4.1. Global Evaluation of Results
Having reviewed the major phonological, pronominal, verbal inflectional and syntactic characteristics of the Arabic dialects of Cairo, Qift, il-Biʿeṛāt, Khartoum, the Šukriyya, and Nigeria, we will now direct the information adduced toward a linguistic evaluation of existing proposals of an Egypto-Sudanic dialect classification, as has been repeatedly asserted on the nonlinguistic basis of shared genealogical history uniting the region’s Arabic speakers. In the event that such nonlinguistic factorsas migration history and common descent prove viable grounds for the classification and grouping of language varieties used in the region, expectation is that a substantial number of shared linguistic features will arise to characterize the varieties in question. This would justify the prediction of a meaningful degree of dialectological similarity as a consequence of the historical and demographic unity ascribed to their speakers by extra-linguistic lines of research.
This expectation, however, is not substantively met by the linguistic data gathered through the process of this inquiry. Of over fifty phonological, morphological, and syntactic features identified and discussed in the preceding subsections, only seven may be recognized as uniformly present across all Egypto-Sudanic varieties sampled. These are:
Proximal demonstrative paradigm on the pattern *dā, dī, dōl, dēl (Section 3.2.2
1.sg/2.m.sg perfect in -t
, or further evolution thereof (Section 3.3.1
To these, we might, for the sake of consideration, generously add six more—those features which proved characteristic of all but one of the surveyed dialects, and whose incidence may thus have been proved broader in a different sampling. These are:
The question, then, stands: Are these features sufficient to corroborate the existence of a linguistically significant Egypto-Sudanic dialect classification, proceeding from a common dialectal input carried by those historical communities who introduced Arabic first to Egypt, then to the Sudanic area via subsequent migration?
Though no conventionalized, objective threshold exists by which to make such a determination, the evidence in the Egypto-Sudanic case is not compelling—neither in terms of its quantity nor, critically, its quality. Of the thirteen isoglossic features identified as uniform or near-uniform across the six varieties examined, two—3.f.sg -at, and distinction of */a, i, u/—are clear retentions from a common Old Arabic inheritance, not innovations distinctive of further dialectal diversification. While thus not contradicting a narrative of dialectal relatedness due to shared migration history, neither do they positively support one: rather, they simply reflect the fact that dialects of the Egypto-Sudanic area have remain largely unimpacted by the mergers of */a, i/ emanating from the west of the modern Arabic-speaking world and */i, u/ associated with its north and east, as well as the change -at > -it typical of a number of Eastern Mediterranean varieties. None of these facts are surprising, and do nothing to indicate a shared developmental history of Arabic varieties in the region—simply a shared, central geography.
Of the remaining features which may be considered genuinely innovative, some are so ubiquitous across modern Arabic as to hold little meaningful value in establishing an identifiable Egypto-Sudanic dialect classification based in shared demographic heritage. Among these are the monophthongization of */ay, aw/, retained as diphthongs only in scattered relict zones; the use of a 1.sg/2.m.sg perfect suffix -t
(< *-tu), typical of virtually all modern Arabic varieties save those of the northern Fertile Crescent and parts of Yemen; and the change of initial */a/ > /i/ in the second person independent pronouns, identifiable in the vast majority of dialects outside the Arabian Peninsula (and many within it). These traits do not serve to differentiate dialects of the Egypto-Sudanic area from their immediate geographic neighbors in eastern Libya, the Hijaz or the Sinai (Owens 1984
; Schreiber 1970
; de Jong 2000
), nor from the bulk of modern Arabic more broadly. A further number of features are not quite so universal in attestation, but still spread far beyond the bounds of the Egypto-Sudanic region. Fortition of interdental fricatives to corresponding stops, though not typical of the Egypto-Sudanic varieties’ closest orbit of northern neighbors in eastern Libya or the Sinai (Owens 1984
; de Jong 2000
), is shared with the majority of varieties (both “sedentary” and some traditionally “Bedouin”) of the remainder of North Africa and the Levant, as well as urban Hijazi speech across the Red Sea (Schreiber 1970
Elision of /i, u/ (but not /a/) in unstressed, open-syllable environment is well known outside the region and is present in the Egypto-Sudanic varieties’ easterly dialectal neighbors in the Sinai and Mecca; the same is true for the voicing of */q/ > /g/, which is commonplace westward into Libya as well (de Jong 2000
; Schreiber 1970
; Owens 1984
). Reflexes of the verb-modifying prefix *bi- extend beyond the Egypto-Sudanic zone’s eastern edges into the urban Hijaz and the Sinai (Schreiber 1970
; de Jong 2000
), and further into Arabia and the Levant. Though absent from eastern Libya, their presence resumes in that country’s west (Owens 1984
). These features, then—while of obvious descriptive relevance—do not much contribute toward the definition of a classificatory unit which interprets the Egypto-Sudanic varieties as a discretely identifiable group, distinguished from other, neighboring dialects by the products of a separate developmental history.
