CC BY
13ФИЛОЛОГИЧЕСКИЕ НАУКИ
1x11
PHILOLOGY
УДК 811.411.21
Original Paper Оригинальная статья
MYTH OF CREATION AND CONTINUITY: A CRITICAL VIEW OF THE LINK BETWEEN ARABIC AND SEMITIC
Al-Sharkawi M.
Wayne State University
Received: November 20, 2021 Поступила в редакцию: 20 ноября 2021 г. Reviewed: November 30, 2021 Одобрена рецензентами: 30 ноября 2021 г. Accepted: December 10, 2021 Принята к публикации: 10 декабря 2021 г.
Abstract
This article proposes that the Semitic ancestry of Arabic is neither beneficial to the study of the history of Arabic; nor is it methodologically feasible in light of the current status in so far as the Semitic and Arabic linguistic data is concerned. In order to make the Argument, the article discusses the current status of data on the Semitic languages. Then, the article also introduced structural (from Peninsular dialects) and ecological (demographic and historical) data from the Arabic language in late antiquity. The article concludes that despite the indisputable Semitic origin of the Arabic language, the relationship between different varieties of Arabic and a specific Semitic branch is methodologically vague and does not cast any explanatory light on the history of Arabic.
Key words: Semitic, Arabic, Peninsular dialects, the historical comparative method, evolution.
For citation: Al-Sharkawi M. (2021). Myth of creation and continuity: a critical view of the link between Arabic and Semitic. Eurasian Arabic Studies, 16, 16-43.
МИФ О СОТВОРЕНИИ И ПРЕЕМСТВЕННОСТИ: КРИТИЧЕСКИЙ ВЗГЛЯД НА СВЯЗЬ МЕЖДУ АРАБСКИМ И СЕМИТСКИМИ ЯЗЫКАМИ
Алъ-Шаркави М.
Университет Уэйна
ev7829@wayne. edu
Аннотация
В этой статье выдвигается предположение, что семитское происхождение арабского языка не является ключевым фактором в изучении его истории. Это же касается и методологической точки зрения в свете нынешнего состояния семитских и арабских лингвистических данных. В качестве аргумента в статье обсуждается текущее состояние данных о семитских языках. Далее в статье представлены структурные (из диалектов Аравийского полуострова) и экстралингвистические (демографические и исторические) данные из арабского языка поздней античности. В статье делается вывод о том, что, несмотря на бесспорное семитское происхождение арабского языка, взаимосвязь между различными его вариантами и определенной семитской ветвью является методологически расплывчатой и не объясняет его историческое развитие. Ключевые слова: семитские языки, арабский язык, полуостровные диалекты, исторический сравнительный метод, эволюция
Для цитирования: Аль-Шаркави М. Миф о сотворении и преемственности: критический взгляд на связь между арабским и семитскими языками // Арабистика Евразии. 2021. №16. С. 16-43. (На английском языке)
INTRODUCTION
This essay is an apology for suspending the Semitic historical depth of the Peninsular varieties in late antiquity in the process of studying and writing the history of Arabic.
In this respect, this attempt here is both preliminary in its theoretical framework and quite general in its treatment of the linguistic data.
An analogy is in order to start with here. Imagine you are facing a big and tall multi-trunk tree, with many branch attachments and clear morphological signs of bifurcation. While you know for a fact that the roots to that tree feed the stem, branches, twigs and leaves, it is not feasible in your human sense of vision and even foresight to know which particular root or root area feeds which particular stem, branch, twig and or leaf. You may be acquainted with the general morphology of trees. Indeed, you know that the roots are essential to the growth of the tree, indeed you know they exist under the soil although you do not see them, and you may indeed know approximately how long these roots may theoretically extend buried under the earth. However, it is not yet scientifically possible to make particular sequential connections in the growth of the tree in sight. The modern historian of Arabic is like the beholder in this analogy and their subject matter is the tree. The Semitic languages are like roots buried under the soil. However, in contemplating this analogy, it is important to remember that it is from the perspective of the Arabic language.
That the Arabic language emerged from a Semitic genetic background is a fact supported by the aggregate of its structures and the totality of its lexicon. Also, that the Arabic language was probably a member of the Central Semitic branch of the family in general is a matter of common belief among most of the scholars of Semitic [Huehnegard, 2017, p. 10], and that Ancient North Arabian varieties particularly close to Arabic, though not Arabic, [Macdonald, 2000, p. 29-30] is a statement of the genetic and geographical obvious. These pillars of the field are probably important in the study of the Semitic languages in general and their evolutionary status in particular. Their value, however, and their explanatory power in understanding the evolution of Arabic as a language both from the Semitic parenthood and on its own as of the 6th century is not as salient or satisfactory. I am going to argue in this article in favor of a practical although admittedly potentially controversial opinion among Arab and Western scholars alike, namely that the study of the development of Arabic within the Semitic family must overcome two types of challenges before it begins to be a part of the discussion of Arabic from a historical developmental perspective. In other words, to find a direct Semitic ancestor for Arabic feasible and or useful, two issues need to be addressed. Short of overcoming these hurdles, explaining a Semitic depth for Arabic is a fool's errand. One of these challenges is the methodological linguistic challenge and the other is the non-linguistic ecological prohibition. This
perspective, again, must be seen only as Arabic-centric, and its focus is limited to the evolution of Arabic individually.
The methodological issue referred to in the previous paragraph is twofold. First part, the position of Arabic within the Central Semitic branch is not so comfortable to allow building an ancestry line. It seems that Arabic also shares with the Ethiopic and Ancient South Arabian languages (as members of earlier subgrouping attempts of the Semitic family) some key features. In addition, the features common in Arabic and these two branches are not comprehensive, and some of them may in fact be explained as areal features that may in part have resulted from borrowing. The second part is that comparing Semitic to Arabic seems to be difficult if not impossible. Research in pre-Islamic Arabic is in its infancy, data from the period is not forthcoming. The data available from the books of medieval Arab grammars are not a representative of the pre-Islamic situation. We do not have a structural makeup of Arabic in late antiquity which determines its structural features and contours that we can use for any comparative research. The first aspect renders the genetic affiliation blurry at best and betrays a strong contact situation. The second aspect makes the comparison between any form of Arabic and its Semitic structural affiliates vague and less productive.
These methodological difficulties are espoused to a set of prohibiting ecological factors in Arabia in late antiquity that, I assume, render the reconstruction of an ancestral identity for Arabic through the historical comparative method with any degree of soundness nearly impossible in the current state of research. Although this article does not argue for a causal relationship between adverse ecological reasons and the ambivalence of the position of Arabic in Semitic, there is a correlation between the two elements. The socio-historical situation in Arabia in late antiquity was geared towards contact: human and language contact. Among these adverse ecological factors, I will discuss the contact situation, tribal structure, and the identity of the Arabs as a social group in late antique Arabia. It is my understanding that the combination of these factors renders a direct line of descent of Arabic from Semitic opaque and in deed not useful. These factors made the Arabian Peninsula a language area and a contact zone. While genetic relations may very well have played a role in the shaping of Arabic, this role will remain opaque. For a student of the evolution of Arabic, therefore, an analogy of Arabic as a tall branching tree is an accurate expression of the general evolutionary situation.
