Linguistic Variation, Social Meaning and Covert Prestige in a Northern Moroccan Arabic Variety

Abstract

This paper addresses how gender and age, as macro-sociological factors, influence variation and change in the Northern Moroccan Arabic variety of Ouezzane, and how social meaning plays a role in this variation. To do so, it examines the high degree of variability in the realization of two phonetic variables, the voiceless alveolar plosive /t/ and the voiceless uvular plosive /q/, in a corpus of semi-scripted interviews with 20 local informants. The data for the study was gathered during several fieldwork campaigns carried out between 2014 and 2021. The analysis combines quantitative and qualitative methods. Quantitative comparisons are drawn across gender and three age categories (under 30, between 30 and 50, and over 50) to search for gender and/or age markers, while the data are qualitatively analyzed with regard to the increase in the use of certain allophones, attrition and loss of other variants, and metalinguistic comments made by informants on those traits. These two methods make it possible to identify how the phonetic variables analyzed contribute to the construction of various identities, such as an “older person” identity, as well as self-affiliation with particular social groups, such as “artisans” or “rural women”, from which other groups, such as male university graduates, are keen to distance themselves.

1. Introduction

Ever since the publication of the foundational studies on linguistic variation and change (Labov 1963, 1972), it has been generally accepted that macro-sociological categories such as gender, age, social class or level of education play important roles in triggering language change and then shaping its direction. Later studies demonstrated that stylistic variation and a speaker’s awareness of a particular trait’s social value could be equally critical—or even more so—to the spread of such changes than macro-sociological categories (Trudgill 1986). More recent research on Arabic-speaking communities, in particular, has highlighted that other extralinguistic factors, such as context, urbanization, style or the social meaning of certain features can also play a decisive role in language variation and change.

This paper considers the impact on language variation of two social factors, gender and age, together with social meaning in the vernacular variety of Arabic spoken in Ouezzane, a middle-sized, recently urbanized city in northern Morocco. Linguistic variation is therefore explored here in two different ways. On the one hand, I will apply a quantitative analysis to look for measurable parameters—if any are present—where gender and age are significant and might be driving the variation. On the other, I will apply a qualitative analysis to examine the social meaning that some linguistic forms index. My two-pronged approach centers around two phonetic features, /t/ and /q/, of the local Arabic variety that exhibit a high degree of variability in this region, ranging from [t], [t^s], [t^š], [θ], [s] or [d] in the case of /t/, to [q], [g], [Ɂ] or [x] to the case of /q/. The corpus analyzed, which consists of 20 semi-scripted recorded interviews with local informants, was gathered during fieldwork conducted between 2014 and 2021. This research aims to examine the extent of variation in this community, why it exists and where it is going.

1.1. Theoretical Framework

The focal concepts in this paper are linguistic variation—meaning the geographical, social and/or contextual features that particularize a language or variety—and sociolinguistic representations—understood as the social meaning that speakers ascribe to a certain feature—since both can reshape vernaculars and local or regional identities.

This paper seeks to find links between macro-sociological categories and “the more concrete local categories and configurations that give them meaning on the ground” (Eckert 2012, p. 93). This means that, in keeping with the second wave of language variation studies, I examine gender and age categories from the perspective of native speakers, who assign social meaning to the features analyzed. At the same time, however, I also adhere to the third wave of sociolinguistics research, which locates “meaning” in stylistic practice rather than in the spoken output of individuals and considers style to be an intrinsic element of language that helps identity construction, group belonging and social differentiation (Eckert 2012). Within the third wave of sociolinguistics, variation becomes “a social semiotic system” (Eckert 2012, p. 94) that can articulate the social issues that exist within a linguistic community. Because these social issues are mutable, the meaning of the variables analyzed is constantly changing, making the notion of social meaning a dynamic concept1 as well.

Thus, the indexicality (Silverstein 2003) of a feature can vary according to not only an individual’s group membership but also the persona or the style to which the individual aspires or seeks to convey for a particular interlocutor. Likewise, it provides a way of identifying the social meaning attached to different linguistic forms, thus allowing us to postulate the reasons underlying linguistic variation and change when macro-sociological categories are not able to explain it. This theoretical perspective also allows us to analyze how speakers “tailor” their own “style” and construct the multiple identities they may choose to embody in various diverse contexts, depending on interlocutor, topic or context, and during different periods of their lives.

Given that I will be dealing with the macro-sociological categories of age and gender and indexicality, the concept of covert prestige may emerge in the social meaning of specific phonetic variables. By covert prestige, following Trudgill (1972), I am referring to the existence of certain values, not expressed overtly by speakers, which are especially relevant in explaining the preservation of particular non-standard features. Trudgill’s study centers around gender, but it is a concept that can also be considered in relation to age and social meaning.

I will focus, then, on the dynamic social meaning of two particular phonetic features, which can index the speaker’s adherence to or dissociation from one or more social groups.

1.2. Literature Review

Linguistic variation and change is not a new topic in research on Arabic. Soon after the publication of Labov’s groundbreaking work on linguistic variation and social factors in the 1960s, sociolinguists and dialectologists began to apply this methodology in their work on various Arabic-speaking communities. Many of these early studies in Arabic sociolinguistics were interested in exploring the role that the standard variety, Modern Standard Arabic (henceforth MSA), has played in variation and change (Abd-el-Jawad 1986; Ibrahim 1986; and others). However, these studies revealed that MSA was not the only prestigious Arabic variety involved given the emergence of prestigious vernaculars, usually linked to varieties of Arabic spoken in major urban centers. Whether old medina dialects or new urban koines, these vernaculars were endowed with higher social status than rural or mountain varieties (see Miller 2007a, 2007b; Miller and Falchetta 2021). In fact, as Al-Wer (2002) demonstrated, the more speakers master MSA, the more their linguistic innovations tend to go in the opposite direction, towards the prestigious vernaculars.

Social prestige ascribed to the features of a language style is an important facet of social meaning, and as we will see in this study, it can also be closely linked to gender and age factors.

