1Following urbanisation and the development of mass media, a group of “supralocal” spoken varieties1 of Arabic has apparently emerged in many Arabic-speaking countries in the last 50-150 years. Such “varieties” could be defined as codes employed in contexts of communication that involve Arabic speaking compatriots whose regional origins are diverse, or unknown to the participants. This is what is also happening in Morocco, although here, unlike other countries, it is not clear whether the supralocal variet(ies) that seem(s) to be currently spreading have directly originated from any of the Arabic-speaking centres of the country.

2In previous sociolinguistic works concerning Morocco, this phenomenon has been approached under two different focuses: some scholars, such as Youssi (1992), have followed the same line as, for example, Mitchell (1986, cited in Vicente 2008) and dealt with formal and semi-formal contexts of communication, in which, according to their view, an intermediate variety between Modern Standard Arabic (fuṣḥā) and Moroccan Arabic dialect (or spoken Moroccan Arabic, commonly called dāriža) is usually employed. This variety (which Youssi calls arabe marocain moderne), supposedly uniform for all Moroccan Arabic speakers, virtually consists of a mixture of prevalently dāriža grammar and morphosyntax, with frequent usage of fuṣḥā lexical items (Messaoudi 2003: 145). Others, such as Messaoudi (2001, 2003) and El Himer (2015), have moved from the classical distinction operated in studies on Maghrebi dialects between parlers citadins (“old urban varieties”) and parlers ruraux (“rural varieties”) and identified a third type of emerging variety, parlers urbains (“new urban varieties”); these present a mixture of linguistic features drawn from the first two types, although rural features prevail (at least in the context studied by these two authors, i.e. the towns of Rabat and Sale respectively).

3The studies mentioned above have the merit of focusing on phenomena which are central to the understanding of ongoing socially-related language development in Morocco; however, they all attempt to describe a more or less stable variety (or type of varieties) consisting of a co-occurring set of features and a sort of “grammar” rules. In fact, in each of the cases described, the use of such features seems to be too variable to prove the existence of a single “variety”, with its distinctive grammar and structure. Therefore, if one is to analyse the nature of the supralocal “way of speaking” that Moroccan Arabic speakers recur to, it seems preferable to first find out which contexts require a “supralocal” (i.e. “national” and “de-localised”) code, and then focus on the single linguistic features which are employed to communicate and express oneself in such contexts, rather than describe a whole, supposedly homogeneous “variety”. A list of such contexts may be drawn that includes the following:
- Talk to Moroccans of regional origin other than one’s own
- National media, i.e. media addressing a Moroccan audience (radio, television and internet)
- Political and scientific debates
- University lectures

4The list does not imply that local dialects, fuṣḥā, Amazigh or even a foreign language will never be employed in any of these contexts. However, empirical observations suggest that these are the situations where it is most likely that a “supralocal” way of speaking Arabic will be chosen for communication.

1. The field study and the media corpus

5From September 2015 to July 2016, a field study conducted in the Moroccan town of Temara, situated ca 6 km south-west of Rabat, aimed at collecting data in the form of live conversations in Moroccan Arabic among the residents. The purpose was to build a picture of sociolinguistic variation in spoken Moroccan Arabic in the context of interactions among Moroccans of different regional origin (one of the contexts that, as has just been observed, would demand a de-localisation of one’s speech).

6From this point of view, Temara is ideal for the circumstances of its urbanisation: a rural municipality until as close as 1983, it has witnessed a dramatic growth of its population in the last few decades (from ca 20.000 in 19712 to over 300.000 people in 20143) following massive immigration from other Moroccan areas, both rural and urban. Previous sociolinguistic studies on Moroccan towns (particularly: Messaoudi 2001; El Himer 2015) have described dynamics concerning cities with an historical urban population that was characterised by an ancient urban culture; on the other hand, Temara has a sort of “new-town” status as its original nucleus was rural rather than urban, with possibly different implications on the town’s inter-group dynamics. Besides this, the mere fact of hosting a great number of immigrants from many areas of the country raises questions about whether this socio-demographic situation will affect the way of speaking of the people that grow (and learn to speak) in this new reality, with Moroccan families of different regional origin living side by side in the same neighbourhood. Similar cases have already been widely studied in the domain of new-dialect formation in English-speaking areas (cf. Kerswill 1996; Trudgill et al. 2000, among others), although the same kind of phenomena will not necessarily produce the same linguistic outcomes in Temara, given the social, cultural and linguistic differences of the area.

7In order to answer these questions, a corpus of approximately 30 hours of recordings (before selection for the analysis) was collected during the field study, so as to obtain a sample of the actual speech of Moroccan Arabic speakers residing in Temara. Since the aim was for the informants’ speech to be as spontaneous as possible (acknowledging to have been influenced in this by a Labovian concern for the Observer’s Paradox; cf. Labov 1972: 208-209), it was decided to record either spontaneous conversations whose participants were not aware of being recorded in that specific moment, or focus groups made up of people well acquainted with each other. However, as this data collection “policy” only allowed to collect the speech of youth, ordinary interviews were also done to extend the sample of informants to elder elements of the community: as a result, about 40 informants were included in the final sample.