A word of caution is in order here. While this article does not make direct claims to any arguments for a historical dating of the beginnings of the Arabic language, naturally it advocates indirectly an artificial date selection, one whose criteria are not
linguistic, but are not discussed here and are not yet widely discussed in the field at large as well. This being the case, however, this article does not argue that the Arabic language cannot or may not permanently be placed in a nexus from a particular Semitic branch. It certainly can. Before scholars make such a claim they must overcome the methodological and ecological issues I will discuss in brief below. At the current state of the research, any such claims remain only logical but without sound admissible linguistic evidentiary basis that can be used to write a history of the Arabic language. Finally, this article being an apology of a sort is not in a position to read critically the scholarship on the Semitic languages. It will merely survey and report on the available research to make the aforementioned point. Two sections and a conclusion will, therefore, follow. In the following section, the methodological issue will be discussed in three paragraphs. In the last section, the ecological prohibition will be introduced. THE POSITION OF ARABIC IN SEMITIC
In this section, I will argue for the pre-maturity of attempting to link Arabic genetically to an ancestral Semitic branch, and the futility, therefore, of establishing a historical ancestral depth beyond the point of mutual intelligibility. I will briefly deal with two main points: one relates to Semitic and the other relates to Arabic. As far as the Semitic languages are concerned, I will highlight two points: the position of Arabic in the Central Semitic sub-branch and the features it shares with other branches in its geographical area through a discussion of case. I hope to show that the position of Arabic in the family is not indicative of or conducive to a direct descent or sub-categorization. The issue of genetic affiliation is fundamentally an issue of data. As far as Arabic is concerned, however, the issue is purely methodological. I will discuss the tendency among historians of Arabic and Semitists alike to include in the discussion data from modern dialects of Arabic. A tendency, I assume, is anachronistic and problematic. 1 ARABIC IN CENTRAL SEMITIC
The genetic affiliation of Arabic with the Central Semitic branch is not yet final or without dispute. Although the persuasive work by Hezron and Huehnergard places Arabic into Central Semitic more firmly than elsewhere in the family, there are strong claims pulling Arabic towards an earlier Southern Semitic language classification [Rubin, 2008, p. 66]. This alternative claim also finds support in non-linguistic geographical and historical ecological factors, which I will discuss in the following section. In any case, the Central Semitic features in Arabic are problematic for our purpose in two respects: one is productivity and the other is the sheer comparative
number of these features. Some of the Central Semitic features in Arabic are summarized here as follows:
1. Geminate verb phonetics reflect a proto-Central Semitic sound rule which metathesized the second root consonant and the theme vowel, thus producing yaruddu instead of yardud [Huehnergard, 2017, p. 15].
2. There was another sound change from w to y after i [Huehnergard, 2017, p. 16].
3. Arabic and some other Central Semitic languages developed a definite article, which did not exist in both proto-Semitic and proto-Central Semitic [Huehnergard, 2017, p. 17].
These three Central Semitic features are shared by all of the Arabic varieties. I consider them, therefore, solid features, which do not require explanation or defense, except probably number 3 above. Not all the Central Semitic languages developed the definite article. There exist some other less solid and problematic Central Semitic features in some of the Arabic varieties and not in others. These are as follows:
4. Taltala is a Central Semitic feature that is common in the overwhelming majority of the Peninsular dialects on the eve of Islam. It was not a feature of the Hijazi dialects. It was therefore not a feature of Classical Arabic [Al-Sharkawi, 2017, p. 84-88].
5. The use in some Peninsular dialects of du/dl as a relative pronoun, which the rest of the dialects and Classical Arabic did not [Al-Sharkawi, 2016, p. 83-84].
These features exist in some early Peninsular dialects, but do not exist in Classical Arabic and/or other contemporary Peninsular dialects. Taltala exists in most of the Peninsular dialects except the Hijazi dialects. The relative pronoun du/dl exists in northern Yemeni dialects and in Tayyi' only. There are also Central Semitic features that exist in Classical Arabic but do not exist in the early Islamic Peninsular dialects. Or, at least, the medieval grammatical literature is mute about them both in token data and in testimonial anecdotal evidence. Some of these features are as follows:
6. Classical Arabic derives its aspectual and mood systems from Central Semitic.
7. Classical Arabic shares with Central Semitic the independent first person common pronoun nahnu 'we'.
In terms of coverage, judging by the Peninsular dialects in the 2nd/8th and 9th centuries, three of the above features exist in the language of the Qur 'an and in the dialects. Four Central Semitic features exist in one or the other but not in both. The taltala in number 4 above is especially interesting. It exists in the Najdi, northeastern and northwestern Peninsular dialects but not in the Hijazi dialects or in Classical Arabic. There is emerging textual data that point towards the Classical language
borrowing its sound system from the prestigious dialects of the Hijaz region. In terms of number, we will see in the following paragraphs that Arabic shares with other branches of the family relatively fewer features, but the numerical difference is not so wide to convince of a solid affiliation. Advocates of tracing Arabic to an earlier ancestor, namely Central Semitic, must explain the discrepancies in features. In addition to a Central Semitic claim, Arabic shares with other branches of the Semitic family some structural features, some more than others. The manner by which Arabic came to share with these other languages some of its structural features is not relevant for our purpose here. It is important to know that these common features cloud further the genetic affiliation. As to be expected, Arabic shares with the Ethiopian languages some features. They are as follows:
1. The change from proto-Semitic p into f.
2. The existence of the fa 'ala form III stem.
3. The use of broken plurals.
Numbers 1 and 3 exist in both the dialects and Classical Arabic, while number 2 exists in Classical Arabic but its status in the Peninsular dialects is unclear. These three isoglosses make the position of Arabic in the Central Semitic branch less secure. Number 2 again brings up the issue of coverage of the features in question in the Arabic spectrum. Yet, classification is not the purpose of these paragraphs. But both the shared innovations that put Arabic in the Central branch and the shared features with other languages (regardless of their origin, be it drift in 1 or retention in 2 and 3) cause the link between Arabic and an earlier branch of the family tree to be opaque.
Finally, Classical Arabic retains ancient features of West Semitic in a productive fashion. These features were not shared by contemporary tribal Peninsular dialects, or the medieval sources were mute about them in the 3rd/10th century. These features are:
1. The use of the yaf'al present tense form in the past tense after the use of the negative particle lam.
2. The original West Semitic tense-less function suffixed form remains in Arabic as an optative and epistemic.
In the case of number 2 above, the medieval sources are mute as to its situation in the dialects. In addition, optative and epistemic functions exist in the modern dialects in different nominal forms. But active participle forms are used to express these functions. As far as number 1 above is concerned, lam yaf'al is not a productive structure in the modern dialects. So far, I mentioned shared innovations and shared retention between Arabic and one or more of the branches of the Semitic family. In
the coming section, I will mention different projections of the same feature in different branches, namely case, in one variety of Arabic, namely Classical Arabic. Since the purpose of this discussion is not the genetic classification of Semitic languages, it is not relevant for our purposes of the history of Arabic whether the features listed above are shared innovations or retentions. They draw a picture of late antique Peninsular varieties that borrow from different sources. The case system shows that some of these sources are chronologically far detached from Arabic. Any attempt to trace the history of Arabic to a Semitic origin will have to account for the following: the typological similarity between Arabic and historically non-contemporary languages (Akkadian), the retention in Arabic of two different case systems (diptotic and triptotic), the existence of features in Classical Arabic and not in the Peninsular varieties and vice versa, and explain whether similarity with Ethiopian and Ancient South Arabian were areal features or shared retentions. The above listed features also beg the question: to what extent is this opaque image the product of the condition of data in Semitic and Arabic? Are there relevant gaps in the available data?
2 THE POSITION OF CASE
Finally, it is beneficial to address one structural feature by way of illustration. Even when we leave the structural generalization of Arabic behind and limit our focus to a particular variety and to a particular structural feature within that variety, heterogeneity of structures still shows and causes the position of Arabic remains vague at best. Case in Qur 'anic Arabic is one good such structure. Arabic shares with Akkadian, Ugaritic, Early Canaanite, and probably Old South Arabian a generally triptotic case system [Hasselbach, 2013, p. 16]. It is a productive system (albeit with a less than productive functional yield) with three cases in the singular and two in the dual and plural nouns. This distribution was productive in Old Akkadian, Old Babylonian and Old Assyrian, in addition to Classical Arabic of course. In the case of bound nouns without pronominal suffixes, Old Akkadian only maintained a distinction between the genitive case and nominative and accusative nouns, which were not marked overtly for case in both the singular and feminine plural. In a few cases, the nominative noun is marked with a final -u short vowel, but the accusative was not marked. However, after the Old Akkadian period, singular construct and feminine plural nouns were not marked for case, while the masculine plural and dual construct nouns correspond to their unbound counterpart nouns [Hasselbach, 2013, p. 17]. Mimation was lost by the end of the Old Babylonian period by the end of Old Babylonaian and Old Assyrian, case on the unbound nouns was gradually lost
[Hasselbach, 2013, p. 17]. Qur'anic Arabic retained case on unbound and construct nouns equally and on duals and plurals as well.