The original thirteen features which might have been invoked in this regard, then, have fallen to four: a proximal demonstrative paradigm on the pattern *dā, dī, dōl, dēl, unmarked and obligatory post-nominal demonstrative order, in situ WH-question formation, and use of the genitive exponent *bitāʕ. These traits, held in common across all or nearly all members of the sampled group, are both innovative and, largely, distinctive—not generally encountered beyond these dialects’ immediate environs, neither are they typical even of closely neighboring varieties. A similar demonstrative paradigm is reported for Mecca alongside more common variants with an initial *hā- element, and *bitāʕ is variably attested in some dialects of the Sinai, but neither trait dominates in either region (Schreiber 1970
; de Jong 2000
). Both Meccan and eastern Libyan Arabic allow in situ WH-question and post-nominal demonstrative orders, but these are not unmarked or obligatory to the degree identified among the Egypto-Sudanic dialects considered here (Schreiber 1970
; Owens 1984
). From a synchronic descriptive standpoint, then, these four isoglosses stand as strong candidates to delineate linguistically meaningful boundaries between dialects of the Egypto-Sudanic area and adjacent Arabic varieties.
Such does not automatically, however, render these four features supportive of an Egypto-Sudanic dialect classification of the form so often proposed, predicated on the shared genealogical history of the Egyptian and Sudanic Arabic speech communities. Under such a framework, the claim advanced is that the migration of Arabic speakers from Egypt to the Sudanic region from the early Middle Ages onward carried to the latter a linguistic input characterized by recognizable dialectological features which may be observed to meaningfully describe and unite Arabic varieties of the Egypto-Sudanic zone to this day. There are clear reasons to doubt, however, that three of the four diagnostic features remaining to us represent the products of such a history. The *bitaʕ-type genitive exponents, for example, may be of reasonable antiquity—possibly attested as early as the eleventh century (Lentin 2018
)—yet at the same time show every indication of representing a (Lower) Egyptianism only much later adopted by Arabic speakers of Upper Egypt and the Sudan. In the present sample, reflexes of *bitaʕ exist below Qift only in variation with other, heterogeneous genitive exponents, and are consistently identified by researchers and speakers alike as carrying urban and Egyptian sociolinguistic valuation (for empirical investigation of this sociolinguistic dimension, see Miller and Abu-Manga 1992
; Miller 2005
). These facts, combined with the relative novelty of *bitaʕ forms noted by Hillelson
) and their absence from Nigerian, would support a scenario of spread accompanying the colonial expansion and consolidation of Cairene political influence throughout the region under the Ottoman/Khedival and Anglo-Egyptian state apparatuses (ca. 1820–onward), rather than as part of an original linguistic input carried southward during the first waves of Arabization several centuries earlier.
