An Innovative Copula in Maghrebi Arabic and Its Dialectological Repercussions: The Case of Copular yabda

: Research on copulas in Arabic dialects has hitherto largely focused on the pronominal copula, and has also mostly ignored Maghrebi dialects. Drawing on published literature as well as ﬁeldwork-based corpora, this article identiﬁes and analyzes a hitherto undescribed verbal copula in dialects of Tunisian and northwestern Libya deriving from the verb yabda (“to begin”). We show that copular yabda occurs mostly in predicational copular sentences, with time reference including the habitual present and generic future. It takes nominal, adjectival, and locational predicate types. We also argue for broader inclusion of syntactic isoglosses in Arabic dialectology, and show how copular yabda crosses the traditional isogloss lines established on the basis of phonology, morphology, or lexicon, and therefore contradicts established dialect classiﬁcations such as Bedouin/sedentary or Tunisian/Libyan.


Introduction
Arabic dialectology has largely focused until now on understanding the geographic distribution of varieties through socio-historical parameters. The traditional dialectological approach to the Arabic varieties of northern Africa ("Maghrebi" varieties) foregrounds a classification scheme which is organized not only along geographical lines, but one which also depends on ecological categories ("Sedentary" vs. "Bedouin") as well as sociohistorical ones ("pre-Hilali" vs. "Hilali") (Caubet 2001;Palva 2006;. While certain categories used for classifying Maghrebi Arabic varieties have recently been subject to critique from historical perspectives (Kosansky 2016 on "Judeo-Arabic"; Benkato 2019 on "Bedouin"), it has also been shown that the existing linguistic evidence does not necessarily support the utility of other categories. 1 Similarly, it can be pointed out that the existing classifications rely almost exclusively on phonological and morphological isoglosses, and to a lesser extent on lexical ones. Though neglect of morphosyntax for drawing isoglosses is typical of dialectology in general, the problem is particularly astute in Arabic dialectology in northern Africa. This is not only because morphosyntax is almost entirely ignored, but because regional variation in phonology and morphology can often be rather limited, meaning that dialect boundaries drawn on the basis of a handful of such isoglosses are not strong.
The dialectology of Maghrebi Arabic, therefore, could benefit not only from the continued interrogation of the traditional classification system but also from drawing on a broader set of data that includes previously unexamined linguistic features, particularly morphosyntactic ones. This study, by describing a syntactic feature and examining its consequences for dialectology, aims to show how such work has the potential to change the traditional map of Arabic in northern Africa. It opens by giving a brief overview of the copula in Arabic dialects (Section 2), before proceeding to the description of a hitherto unidentified copula in varieties of Tunisia and Libya (Section 3). The study then discusses the neglect of syntax in Arabic dialectology and shows that syntactic isoglosses may conflict with isoglosses based on other linguistic features (Section 4).