As far as unbound nouns are concerned, Ugaritic shares with Qur'anic Arabic and Akkadian the same triptotic case marking system in the singular and diptotic system in duals and plurals. Like Arabic, it may have preserved case in construct nouns with pronominal suffixes [Tropper, 2000, p. 306-7 and 339]. Similarly, the earliest attestations for Canaanite exhibit a productive case system that is similar to the case systems of Akkadian and Classical Arabic [Hasselbach, 2013, p. 25]. Case in Old South Arabian can be constructed similarly to the previously mentioned languages [Hasselbach, 2013, p. 26-27]. The system that was attested in Arabic for the first time in the pre-Islamic poetry and the Qur 'anic text is echoed by some languages from East Semitic, West (non-Central) Semitic and by South Semitic languages. None of the other Central Semitic languages shares the same or even remotely similar case system. In addition, the case system is shared by these languages whose historical periods differed widely. Old Akkadian was attested until 2000 BCE. The other languages come from the first millennium BCE. The earliest attestation of the system in Arabic was in the beginning of the 1st millennium AD.
Arabic, moreover, shares with individual members of the non-Central Northwest Semitic group particular interesting features of its case system. The language of the Qur'anic text was ambivalent in the use of the accusative case after 'inna and its sisters, but Classical Arabic used it consistently. Similar behavior can be found in Hebrew. Any noun after the particle hinne 'behold' takes an accusative case suffix, as is the case with the negative existential particle 'en and the particle cod. The preposition kd- with pronominal suffixes also takes the accusative marker suffix. Neither of these Arabic and Hebrew particles are inherently transitive [Hasselbach, 2013, p. 46-47]. This phenomenon is common only to Arabic and Hebrew, a language that otherwise does not have a productive case system and one that was not a spoken vernacular at the time of the earliest attestation of case in Arabic. Arabic also exclusively shares with another Northwest Semitic language, Ugaritic, a feature it shares with no other language, namely the diptotic declension of some nominal morphological unbound categories [Hasselbach, 2013, p. 44-45]. I will elaborate on these two features further later in this section. It is enough now to say that both of these features express a great deal of variability between the peninsular dialects and Classical Arabic.
In addition to all this, Classical Arabic behaves differently from the triptotic case languages and similar to the diptotic case languages within the Semitic family in the use of the accusative case after kana and its sisters- verbs of existence. In Classical
Arabic, some Peninsular dialects and Ge'ez, verbs of existence take a predicate in the accusative. The Hijazi dialects differ from the rest of the Peninsular dialects in the treatment of these verbs of existence in a way I will elaborate on later in this section [Hasselbach, 2013, p. 45]. Classical Arabic also uses case in ways that differ from both its triptotic neighbors, diptotic neighbors and even the Peninsular dialects. Some of this behavior is as follows:
1. La of absolute negation does not have an inherent accusative function. It however bestows one on the following noun.
2. The exception particle 'illa when followed by a positive clause gives an accusative case to the following noun. The excepted entity also receives the accusative case when it comes before the noun it refers to or when it belongs to a different category.
3. One more instance of the unique use of the accusative case in Classical Arabic occurs after the vocative particle ya. After this particle, nouns that are in the construct state take the accusative case marking, while unbound nouns after the same vocative come in the nominative case without nunation.
It seems that diptotsis existed in Classical Arabic and the Hijazi dialects as a full and productive system. In the Najdi dialects, the system was equivocal. Medieval grammars furnish us with the following remarks:
1. Examples of diptotic noun categories declined for case regularly in Asad are limited to the adjectival pattern falan whose feminine is fala such as 'atsan/'atsa 'thirsty' [Al-Zagag, ma yansarif, p. 35 and Hasiya, vol. ii, p. 98]. This pattern of nouns receives an -u in the nominative, an -a in the accusative, an -i in the genitive and tanwin when it is indefinite.
2. In TamTm, nouns on the pattern offu al are declined in full for case [Al-Farra', maanT, vol. i, p. 345 and Ibn AqTl, sarh, vol. ii, p. 326].
3. In unidentified dialects, broken plurals were declined for case. This phenomenon is especially frequent in the qira at literature. NafT, 'Asim and al-Kisaiyy read Q76/4 inna a tadna lil-ka+rina salasilan wa- aglalan wasa Iran with tanwin on salasilan and glalan [Al-Farisiyy, Al-Hugga, vol. vii, p. 216]. The same readers also read Q76/15 kanat qawariran with tanwin [Ibn Al-Gazriyy, al-Nasr, vol. ii, p. 395].
4. There is also some data; admittedly few, that indicate that foreign proper noun words were marked for case. Al-'a'mas read Q71/23 as wala yagutan wa ya uqan with the masculine singular indefinite accusative tanwin [Ibn Xalawayhi, al- Hugga, p. 162].
5. Some nouns ending in the long -a are declined in full for case and are not dealt with as mamnuc min al-sarf. Slbawayhi [Al-Kitab, vol. iii, p. 211 and ma yansarif, p. 29] discusses difra, and makes the claim that some Arabs that he did not identify give it tanwin when it is indefinite. Again, Slbawayhi does not ascribe this feature to any particular tribe. But Al-Zagag makes the claim that most of the Arabs deal with such nouns as mamnuc min al-sarf. The same goes for tatra, which very few of the Arabs decline for case [ma yansarif, p. 28 and 'i rab al-Quran, vol. iii, p. 114]. Most of the Arabs do not decline it for case [ma ani al-Qur an, vol. ii, p. 236].
6. Nouns ending in a long alif followed by a hamza are also dealt with in some unidentified pre-Islamic dialects as fully declined. gawga' 'mob' is one of those words that Slbawayhi claims are sometimes mamnuc min al-sarf and some other times not [Al-Kitab, vol. iii, p. 215 and ma yansarif,, p. 34].
7. Some foreign place names are declined fully for case. Slbawayhi claims that most of the Arabs decline place names such as Qiba \ Hagar and Wasit [ma yansarif, p. 54].
8. Slbawayhi makes the claim that nouns referring to time units such as bukra 'tomorrow' and gudwa 'tomorrow' are dealt with in two ways. The majority of the Arabs consider these nouns mamnu c min al-sarf, while few an Arab tribal dialects decline it in full [al-Kitab, vol. iii, p. 293-294].
Interestingly, in the cases where Classical Arabic is unique in its use of the accusative case, the dialects differ from one another. In Hijaz, after the alleviated 'in and 'an, the subject stands in an accusative case, while in the pre-Classical and Classical language and in the Najdi dialects, alleviated particles lost their effect on the following nominal clause.
After 'inna and its sisters, Hijazis put the subject and predicate of the sentence in the accusative case. Ibn Hisam [mugniyy al-labib, Vol. I, p. 35] explains the agreement between the subject and predicate in a nominal sentence after 'inna, following the structure of the saying of the prophet 'inna qa'a gahannama sab'ina xarif-an, asserting that Hijaz did not distinguish between the subject and predicate in case endings after 'inna.
The predicate of the nominal sentence after kana and its sisters was given a nominative case in Hijaz, while an accusative case is assigned to it in pre-Classical Arabic.
In Hijaz, ma, la, and 'in had the same effect as the pre-Classical laysa in assigning to the subject the nominative case, and to the predicate the accusative case.
To conclude this section, case in Arabic fits typologically non-Central Semitic and ancient languages. So far, Classical and Peninsular varieties agree. However, Classical Arabic shares with Ethiopian and Ancient South Arabian languages diptotic nominal case marking and anomalous use of the accusative case. Here, the varieties that we take in aggregate to be Arabic differ. Where Classical Arabic has productive diptotic nominal categories, the Najdi dialects vary extensively. The sources are mute as to the status of the diptotes in the Hijazi dialects. Where Classical Arabic uses the accusative case marker uniquely the Hijazi dialect differs. The sources are mute as to the distribution of these instances in the Najdi dialects. To draw an evolutionary route for the case system in particular, and to the Arabic language in general, we need not only to explain this heterogeneity, but also to determine whether this heterogeneity is a byproduct of a potential gap in the data, or the effect of contact induced change, or a plurality of sources. Until then, a linear chain of texts from modern Arabic to an ancient Semitic source as a motif is difficult to establish. The reader will also note that I have not discussed the status of the inscriptional data commonly known as Ancient North Arabian. It is my assumption that linearity as a historical assumption is substantially weakened by the status of the data in medieval sources. 3 MODERN ARABIC
Despite the fact that modern dialects of Arabic by default offer qualitatively and quantitatively better data than Peninsular varieties from late antiquity, and despite the fact that these varieties are undoubtedly perceptually and structurally Arabic, using them towards a generalization about the history of Arabic or about the relationship between Arabic and Semitic is methodologically problematic. Two points are worth mentioning here: one is data related while the other one is process related. First, the lack, ambivalence, variability, and de-contextualization of the Arabic data in late antiquity that can be used as baseline data for a comparative study between modern varieties and older forms of Arabic. Structural development, is, therefore, difficult to assess. Second, the processes by which Arabic became the vernacular language of the Middle East and north Africa after the diaspora in the middle of the 1 st/7th century are not yet well known to us. Although there are some attempts to explain the language shift to Arabic, they remain partial, speculative and their explanatory power not quite inclusive. Linearity is opaque at best.