1.2.1. Gender

Regarding the influence of gender on linguistic variation, various noteworthy contributions have pushed forward our understanding of this area in relation to Arabic sociolinguistics. For instance, Al-Wer (2009) states that in general Arabic-speaking women prefer the use of supralocal features, while men use the more localized traits that characterize a particular dialect. Albeit not centered around social class, her findings are consistent with Trudgill’s (1972) conclusions about speech styles in Norwich, England, namely that women produce prestigious features more often than do men. This claim has been shown to hold true for Amman Arabic (Al-Wer 1999), Cairo Arabic (Bassiouney 2011), and varieties of Arabic spoken in Korba, Tunisia (Walters 1989), and Tlemcen, Algeria (Dendane 2013). However, the greater use of prestige varieties among women has proven less clear-cut in studies carried out at Balyana in Upper Egypt (Miller 2003) and the mountains of northern Morocco (Vicente 2002), where sexual segregation, lower mobility and reduced access to education have meant that the Arabic varieties spoken by women can be characterized as conservative or even archaic. These differences confirm what Vicente (2009), in an exhaustive state-of-the-art overview, has claimed about the complexity of analyzing gender variables across a geographic area that is as broad as the Arabic-speaking world, where women’s circumstances and linguistic situations are by no means homogenous. Vicente shows that women can act as leaders of linguistic innovation in urban contexts, as noted above, but in rural, isolated and segregated contexts they can also be conservative linguistically; thus, their agency in innovation or conservation is related to the context.

1.2.2. Age

Concerning the age factor, previous research has pointed in two different directions. On the one hand, as Al-Wer (2009, p. 630) pointed out, differences in the linguistic behavior of different age groups represent different stages of evolution in the language itself and thus serve as an indicator of linguistic change. This set of studies shows that young people tend to be more innovative at all linguistic levels than older generations. On the other hand, Miller and Caubet (2010) have pointed to the emergence of “young person speech” as a sociolinguistic category in itself, which is extensively analyzed in the compilation by Trimaille et al. (2020). Miller’s (2022) summary is particularly valuable, but more localized studies must be mentioned, too. For instance, Ziamari et al. (2020) analyzed Arabic youth style in different areas of Morocco; Pereira’s (2010) work on Libya claims that young people make more frequent use of words related to sexuality and injuries, for example, than do older adults; Abdulaziz’s (2015) research examines street art as a multilingual linguistic landscape in Tripoli, Libya; and Guerrero (2019) and De Blasio (2022) explore the use of language in rap music in Morocco and Egypt, respectively. Therefore, besides being more innovative, youngsters tend to be more creative.

2. Ouezzane

As I have mentioned above, my research has largely been conducted in Ouezzane, a medium-sized city located in northern Morocco at the southwestern foot of the Jbala Mountains. This area is considered to have been Arabized in the earliest wave of Arabization and people there speak a sedentary-type Arabic vernacular.2 As I have discussed elsewhere (Benítez Fernández 2016, 2022), the location of this city between the mountains and Atlantic coastal plain means it straddles two different linguistic varieties, the sedentary, or pre-Hilali, varieties found in the old medinas and mountains of northern Morocco, and the Bedouin, or Hilali, varieties spoken some kilometers to the south.3 These varieties differ at every linguistic level—for example, in the realization of /q/, which will be relevant in this study—and in terms of the social values attributed to different linguistic features, such as affrication, which will also be a key point in this study.

In the 17th century, Ouezzane was a rural hub noted for its Sufi tariqa and zawiya,4 both founded by Moulay Abdallah Cherif. Although there are no linguistic studies focusing on Ouezzane speech from that period, it may be assumed that its linguistic features are consistent with other Jebli varieties—that is, those varieties characteristic of the Jbala Mountains.5 The importance of the annual zyara,6 revenues sent by other related zawiyas to the main one at Ouezzane, prominence that Ouezzane acquired during the French Protectorate, and large-scale exodus from the countryside to cities that Morocco has experienced since the 1970s have all been factors in the development of Ouezzane into an urban area, a home currently to some 61,583 inhabitants.7 This process of urbanization has had an impact on the speech of the local population, whose vernacular has become an urbanized variety. Furthermore, Ouezzane’s geographical situation within the transition zone between the mountains and the coastal plain has facilitated contact with the adjacent Bedouin varieties.8 This background needs to be borne in mind, since it is one of the main factors triggering the variation in some of the variants examined, as I will show in Section 4.

Linguistically, Ouezzane has attracted the attention of researchers since the late 20th century. The first study focusing on this variety was a master’s thesis by Jaouhari (1986) in which the author used Ouezzane speech as a case study to analyze the Moroccan verbal system. The first description of the dialect of the city and its hinterland took the form of an undergraduate thesis by Khoukh (1993). This study consists of a collection of transcribed texts, gathered during fieldwork in 1992, with the most salient linguistic features pointed out in the footnotes. More recently, Ech-Charfi (2016) analyzed the standardization process that Moroccan Arabic is undergoing and explored how traditional northern varieties, such as the dialects of Ouezzane and Tangier, are adapting to—or resisting—the emerging supra-regional variety, which he labeled “Synecdochic Moroccan Arabic”. El Khomssi (2017) offers a preliminary descriptive study of the speech of a market vendor who lives in the countryside but sells her produce weekly in the city of Ouezzane. In 2020, El Khomssi defended her doctoral thesis (El Khomssi 2020), which dealt with linguistic particularities and social representations in the speech of two tribes from Ouezzane province, the Rhouna and Beni Messara [sic]. Her informants were living either in the tribes’ original rural locations or the large urban centers outside Ouezzane province to which they had migrated. Finally, in an ongoing series of papers, Benítez Fernández (2016, 2019, 2022) analyzes variation related to age, gender and rural versus urban speech styles in Ouezzani vernacular.

In this article, I focus on linguistic variation induced by social factors and linked to social meaning to analyze the relative impact of each of these variables.