Copulas in Arabic
Copula constructions are to be understood as constructions used to encode the identity of two participants and to express group membership, classification, location and the ascription of a range of properties to a participant and the element linking these is a copula. It is common to assume that the copula is lexically-semantically empty (Pustet 2003, p. 5) and that its main role is in semantic composition and in carrying tense/aspect . The generally-accepted major types of copula construction are predicational, equative, specificational, and identificational (Higgins 1979, pp. 204-93;. For our purposes, an equative copula construction is one which equates the referents of the two elements besides the copula (Mikkelson , pp. 1807, while a predicational copula construction is one whose subject is referential and whose predicate is some non-verbal element, whether nominal, adjectival, or prepositional (Mikkelson , pp. 1808. While languages vary greatly in terms of what elements provide copulas, and which constructions require overt copulas, we can state the following regarding how copula constructions in Arabic are typically viewed. 2 Predicational constructions with present reference usually use a zero copula (1-2). In Tripoli Arabic, copula constructions with zero copula describe facts and express general truths in thetic utterances, serving to present an entity, a proposal or a state of affairs that is new information. 3 In such utterances, a state of being (an inherent or permanent characteristic of a being, as in the first example) or a current activity (including location, as in the second example), considered true by the speaker, is expressed. (1) Tripoli Arabic (Pereira corpus) 4 šaxs .ī yt-a Ø d . Qīf-a personality-3SG.M Ø weak-F "His personality is weak" (2) Tripoli Arabic ( In examples (1-2) above, the zero copula is employed in phrases in which the speaker validates the predicative relation. The zero copula thus expresses realis/indicative. Copula constructions of all types which have temporal reference to the non-present, however, require an overt copula, usually provided by a form of the verb kān/ykūn "be" (3a). Moreover, if epistemic modality is to be expressed, the overt copula ykūn is required (Pereira 2010, pp. 453-67): in copula constructions with ykūn, the predicative relation is to some extent uncertain and the construction thus expresses irrealis/potential (3b). (3) Tripoli Arabic (Pereira corpus) (a) Qadnān kān f-@s . -s . ah . r . a Adnan be\PFV[3 SG.M] in-DEF-desert "Adnan was in the desert" (b) Qadnān y-kūn f-@s . -s . ah . r . a Adnan 3M-be\IPFV in-DEF-desert "Adnan may be in the desert" To explicitly situate the copula construction in the future, the preverb h .āprecedes the verb ykūn (Benmoftah and Pereira 2019). In the following example (4), we can compare the use of the zero copula with a generic present reference and the verb ykūn with a future reference. Moreover, many dialects make use of copula forms in addition to the zero copula and kān/ykūn copula. For example, in some dialects, such as those in Egypt or Lebanon, presenttense equative constructions in which the complement is a definite noun phrase optionally use a copula based on the 3rd-person independent pronoun (Choueiri 2016) (example 5). Peripheral Arabic dialects go farther and employ the full range of the independent personal pronouns in these constructions (Akkuş 2018, pp. 459-62)  It is worth pointing out that essentially all literature on the copula in Arabic, theoretical or descriptive, has been devoted to either the "typical" copula situation or to the pronominal copula. 5 Other types of copulas in Arabic dialects, especially ones which derive from verbs, have hardly been described. Only very recently have scholars begun to address the existence of other copulas, in particular the use of gāQid, formally the active participle of "to sit/to stay", as a present-tense predicational copula in varieties such as Maltese and others (Camilleri and Sadler 2019, 2020) (example 6). (6) Maltese (Camilleri and Sadler 2019, p. 10) omm-i qieghd-a d-dar mother-1SG COP\PTCP.ACT-F DEF-house.F "My mother is at home" Here, we describe for the first time the existence of an additional copula occuring in Arabic varieties of Tunisia and northwest Libya. This copula, supplied by the verb yabda (lexically "to begin") occurs in certain types of predicational constructions. In Section 3, we will analyze copular yabda on the basis of representative examples from the welldocumented varieties of Tunis (northern Tunisia), Douz (southern Tunisia), and Tripoli (northwest Libya). Since the goal of our study is dialectological in nature, deeper discussion of the grammaticalization path undergone by yabda to become a copula will be left aside, and we will concentrate on describing and comparing its function in these three dialects. 6 The geographical range of copular yabda and its importance for dialectology will then be discussed in Section 4.

Copular yabda in Maghrebi Dialects
A copular element consisting of the verb yabda in the imperfective conjugation occurs in predicational constructions, mainly those where an overt copula is required. Copular yabda mainly occurs in narrative or descriptive contexts to refer to a habitual action, event, or description. It can also be used to refer to a present state and to describe an event which is happening at the moment of speech. Moreover, yabda can have a future value. Finally, it is used in addition to ykūn as the auxiliary of the future perfect. So far as we can tell, yabda never has a past reference, that is, in the perfective conjugation it is used only as a lexical verb and not as a copula. From a modal point of view, copular yabda seems to be used, as opposed to ykūn, when the speaker considers the states and the situations to be true or when the speaker believes that the content of the interrogative sentence can be validated by the interlocutor.
While the dialects under discussion all have parallels in the syntax of yabda, it should be noted that the phonological or morphological particularities of each dialect do apply to copular yabda, without affecting its meaning. For example, the variety of Douz marks gender in the plural verb while that of Tripoli does not; hence Douz has both a 3PL.M yabdu and a 3PL.F yabdan while Tripoli has only 3PL yabdu. Or, the morphophonology of the 3PL may differ: yabdāw in Tunis but yabdu in Tripoli.