I start here with the issue of data. Owens [2003, p. 715-20] refers to one part of this issue when he stated that any reconstruction based on the modern dialectal forms must be checked against the oldest citation of the same form in Arabic. Tracing modern forms to their older, let alone the oldest, distributions is problematic. There is no source of dialectal data that can reflect in any degree of clarity the structural shape
of that particular dialect at an earlier period. Taking Middle Arabic texts in some parts of the Arabic-speaking world as data mines provides a post hoc interpretation. Middle Arabic also faces the question: is the feature in question a dialectal feature, a dialectal rendition of the Classical Arabic form, interlanguage, or a celebrate use of a dialect form? There is of course also the pre-Classical and Classical Arabic. The relationship between Classical Arabic and the Peninsular varieties is not yet conceptualized. Early scholars of Arabic up to the time of al-Farra' considered variation only, and did so to explain and/or apologize for the language of the Qur 'an. From the 4th/10th century Classical Arabic became almost synonymous with the spoken Peninsular vernaculars. Any variation was considered merely mannerisms. From the times of Ibn Faris, historical depth of Arabic meant the Classical form and variation meant contemporary linguistic inadequacy on the part of the user. If we assume, and I certainly do, that Classical Arabic is not the oldest form of Arabic but a later development than older less known varieties betrayed by variable forms littering the medieval Arabic grammatical sources, there follows a need to reconstruct these earlier varieties in order to connect Arabic to its Semitic ancestry. Here, the issue of the linguistic data in the medieval grammatical sources warrants a closer look at the allocation of tokens and consistency of the token allocation. Arab grammarians introduce data in ways that generally do not render reconstruction to an earlier form feasible. In some cases, the sources are totally silent on a particular phenomenon or on the existence of a phenomenon in a particular tribe. There is, for instance, no reference in the grammatical literature for the declension of the dual and sound plural suffixes for case in the Peninsular dialects. All direct data on the declension of dual and plural suffix comes from the language of the Qur 'an and pre-Islamic poetry, which can indeed be an indirect indication for the existence of a diptotic declension of a sort. However, the system remains inaccessible. Judging by the example in the previous section, the diptotic system in the Najdi dialects seems to have been in development in the 8th and 9th centuries. The same is true in the production of the internal passive. In fact, the literature is mute about all linguistic phenomena in some tribal dialects. We, for instance, do not know any structural features of the 'Azd of Oman and 'Abdul-Qays of Bahrain.
In some cases, a particular feature is ascribed in the literature to only one tribe or a region. Case marking, as we saw above, after kana and its sisters and after 'inna and its sisters behaves in Hijaz in a way that differs from the Classical language, with no information about the rest of the dialects. There is in the medieval literature also 'an'anat tamim, 'agrafiyat qays and kaskasat 'sad [Ibn Faris: al-sahibiyy, p. 29]. Each of these phenomena is ascribed to one group only. In some cases, the allocation
is to a particular group but tokens are found in other groups as well. A good example of this ambivalence is the taltala phenomenon, which all grammarians from the time of Ibn GinnT onwards ascribe to Bahra', although there are numerous examples from all over the Arabian Peninsula. In addition to the logical questions of productivity that differential allocation poses, there is the question of originality. Let us consider the taltala phenomenon as an example:
Taltala in STbawayhi's (148/768-180/796) Kitab is not defined in a hadd, but merely introduced by example and instructions as follows:
Wa-dalika qawluhum 'anta ti- 'lam dalika wa- 'ana 'i- 'lam wa-hiyya ti- 'lam wa-nahnu ni- 'lam dalika [al-Kitab, IV, p. 110].
STbawayhi [al-Kitab, IV, p. 110] further makes the claim that the phenomenon is shared among all the tribal dialects of the Peninsula except those of the Hijaz region without naming members of each group individually, which may indicate that it was an areal phenomenon more than merely a tribal one:
Wa-dalika fi luga gami '-l- 'arab 'illa 'ahl-l-higaz
And that takes place in the dialect of all the Arabs except the people of Hijaz.
A century and a half later, Ibn GinnT (322/941-392/1002) [al-Xasa'is, vol. II, p. 11 and Sirr, vol. I, p. 230] similarly defines the phenomenon by examples. However, he differs from his predecessor as he remarkably ascribes its use only to Bahra' of Quda'a in northwest Arabia of all the Arabs. Few centuries later still, Ibn Manzur (630/1233-711/1311) defines taltala in the same way as follows:
Wa-taltalat bahra' kasruhum ta' taf'aluna
He also ascribes it only to Bahra' among all the Arab. Bahra' was a small clan in northwest Arabia that belonged to the larger 'Adnani Quda'a tribe [Gamhara, p. 412413]. The data prayer to the 4th/11th century starting with the vague and sweeping statement of STbawayhi in the following section show a wider tribal use of taltala. Between the 2nd/9th and the 4th/10th then, the use of the phenomenon presumably in the production of poetry and the Qur'anic text was diminished considerably. Bearing in mind the prevalence of the phenomenon in the modern
dialects of Arabic in northwest Arabia, Iraq, and the non-Peninsular modern dialects, I feel justified in limiting the previous discrepancy in the phenomenon to the production of poetry and the Qur 'an.
In addition to productivity issues, the previous discussion of taltala begs another question, that of historical depth. Is the difference between the definitions of STbawayhi and Ibn GinnT a difference of historical development? Did the phenomenon recede to one tribe only between the 2nd and the 4th centuries? Or, was it simply ascribed to Bahra' for titular reasons since the phenomenon fell to social disfavor in the 4th century? There is, in addition, one more question of domain these tokens pose. I will again use the taltala example cited above:
In the Qur'anic data, taltala shows several interesting characteristics: a) The data confirms the taltala realizing morphological categories mentioned by grammarians. Every token is attested in the sad readings, b) The total number of tokens is 8, and c) All the eight tokens share the same two authorities as the source of their chain of transmission, Yahya Ibn Watab (d. 103/721) and al-'A'mas (61/681-147/764). Two points are worthy of note here. First, the Qur 'anic data, judging by the lives of
the readers, extends from the second half of the 1st/7th to the first half of the 2nd/8th century, approximately a half century before the career of STbawayhi. This indicates that the structural scope of the taltala phenomenon was wider than the data from STbawayhi onwards. The second observation is that both Yahya Ibn Watab and al-'a'mas did not belong to Bahra', but rather to 'Asad and lived in Kufa. We also know from the biography of Yahya Ibn Watab in the tabaqat books that he has been mentored and guideline the art of Qur 'anic recitation by many authorities who were themselves among 'ahli-l- higaz who migrated to this metropolitan city such as Ibn Mas'ud's (d. 32/654) students [Siyar, vol. I, p. 462 and vol. IV, p. 380-381]. The use of taltala was therefore acceptable in the recitation of the Holy Book. In addition, TamTmi readers were tagged with the use of taltala.