Certain data likewise complicate the identification of two further syntactic features, post-nominal demonstrative order and in situ WH-question formation, as having arrived to Sudanic territory as part of a founding in-migration of Arabic speakers from Egypt. While post-nominal demonstrative ordering is normative throughout the Egypto-Sudanic region today (as the sampled dialects attest), this is known to not always have been the case. Doss has demonstrated that pre-nominal demonstrative ordering in Egypt long existed as a historical alternative alongside the presently familiar post-nominal, and was “alive and productive” (Doss 1979, p. 356
) in direct historical attestations dating as late as the seventeenth and eighteenth centuries; the pre-nominal structure, in fact, still exists in modern Cairene in a number formulaic usages and fixed expressions, including the grammaticalized dilwaʔti
‘now’ (< *di l-waʔt ‘this time’). Though lacking a pre-modern textual record to provide comparable direct evidence, similar synchronic clues (e.g., Šukriyya and earlier Khartoum daħīn
‘now’ < *dal-ħīn ‘this time’) indicate that exclusively post-nominal demonstrative order has likewise not always been uniform in the Sudanic area (Reichmuth 1983, pp. 122–26
). In this light, the present-day regime of obligatory post-nominal demonstrative ordering becomes a far less viable candidate to have been imported to the Sudanic area from Egypt as part of the latter region’s initial Arabicization—not only because it does not appear to always have existed in Sudanic Arabic varieties, but also because it would not seem to have been so established in Egyptian varieties of the relevant era to begin with.
Direct historical attestation of WH-question formation is unfortunately less forthcoming, but internal reconstruction of the multimorphemic Sudanic interrogative pronouns *šinu and *minu may prove similarly revelatory. In contrast to their Egyptian counterparts of the types *ē(h) (< *ēš) and *mīn, these forms incorporate a reflex of a personal pronoun, which in some varieties still inflects to demonstrate agreement with the interrogated noun phrase. This difference is a critical one, in that it points to a structural dissimilarity in the diachronic source constructions that have given rise to the respective sets of interrogatives. Namely, the presence of the incorporated pronoun in the Sudanic varieties indicates the (historical) presence of a syntactic transformation in WH-questions, by which the noninterrogative element undergoes movement and is resumed by a third person pronoun in its deep-structure position. The following alternation of interrogatives with/without incorporated pronouns in the dialect of the Šukriyya is instructive:
The pronoun-incorporating structure in (18) would, presumably, have originally had its roots in a more complex, cleft-like structure on the order of (19), which has subsequently been subject to syntactic reanalysis/rebracketing:
|19.||*[al-ħaddas-ak]i min [hū]i?|| |
| || rel-told.3msg-you who he|| |
| ||‘He that told you, who is he?’|| |
While sentences like (15), above, make it demonstrably clear that pronoun-incorporating interrogatives in present-day Sudanic varieties do not (or do not necessarily) carry a synchronic clausal interpretation of this type, the diachronic implication of this developmental pathway should not be overlooked. While questions formed in the manner of raʔy-ak ē?
and raʔy-ak šinu?
‘What’s your opinion?’ (Cairo and Khartoum, own knowledge) may both be validly described synchronically as displaying in situ formation, the latter presupposes an earlier cleft structure (*[raʔy-ak]i
‘Your opinion, what is it?’), which in turn presupposes the existence of a once-productive, WH-fronted, pronounless šin
(cf. older Sudanese šin gōl-ak
‘What do you say [lit. What’s your saying]?’; Hillelson 1935, p. 62
). The former does not, and the congruous modern products are thus assigned to two demonstrably incongruous developmental paths.
In the cases of WH-questions and demonstrative order, then, we must heed Pat-El’s warning that “syntactic reconstruction based on cognate patterns may conflate genuine inherited syntactic material with cases of parallel development” (Pat-El 2020, p. 332
)—or, we may add, cases of contact-induced convergence. Either or both of these syntactic patterns may have emerged in dialects of the Egypto-Sudanic area independently, or either or both may be the products of mutually influenced development through centuries of intra-regional contacts. In light of the historical and internally reconstructed data, however, neither appear to have been imported intact from Egypt to the greater Sudan with the onset of Arab settlement.