Habitual Present
In the majority of our examples, copular yabda indicates the usual occurrence of a state or situation. In all the following utterances, copular yabda has a habitual present value. It can also be used to express a general truth. It appears in main clauses as well as temporal clauses and can occur with adjectival, nominal, or locational predicates. 7

Adjectival Predicate
The predicate can be adjectival (including passive participles). In the following examples, yabda refers to actions or events that take place habitually: indeed, in the first example, it describes an event that takes place every year because of the change of seasons; in the second one, every time a meat dish is cooked according to a particular method; finally, in the third example, every time the family gets together.

(7)
Tripoli (Pereira 2010, p. 361) 8 f-fas . l @r-rbīQ ta-bda d-d@nya xad . r . a in-season DEF-spring 3F-COP\IPFV DEF-world.F green.F "In spring (i.e., every spring), nature is green" kāda so "And of course (each time we go to my grandfather's house) the family is reunited and we begin to talk about general topics and topics concerning the family and so forth" It is also the case in Tunis Arabic where yabda allows the expression of habitual facts. Without yabda, examples (10-11) would have an actual present meaning. 9 (10) Tunis (Singer 1984, p. 317) el-far . š y@-bda Qāli Qa-l-ar . ḋ DEF-bed 3M-COP\IPFV high above-DEF-ground.F "The bed is high above the ground" (11) Tunis (Singer 1984, p. 317) t@-bda d-denya bard-a 3F-COP\IPFV DEF-world.F cold-F "Nature (i.e., the weather) is cold" Copular yabda also appears in temporal clauses with a habitual present value. The conjunctions (lamma, kīf and kī) refer not to a single, but rather to the habitual, occurrence of an event. The three representative dialects agree in this usage. (12) Tripoli (Pereira corpus) al lāh lamma ya-bda xrayy@f kīf god when 3M-COP\IPFV lamb\DIM just madbūh .
al lāh slaughter\PTCP.PASS.SG.M god "Oh god (i.e., what a delight) when a lamb is just slaughtered!" When the door is like this white and (with) these colors on it, it means that the owner has gone to the pilgrimage" Douz (Ritt-Benmimoun 2014, pp. 67, 195 "When they are in a place which isn't good"

Nominal Predicate
The predicate can also be a nominal phrase and copular yabda allows to provide a comment on an event or a fact as they habitually occur.
But when it is a good year, the male or female camel sniffs the plain with its nose" "He is a minister, and his doorman, they sit at lunchtime and eat lunch together"

a divorced woman) is back with children and this is the other big problem"
In the following utterance, contrary to the previous examples, copular yabda is used in Tunis Arabic in an equative construction. (21) Tunis (Sellami corpus) fi š-šmāl @l-m@sfūf y@-bda kisiksi žwayy@d abyaą in DEF-north DEF-masfūf 3M-COP\IPFV couscous fine white "In the north, masfūf is fine white couscous" This example shows that yabda is required because there is a semantic constraint, in this case the circumstantial fi š-šmāl "in the north", which limits the applicability of the claim about what masfūf is to a particular region. Otherwise, the equative construction with no overt copula would be used: @l-m@sfūf Ø kisiksi "masfūf (is) couscous".

Locational Predication
Locational predication can also be expressed with copular yabda. In this case, the copula complement consists of a prepositional phrase or a locational adverb. The locational predication can have a habitual value (22-23). In the following example from Tunis, copular yabda expresses locational predication in a temporal clause. (24) Tunis (Singer 1984, p in wife-3SG.M "Every time he is at home, he keeps criticizing his wife".