3. Materials and Methods

My analysis is based on a corpus consisting of recorded semi-scripted interviews with 20 informants collected during my fieldwork in Ouezzane between 2014 and 2021. The fieldwork was conducted mainly in the old city center, or medina, but also in the populous Quššāriyyin neighborhood northeast of the medina. The informant sample was made up of nine men and eleven women. Of the males, four were under 30, four were between 30 and 50, and one was over 50. Five of them had received only a primary education (two of the young adults, two of the middle-aged adults, and the elderly man), one had a secondary education and three were university graduates. As for the women, two of them were under 30, five were between 30 and 50, and four were over 50. Two of the women had a university education, two had a secondary education and the remaining seven—all the older women, two of the middle-aged women, and just one young woman—had only a primary-level education. All of the informants had either been born or spent much of their lives in the city of Ouezzane, but those not born in the city had roots in the immediate mountain hinterland.9 This is important to know since the Jebli varieties of Moroccan Arabic are heavily stigmatized because they index rurality or backwardness. On the other hand, these informants had experienced a certain degree of mobility. Since Ouezzane has no university, this mobility was most evident among the more educated informants. Several of the men had spent some years in Kenitra pursuing their education while some of the women had done the same in Tetouan or Larache. However, in the case of other informants, it was the greater work opportunities that had brought them to live elsewhere. In particular, one woman and one man, each with just a primary education, had lived in Tangier and Casablanca, respectively, the former to work in a factory and the latter to be trained as a craftsman. Finally, two other informants, one female, the other male, were shopkeepers who traveled regularly to Tangier, Rabat or Fez to buy supplies for their businesses.

Table 1. Participants according to age and gender.

Age	Women	Men	Total
18–30	2	4	6
31–50	5	4	9
>50	4	1	5
Total	11	9	20

breaks the informants down by gender and age. The group of elderly informants is not balanced by gender, there being four elderly women but only one elderly man. Therefore, it is difficult to draw gender-based conclusions from this cohort.

4. Description of Variables

In this section, I will examine the behavior of two phonetic variables chosen because they are the ones that show the highest degree of variability in the corpus. I organize the information as follows. First, I offer a brief overview of the relevant literature that has previously described these phonetic forms (Section 4.1 and Section 4.2). Second, adopting a variationist perspective, I use quantitative analysis to explore the extent to which gender and/or age are key social factors in the variation seen (Section 5). Finally, shifting to a qualitative approach, I examine the sociolinguistic representations that the data reveal (Section 6).

The variables I will focus on are the dental plosive */t/ and the voiceless uvular */q/. In my corpus, I have inventoried six allophones of the dental plosive /t/: the voiceless alveolar plosive [t], voiceless alveolar affricate with sibilant appendix [t^s], voiceless alveolar affricate with palato-alveolar appendix [t^š], voiceless interdental [θ], voiced dental plosive [d] and voiceless alveolar sibilant [s]. Regarding the voiceless uvular plosive */q/, five allophones can be found in different Moroccan vernaculars, all of them—with the exception of the voiceless velar plosive—present in the corpus. These allophones are the voiceless uvular plosive [q], which is the most common realization, voiced velar plosive [g], voiceless velar plosive [k], voiceless velar fricative [x] and glottal stop [Ɂ].

4.1. Dental */t/ and Its Variants

In his Textes arabes de Tanger, William Marçais (1911, p. XIV) already describes two variants of the dental occlusive: the voiceless alveolar affricate and the voiceless alveolar plosive. He notes that the latter appears in loanwords, in the first element of a gemination [tt^s]; in place of the voiceless dental velarized /ṭ/ whenever the context includes /i/ or non-emphatic /r/; and in place of [t^s], whenever this allophone is in contact with sibilants (/s/, /š/, /z/ and /ẓ/), liquids (/l/ and /r/) or the nasal /n/. Lévy-Provençal (1922, p. 19) and Vicente (2000, p. 40) identify these contexts in Ouargha and Anjra varieties, respectively. In connection with the dental occlusive /t/ form, Cantineau (1960, p. 37) describes occurrences of [t^s] as “unconditioned curious alterations in Algerian and Moroccan sedentary vernaculars”.10 Comprehensive studies of Northern Morocco varieties, like those carried out by Colin (1918) or Lévy-Provençal (1922) for Southern Jebli or more recent ones focusing on Northern Jebli (Vicente 2000; Moscoso 2003) reach similar conclusions, noting an affricated realization of the phoneme /t/, one with a sibilant element (>[t^s]) and also an interdental one (>[θ]).11

Other authors exploring Moroccan vernaculars claim that the most common pronunciation of this phoneme is the voiceless alveolar affricate [t^s] allophone. For instance, Heath (2002, p. 135) states that, “unpharyngealized t typically has a noisy assibilated release pre-pausally or before a V[owel] or noncoronal C[onsonant] in M[oroccan]A[rabic] dialects […] Specialists who pay special attention to phonetic transcription often comment that the assibilation feature is unsystematic (especially for southern dialects) or is largely predictable from phonetic context.” Stillman (1988), as cited in Heath (2002, p. 135) adds that this realization characterizes all the old urban dialects of Morocco, and Aguadé (2003, p. 66) points out that the degree of affrication found in mountain varieties in northern Morocco can vary greatly depending on the speaker. Most of these authors connect this type of affrication to the Berber substrata and adstrata.

It is thus not surprising to find such variability in vernaculars of the Jbala region and the cities of northern Morocco, and Ouezzane is no exception. In fact, Caubet’s analysis of different corpora from northern Morocco collected in the 1990s shows a frequent, but not exclusive, use of [t^s] in both men and women and all generations (Caubet 2017). One of these corpora is that collected by Khoukh in 1992 (Khoukh 1993). In his analysis, he focuses on Ouezzane and its hinterland variety and shows a far higher use of [t^s] over [t]. However, what we can see in my data is a more diverse variability. Along with the voiceless alveolar affricate [t^s] and the voiceless alveolar plosive [t], the data show some occurrences of other realizations such as the voiceless alveolar affricate with palato-alveolar appendix [t^š], the voiceless interdental [θ] and even the voiceless alveolar fricative [s], by an assibilation process, and the voiced dental plosive [d] when /t/ is in contact with the voiced velar [g]. These last two variants will not receive any further attention in my analysis, since I have included the former among the [t^s] items and the latter is a phonetically-induced change.

