A Companion to Late Antique Literature

by Scott McGill

  By son of Ġ son of son of son of Ġ son of Wdm son of S1r son of Ṣbḥ: he wept and grieved for his father, who was murdered, so, O Lh, he will have vengeance; and he longed for his paternal uncle and all of his companions; and may he who would efface (this inscription) go blind. (LP 243; Al‐Jallad 2015, p. 266)

  By ʿbs1ʿd son of Ḥny son of ʾbd son of S1ʿd: he found the traces of his companions, for those who remain despair, and then he was sad on account of the sheep which had gone hungry. (C2713; Al‐Jallad 2015, p. 237)

  By Mty son of Ḥzn: he rebelled against Rome the year the Persians came to Bosrā, so, O Lt, let there be security. (SIJ 78; Al‐Jallad 2015, p. 281)

  The ancestor of classical Arabic, which specialists call Old Arabic, emerged in a historical context and by a historical process that, it must be said, scholars have not fully understood. Different approaches have been taken to explain the origins of classical Arabic and the people who spoke it. For much of the twentieth century and still today it has been widely held that, essentially, the Arabs have effectively always been in Arabia, speaking an undocumented, very early form of Arabic. This view appears often in works on the origins of Islam in which “the Arabs,” the default agents in the narrative, are united by Islam and undertake a vast conquest. The focus of such explanation is not on Arab ethnogenesis, because a uniform Arab ethnic group is taken for granted, but on the organization of Arabs into a state or the like.

  A second approach emphasizes the relative newness and changeability of Arab identity and seeks to discover the history of Arab ethnogenesis, while taking seriously the traditional literary Arabic narratives of pre‐Islamic times compiled by genealogists, Arabic language scholars, and antiquarians from the eighth century onward. Selected epigraphical and archaeological materials are used in this theory to support the account in which the genesis of Arab tribes who carried out the conquests of the seventh century occurred in the wake of a long period of staggered migrations of pastoralist peoples from southern Arabia into the highland (Arabic Najd) interior of the Arabian Peninsula and still further toward the north into Syria and Iraq, starting in the third and fourth centuries and continuing – and, in a way, culminating – in the Arab conquests of the seventh and eighth centuries (Hoyland 2001, pp. 229–247). As Michael Zwettler (2000) and Robert Hoyland (2009) have argued, the family and tribal names attested in the thousands of Ancient North Arabian inscriptions of the period 200–400 mostly disappear from use, only to be replaced eventually by tribes first mentioned in the south in Ancient South Arabian inscriptions and whose names – such as Tanūkh – are all attested and made famous later in the period from the seventh‐century conquest. They take these epigraphic data to corroborate independent literary sources of later times: the massive genealogical compendia of the Arabs after the conquest, surviving from the ninth century onward, as well as a number of detailed collections of pre‐conquest Arabian lore aggregated by specialists such as Ibn al‐Kalbī (d. ca. 820) and al‐Aṣmaʿī (d. 828). These collections of initially orally transmitted Arabian tribal and family tradition describe just such migrations from south to north. For now, however, this account remains to be tested by archaeology and careful source analysis. A new analysis, presented by Webb (2016), argues that the Arab identity was generated in the midst of or in the wake of the Islamic conquests and that it did not exist as such before.

  A third approach ignores the later Arabic traditions and emphasizes the comparative method of historical linguistics as applied to the Semitic languages; it adheres most closely to the epigraphic evidence and their precise distribution, studying also the genesis of the Arabic script. The view from this approach suggests Arabic was an ancient language that was long prevalent in and local to northwestern Arabia, including the Ḥijāz mountain region. Arabic comes into our view when speakers of Old Arabic who had access to the Nabataean script began to use it. Scholarship in this vein has largely not addressed the development of classical Arabic during and after the conquest but regards isolating and defining the earliest form of Old Arabic as a precondition to determining the later history of the language.

  Still other views on the problem of Arab ethnogenesis can be found, holding various elements in common with the approaches just described. Suffice it to say here that the demographic and linguistic changes in late ancient Arabia are the subject of intensive ongoing research and have been explained in mutually incompatible ways by different scholars. It is a controversial subject and the picture may change radically or, it is hoped, come into focus sharply. Several factors have inhibited research into the topic, but a shortage of unambiguous sources and the default assumption of ethnic homogeneity across much of Arabia have been some of the chief ones. A recent collaborative volume edited by Greg Fisher (2015) usefully gathers and introduces many of the relevant materials with the help of experts in subfields.

  The first signs of what would become classical Arabic occur in a small number of inscriptions over several centuries written in Old Arabic. These appear first in the vicinity of modern Jordan and Syria. There are several ancient inscriptions of the first centuries CE that may lay claim to being partly or wholly Old Arabic (MacDonald in Fisher 2015, pp. 395–417), but the Namāra inscription of 328, from southeastern Syria, presents the oldest uncontroversially entirely Old Arabic text. It is a grave inscription of a King Imruʾalqays, and it commemorates his victory over Arabian peoples as far as Najrān, near the present‐day Yemeni–Saudi border. From the sixth century several more Arabic inscriptions appear on a Christian martyrion, on a church doorframe, and in a graffito. These late ancient Old Arabic texts were not written in a variety of the script used across ancient Arabia but in a late cursive variety of the Aramaic script used in the Nabataean Kingdom and in the Roman province of Arabia, a territory roughly corresponding with present‐day Jordan.

