. London: Kegan Paul International. Ibn-Khaldoun, Abd el Rahman Ibn Muhammad (1997) Discours sur l'histoire universelle (Al- Muqaddima), traduit par Vinvent Monteil, Sindbad. Jong, R. E. (de). (2000). A Grammar of the Bedouin Dialects of the Northern Sinai Littoral. Bridging the Linguistic Gap between the Eastern and Western Arab World. Leiden: Brill. Kerswill, P. (2002). Koineization and accommodation. In John, K. Chambers, Peter Trudgill and Natalie Schilling-Estes (eds) The handbook of language variation and change. Oxford: Blackwell. 669-702. Labed, Z. (2014). Genealogical Koineisation in Oran Speech Community: The Case of young University Oranees. Unpublished PhD Dissertation. University of Oran. Labov, W. (1972). Sociolinguistic Patterns. Philadelphia: University of Pennsylvania. Labov, W. (1994). Principles of Linguistic Change, vol 1: Internal Factors. Wiley-Blackwell. Labov, W. (2001). Principes of Linguistic Change, vol 2: Social factors. Wiley-Blackwell. Labov, W. (2007). Transmission and Diffusion. Language.Volume 83, Number 2, June 2007, 344- 387. Linguistic Society of America. Madouni, J. (1996). Lexique et description d'éléments grammaticaux d'un parler de l'Ouest algérien (région de Sidi-Bel Abbés), thèse pour le Doctorat d'Etat, sous la direction de Dominique Caubet. Université Paris III, Paris.

185 Map of Sidi Bel-Abbes: http://www.mapnall.com/fr/Carte-géographique-Sidi-Bel-Abbès_ 1135726.html

Marçais, P. (1940). Algérie. Langues. Encyclopédie de l’Islam, Tome I. Leiden: Brill. 384-390. Marçais, P. (1956). Le parler arabe de djidjelli (Nord Constantinois, Algérie) [= Djidjelli], Publications de l’Institut d’études orientales d’Alger, XVI. Paris: Adrien-Maisonneuve, XVIII + 648 p. Marçais, P. (1977). Esquisse Grammaticale de l’Arabe Maghrébin. Librairie d’Amérique et d’Orient. Paris: Adrien Maisonneuve. Marçais, W. (1902). Le dialecte parlé à Tlemcen: Grammaire, textes et glossaire. Paris: Leroux. Marçais, W. (1908). Le dialecte arabe des Ulad Brahim. Paris: Honoré Champion. Marçais, W. and Guida, A. (1925). Textes arabes de Takroûna, (Textes, Transcription et Traduction annotée). Paris: Leroux. Marçais, W. (1938). Comment l’Afrique du Nord a été arabisée. Annales de l’Institut d’études orientales de l’Université d’Alger [AIEO], IV, 1-12, and XIV, 1956, 5-17.(Repr. William Marçais Articles et conférences. Paris : Adrien Maisonneuve, 171- 192, 1961). Messaoudi, L. (2002). Parler citadin, parler urbain. Quelles différences? 5th Aida Conference, Cadiz. Miller, C. (2004). Variation and Change in Arabic Urban Vernaculars. In Haak, M., Versteegh, K. and de jong, R. (eds) Approaches to Arabic Dialects: Collection of Articles presented to Manfred Woidich on the Occasion of his Sixtieth Birthday. Amsterdam: Brill.177-206. Miller, C. (2007). Arabic Urban Vernaculars: Development and change. In Catherine Miller et al. (eds) Arabic in the City: Issues in Dialect Contact and Language Variation. London: Routledge. Milroy, L. (2002). Social Networks. In J.K. Chambers, Peter Trudgill and Nathalie Schilling- Estes, The Handbook of Language Variation and Change. Oxford: Blackwell, 549-572. Milroy, L. and Gordon, M. (2003). Sociolinguistics: method and interpretation. Oxford: Blackwell. Morsly, D. (1996). Alger plurilingue. Plurilinguismes 12, 47-80. Morsly, D. (2012). La sociolinguistique en Algérie: état des lieux et perspectives. Réflexions et perspectives, Revue Scientifique et académique de l'Université d'Alger. 186 Owens J. (2001). Arabic Sociolinguistics. Arabica, 48, 419-469. Siagh, Z. (1976). Le système phonologique de quelques parlers arabes de l'ouest algérien (Ghazaouet-Oran-Tlemcen: essai d'analyse). DEA, Alger. Taleb-Ibrahimi, K. (1997). Les Algériens et leur(s) langue(s). Éléments pour une approche sociolinguistique de la société algérienne. Alger: El Hikma. Taleb-Ibrahimi, K. (2004). L'Algérie: coexistence et concurrence des langues. L'année du Maghreb I, 207-218. http://anneemaghreb.revues.org/305?lang=ar. Trudgill, P. J. and Hughes, A. (1979). English Accents and Dialects. London: Edward Arnold. Trudgill, P. J. (1986). Dialects in Contact. Oxford: Basil Blackwell. Trudgill, P. J. (1998). The chaos before the order: New Zealand English and the second stage of new-dialect formation. In Ernst Hâkon Jahr (ed.). Advances in historical sociolinguistics. 1- 11. Berlin: Mouton de Gruyter. Trudgill, P. (2004). New-dialect Formation: The Inevitability of Colonial Englishes. UK: OUP. Van Herk, G. (2012). What is Sociolinguistics? Wiley-Blackwell.

187