8The analysis of the corpus will aim at identifying which linguistic features vary by effect of being charged with any kind of social value, and how speakers exploit the social significations of these variables in the frame of their interactions in contexts that require a de-localised linguistic code. This approach is in line with the choice, stated above, of focusing on single features rather than looking for what constitutes a Moroccan supra-local “variety”; it also rests on the scientific works of Eckert (2003; 2008) and Silverstein (2003), which bring forward a view of the language–speaker relationship that is based on the speaker’s active work of interpretation and assignment of meaning to form, as well as on a comprehensive view of language variation which goes beyond the idea that language is a mere dependent variable of social categories.

9Following what has been said, the next question to be answered in order to proceed with the analysis of the data is: how can socially-marked linguistic features be identified, and how can their social meaning be defined?

2. The media corpus

10To partially solve these issues, another corpus consisting of media products of the fiction genre has been tentatively added to the field corpus; this second corpus includes:
- Foreign soap-operas dubbed in Moroccan Arabic
- Moroccan sit-coms
- Youtube animated series (e.g. Bouzebbal)4

11The usefulness of the media corpus as a term of comparison is based on three assumptions:
- Speech in this kind of products is script-based, and therefore not only “thought in advance”, but also (presumably, or in most cases) faithfully reproduced by the performer (be it a dubber or a real actor) in the same form conceived by the scriptwriter.
- The scriptwriter’s own use of language in writing the script (i.e. in choosing how to make each character in the plot speak) should reflect the connotational values that he/she assumes their audience assign to linguistic features. This way, it seems that the texts could at least provide an interpretation (made by native speakers, i.e. the scriptwriters) of the link between certain linguistic features and the value that Moroccan Arabic speakers charge them with.
- The advantage of the texts of media fiction is that the connotation of a feature is usually straightforward: if the scriptwriter does associate a given trait to a certain social group, then that trait will be equally found in the speech of all the members of that group who are in the storyline; conversely, if he/she does not associate that trait to the same group, then its members will not show it in their speech, or at least not consistently.

12The problem of the interpretation of sociolinguistic variation offered by the media is obviously that it will not be necessarily representative of that found in the community. On the other hand, the shortcoming of the data obtained from the field corpus is that the speakers’ interpretation of the sociolinguistic meaning of a feature is less explicit (whereas in the media texts it is more easily inferable from the pattern of association between a character and its way of speaking); conversely, the utility of the field corpus is that it constitutes a sample of the real everyday speech of the community. In sum, if the pictures given by the two corpora are by no means complementary, they allow at least a double perspective: features that are variable in both of them can be picked out for analysis, differences in their distribution in the field and media corpora can be examined and, finally, an interpretation of the sociolinguistic value that speakers charge those features with can be made on the basis of such analysis.

3. The features chosen for analysis

13A first analysis of the two corpora has led to the individuation of a great number of relevant linguistic features; however, only four have been selected for analysis:
- The phoneme /t/, which presents a wide number of allophones that go from [t] to [tˢ] / [ts] (affricated “t”) or [t^š] / [tš] (palatalised “t”, sometimes going as far as [š])
- The alternation between the phonemes /q/ and /g/ in a selection of lexemes where such alternation is found to be socially relevant both in the media and in the community studied
- The alternation between the seemingly semantically equivalent verbs /hdəṛ/ and /dwa/, both approximately combining the meanings of the two English verbs “to talk” and “to speak”
- The deceleration of the end of sentences, also affecting the length of vowels, in what could be called a prosody-affected phonetic phenomenon (e.g. š-bāl lī-k f hād-əl-fīləm [ʃ-bæ͜l͜ lik f-ɦæ:d l-fi::ləm], where the decelerated part is underlined: in it, the word [film] undergoes an ultra-lengthening of the already long [ī] and the insertion of a [ə] where it is usually not pronounced5)

14A schematic summary of the use of these four variables (as it emerged from this first analysis) is reported in Table 1. The media products analysed so far (with more to be included) are three foreign soap operas dubbed in Moroccan Arabic (Diablo, Mar de Amor and Samhini), two Moroccan sit-coms (L’Coupl and Loberj) and a Youtube animated series (Bouzebbal); two hours have been analysed for each series, except for Diablo (for which only one episode was available on Youtube) and Loberj (for which analysis is not complete). In the column “Occurrences in field study”, a confrontation with Casablancan speech is reported when relevant (a group of Casablancan informants was included in the sample for inter-urban comparison purposes).