4.2. Voiceless Uvular */q/ and Its Variants

Traditionally, dialectologists working on North African varieties, and on Moroccan varieties, in particular (Cantineau 1960, p. 68; Marçais 1908, pp. 12–13), have regarded dialects in which old *q is represented by a voiceless consonant—for instance, the uvular realization [q] or the glottal stop [Ɂ]—as a feature that characterizes “sedentary varieties”, meaning those spoken in the old medinas (Vicente 2021) and the Jbala region (see, for instance, articles in Vicente et al. 2017). Meanwhile, the voiced realization, [g] or [k], is associated with “Bedouin varieties”.12 However, this diachronic distribution is now much more complex, since processes like urbanization, the rural exodus, increasing levels of literacy and education, and a more frequent use of vernaculars in the mass media has meant that the [q] realization can now be found everywhere in Morocco, not only in sedentary varieties. In fact, it could be considered the most widespread realization in the country at the moment. Therefore, this realization can also be found in recently urbanized areas such as Casablanca, as Moumine (1995) and Hachimi (2007) note, as well as in the Saharan and central type vernaculars (Heath 2002, pp. 6–10). As aforementioned, the other feature linked with sedentary varieties is the glottal stop [Ɂ], which is clearly undergoing a regressive process and mostly related to rural varieties.13 In this regard, Colin (1918, p. 40) reported that /q/ was realized as [Ɂ] among illiterate children and most of the chorfa groups in the northern area of the Taza region. More recently, Caubet (1993, p. 15) documented this realization among certain Muslim shopkeepers in Fez. For his part, Heath (2002, p. 139) claimed that it “is not broadly characteristic of any ‘dialect’”. Hachimi (2011)—writing about the Fessi variety in Casablanca—and Vicente (2021)—analyzing various linguistic situations in northern Morocco—agree with Heath’s assertion, but they, as well as Benítez Fernández (2019), argue that the [Ɂ] realization is a sociolectal variant. This is why Heath (2002, pp. 141–47) did not include it among the features characterizing the Ouezzane Muslim or Jewish dialects, while Khoukh (1993, p. 12) and El Khomssi (2017, p. 160) note the realization of [Ɂ] exclusively among women. However, perhaps this should be qualified as referring more specifically to “rural women” since the informants for those two authors were an illiterate elderly woman coming from the mountains and a female vendor at the weekly market who did not live in the city, even though she owned a house in the Quššāriyyin neighborhood.

As for the voiced velar realization [g], it has been traditionally linked with Bedouin varieties, or the central and Saharan type in Heath’s (2002) terms. The domain where it is common includes recently urbanized hubs such as Kenitra, Casablanca, El Jadida or Agadir in the Atlantic coastal plain, and other old cities that received a heavy influx of population from rural areas during the last part of the 20th century, such as Rabat or Fez, urbanization processes that have affected its current distribution. Thus, it can be found in lexical borrowings that sedentary varieties have taken from the Bedouin vernaculars (Hachimi 2007; Vicente 2021), maintaining the voiced velar [g] pronunciation, such as the word ḥārrāga ‘illegal emigrant’.

Finally, the two other realizations, the voiceless velar plosive [k] and the voiceless velar fricative [x], are either a sociolect feature or a conditioned variant. The former used to be common in Jewish dialects (Tafilalt, Dra Valley, Atlantic-strip, Debdou) of Moroccan Arabic (Heath 2002, p. 142) but most members of this community left Morocco some decades ago. The latter normally occurs in contact with the dental /t/—especially in expressions that incorporate the word waqt ‘time’, as pointed out by Colin (1918, p. 43)—and thus constitutes an intralinguistic change. It could be considered a kind of spirantization of /q/ that is present in various rural northern dialects, such as those spoken in the Jbala Mountains (Vicente 2022b).

This variation I have just described can be seen in the following examples from the corpus. As observed, the corpus contained no instances of the [k] variant.

• qăllǝt n-nĭyya ‘lack of goodwill’, naqqi-ha ‘clean it!’, ġa-ylqā-ni muwwžda ‘he will find me there’, ma kāyna š farq ‘there is no difference’.
• kīma dgūl māma ‘like my mother says’, ka-ngūl l-a nūdi ‘I tell her: stand up!’, k-ygūl l-um ‘he says to (or tells them) them’.
• ʕa nɁūl l-әk ‘I will tell’, bāɁi ‘still’, zuwwāɁ ‘embroiderer’.
• fūyaḫ (>f-ayyi waqt) ‘when’, ka-yddi mʕa-ha l-wăḫt ‘he spends time with her’.

In sum, the literature attributes the high variability shown by these two phonemes to either the context (/t/ in contact with sibilants and approximants), intra-linguistic changes ([d] and [x]), historical reasons linked to the Arabization process (the voiceless or voiced realization of */q/) or, to a lesser extent, sociolinguistic factors (prestige in the case of [q], stigma in the case of [Ɂ]). The reason why I come back to the analysis of these two features is the meager bibliography that considered macro-sociological factors or social meaning (Hachimi 2011; Falchetta 2019; Vicente 2021; Benítez Fernández 2019) as the catalyst for this variation. In this study, I assume that such a high variability is due to the fact that some allophones are either gender or age markers, or index social meanings that differentiate social groups.

5. Quantitative Analysis

5.1. Sociolinguistic Distribution of */t/

I identified in the corpus 2561 tokens of the voiceless dental phoneme */t/, including all the allophones ([t], [t^s], [t^š], [θ]). Of these instances, 1155 (45%) correspond to the voiceless alveolar plosive and 1328 (52%) to the alveolar affricate. This can be seen in

Table 2. Dental variants distribution.

	[t]	[ts]	[tš]	[θ]
N	1155	1328	73	5
%	45	52	2.8	0.2

below.

Consistent with what has been already shown in previous research, my data reveal a preference for the alveolar affricate allophone [t^s], which is slight compared with the plosive allophone but quite strong relative to the other allophones. Nevertheless, the percentages of these two allophones do not represent, in overall terms, a difference that is significant.

In what follows, I will limit my focus to these two main allophones ([t] and [t^s]), so the total number of items analyzed is 2483. Though I will come back to the other dental allophones later, in a qualitative analysis (Section 6), altogether they represent a mere 3%.

Thus, the informants overwhelmingly use both the affricate [t^s] and the plosive [t]. As an example, one individual utters the exact same word interchangeably using the affricate and the plosive. Broadly speaking, the voiceless alveolar affricate and the voiceless alveolar plosive allophones seem to alternate randomly, which matches the “unsystematic” phenomenon already described (Heath 2002, p. 135).