In terms of common Egypto-Sudanic features identified by this investigation which do in fact support such a narrative, we are subsequently left with a single linguistic trait: a proximal demonstrative paradigm on the model *dā, dī, dōl, dēl. This commonality is a genuinely striking one—being both innovative and distinctive—and demonstrative pronouns are undoubtedly a substantial feature of relevance to any serious attempt at Arabic dialect classification (see Magidow 2013
). Yet, most would agree that they do not, in isolation, provide a viable solitary basis for the formulation of such groupings. This remaining commonality is thereby rendered less proof positive of classificatory relationship and more enigmatic isogloss to be marked for future investigation in light of broader Arabic demonstrative typologies. The traditional Egypto-Sudanic classification of the Arabic dialectology literature, predicated on the nonlinguistic genealogical relatedness and shared migration history of the region’s Arabic-speaking communities, is thus left roundly unsupported following focused linguistic review.
4.2. Whence from Here? An Excursus in Historical Glottometry
Rejection of the traditionally formulated, genealogy-based Egypto-Sudanic dialect classification at a macro-level does not, however, refute or diminish the multifarious and noteworthy dialectal commonalities linking and cross-cutting smaller subsets of Arabic varieties spoken in this region, in varying combinations. These isoglosses, and the linguistic relationships they identify, are real and significant, and merit further study and elaboration—more than can be accomplished in a single contribution, by a single researcher, or, perhaps, via a single perspective on the information at hand. In cases like the present one, in which a long-standing hypothesis has been determined to lack fit, a fresh view on existing data is often as essential, and as conducive to progress, as the gathering of new. Here, one such opportunity (among many) comes in the form of “Historical Glottometry,” a novel approach to linguistic subgrouping recently elaborated by François and Kalyan (François 2014
; Kalyan and François 2018
Historical Glottometry was developed by its creators for application in scenarios in many ways analogous to the Egypto-Sudanic case described heretofore, in which the potential for “tree-like” relationships between once-unitary dialectal entities and “wave-like” patterns of convergence between previously more distinctive groups both loom large, and need both be considered in any comprehensive interpretation of the data. The method accomplishes this by integrating the key dialectological notion of the isogloss with the comparative method’s focus on the common innovation, and labors to produce a diachronically interpretable measure of the relative strengths of multiple potential classificatory units revealed by analysis of a given dataset. Such an approach has been called for previously in the study of Arabic dialects (for a forcefully argued articulation, see Magidow 2017
), and Historical Glottometry in particular has fruitfully filled this role in the examination of Boni dialect linkages (Elias 2019
) and the Sogeram language family (Daniels et al. 2019
), among others. I offer a preliminary application of the method here not as a route to a definitive classificatory model, but instead as an exploratory exercise into new views which may inform future analysis of the Egypto-Sudanic data, failing the identification of a meaningful macro-level relationship based in shared migration history. For example, review of the isoglosses presented in Section 3
offered numerous examples of two-way divisions separating northern dialects of the Egypto-Sudanic area from southern, but the precise positioning of isoglosses within this general pattern was observed to frequently shift on the basis of individual features, and to display a number of variable exponences. Can a technique like Historical Glottometry offer additional, informative perspective which might lead to clarity in the comprehension and description of cases like these?
The tradition of quantitative dialectometry of which Historical Glottometry is a part is not alien to the Arabic dialectological tradition (see Behnstedt and Woidich 2005, pp. 106–35
, for discussion), yet similarly has not been widely embraced by the field’s practitioners—for a host of valid critiques. As an analytical tool, Historical Glottometry joins these approaches in the effort to produce a linguistically meaningful yet condensed mathematical summation of data researchers “already know” (Daniels et al. 2019, p. 124
) but which is copious and complex enough to defy ready intra-set comparability without transformation. Historical Glottometry accomplishes this via the production of two related values, each attending to a different aspect of linguistic classification generally agreed to hold significance in the field: “cohesiveness,” a measure of the proportion of relevant isoglosses held in common by the members of a potential classificatory unit, and “subgroupiness,” a measure of the number of isoglosses unique to the members of a proposed grouping. For a fully elaborated discussion of these measures’ conception and justifications, see Kalyan and François
(2018, pp. 68–71
); to summarize, cohesiveness is calculated as the number of innovative isoglosses shared by all members of a proposed grouping divided by the total number of isoglosses attested by any member of the group, thus taking into account both the quantity of isoglosses supporting a group and those conflicting with it; subgroupiness is derived by multiplying a grouping’s cohesiveness value by the number of exclusively shared isoglosses unique to the members of that group, thereby recognizing the importance of distinctiveness to most models of dialect classification while weighting the value of such features to reflect their position in broader dialectological context.