The purpose of these questions is in fact not to discount the medieval sources. But until some of these questions are addressed, using medieval data sources in a reconstruction of a proto-Arabic remains methodologically questionable. THE ECOLOGY OF LATE ANTIQUITY
The previously discussed opaque linguistic picture of the Arabic language's genetic affiliation and lineage in late antiquity coincides with, or is even caused by, equally adverse socio-political ecologies. Ecological features, I will argue, make it difficult to accept any linear connection between the pre-Classical variety and any Semitic ancestry as solid. These factors are conducive to contact induced innovation and language change; which processes render structural affiliation opaque. In this section
I will briefly introduce the adverse demographic (tribal structure and conditions), the political ecology (alliances and power shifts) and linguistic and literacy plurality. I hope to convey at the end of this section that non-linguistic ecological factors indicate a language situation that was more conducive to a mix of Semitic languages in the formation of Arabic rather than a direct heritage. It is important to reiterate here that I will not delve into these issues in great detail for space reasons. I will refer the reader to relevant sources, however. 1 DEMOGRAPHIC ECOLOGY
As far as the demographic ecology is concerned, I will focus on the definition of Arab-ness (who was an Arab?), tribal size and structure, flexibility and mobility. I assume that these factors further emphasize the opaqueness of the linguistic situation of pre-Islamic Arabia. The definition of 'Arab' in medieval Arabic sources evolved from the 2nd/8th century to the 6th/12th century, from an identity based on clarity of speech to one based on descendance [Ulrich, 2019, p. 1-2]. In any case, it did not refer to the same social groups that lived on the Peninsula and were referred to by earlier Greek, Roman and other sources. Both sources are not clear as to the designation.
Late antique Greek sources used the term Sarakenoi and Skematai 'dwellers of tents'. The term Saracen appears first time in the geography of Ptlolemy in the 2nd century of the common Era. The term comes from the Ancient North Arabian root s-r-q and the word sarq 'east'. The word was used in the Greek sources to designate seasonal movement in the desert area. In Syriac the term used for the inhabitants of the Arabian Peninsula was Tayyaye. The word 'Arab' was almost nonexistent in these sources [Al-Azmeh, 2014, p. 104]. These designations may in fact have been a reflection of the social groups these Greeks and Syriacs came in contact with in the 2nd century. Arabs, on the other hand, were probably the inhabitants of the Roman province of Arabia. Arabs, in these ancient writings were described as savages and almost necked. At the same time, they were depicted in the Greek literary sources as merchants [Al-Azmeh, 2014, p. 104-105]. In addition, South Arabian, North Arabian and Syriac sources used another interesting designation, that of Ma'add/Ma'addaye/Maddenoi. This term does not designate distinct social groups that we can identify on any ethnic basis. It rather refers to a wide range of groups who lived in Central and West Central Arabia, and who did not fall under any organized political or state authority [Zwettler, 2000, p. 254-256]. These were generally camel breeders. It seems, therefore, that among the inhabitants of the Peninsula themselves, there was a distinction between those who were rule-followers and those who were Ma 'add.
Those Arabs who were not Ma'add were most probably organized not according to a bloodline or otherwise social factor, but rather made into groups by the sheer force of an external political power. The Tamud in the 2nd and 3rd centuries of the Common Era, for example, were designated as a srkt, a designation that only indicated a gathering in relation to the Roman empire [Al-Azmeh, 2014, p. 109-110]. After the
rc\
end of Palmyran and Harran authority in the 3 century, Arab groups that were governed by proxy before gravitated directly towards the two major empires and in turn formed client principalities and established their own networks and internal political affiliations. The names that these affiliations adopted since their inception became the genealogical titles given to them as tribes later by medieval Arab scholars. But they were all designated in the late antique Roman sources as Saracens. These newly formed groupings took continuous spatial patterns that are not determined by geography only, but also by trade routes and military force. These units, therefore, were under the supervision of the respective empires [Al-Azmeh, 2014, p. 114-118]. Trade and security changes may have contributed to the instability of these alliances that can be seen in Arab folklore after the emergence of Islam. What gives these alliances the semblance of tribe is that affiliation usually happened between ruling elites who were generally from a simple line of descendants. In late antiquity, therefore, there was no textual designation of Arab except in reference to the inhabitants of the Roman Arabia colony. There, however, were groups in the interior of the Peninsula of unregulated and probably unruly ma 'add and srkt. The duty of the latter, among other things, was probably to control the earlier. The Designation of Arab appears in the medieval sources. There, it is vague at best. Medieval Arabic sources, genealogical, linguistic, literary and geographical, are equally unsatisfactory. Not all the Semitic tribes/groups in the Arabian Peninsula were designated Arabs. The term covered less of them than it did in the 1st/7th and 2nd/8th centuries. Some of these Semitic groups were given different titles. At least in the eyes of later genealogists, not all of these social groups were equal in social status. We will talk here about the Arabs as they were depicted in medieval native sources immediately before Islam and immediately thereafter. In addition to the unspecific referent of the designation "Arab", it seems that the inhabitants of Arabia before the emergence of Islam did not see themselves as members of the same overarching nation or social group [Al-Sharkawi, 2017, p. 4]. It also seems that historical depth did not furnish clarity of designation. It seems that the concept developed with time and the change of the political situation. Ibn Xaldun (723/1332-808/1406) defines the Arabs as a social group according to a combination of criteria, none of which is a supposed ethnicity, bloodline or even linguistic
homogeneity. He classifies al- 'Arab as a subcategory of a larger social group, the badw, This larger group in turn is juxtaposed to the hadar, 'urban people.' The overarching category of the badw, 'Bedouin,' is defined as everybody living outside the walled and gated cities and urban locations in tents and/or in villages. The badw may, therefore, be nomads or farmers or even cattle breeders. Among the badw, al-'Arab are the most primitive as far as their style of life is concerned and pure in lineage due to their harsh lifestyle. Arabs intermarried only with Arabs. In addition to the pure lineage criterion, the Arabs are among the cattle breeding badw, those also who breed camels [Retso, 2003, p. 20-21].
The Arabs were not always so defined. Few centuries before Ibn Xaldun, al-Gahiz describes the Arabs, in one of his shorter prose essays, Risala 'an fada'il al-Turk, 'a treatise on the advantages of the Turks'. He diminishes the importance of the purity of lineage and stressed the lifestyle aspect of the definition. He describes the group of al- 'Arab as one that was not originally descending from one single ancestral family or bloodline. Rather, it was to him the product of a process of cross-breeding and intermarriages among different extended families that lived in close territorial proximity, under the same geo-political circumstances and largely similar styles of life. The result of this process of mixture was a new breed of people who claim an identity based on kinship [Al-Gahiz, Rasa'il, p. 10-12]. Although al-Gahiz does not state it verbally, it seems that the creation of a social group through marriage met coincidentally with the rise of Arabic as a language in the collective consciousness of this social group.
There was in the medieval sources a distinction between two social groups on the Peninsula, which allows us to think that not all who lived on the Peninsula were Arabs. This distinction was also based on lifestyle, namely the distinction between al-'Arab and al- 'a 'rab. This distinction is important for the Arabian Peninsula in late antiquity and after Islam. Both terms are derived from the same lexical root '-r-b. The latter term is not only distinct from the earlier based on the style of life, but it is also socially and emotionally more complex than the earlier. In later lexicographical sources, the 'a'rab are a subcategory of the people of al-'Arab. But we have to emphasize here an important point, namely that this classification is not prevalent in earlier sources. The 'a 'rab, in the late medieval lexical sources, were the people who settled in the badiya, 'desert,' or joined those who live in it moving with them and living like them (Lisan, '-r-b). This meaning puts the 'a'rab in the category of nomads. According to the definition just given from the thirteenth century, one can make the claim that the 'a 'rab were merely the nomadic Arabs who lived on the
fringes of a settlement and moved between settlements. It seems that the term carries a derogatory sense in addition to its classificatory one in the earlier imagination. In the same lexical entry, Lisan al- 'Arab claims that the 'a 'rab would rejoice if they were called 'arab, indicating a prestigious status for the Arabs and a since of inclusion among the 'a 'rab in this exclusive club. The opposite is also true; al- 'Arab would feel insulted if they were to be designated as 'a'rab, indicating a less prestigious status for the latter. This prestige differential attitude is probably the result of a difference in acquired status, with the tribesman being superior to the nomad due to particular historical events and/or social stigma in early Islamic times. The difference in prestige does not only come from the status of the tribes, but also from early Islamic religious attitudes.