This can be seen in the following examples: mātālan vs. māt^sālan ‘for example’; ntūma vs. nt^sūma ‘you (pl.)’; ida bġīt^si tanžaḥ ‘if you want success’; nta žāy t^sākul, t^sbāt ‘you come eat, spend the night’; ka-tḥass ‘you feel’; tta dīk əs- sāʕa ‘until that time’; tḥarras ‘it (masc.) is broken’; ġnāt-ək ‘she got rich’; tkūn tqīla ‘it (fem.) is heavy’; ḫūt^s-a ‘her sister’; t^sa-ybīʕ ‘he sells’; mt^sāl ‘sayings’.

Bearing in mind that gender or age can affect the choice of these allophones, I have carried out an analysis of the data according to these two factors, with the idea of identifying any social patterns.

Concerning the gender variable, as can be seen in

Table 3. Gender distribution of dental variants.

	[t]	%	[ts]	%	Total
Women	694	45	835	55	1529
Men	461	48	493	52	954
Total	1155		1328		2483

below, my data analysis shows that the sample is unbalanced, since the data produced by the subgroup of women contain many more tokens of both variants (1529 occurrences) than the data produced by men, which include just 954 items. Even so, the use of the two allophones in both subgroups is quite evenly distributed. Both subgroups show a slight preference for the alveolar affricate [t^s], a tendency that is somewhat stronger among the women.

To better analyze the age variable, I divided the sample into three age ranges: the first covers from 18 to 30 years old, the second from 31 to 50 and the third comprises people over 50. The distribution of allophone use across these three age groups is shown in

Table 4. Age distribution of dental variants.

	[t]	%	[ts]	%
18–30	197	37	335	63
31–50	392	40	598	60
>50	566	59	395	41
	1155		1328

.

As can be seen, up to age 50, there is a clear predilection for the alveolar affricate [t^s], while the preference of the over-50s is for the dental plosive allophone [t]. Even if the rate of dental plosive use represents only 59%, this finding suggests a dynamic trend or at least some kind of evolution.

If we did not know the area, we might be tempted to postulate that the affricate allophone embodies an innovation introduced at some unknown recent moment. However, thanks to the body of previous research concerning the Jbala region, particularly Khoukh’s (1993) study on Ouezzane and its hinterland, we know that this region favors affrication. In fact, Khoukh’s data show overwhelming use (87%) of the alveolar affricate [t^s], and informants who were middle-aged in 1993 logically would constitute the over-50s in my sample. So how can we explain this change in preferences?

5.2. Sociolinguistic Distribution of */q/

I have identified 1201 tokens of */q/, whose variability is shown in

Table 5. Voiceless uvular and its variants distribution.

	[q]	[g]	[Ɂ]	[x]
N	1077	18	96	10
%	89.7%	1.5%	8%	0.8%

.

It is immediately apparent that the voiceless uvular [q] is the most widespread realization among informants in my sample, which is consistent with the literature. This data show that the other variants are clearly recessive; but again, these results are also to be expected given previous research. Since the voiced velar is largely restricted to borrowings taken from other varieties, its low rate of occurrence is not surprising. Building on what has been already reported about the glottal stop in Moroccan varieties, one could hypothesize that it might characterize the idiolect of a particular social group, such as “females”, “older females” or “speakers of rural origin”, etc. An analysis of the gender and age distribution might offer insights in this regard, even though the prevalence in the sample of the other variants is quite low.

To this end, in

Table 6. Gender distribution of voiceless uvular and its variants.

	[q]	%	[g]	%	[Ɂ]	%	[x]	%	Total
women	551	93.2	0	0	38	6.4	2	0.3	591
men	526	86.2	18	2.9	58	9.5	8	1.3	610
Total	1077		18		96		10		1201

the data have been broken down by gender and variants. The table makes it clear that the uvular [q] is the most popular realization among both women and men in Ouezzane. This should come as no surprise given that, as I have already noted, this is also the case all over Morocco. However, my personal experience suggests that the use of the voiceless uvular [q] is more consistent among women than men, who use a wider number of variants, albeit with the voiceless uvular predominating. It is noteworthy that the voiced velar [g] is not used by women at all, but more so the fact that they use the glottal stop variant [Ɂ] only rarely, despite its being identified in the literature about Ouezzane as a characteristic feature of women’s speech. Their negligible use of the voiceless velar fricative [x], usually regarded as a sign of intralinguistic change, is also striking.

I have also analyzed the data according to the age variable, again divided into three different categories: young people (under 30), middle-aged adults (between 31 and 50) and older adults (over 50), as can be seen in

Table 7. Age distribution voiceless uvular and its variants.

	[q]	%	[g]	%	[Ɂ]	%	[x]	%	Total
18–30	219	88.7	17	6.9	4	1.6	7	2.8	247
31–50	240	73.6	1	0.3	84	25.8	1	0.3	326
>50	618	98.4	0	0	8	1.3	2	0.3	628
	1077		18		96		10

.

Clearly, the voiceless uvular allophone [q] is the most representative feature in every group, but its use is especially overwhelming among the older speakers, despite what might be expected, since the glottal stop [Ɂ] and the voiceless velar fricative [x] are conservative traits14 in other varieties. Though the glottal stop is quite regressive, it is also present among every age group. Its lower prevalence among young people is expected since most of them are highly literate, lending support to the idea that the glottal stop [Ɂ] remains an old-fashioned feature.15 Taking into account Khoukhs’s claim (1993) that this feature is mostly realized by elderly women who are illiterate, the presence of [Ɂ] among middle-aged adults that we see here is bewildering. According to Table 5, the glottal stop is more available among men than women.

The data also make it clear that voiced velar [g] identifies young people, since here it is among that subset of informants that it mostly occurs. Thus, voiced velar [g] can be considered an innovative introduction coming from young men, and this is why previous studies did not include it in the phonetic repertoire of Ouezzane.

Contrary to what was noted about the dentals above, gender and age are significant variables when it comes to the voiceless uvular and its variants. For instance, the glottal stop is mostly a feature found in the speech of elderly men, while the voiced velar is mostly used by young men. However, why are these variants more prevalent among men? What are some of the decisive factors in those choices? These questions will be addressed next in the qualitative analysis.

6. Qualitative Analysis: Social Meaning

In what follows, I will relate the results of my quantitative analysis of the two features analyzed to previous research and also data obtained in sociolinguistic interviews, as well as my own participant observations.