To apply this approach and calculate cohesiveness and subgroupiness scores for the array of dialect linkages attested by the Egypto-Sudanic data, I have accumulated the combined set of isoglosses considered in Section 3
, focusing on those features which are clearly identifiable as innovative which are attested in a minimum of two varieties, and determined their presence/absence in each of the six dialects sampled. This tabulation of 342 values (6 dialects × 54 isoglosses) is included in Appendix A
. I then calculated cohesiveness and subgroupiness scores for each of the subgroupings attested in the collected data, summarized in Figure 1
, below. Cohesiveness scores are shaded in black, subgroupiness scores in white. Acronyms identify the composition of each classificatory group supported in the data by at least one exclusively shared feature (e.g., CBQK is a group consisting of the dialects of Cairo, Qift, il-Biʿeṛāt and Khartoum; KSN those of Khartoum, the Šukriyya and Nigeria, etc.).
The first and most evident take-away from the Historical Glottometry analysis of the Egypto-Sudanic dialects is that two potential classificatory units stand out as particularly strong and “subgroupy”: these are CQB and KSN—in other words, the three dialects of the Egyptian area taken as a group, and the three of the Sudanic. Not only do these respective sets of varieties share a meaningful proportion of their total features, but they also display a high number of exclusively shared features not identifiable outside the confines of the grouping (9 for each group). This geographical polarization of the dialect region, divided into groups representing the three northernmost and the three southernmost varieties of the sample, is replicated in the four-way groupings that emerge, which, with one (weaker) exception, consist of all three members of CQB or KSN in addition to one member of the other triad—the substantial diminution of both cohesiveness and subgroupiness incurred via such additions, though, reinforces the interpretation of the Egyptian/Sudanic split as a primary faultline in the data, rather than a single stage in a more gradual fading between northern and southern features. Indeed, turning to pairwise relationships, we similarly see that, excepting two linkages involving Nigerian, all other two-dialect groupings attested are internal to the CQB or KSN headings. Despite high cohesion, these are on the whole substantially weaker than either of the three-way groupings in terms of subgroupiness. Even the most significant pairing, KS, emerges as notably less strong than its superordinate KSN. These are key indications that the pan-Egyptian and pan-Sudanic dialect entities KSN and CQB represent are not illusory extracts of a gradated continuum, nor secondary linkages of core plus orbit, but rather demonstrable, classificatorily significant units across which multiple distinctive, innovative features obtain. Historical Glottometry, then, has offered incisive, actionable insight to be further pursued in reshaping understandings of what dialect classifications may succeed the macro-level Egypto-Sudanic hypothesis: a scenario under which an Egyptian and a Sudanic group, though sharing a few broad characteristics and more numerous partially cross-cutting trends, stand out as robustly and independently definable in the absence of overarching linkage.
The exception to this pattern is, as mentioned, Nigerian, which the four-way grouping CQBN shows to pattern more closely to the body of Egyptian varieties than does any other Sudanic dialect sampled; the still stronger pairwise grouping BN shows this affinity to exist more precisely with the dialects of Upper Egypt, particularly those represented by Bʿēri. This finding is notable in light of the well-described linguistic and demographic linkages between Upper Egypt and the Western Sudanic area detailed by Owens
), including isoglosses beyond those considered here and a set of thoroughly sketched population movements from north to south occurring most prominently in the years leading up to 1500. The significance is thus twofold, serving as: (a) corroboration (admittedly circumstantial) of Historical Glottometry’s compatibility with otherwise-derived understandings of the region’s linguistic interrelationships, and (b) a reminder that the impact of migration events and shared genealogical history is not to be ignored in the interpretation of linguistic classificatory relationships. This last, then, underlines the urgency of the question of how a linguistically meaningful Egypto-Sudanic classification at large, girded by similar nonlinguistic factors, could fail to emerge in our broader analysis?