The word 'arab is mentioned few times and vaguely in the sources contemporary to the tenure of the Prophet Muhammad. However, some characteristics of this social group can be harvested easily from antique and late antique sources [Retso, 2010, p. 285]. Through these early writings we can see that the Arabs did not drink wine, did not live in stone houses, shaved their foreheads and worshipped or venerated two deities that were merely a medium to the almighty God Almighty of the Abrahamic tradition [Retso, 2010, p. 286]. The statement of the Q41:44 verse indicates clearly that the divine message was sent to those groups who have been exposed to the belief before in some form or another. To the experienced, it is a message of healing and guidance. Revelation does not come to other groups who did not witness revelation before because they are not receptive.
There is a lot of supporting evidence for the idea that the 'a 'rab were the nomads among the people who live outside gated cities in general, and are not an exclusive group of nomads living around MadTna or just a religiously condemned group of deviant people. It is unfortunately not possible to equate them with the Ma'add of the late antique sources, although this remains a possibility to be studied. Al-Zamaxsariyy [Al-Kassaf, Vol. II, p. 300] lists the names of some places in Hijaz and away from MadTna in which 'a 'rab can be found. In fact, we can find in Mecca a demographic situation similar to that in MadTna, where the tribesmen lived inside and the 'a 'rab occupied the outside unplanned and unbuilt open space (Lisan, d-h-y). The same spatial and demographic arrangements of 'a 'rab around other settled communities are mentioned outside the Western region of the Arabian Peninsula all together as well. In Eastern Arabia also, they were mentioned as early as the fourth century, and in Yemen, they appear in historical Arabic and non-Arabic sources in the sixth century [Retso, 2003, p. 92]. It is important to note that in all these contexts there is no negative religious or political connotation to the term. There is, however,
an element of prestige either in the case of the 'a'rab around Mecca, where the inhabitants of the settlement are described in Lisan to be more prestigious than the inhabitants of the open lowland deserts outside it.
The term 'arab, the early Arab lexicographical and general sources, on the other hand, seems to have been demographically less inclusive. It excluded the 'a 'rab; they were not dealt with as a subcategory. The two terms denoting the two social groups were sharply differentiated in these early works such as the fourth/tenth century al-'Iqd al-Farid, where Ibn 'abd Rabbihi, the author, (d. 328/939) lists the features, history and stories of the Arabs in several chapters and dedicates a single independent chapter for the 'a'rab. The early ideological differentiation may probably have been the result of a generally negative religious reputation for the 'a 'rab in early Islam [Retso, 2003, p. 82-86]. Al- 'a 'rab are mentioned several times in the Qur'an in four different suras 'chapters': al-'Ahzab, 'The Confederates,' Q/XXXIII-20, al-Fath, 'Victory,' Q/XLVIII-11 and 16, al-Hugrat, 'Chambers,' Q/XLIX-14 and al-Tawba, 'Repentance,' Q/IX-90, 97-99, 101 and 120. In Q/XXXIII-20 Al-'a'rab are depicted as a group of people living in the desert surrounding the al-MadTna and were neutral insofar as taking sides in the war between Muslims and the confederates in the battle of the Trench. The verse indicates that this is a negative neutrality, as they do not side with the true and righteous side of the Muslims.
In Q/XLVIII-11 and IX-90, al-'a'rab are described as those who do not stand up to their oath of allegiance to the prophet and support him against his enemies An oath with the messenger of God is effectively an oath with God himself. Therefore, breaking an oath with the prophet is like breaking it with God himself. But unlike Q/IX-97-98 which dooms them forever as most disbelieving and hypocritical, verse 16 of the same Q/XLVIII sura the same meaning of al-'a'rab as breakers of oath and allegiance is stressed, but the verse also does not declare them as doomed forever, as it gives them an ultimatum. If they obey God in the future, they will get the support and reward of God, and if they do not, their enemy will defeat them. In Q/XLIX- 14 the prophet is commanded to correct the assumption of the al-'a'rab about themselves that they are true believers, as true believers are the ones whose hearts accept belief in God and his messenger and who strive to fight for this belief with their property and even their souls. They are, rather, merely Muslims. In Q-IX-99 and 101 the picture is more interesting. The earlier verse declares that among the 'a 'rab there are those who believe in God and the last Day of Judgment on the level of the belief. On the level of work, they do not consider that which they spend for God to be a costly fine, but rather a merit to themselves. In the latter, the Holy Book declares that the
'a'rab are not the only group of people (it did not specify whether the verdict is open or limited to the peninsula) who include unbelievers and hypocrites. It is understandable that the negative connotation for the term from their adverse image in the Holy Book should force early Arab scholars of genealogy to dismisthe notion that the 'a'rab are a subcategory of the 'arab. Late lexicographical sources were probably not under the same moral pressure, and therefore, categorized 'a'rab under 'arab. However, the importance of this categorization comes not from the stigma of the Holy Book, but from the fact that the 'a 'rab were not tribesmen of known genealogy. But they lived outside gated and walled cities. This status must have affected their value as sources of linguistic data and informants from the point of view of the medieval Arab grammarians. 2 THE TRIBE IN CONCEPT
QabTla, 'tribe,' is the commonly believed socio-political unit of pre-Islamic/early Islamic peninsular Arab society. Although the term is technical in the medieval literature on genealogy, its popular and scholarly use alike was ambivalent. It was almost always generalized to include smaller subcategories. According to al-Qalqasandiyy (756/1355-821/1418) the Arab scholars categorized al-'Ansab (sing. nasab), 'genealogies,' into six main descending divisions. The overarching category is Sa'b, 'people.' The inhabitants of the Peninsula generally descended from two main distinct peoples: the Qahtan and 'Adnan, Qahtan being al- 'Arab al- 'Ariba, 'the real Arabs', whose origin is Yemen [Nihayat al- 'Arab, p. 211] and 'Adnan being al-'Arab al-Musta'riba, 'the Arabized Arabs' [Al-'A 'laq al-Nafisa, p. 59]. Each sa'b is branches further into two qaba'il (sing. Qabila), 'complex tribe.' The Qahtan begets Himyar and Kahlan. From Kahlan descends Tayyi' and Kalb [Subh, Vol. I, p. 315]. 'Adnan begets RabT'a and Mudar [al-'A 'laq al-Nafisa, p. 59]. Al-Hamadaniyy states that to RabT'a belongs Asad [Sifa; p. 171], Qays comes from Mudar [Mu'gam ma Ista 'gam, Vol. I, p. 87].