6.1. Social Meanings Associated with */t/

As mentioned above, though previous studies reported a fairly generalized use of the alveolar affricate [t^s], the quantitative analysis carried out in this paper shows some mismatch with these references. In my sample, only women choose—albeit rarely—the alveolar affricate. On the other hand, the group as a whole shows a balanced choice between the two allophones ([t] and [t^s]), while the elderly group prefers the dental plosive.

Concerning this balanced distribution of the two allophones, I should again point to Ouezzane’s geographical location between two different vernacular varieties, the Pre-Hilali and Hilali. In the Pre-Hilali or sedentary varieties of Moroccan Arabic, the usual way to pronounce the dental phoneme is as an alveolar affricate, as attested in the previous literature, but in other varieties, this realization is a stigmatized trait. In fact, in a recent study of the slang and style of young Moroccans Ziamari et al. (2020) claim that “in big cities like Casablanca or Meknés, affrication connotes popular, even vulgar speech, and is considered a rather masculine trait.”16 Thus, the social representation of this trait could explain the balanced distribution of the two allophones across my data.

Besides geographical location, or rather in addition to it, urbanization must be mentioned as a factor. In the past five decades, or perhaps even longer, the city of Ouezzane has undergone considerable growth that has left clear footprints in the speech of the community (Benítez Fernández 2022). As I have noted, this expansion of the city has been due to immigration from the surrounding rural areas. In fact, most of my informants either had a Jebli heritage or were themselves of Jebli origin, so they were far from unfamiliar with affrication. Since it is a marked trait that can betray a speaker’s origin, the urban growth process has tended to favor a decline in its use, a certain attrition of affrication. The same can be said of the voiceless interdental fricative [θ], since it is also a salient trait that is identified with the rural Northern Jebli vernaculars spoken in the mountainous areas adjacent to northern Moroccan cities like Tangier, Tetouan, Larache, Chefchaouen and Ouezzane (Cantineau 1960, p. 37). In a study that compares the vernacular variety of Moqrisset—a village in the mountains northeast of Ouezzane—with the vernacular variety of a group of informants living in the city but having Moqrisset origins, Benítez Fernández (2022) reported that the voiceless interdental fricative [θ] allophone appeared, both in intervocalic and final position, among the informants living in Moqrisset but had almost completely disappeared among the informants who had migrated to the city (like the informants in this paper). The Jbala area is an impoverished region with a high rate of illiteracy, so those informants displaying the voiceless interdental fricative [θ] in their repertoire may tend to suppress it to avoid the implicit stigma associated with Jebli traits. So, again, the intersection between Northern Jebli linguistic identity and “urbanization” could also be responsible for the lower prevalence of this feature among urban informants. Affrication and interdental may well reflect a conscious choice on the part of speakers to dissociate themselves from their origins, a kind of “de-Jeblization”.

As for gender, even if the difference in percentages seems too small to justify interpretation in terms of gender-specific preferences, I am tempted to explain the fact that the affricate allophone occurs slightly more among women (55%) than men (52%) in my sample in two different and perhaps contradictory ways. One is to play the card of the supposed linguistic “conservatism” of women’s speech, which is very often pointed out, especially regarding the varieties spoken by rural women. This might explain the behavior of some of my female informants, but not all of them. The alternative explanation, I suggest, should be sought among the universals of sociolinguistics (Labov 1972), whereby women tend more frequently to use supra-local variants while men lean towards more localized traits, as has already been shown for the Amman, Jordan, context by Al-Wer (2009). This does not mean that in Ouezzane the dental plosive is necessarily the local variant. Rather, I merely wish to point to the accepted fact that there are two different prestigious varieties in competition. On the one hand, there is the new regionally prestigious “Northern Moroccan” urban standard, which Sánchez and Vicente (2012) already mentioned concerning the Tangier–Tetouan axis, and on the other, the new urbanized varieties, which have Casablanca as a particular focus but are widespread all over Morocco in cities such as Kenitra and Meknés as a result of television and radio broadcasting. The women of Ouezzane may prefer to identify themselves as speakers of Northern Moroccan, since it is a prestigious variety throughout the region and is furthermore especially linked to “femininity”, as some current linguistic stereotypes indicate (Vicente 2022a).

Regarding the age factor, we have already seen two opposite patterns, with those under 50 using slightly more frequently the affricated allophone while those over 50 favor the plosive one. This preference among the older cohort might simply point to tooth loss, but other possibilities should be considered. Again, Ziamari et al. (2020) claim that affrication “seems to have become one of the main phonological markers of a ‘young’ style,”17 and, together with palatalization, “is a trait that seems to be increasingly shared among young people, even if the values associated with this variable vary depending on the region”.18 This feature has gained particular prestige among young people because of its use in TV and radio programming aimed at young people all over Morocco, in general, and in the underground music scene, in particular. This rising prestige has meant that the trait is gradually losing its stigma among younger speakers. This may be affecting patterns of usage among not only younger but also older speakers, since if affrication has become a marker of “young person” style, older speakers who once used it may choose to drop it in favor of the plosive as an index of their self-perceived greater maturity. Further research should be carried out, for instance, to enlarge the number of male informants over 50, but the use of the voiceless alveolar plosive may signal a reconfiguration of the “old person” style as a marker of group affiliation or social differentiation, as Eckert (2012) has claimed.

Let us now return to the other much less common dental allophone, the voiceless alveolar affricate with palato-alveolar appendix [t^š]. As noted, although it is rarely found in our data, when it appears it is typically coming from young males. This allophone could well become a gender or age marker in the future, but given that the number of occurrences is quite limited in my sample, we must restrict ourselves to merely noting its presence. However, this trait is noteworthy in the case of one particular informant because it is the allophone that predominates in his speech. This informant was male, under 30 years of age, with only a primary education, possibly economically independent but a bit of a rebel, given that his parents complained that he did not support them financially despite his being single, living at home and having a job. His use of this feature might therefore signal his self-identification as a modern “young person” who follows Casablanca fashions, underground music, and so on, similar to Falchetta’s (2019) and Ziamari et al.’s (2020) participants. As they point out, though this trait is stigmatized in Central and Southern Moroccan varieties, [t^š] indexes not only youth but also success, modernity and popularity for younger speakers. The young informant in question claimed to possess these qualities, since he talked with pride about being very skilled at his job, but also boasted of having certain social contacts of questionable reputation, highlighting his transgressive “coolness”. This shows that, for him, the trait had gone from being stigmatized to signaling covert prestige, and he uses it to include himself in the “young, resourceful, modern male” crowd. When this informant uses this trait as part of his own idiolect, we may assume that it is a way of transgressing norms, just as he transgresses other social norms. By using [t^š], he takes a stand as a member of an exogenous group, breaking away from his own origin group of Ouezzane speakers.