Future
Depending on the context, copular yabda situates an event or a state in the future, whether it is a question about location or state or a wish or hope about a situation. The three representative dialects agree in this usage. (25) Tripoli ( 2-see\IPFV "God-willing you will be near us and you will watch, you will see"

Future Perfect
Finally, followed by a verb in the perfective, yabda is also used as the auxiliary verb of the future perfect, indicating a state or situation that is expected or planned to occur in the future. Here, however, yabda and ykūn can both be used with a variation in meaning that requires further study. This usage only exists in two of the representative dialects: Tripoli and Tunis. (29) Tripoli ( "On the 21st I can't I will have gone back home". (32) Tunis (Sellami corpus) s . h .ā b-@k l-qdom l-koll y-kūn-u Qarrs-u w friend\PL-2SG DEF-old\PL DEF-all 3-be\IPFV-PL marry\PFV-3PL and @nti lā 2SG.F no "All of your old friends will have gotten married but not you".
In Tripoli, the use of yabda or ykūn for the future perfect seems to break down along the following lines: yabda is used when the speaker considers the future state or situation as certain to occur, while ykūn in contrast allows for the addition of modality, expressing a supposition or a fictional or probable hypothesis. This aligns, in fact, with the use of ykūn for expressing epistemic modality in the present (Pereira 2010, pp. 453-67 The distinction between yabda or ykūn seems to be similar in Tunis as well, though this requires further study.