These names are technically designated 'imara, 'structure.' The 'imara as a technical term is often ignored by the Arabs, scholars and laypeople alike, in favor of the broader term qabila. This use of a broader term to designate the subcategory can suggest a diversification of the clans a single tribe has and an expansion of their numbers. A 'Imara further begets batn, 'belly,' and faxid, 'thigh', respectively. Al-Azd is one 'imara of Kahlan. One batn in it was gassan. This batn was spread all over western Arabia, both in the northeast and in Hijaz proper, namely in MadTna [Subh, Vol. I, p. 319-320]. The smallest tribal subdivision was a fasila, 'group.' Technically the Aws and Xazrag of MadTna were each a fasila of gassan. We know from medieval genealogical sources that these two social groups were called tribes,
which can be an indication of their large size. Another interesting indication not only of the size a social entity can be, but also of the widespread moving potential it can theoretically have is Azd. Some of the Azd clans were Bedouin and lived to the northwest of Yemen [Al-Sharkawi, 2010, p. 45]. Parts of were urban, such as the Aws and Xazrag of MadTna. Not all tribes were that large or diversified in lifestyle. It is fair to note here that this neat subcategorization of the Inhabitants of Arabia in late antiquity does not match the linguistic and epistemological evidence we discussed earlier. Which casts doubt on the genealogical data of the medieval Arabs as reconstruction and rearrangements. Nasab, as a discipline, was first formulated in late Ummayad times. It emerged to organize and charter a huge set of raw data of families and lineages. The emergent picture of the Arab grid was in fact not known to the pre-Islamic inhabitant of the Peninsula [Szombathy, 2003, p. 12-13]. Therefore, a sense of the network of the tribe as a social unit and its relative power and standing was an alien construct to pre-Islamic Arabia. The grammarian linking of structural linguistic features to tribal units may have worked at their time. Assuming the same linkage for the pre-Islamic context, assuming the stability and continuity of the social unit is a methodological failure. In addition to these anthropological considerations, there are methodological issues that render the concept of a tribe and its anecdotal and literary sources doubtful. On top of the copious and confusing style of the genealogical writings Lecker [2014, p. 153-156] introduces some socio-political issues that cast doubt on the accuracy of the data if not also on its veracity. The Peninsula hosted nomadic, semi-nomadic and settled clans. The settled ones had similar genealogies as these of the others. They are either 'northern' or 'southern.' Genealogy also defined and recorded their links with other tribes within clusters or federations. Qurays, for example, was member of Kinana, and other members were, therefore, its closest relatives. The settled tribes, did not receive their due share of scholarly attention, because it was given to the nomadic clans, more precisely on their military activities, no matter how insignificant. Tribal genealogists focused on the military achievement since the performance of town dwellers in the realms of trade and agriculture were less spectacular, and hence less contributive to tribal solidarity. The livelihood of the social group depended less on the economic activity than on the military force defending it. Despite the general changes that betook Peninsular tribes in the 1st/7th century, they kept their rich oral lore which included genealogy. The political influence of the tribe may have created a proportionate echo chamber of that lore. The coverage of individual tribes was uneven, as it relied on their activity in support or in defiance of the project of Islam in its formative years. Tribes such as Gifaar, Muzayna, Guhayna surrounding around Mecca and Medina are better known
to us than stronger tribes as Asad and Gassan [Lecker, 2014, p. 153-154]. Disproportionate attention led to a distorted perception of the pre-Islamic tribal society and its lineages. For the purpose of writing a history of Arabic, genealogical data, based on the above, is both defective and partial at best. 3 TRIBAL MOVEMENT
Between the end of the 2nd and the 6th centuries of the Common Era, Peninsular tribal names and collective designations, that were written in Epigraphic South Arabian inscriptions, can largely be identified with the names of major tribes or confederations featured in the ayya^-literature and genealogical studies later in early Muslim times. In contrast, very few of the tribes named in the Epigraphic North Arabian graffiti and inscriptions can be linked with any degree of certainty to a correspondingly named group in the Arab-Islamic historical and/or genealogical traditions. A number of the groups tagged as Arab recorded in Epigraphic South Arabian texts in southern/central Arabia appear later in north Arabia and Syria/Iraq as though they have migrated to this latter region. Ghassan lived in central Arabia in the period 260-360. But then they defeat the SalTh tribe to become the chief allies of Rome in Syria by the sixth century. So it seems that these 'Arab' groups that entered into imperial service in Syro-Mesopotamia in the fourth-sixth centuries were coming from southern/central Arabia [Hoyland, 2009, p. 384-385]. There must have been more frequent movements of smaller groups within the arid areas of the Syro-Arabian area for a lot of different potential reasons, such as pasture, water, trade, booty, employment [Hoyland, 2009, p. 387].
Tribal territorial domains were generally, albeit loosely, acknowledged. Tribesmen were aware at which point they may have left their tribe's territories. But just like nasab, tribal boundaries fluctuated in response to changing actual political or environmental circumstances. A tribe's territory often included enclaves or corridors inhabited by members of other tribes, which imposed constant contact and cohabitation between the tribes involved. In addition, the genealogical diversity of the settled populations may very well have been bigger than that of the nomadic or semi-nomadic tribes. This was the case of the Christian tribal groups living in northeast Arabia. They were labeled collectively al-'Ibad, although they each preserved their original tribal affiliations. The same phenomenon also existed inside the Peninsula, namely in Hijaz. The towns of Mecca and MadTna contained diverse multiple tribal affiliations who lived in close proximity [Lecker, 2014, p. 157]. Even larger Nomadic and semi-nomadic tribal bodies such as TamTm did not live in isolation. Their livelihood depended to a large extent on the surrounding urban and semi-urban settlements in northeast and eastern Arabia. Immediately before Islam,
TamTm frequented Hagar in northeast Arabia for supplies and other economic activities. Hagar was the largest date producing center in this part of the Peninsula. Co-dependence was not a one-way affair in late antiquity. Settled tribes also relied on nomadic clans for economic duties and logistics. For instance, wh en Banu an-NardTr were expelled from their quarters in Medina in the early years of Islam, they hired a large number of camels from the nomad clans around MadTna. These camels were readily available for immediate hire because the livelihood of these nomads was based on transporting goods for and among the different MadTna settlements, especially produce and harvests [Lecker, 2014, p. 159]. However, the most important aspect of codependence between settled and nomadic tribes was military support. In several exchanges between the Prophet and some of his contemporaries, we understand that each of the two groups provides support for the other in need [Ibn Sallam, Kitab al- 'amwal, p. 280].
In addition to the codependence of settled and nomadic Peninsular Arabs, settled groups also depended on one another and were, therefore, in constant contact. There is historical evidence that the wealthy merchants of Qurays owned farm lands and orchards in at-Ta'if and neighboring areas, because of the adverse natural circumstances in Mecca. The main consumers of the products of these farm lands and orchards were other settled areas. Instrumental in the transportation of the produce were the surrounding nomadic clans [Al-WaqidT, Magazi, vol. I, p. 16]. To conclude this section, it is important to remember that I just put forward topic heads for discussion. It is also important to repeat that recent historical research may shed more light on the position of the Peninsula and its inhabitants in late antiquity. We do not know the exact identity and/or the location of the Arabs within the Peninsula in late antiquity. Certainly, they were not the only inhabitants of the area during the tenure of the Prophet. Moreover, the social structure and the demographic distribution of the peninsula was not conducive to the separate development of independent tribal groupings, let alone separate linguistic varieties. Contact was probably the norm of the Peninsular life. These non-linguistic ecological points correspond very well with the mixture of structural features among Peninsular varieties and within the same variety. They, however, do not work well with a process of reconstruction. Any assumption of historical depth is, based on all of the above, methodologically unacceptable. A FINAL WORD
To understand the development of Arabic, establishing a historical depth or ancestral lineage beyond mutual intelligibility is not only a futile endeavor. But it is also one that is fraught with methodological and epistemological difficulty. It is important to
repeat here that this purposeful attitude towards the writing of history is centered around Arabic. The issue is not if Arabic is a Semitic language. The issue rather is if connecting Arabic to a particular Semitic branch is relevant to the study of its development or even feasible. It is, therefore, a part of the duty of the historian of Arabic to determine a historical and structural point of departure for their historical research. Earlier structural layers become in light of this quest only relevant as they can shed light on the period and purpose in question. However, Arabic is not the object of study for only Arabists. The genetic affiliation and historical depth of Arabic, its varieties and its structural features are in fact relevant to the development of the Semitic languages and students of other disciplines and time periods. Those disciplines may use Arabic data towards understanding their subject matter with their methods pursuant to their goals separate from those of Arabic. In this essay, I have not discussed the results of the promising research in both the fields of late antique history and Arabic historical sources. Limitations of space and the desire to be guided first and for most by the structural material guided my discussions. However, this is essay argued against a nexus between the Semitic languages and Arabic. However, I did not propose a starting point for the history of Arabic, one which I consider methodologically sound. Again, considerations of space guided the decision. It is now my intention to propose in a following article mutual intelligibility as a starting point for writing a history of the Arabic language. In the process, I will be able to discuss the historical research and the new insights on the medieval Arabic sources and pre-Islamic poetry.
BIBLIOGRAPHIC SOURCES PRIMARY SOURCES
1. Huehnergard, J. 2017. "Arabic in its Semitic Context," Arabic in Context, Celebrating 400 Years of Arabic at Leiden University, ed. Ahmad Al-Jallad. Studies in Semitic Languages and Linguistics 89. Leiden / Boston: Brill, p. 334.
2. MacDonald, M. 2000. "Reflections on the Linguistic Map of Pre-Islarnic Arabia," Arabian Archaeology and Epigraphy, 11, pp. 28-79.
3. Rubin, A. 2008. "The Subgrouping of the Semitic Languages, " Language and Linguistics Compass 2, p. 61-84.
4. Al-Sharkawi, M. 2017. History and Development of the Arabic Language: from Pre-Islamic Times to the Age of Conquest, London: Routledge.