6.2. Social Meanings Associated with */q/

In this section, I explore the social meanings ascribed to the voiceless uvular, glottal stop and voiced velar. I will not focus on the voiceless velar fricative [x] since, as noted, it represents an intralinguistic change. The analysis of these features will attempt to explain why they are either widespread or not completely abandoned, and, in the case of the voiced velar, why it has entered this variety.

The voiceless uvular is the most widespread of these allophones. As we have seen, it was used by all the participants in this study, with most women and highly educated men using it exclusively and the remaining participants using it in alternation with other variants. It could thus be posited that [q] is a neutral trait of this variety with no social meaning attached to it. However, the fact that the other variants seem to be part of an attrition process or index some kind of style could indicate that this feature is a marker of urbanization in the Ouezzane variety.

According to Vicente’s recent work (2021) on the variants of /q/ as used in the old medina of Tetouan on the northern coast of Morocco, the social meaning of the glottal stop is changing. This realization has long been used by the ancient community of inhabitants claiming Andalusi origins, who belong to the bourgeoisie and are regarded in Tetouan as a socio-cultural and economic elite. For this reason, the glottal stop has traditionally been considered a prestigious trait. However, nowadays, this attitude is present only in the older generation, whereas for younger speakers it is stigmatized as a sign of being old-fashioned or even snobbish. Despite these qualifications, the general consensus on Moroccan Arabic is that besides being the most widespread realization, [q] is currently the most prestigious realization or at the very least a non-marked feature.

In Ouezzane, on the other hand, according to the informants on which I have based my research, the glottal stop realization is also stigmatized for most of the population, albeit for a different reason than that reported in Vicente (2021). In this case, the glottal stop is seen as a sign of rural backwardness and is therefore avoided by the long-time city-dwellers, who regard themselves as more sophisticated than country-dwellers. This attitude was exemplified by one of my participants, a 40-year-old woman at the time of the interview who was born in the city but whose parents came from the Jbala Mountains. For family reasons, she had not been able to receive more than a primary school education. Furthermore, though she had been born and raised in Ouezzane, a middle-sized city, she had also spent five years in Tangier, a large city, qualifying her as an “urban” Moroccan woman. In response to my query about the glottal stop, she identified the use of this allophone with ǝn-nās dyāl l-bādiyya ‘people from the countryside’. The overwhelming preference of informants for the voiceless uvular appears to confirm that this metalinguistic judgment is not an isolated opinion, but rather a generalized one.

As shown by the data, the glottal stop is mostly used by men but also, to a smaller extent, by women. Analyzing what kind of informants still use this feature, I note that categories other than gender and age, such as education level, place of residence, profession or family background may play a role in explaining why this feature is still in use even if it is stigmatized and identified as rural. One might assume that the glottal stop is mainly used by informants of rural origin, but this is not always the case.

On the one hand, in my corpus, the glottal stop realization was used by men born in the city, with primary-level education, who were working as craftsmen—chiefly as weavers, embroiderers or tailors, the traditional wool-working occupations in the city—and, in most cases, working in neighborhood workshops. These informants could be classified as members of a specific social group sharing gender, level of education, profession, place of work, linguistic features—at least with regard to [Ɂ]—and other practices. The use of this feature could therefore index the style of a particular social group, the craftsmen of Ouezzane, since—despite the alternation between [q] and [Ɂ] in their speech—they choose it without hesitation or self-correcting, in a way that belies the fact that this allophone is a stereotyped and even stigmatized trait for other groups. The strength of this feature among them, otherwise recessive, signals a kind of pride in belonging to a particular guild that was once the economic powerhouse of the city.19

On the other, my female informants shared something besides their gender. One of them was learning embroidery under the guidance of one of the male craftsman participants. It is, therefore, conceivable that for this woman the glottal stop indexed the embroidery profession rather than gender. However, she should be regarded as an exception in this respect given that all the other female informants were either recent immigrants from the country who had settled in the Quššāriyyin neighborhood or were still living in the rural hinterland near Ouezzane. I observed that this allophone was not the one used by their daughters, or by their sons and husbands. This suggests that for these women the glottal stop was a marker of their “rural” identity. In this regard, my data is in line with previous studies which describe the glottal stop as characterizing the sociolects of middle-aged craftsmen (Caubet 1993, p. 15) and rural women (see studies in Vicente et al. 2017; Khoukh 1993, p. 12; Caubet 2017, p. 114; El Khomssi 2017, p. 160; and Benítez Fernández 2022).

Regarding the voiced velar [g], I have already mentioned that it belongs to the central and Saharan varieties of Moroccan Arabic. Identifying it as a new urban trait may indicate that it is becoming another prestigious standard feature and may also explain why it has become widespread. This realization was found among the young men of Ouezzane, who were among the more educated informants. As can be seen in the examples above, this variant appeared in the corpus in different forms of the verb gāl–ygūl ‘he said–he says’, while other instances were realized with the voiceless uvular. Thus, it is a new feature introduced by young men who have had contact with other vernacular varieties.

As we have noted, this velar realization [g] is a trait typical of the Bedouin varieties of Moroccan Arabic. Speakers of sedentary varieties have traditionally stigmatized this realization because it—and the Bedouin varieties in general—have been perceived as rural or uneducated. So how has it come about that young male speakers of sedentary varieties now include this traditionally stigmatized trait in their repertoire? To attend university, some of these young and highly educated men move to the city of Kenitra on the Atlantic coast, a much bigger city than Ouezzane and a Bedouin variety is spoken there. However, even if they spend the academic year in Kenitra, they regularly come back to Ouezzane for weekends and holidays. Kenitra represents for them, and most of the young inhabitants of Ouezzane, the modernity of the big city and the hope of upward social mobility because of the university and greater job opportunities. Furthermore, young people are continually exposed to the velar realization in the mass media, where new urban music stars generally perform in Casablanca vernacular or other Bedouin-origin varieties. This finding agrees with what Hachimi (2011, pp. 32–35) shows for young men and women from Fez who are living (or even were born) in Casablanca, whose use of [g] in [gāl] (“he said”) has been reinterpreted as “normal”. It is the only feature that shows for them the positive aspects of ḥrušiyya (“roughness”), meaning the resourcefulness, competence and independence that characterize the real urban Casaouis (natives of Casablanca).20 In other words, what is occurring in Ouezzane is a kind of re-evaluation process affecting this formerly stigmatized trait, with the result that young men use it in particular lexical items like the verb qāl–yqūl, which has become in their speech gāl–ygūl.