15A few, partial observations can be made at this stage; further analysis of the data will allow a clearer view of the phenomena at hand and, consequently, confirm or relativise the following remarks:
- /t/ allophones: both in the field and in the media corpus, the general trend is for palatalised allophones ([t^š] and [tš]) to be produced by young men and fricatised allophones ([tˢ] and [ts]) by elder people6. However, as far as the palatalised ones are concerned, their frequency and degree of palatalisation are much lower in the media than in the field corpus, where almost all young male informants constantly palatalise in linguistic environments allowing this, with many of them almost going as far as full fricativisation (thus pronouncing /t/ very close to [š]). This suggests that the media’s policy may be that of dealing with palatalisation of /t/ in a somewhat “moderating” way, which could possibly be due to a negative connotation of this phenomenon. It is therefore interesting to observe its widespread use in the community studied, as well as among Moroccan youth in general (cf. Caubet et al. forthcoming)

16- /q/ vs /g/: the lexemes listed in the table are those for which phonemic variation has been observed both in the media and on the field. However, sociolinguistic distribution in the media is much more “straightforward”, with rural characters invariably producing all these lexemes with /g/ and all other characters (with a few exceptions such as suburb residents in Bouzebbal) producing them with /q/. On the other side, field informants show a much greater instability in the selection of /g/ vs /q/, with some invariably choosing either of them and some alternating between the two of them even in the same sentence (obviously, and in addition to that, each lexical item tends to be treated differently by each speaker). However, /q/ is in general much more frequent than /g/. This could signal that, for some reason to be enquired, recently urbanised (also young) Témarois are “late” in abandoning rural features (such as /g/), but it could also indicate that young speakers are reinterpreting either the meaning of traits once considered as merely urban or rural, or the concepts of urbanity and rurality themselves.

17- /hdəṛ/ vs /dwa/: the inter-corpus comparison seems to suggest that /dwa/ is a word mainly used in Casablanca, whereas /hdəṛ/ may be more diatopically “neutral”. In fact, /hdəṛ/ is the most frequent (but not exclusive) choice in both corpora, with three remarkable exceptions: a group of Casablancan youth included in the field sample, Bouzebbal (which is set in Casablanca) and Samhini, a Turkish soap-opera dubbed in Moroccan Arabic and still airing on Moroccan TV to this date; in these three domains, /dwa/ is almost categorically used instead of /hdəṛ/. Indeed, in all three of these contexts, one would expect to hear people speak “like Casablancans”: the youth are from Casablanca, as well as the Bouzebbal characters. As for Samhini, why should members of high class Turkish families speak like Casablancans? In fact, Moroccan Arabic speakers have already been observed to perceive translations of foreign soap-opera scripts to Moroccan Arabic as closer to the Casablancan dialect: this, at least, has been the case for the Mexican soap-opera Las dos caras de Ana (Miller 2012: 172; Barontini & Ziamari 2013: 121). Coincidentally (or not?), the translation and dubbing of Samhini has been taken charge of by Plug-In, the same company that managed the translation and dubbing of Las dos caras de Ana. This may indicate that /dwa/ is truly closer to Casablancan speech. If this is the case, then the analysis can be directed at evaluating the occurrence of either lexeme when we would not expect it, i.e. of /hdəṛ/ in supposedly Casablancan speech and of /dwa/ in supposedly non-Casablancan speech.

18- Sentence-final deceleration: this is a supra-segmental feature involving both a decrease in articulation rate before a pause and a lengthening of the vowels involved (usually those in the last word preceding the pause). It seems to serve the purpose of sounding aggressive, “dangerous”, like “gangsters” or “criminals” (but not necessarily with the intent of making other people believe that one is a criminal). To use a lexicon more familiar to Moroccans, this feature allows the speaker to sound like a mcharmel (/mšəṛməl/, approximately “petty criminal”) in order to give the impression of being mqawed7 (/mqəwwəd/, approximately “badass”8). While the characters that use this feature in the media corpus (petty criminals in Bouzebbal and two convicts in the Mar de amor dubbed soap-opera) seem to confirm its original association to people of low morality, the tropic use (Agha 2007) of it seems to be only found among a part of the youth in the field corpus. Further analysis will shed light on whether young speakers “speak like mcharmlin” in order to “get a mcharmel attitude” or are just used to speak like that, possibly by effect of their social origin (e.g. having grown up in an environment “full of mcharmlin”)9.

Conclusion

19This brief article aimed at giving a general presentation of an ongoing research project on sociolinguistic variation in Moroccan Arabic as it is spoken in the city of Temara, and mainly focused on the theoretical motivations underlying the choice of the object of study and the method of analysis adopted, with the addition of a few observations drawn from a first, partial analysis of the data. The completion of the compared analysis of the field and the media corpora, which constitutes the next step in the treatment of the data, will be guided by the following questions:
- How do the media and the informants of the sample differ in interpreting sociolinguistic variation in spoken Moroccan Arabic?
- How do the speakers give a sense to the use of alternative linguistic features, in the frame of their language practices and the urban context examined?
- Do media-vs-society differences in sociolinguistic choices exclude, or affect, the possibility that a shared, pan-national spoken variety of Moroccan Arabic may be taking shape?

20Answering these questions will hopefully shed light on how Moroccan Arabic speakers not only make use of their linguistic repertoire in order to express themselves and be understood, but also on how they interpret the linguistic variation they are faced with in their social environment, in the frame of the development of national contexts of communication and the ongoing elaboration of the Moroccan national identity.