5. Hasselbach, R. 2013. Case in Semitic: Roles, Relations, and Reconstruction, London: Oxford University Press.
6. Tropper, J. 2000. Ugaritische Grammatik, Münster: Ugarit.
7. Al-Zagag, Mä Yansarif wa mä lä yansarif. Ed. Huda Qurra'a. Cairo: Matba'at al-Ahram.
8. Al-Farra', Ma äni al-Qurän. Ed. Ahmad Yüsuf NagatT. Cairo: Dar al-Kutub al-Misriyya.
9. Ibn 'AqTl, Sarh alifyyat Ibn Mälik. Cairo: Matba'at as-Sa'ada.
10.Al-Farisiyy, al-Hugga f ilal al-qiräät. Ed. 'Aliyy Nasif. Beirut: Dar al-Kitab al-'arabiyy.
11.Ibn Al-Gazriyy, aN-Nasr f al-qirä 'ät al- 'asr. Ed. 'aliyy Muhammad ad-dabba', Cairo: al-Matba'a at-Tigariyya.
12.Ibn Xalawayh, al-hugga f al-qirä'ät as-sab'. Ed. 'abd-al-'al Saalim Makram, Cairo: Dar a s- surüq.
13.Abü 'Ubayd al-Qasim Ibn Sallam, Kitäb al-'amwäl. Ed. Muhammad XalTl Harras, Cairo, 1976.
14.Ibn Hisam, Mugni al-Labib 'an kutub al-'a'ärib. Ed. Muhammad Muhyi ad-DTn 'Abdul-hamTd. Cairo: Matba'at al-Madaniyy.
15.Owens, J. 2003. Arabic dialect history and historical linguistic mythology. Journal of the American Oriental Society. 123: 715-740.
16.Ibn Faris, Maqäyisu-l-Luga. Ed. 'Abdus-Salam Harün, Cairo: Maktabatil-XangT.
17.Ibn GinnT, al-Xasä'is, ed. Muhammad 'aliyy an-Naggar. Cairo: Dar al-Kutub al-Misriyya.Ibn hazm, gamhara 'ansab-l-'arab.
18.Unrich, B. 2019. Arabs in the Early Islamic Empire: Exploring al-Azd Tribal Identity, Edinburgh: Edinburgh University Press.
19.Al-Azmeh, A. 2014. The Emergence of Islam in Late Antiquity: Alläh and His People, London: Cambridge University Press.
20.Zwettler, M. 2000. 'Ma'add in Late Ancient Arabian Epigraphy and Other Pre-Islamic Sources', Wiener Zeitschrift für die Kunde des Morgenlandes 90, 223305.
21.Retsö, J. 2003. The Arabs in Antiquity: Their History from the Assyrians to the Umayyads, London: Routledge.
22.Retsö, J. 2010. "Arabs and Arabic in the Age of the Prophet," The Quran in Context: Historical and Literary Investigation, eds. A. Neuwirth, N. Sinai and M. Marx, pp. 281-292, Leiden: Brill.
23.Al-Zamaxsariyy, al-Kassäf Beirut, Dar al-Ma'rifa.
24.Al-Sharkawi, M. 2010. The Ecology of Arabic: A Study of Arabicization, Leiden: Brill.
25.Szombathy, Z. 2003. The Roots of Arabic Genealogy: A Study in Historical Anthropology. Piliscsaba: The Avicenna Institute of Middle Eastern Studies.
26.Lecker, M. 2014. "Pre-Islamic Arabia," eds. D. Morgan and A. Reid, The New Cambridge History of Islam, pp. 153-170, London: Cambridge University Press.
27.Hoyland, R. 2009. "Arab Kings, Arab Tribes and the Beginnings of Arab Historical Memory in Late Roman Epigraphy," in From Hellenism to Islam: Cultural and Linguistic Change in the Roman Near East, eds. H. Cotton, R. hoyland, and J. Price, p. 374-400, London: Cambridge University Press.
28.Al-WaqidT, Magazi. Ed. M. Johns, Beirut: Dar al-'A'lamiyy. Al-Xudariyy, Hasiyat al-Xudariyy ala sarh Ibn Aqil. Cairo: al-Matbaa al-Maymaniyya.
SECONDARY SOURCES
1. Al-Jallad, A. 2019. "The Classification of Arabic and Sociolinguistic Variation in the Pre-Islamic Period," in The Routledge Handbook of Arabic Sociolinguistics, eds. E. El-Werr and U. Horesh, p. 15-30, London: Routledge.
2. Al-Zagag 'irab al-Quran. Ed. Ibrahim al-'Ibyariyy, Cairo: Dar al-Kutub al-'islamiyya.
3. Corriente, F. 1971. "On the Functional Yield of Some Synthetic Devices in Arabic and Semitic Morphology," JQR, 62, pp. 20-50.
4. Corriente, F. 1973. "Again on the Functional Yield of Some Synthetic Devices in Arabic and Semitic Morphology," JQR, 64, pp. 154-163.
5. Corriente, F. 1975. "Marginalia on Arabic Diglossia and Evidence Thereof in the Kitab al-Agani," Journal of Semitic Studies, 20, pp. 38-63.
6. Corriente, F. 1976. "From Old Arabic to Classical Arabic through the Pre-Islamic Koine: Some Notes on the Native Grammarians' Sources, Attitudes, and Goals," Journal of Semitic Studies, 21, pp. 62-98.
7. Diem, W. 1973. 'Die nabataischen Inschriften und die Frage der Kasusflexion im Altarabischen,' Zeitschrift der Deutschen Morgenlandischen Gesellschaft 123: 227-37.
8. Graf, D. 1978. "The Saracens and the Defence of the Arabian Frontier," Bulletin of the American Schools of Oriental Research, 229, pp. 1-26.
9. Hoyland, R., 2009. "Late Roman provincia Arabia, monks and Arab tribes: a problem of centre and periphery," Semitica et Classica 2, 117-139.
lü.Huehnergard, J. 2005. "Features of Central Semitic," Biblical and oriental essays in memory of William L.Moran, ed. by Agustinus Gianto, p. 155-203. Rome: Pontificio Istituto Biblico. ll.Huehnergard, J. and Rubin, A. 2011. "Phyla and Waves: Models of Classification," Semitic Languages: An International Handbook, eds. S. Weninger, G. Khan, M. Streck and J. Watson, p. 259-278, Berlin: de Gruyter Mouton.
12.Ibn Faris, a§-§áhibiyy fi Fiqh aL-Luga wa Sunani al-'Arab fi Kalámihá. Ed.
Ahmad hasan basTg, Beirut: Dar al-Kutub al-'ilmiyya. 13.Ibn GinnT, al-Muhtasib fi Tabyin Wuguh Sawád al-Qir'át wal- 'idáh 'anhá> Ed.
'aliyy an-Nagdiyy Nasif, Cairo: wazarat al-'awqaf. 14.Ibn Manzür, Lisán al- 'Arab. Beirut: Dar sadirIbn ginniy, Sirr sina'at al-'i'rab.
15. Owens, J. 2013a. "History," The Oxford Handbook of Arabic Linguistics, ed. J. Owens, pp. 451-471, Oxford: Oxford University Press.
16.Rubin, C. 2012. "Arabian and Ethiopia," The Oxford Handbook of Late antiquity, ed. Scott Johnson, p. 248-339, London: Oxford University Press.
17.STbawayh, al-Kitáb. Ed. 'Abdu s-Salam Harün, Cairo: al-Hay'a al-Misriyya al-'ama lit-Ta'lTf wan-Nasr.
18.Versteegh, K. 1984. Pidginization and creolization: The case of Arabic. AmsterdamPhiladelphia: John Benjamins.
19.Zwettler, M. 1978. The Oral Tradition of Classical Arabic Poetry: Its Character and Implications, Columbus: Ohio State University Press.
Information about the author
Associate Professor
Muhammad Al-Sharkawi,
Wayne State University
Detroit, The United States of America
Информация об авторе
Доцент
Мухаммад Алъ-Шаркави, Университет Уэйна Детройт, Соединенные Штаты Америки
Conflicts of Interest Disclosure: The author declares Conflicts of Interest Disclosure.
Раскрытие информации о конфликте интересов: Автор заявляет об отсутствии конфликта интересов.