Copular yabda as Isogloss and the Problem of Syntactic Isoglosses
As shown in the preceding section, copular yabda exists in both "northern" and "southern" Tunisian varieties, as typified for this study by the areas of Tunis and Douz, respectively. More generally, according to Tunisian colleagues and colleagues working on other Tunisian varieties, it can be considered a pan-Tunisian feature. 11 In Libya, the only location where copular yabda has been documented is Tripoli, though it would be unsurprising if other varieties of northwestern Libya, about which there is little published, also had the feature. The total geographic extent of copular yabda is not yet known; but it does not exist in Benghazi or eastern Libyan varieties generally, and there is essentially no documentation of eastern Algerian varieties available for comparison. It is unknown in areas of central coastal Algeria, such as Algiers or Dellys, however. 12 According to the existing information, therefore, it is a shared feature of the varieties of Tunisia and Tripoli (see Table 1). 13 That these dialects share a linguistic feature, in particular an innovation, is unexpected given the categories and isoglosses typically used in Arabic dialectology. Copular yabda crosses not only national boundaries (Tunisia/Libya) but also the pseudo-typological ones most prominent in Arabic dialectology, in particular the categories of "pre-Hilali/Hilali" or "sedentary/Bedouin". Besides the fact that these categories are outdated and problematic from a socio-historical point of view, it must also be pointed out that the collection of features on which they are based almost never includes syntactic features. In Arabic dialectology, syntax plays very little role in discussion of dialect classification. For example in a recent handbook, the authors note that "syntax will, and we do not constitute an exception in so doing, only be taken into account in a restricted manner, although in this area too significant differences between dialects are present" (Behnstedt and Woidich 2005, p. 68). More generally, recent large projects of regional dialectology, such as the Wortatlas der arabischen Dialekte (Behnstedt andWoidich 2011-2021), include phonology, morphology, and lexicon, but not syntax. Even the most recent overviews of Maghrebi dialects (e.g., Aguadé 2018) do not treat syntax. Syntax has received slightly more attention from sociolinguists and contact linguists, but is typically not used as the basis for regionallyorganized dialect groupings nor has it been studied as part of intra-dialect variation in ways comparable to phonology or morphology. 14 Syntax seems to be neglected in dialectology in general regardless of language. Even recently, scholars have gone as far as stating that "there is no doubt that syntax has been the most neglected linguistic subsystem in classical dialectology" (Berger et al. 2012, p. 93). On one hand, this goes back to the fact that traditional dialectological methods, such as the word list and questionnaires, can be unsuitable for describing syntax; on the other, syntax does not necessarily fit the diachronic documentation goal of traditional dialectology, which concentrated on phonological and lexical criteria (Glaser 1996;. However, this state of affairs has changed quite significantly in certain fields, such as Germanic and Romance dialectology Berger et al. 2012;Glaser 2013).
Arabic dialectology has largely shared the traditional dialectological emphasis on uncovering archaisms, partially due to its goal of answering questions about the historical origin of Arabic dialects. As with other languages, Arabic dialect groupings have been made primarily on the basis of phonological, morphological, and lexical isoglosses. 15 For example, of the 73 isoglosses used by De Jong (2000, pp. 39-48) to group the Arabic dialects of the Sinai peninsula, only 4 can potentially be characterized as (morpho-)syntactic. Meanwhile, some of these traditional non-syntactic isoglosses may not withstand scrutiny: Embarki (2008) argues, for example, that some of the isoglosses traditionally considered to be strong markers of dialect type, such as the interdental consonants, exhibit too much variation within a single dialect to really be useful discriminants (and see again Guerrero, forthcoming). This being the case, attention to syntax as part of dialectology has the potential to complexify and even complicate the typical dialect groupings. Indeed, it has been noted that syntactic isoglosses often cross and contradict the established isoglosses based on phonology or lexicon . As Glaser (2013, p. 204) puts it, "that geographically conditioned syntactic variation indisputably exists does not entail, however, that the distribution of syntactic variants is identical to the distribution of phonological or lexical variants". For the Arabic varieties under discussion here, this crossing and contradiction can easily be illustrated with a quick look at only a few isoglosses (Table 2). The above table considers three phonological, two morphological, three lexical, and one syntactic variable. Each of these categories yields different isogloss lines: in some cases Douz and Tripoli agree (nos. 1, 3, 5, 6), in other cases Tunis and Douz agree (no. 2). An isogloss grouping Tunis and Tripoli can even be found, namely the lack of gender marking on plural verbs (no. 4). Of course, many of these features are shared with dialects beyond these three and so only serve to connect two of the three with each other, but not to separate them out from surrounding dialects. Copular yabda not only is an isogloss connecting Tunis, Douz, and Tripoli, but also one which separates them out from other Maghrebi dialects.
This raises the question of how much weight a syntactic isogloss should have as part of a group of multiple different isoglosses. While phonological and lexical isoglosses are typically more valued by dialectologists, and more frequently available in the published literature, Chambers and Trudgill (2004, pp. 96-100) note that there is evidence that "grammatical variables stratify speech communities much more sharply than do phonological and lexical variables", suggesting that regions delimited by grammatical isoglosses will be more strongly regarded as different dialect areas than regions separated by mostly phonological and lexical ones. 16 Moreover, there seems to be agreement that grammatical isoglosses delimit larger areas than phonological or lexical ones. In this regard, one would think that copular yabda and other syntactic isoglosses should actually have a fair amount of weight when it comes to drawing up-to-date subgroups of Maghrebi Arabic.
Proponents of the traditional dialectological view might note that yabda is relatively new in the history of the Arabic varieties in question and that, as an innovation, only represents the spread of a particular feature in very recent history and therefore does not affect the traditional classification. But we would counter that copular yabda is not necessarily all that new, as it is already attested in Tripoli at the end of the 19th century. 17 Moreover, an innovative feature that is well-attested in a fairly significant region should be the concern of dialectologists and future research should attempt to account for its history and present distribution. For example, did copular yabda jump between urban areas, slowly spreading into the rural areas between them? Or did it radiate out from a particular area where it was first innovated? Why has it, seemingly, not been accompanied by the spread of non-syntactic features?
If we are dealing with the spread of a syntactic innovation in the Arabic varieties of a particular region, then we indeed have to think less about the traditional classifications, which attempt to explain how the distribution of Arabic may have looked centuries ago, and more about processes of inter-dialectal contact and diffusion. And it is here that copular yabda may also make a contribution, since studies of inter-dialectal contact in Arabic have typically focused on what happens when different dialects come into contact in urban environments, rather than looking at the diffusion of a feature over a large region. These studies also typically focus on phonological and morphological variables, rather than syntactic ones. Meanwhile, general studies of convergence do typically focus on morphosyntax, though in most cases they deal with totally different languages rather than different varieties of a language. Copular yabda may represent a case of a syntactic innovation being spread through dialect contact over a large region, giving rise to a dialectal version of a "linguistic area", that is, the "outcome of diffusion of structural 'patterns' across language boundaries" (Matras 2011, p. 146). In that case, it may be one example of area formation in Arabic dialects, and indeed one that does not follow national boundaries but instead crosses them. And again here, syntax is important, since, as is clear from Table 2 above, the diffusion of copular yabda seems, so far as can be seen, not to have been accompanied by the diffusion of other linguistic features. It thus speaks to interaction between Maghrebi dialects that can not be seen simply by looking at areas like phonology or lexicon. Future research should therefore look to morphosyntax in search of other features which (unexpectedly) link Tunisia and northwest Libya, or characterize other dialect areas in general.