7. Discussion

The original goal of this study was to find an explanation for the wide variability in the realization of the two phonemes */t/ and */q/ in the vernacular Arabic spoken in Ouezzane that was linked either to macro-sociological factors or identity construction or social differentiation. In other words, I was looking for linguistic markers indexing gender and/or age, or for other social meaning indexes that might play a role in social differentiation, since, as third wave sociolinguistics propose, such indexes may lie at the heart of linguistic variation and change.

The quantitative analysis carried out here did not yield significant results in this respect, except perhaps the preference for the alveolar plosive [t] among the elderly informants (especially women) and introduction of the voiced velar plosive [g] among young men. Nevertheless, the small number of tokens in my corpus precludes any definitive claims about their being gender and/or age markers.

On the other hand, qualitative analysis has allowed various sociolinguistic representations to emerge. I have discerned an “older person style” of speech in the preference for using [t], which the elderly may view as a way to distance themselves from young people’s predilection for affrication, especially in the case of [t^š]. Attrition—seen in the balanced use of the voiceless alveolar affricate with sibilant appendix—and the near-disappearance of the interdental [θ] and the glottal stop [Ɂ] are clear signs of the construction of a neo-urban identity. Both the low rates of occurrences of these last two allophones and metalinguistic comments made by informants regarding the glottal stop point to a rejection of rurality and a desire to be part of that neo-urban elite, even if elite membership does not necessarily imply any kind of socio-economic advancement. However, as in Labov’s Martha’s Vineyard study (Labov 1963), this neo-urban identity construction is contested by some subgroups—the craftsmen and the “half-rural” women—who still maintain in use, albeit tenuously, the glottal stop [Ɂ]. This suggests that the feature conveys some kind of a covert prestige, signaling a stance of honor of Ouazzane’s past. On the one hand, it used to be a small city, where rural farmers went to sell their produce, as El Khomssi (2017) has shown, constituting both its workforce and source of supplies for the pilgrims flocking annually to the zawiya. On the other hand, the past supremacy of Ouezzane was also tied to the craftsmen of the wool trade, since the jellaba ouazzaniyya had a national reputation for quality and was the main source of income for the city. Sadly, since the 1980s, this craft has dwindled in the face of industrial cloth-making processes and the import of cheaper textile materials. As customers become increasingly reluctant to pay the extra cost of handmade and tailored clothing, the wool-makers guild no longer flourishes, and it is becoming more and more difficult to attract young people to learn the craft.

Besides the neo-urban identity construction, I would like to note another style that combines two different features, the voiced velar introduced by university-educated young men and the voiceless alveolar affricate with palato-alveolar appendix employed by a kind of “urban-hip” youth. This can be seen as conveying an “outsider” social meaning, or rejection, indicating a break with the established order, whether it be a local identity, a social convention, etc.

8. Conclusions

The initial hypothesis of this study was to explore how gender and age, as macro-sociological factors, influence variation and change in the variety of Northern Moroccan Arabic spoken in Ouezzane and how social meaning plays a role in this variation. To do so, I have examined two phonetic variables, the voiceless alveolar plosive /t/ and its variants, and the voiceless uvular plosive /q/ and its variants, based on the great degree of variability these sounds show in their realization.

I have first analyzed the data from a quantitative perspective with gender and age as possible clarifying factors. However, the quantitative analysis in itself was not always decisive. In the case of the dental plosive, gender seems to have no impact on the choice of variants, while age may explain the direction of a possible future change. On the other hand, the quantitative analysis has revealed some gender and age issues related to the uvular plosive, such as the introduction of the voiced velar as an innovation originating in the speech of young men.

In the light of the lack of definitive results in the quantitative analysis, I then applied a qualitative approach to verify what role social meaning might play in the variability observed. With regard to the essentially balanced distribution among informants of the dental plosive and the alveolar affricate allophones for /t/, I have pointed to both the influence of two different prestigious varieties centered, respectively, around Casablanca–the Atlantic coast and Tangier–Tetouan, and the increased urbanization of the city of Ouezzane itself. As for the female proclivity for the alveolar affricate, I have shown a certain awareness on the part of this subgroup of a “Northern” identity. Finally, regarding age, not only have I provided evidence for the construction of a “young person” identity—well documented in the aforementioned 2020 study by Ziamari et al.—but also a social meaning-based analysis has allowed me to identify other factors as triggers of linguistic variation. First, I have shown how the impulse to break with the origin group is manifested in the use of the voiceless alveolar affricate with the palato-alveolar appendix in the idiolect of one young informant, and also in the use of the voiced velar in the sociolect of the young men in general. Second, I have shown the tendency of most informants to try to “de-Jeblize” their speech, apparent in the attrition or loss of some traits. Third, I have documented the construction of an “older person” identity, characterized by a diminished use of the alveolar affricate. Finally, I have shed light on a covert prestige process, related to the speaker’s pride in belonging to a very localized group—of either rural women or craftsmen—which is manifested in a consistent use of the glottal stop.

The results of this study show that the use or avoidance of these features—which are stigmatized elsewhere in Morocco—reveal different social stances that construct social identities by either allegiance to, or rejection of, a group. Additionally, I have proven that Ouezzane vernacular is currently undergoing a process of change. This process aims to leave behind the rural character this variety was seen to display in the past and convert it into an urbanized vernacular. This neo-urban variety is situated in an intermediate position between the Northern Moroccan vernaculars, whose regional standard has been described by Sánchez and Vicente (2012), and the so-called national standard, characterized by its inclusion of Casablanca linguistic elements and its dissemination through the popular media (Hachimi 2012). In short, this study has provided further support for the notion that social meaning can have a significant role to play in linguistic variation.