Conclusions
In this study, we have attempted to describe the occurrence of a verbal copula in certain types of predicational, and less frequently equative, constructions in dialects of Tunisia and northwest Libya. This copula is provided by the verb yabda, lexically meaning "to begin", and occurs in predicational constructions which require an overt copula, both in the present and future, including constructions with temporal or modal implication. This can be illustrated succinctly with a final example, taken from social media, where the generic predicational construction with zero copula (35a) contrasts with the temporal construction requiring an overt copula (35b) which is supplied with a form of yabda.
Tunis Arabic (a) umm-@k ø tūnsīy-a mother-2SG ø Tunisian-F "Your mother is Tunisian" (b) kī ta-bda umm-@k tūnsīy-a when 3F-COP\IPFV mother-2SG Tunisian-F "When your mother is Tunisian..." Importantly, however, the yabda copula is attested in a number of dialects, including three dialects-Tunis, Douz, and Tripoli-which are not closely linked in the traditional dialectological classifications. As a syntactic isogloss, yabda crosses the isoglosses drawn from other linguistic levels, ignoring national and typological boundaries, exhibiting behavior seen in syntactic isoglosses more generally. While our study has only been able to use currently existing material to suggest what the rough area contained by the yabda isogloss may be, additional data from locales in between these three representative locations may be able to help us define that area more precisely, and, in addition, potentially show if there are transitional areas as well. More importantly, copular yabda requires explanations that do not draw on the traditional historical classifications for Arabic dialects, but look to diffusion, area formation, and above all contact. We suggest that syntactic features should play a larger role in Arabic dialectology, and including more of them in the lists of isoglosses drawn on for classification has the potential to complexify and even reshape our understanding of the distribution of Arabic dialects and the processes which continue to shape them. Acknowledgments: The authors would like to thank Zeineb Sellami for sharing and discussing unpublished data with us, as well as Marwa Benshenshin and the anonymous reviewers for their valuable suggestions and corrections.

Conflicts of Interest:
The authors declare no conflict of interest.

1
For example, Taine-Cheikh (2017) argues that linguistic criteria do not necessarily support the existence or utility of the subdivision of "Bedouin" into "Hilali", "Sulaymi", and "Ma'qil" subgroups. Guerrero (forthcoming) argues that one of the traditional isoglosses separating sedentary/bedouin or pre-Hilali/Hilali dialects, the presence or absence of the interdental consonants (t, d, d . ), is not well-founded. In addition, the applicability and characteristics of the category of "village" or "rural" dialects has continued to provoke discussion (Mion 2015;Guerrero 2018 As opposed, for instance, to utterances where the pragmatic marker r .ā is used to focus the predicate or the entire predicative relation (Caron et al. 2015, pp. 105-7), such as šaxs .ī y-t-a r .ā -hi d . Qīf-a "his personality, it really is weak" and Qadnān r .ā -hu f-@s . -s . ah . r . a "Adnan, he is indeed in the desert". 4 The syntax of glosses is largely inspired by the Moroccan and Libyan Arabic list of glosses developed by Dominique Caubet, Ángeles Vicente, Alexandrine Barontini, and Christophe Pereira for the CorpAfroAs project. This list of glosses can be viewed at the following address: http://corpafroas.tge-adonis.fr/fichiers/Moroccan_Libyan_Arabic_Glosses_final.pdf (accessed on 21 October 2021). 5 Such as (Aoun et al. 2010, pp. 35-44;Alharbi 2017;Alotaibi 2018, and many others). 6 The grammaticalization of "begin" into a copula seems unusual, and is not mentioned in the typological literature on grammaticalization or non-verbal predication (e.g., Heine and Kuteva 2002;Hengenveld 1992;Pustet 2003). 7 The copular forms of yabda will be glossed as COP rather than etymologically as "begin" in order to avoid confusion with the lexical usage of yabda. 8 Example 7 can be considered as a general truth and this precisely is why the use of yabda is required. To express a change taking place, the verb ywәlli "become" would have to be employed. 9 Singer (1984, p. 317) provides an additional number of examples which confirm the use of yabda with adjective predicates, but which are hard to discuss because they do not include any context besides the copula and predicate. His translations (e.g., t@bda r .ī ḋ a "sie ist breit" or yibdāw fārh .ī n "sie sind froh") suggest that yibda is to be understood as a normal present copula, but this is not the case as far as we can tell. In contrast, it should be noted that the examples from  are based on her published text corpus (Ritt-Benmimoun 2011), and the context can always be checked. 10 Note that this sentence also means "he has built his house" in Tripoli Arabic. The use of yabda instead of ykūn avoids confusion since yabda + suffixed conjugation only expresses the future perfect.

11
Zeineb Sellami, p.c. It is also attested in the areas of Chebba (Luca D'Anna, p.c.), Mahadhba (Marçais and Viré 1981, p. 375), and Susa (Talmoudi 1984, p. 63), though because the published attestations are very few, it is not known if these areas align completely with the Tunis or Douz usages or not.

12
Lameen Souag, p.c. 13 Interestingly, copular yabda is attested in the materials from the Fezzan collected mainly by Philippe Marçais in the 1950s and published posthumously . These materials are problematic, however, as for the most part it is not clear where or from whom any given text was collected (materials were gathered not only in the Fezzan, but also in Tripoli and in Algiers, and only one text has the name or any personal details of an informant). There is thus no way to be sure that yabda in these materials represents an actual usage of a Fezzani dialect rather than the usage, say, of someone who was originally from the Fezzan but had been living in Tripoli for some time. What also casts some doubt on a copular yabda in the Fezzan is that of the nine attestations, seven are attributed to the "Gwayda tribe", suggesting that they may all come from a single informant (the other two have no attribution). We have therefore left these attestations out of our analysis. This is not to suggest, however, that copular yabda has not now spread to some regions of southern Libya; but there are unfortunately no studies which can confirm this as of yet.
14 See the recent handbook chapters (Camilleri 2019;Choueiri 2019). 15 See (Magidow 2016) for an example of dialect classification based on morphological forms. For some broad comparative studies of syntax, see Ritt-Benmimoun (2017, pp. 324-32) on object and aspect marking with fī, or Bettega (2019) on gender agreement. 16 See also Behnstedt and Woidich (2005, pp. 83-92) for Arabic specifically. Note also that whether or not a given isogloss affects mutual intelligibility is a separate question, and the answer may not actually correspond to the isoglosses valued by dialectologists, see (Čéplö et al. 2016).