  The earlier stages of Old Arabic are evident in ancient inscriptions on the fringes of the Nabataean Kingdom (incorporated into the Roman Empire as a province in 106), in northwestern Arabia, and in Arabic words and phrases used in Nabataean Aramaic texts. These point to the primary language of their authors. A tiny number of short inscriptions in the Nabataean script, in unspecifiable Semitic languages from the first century CE (and perhaps earlier), contain linguistic features otherwise known only from Arabic. These plausibly represent earlier relatives of Old Arabic. What is clear, however, is that Arabic as we know it comes to be employed regularly only in late antiquity, being put to writing in the Nabataean Aramaic letters, the ancestor of the cursive Arabic script. During this period, its use must have been largely on perishable materials. Otherwise, it would be very difficult to explain the continuous use of a late form of the Nabataean script to write Old Arabic, long after Nabataean Aramaic had disappeared from available records.

  The history of Arabic before the seventh‐century conquests is known primarily from these inscriptions. It is a matter of debate whether the language of northern and central Arabia was relatively uniform or highly variegated by dialect. Common sense, along with reports about dialectal variation among Arab tribes in the ninth and tenth centuries, supports the latter interpretation. There does survive, however, a remarkable quantity of poetry, Arabic šiʿr, from this ancient Arab population, dating to the fifth, sixth, and seventh centuries (and beyond late antiquity, of course, in unbroken tradition until the present, constituting one of the bases of Arabic literature throughout its history). Arabic poetry from before the Arab conquest, like most extant ancient poetry, survived only because it was useful and entertaining to later readers and listeners who preserved it. The manner of its oral intonation in the time of its composition is not exactly known today, but it is clear that we are dealing with the lyrics to rhythmic songs of a kind. The verb typically used for the recitation of this poetry, inšād, connotes calling out in a loud voice. After the Arab conquest, šiʿr became recognized as different from the poetry of other nations in two ways. Phonemic vowel quantity (long and short vowels) contributed to endowing
heavy and light syllables in Arabic with meaningful distinctions, just as in classical ancient Greek and Latin verse but on different patterns in their own tradition. Arabic poetry cast in these meters was called mawzūn, or “measured,” as on a scale. Its second major formal characteristic was the rhyme of final consonants, called qāfiya; every verse was to end in the same consonant or consonant‐vowel syllable. We might add a third, important feature: The language of this poetry used a set of case‐ and mood‐endings which in most instances constituted whole syllables (and were thus essential to the meter). Such grammatical endings eventually disappeared almost entirely in all vernacular varieties of Arabic after the conquest but remained part of the classical ideal of the medieval literary language. The grammatical noun endings were, as modern linguists have discovered, highly conservative, in that they reflect the proto‐Semitic system of inflection very closely and find their closest extant match in the Akkadian of the third and second millennia BCE. Presumably such linguistic conservatism reflects the isolation of the population of its users, and specifically indicates that very few non‐native speakers learned this form of Arabic before the conquests.

  There were different sorts of poetry. The rhyming rajaz verse was used for boasts and taunts before battle, laments, and various short folk tunes. The more elaborate qaṣīda or “ode” came to be most prestigious and represented šiʿr proper. Although some short poems were regarded as “snippets,” qiṭʿas, the archetypical qaṣīda was conceived as comprising three parts in about 75–100 long lines in total. The first part, the nasīb, is often called in English the “amatory prelude.” It sets a nostalgic tone of longing as the poet describes a place where he camped and met a beloved, or a place in which people dear to him had been but are no longer. The poet mentions or describes an abandoned site reminding the poet of a loss. (This is a frequent theme, too, in the Northern Arabian graffiti described above, suggesting perhaps that the graffiti reflected a culture of song contemporary with it.) The nostalgic prelude leads to an episode representing a journey, riḥla, typically through the desert and often by night, sometimes involving detailed descriptions of the poet’s mount and elaborate similes. Lastly comes the boast, fakhr, or another topic motivating the composition such as praise of a patron. Throughout, the poet may illustrate his point with a proverb, mathal, or words of wisdom, ḥikma. The tripartite construction came to be remarkably consistent, even if quite a few qaṣīdas actually lacked the neat structure. The three ideal parts of the qaṣīda were quite susceptible to use for various purposes. In later times, for example, “amatory preludes” could refer to longing for God, or despair for a deceased patron, or any number of other sorts of longing. Poets of later ages were ingenious in adapting the ancient standards to new ends.

  Hundreds of late‐ancient Arabic poems survive, entire or as fragments, but the question of their authenticity has rightly been a major concern for scholars. The doubts about it have come from several sources. One is the gap of centuries between the alleged times of composition down to the time of earliest attestation in writing. Another is the large number of textual variations between different attestations of the same verses, suggesting a text so unstable as to be unverifiable. These factors even led a few twentieth‐century scholars to suppose some or all of the early Arabic poetry to be a late fabrication (e.g. Margoliouth 1925). It seems safe to say that few if any now adhere to this highly skeptical view. Another approach follows the theories of “Oral Formulaic” verse fostered by Parry, Lord, and others. On this hypothesis the early Arabic poetry was composed extemporaneously and orally, and differently on each occasion, by professional versifiers who used metrical formulae to fill out lines spontaneously in the same manner as twentieth‐century Serbian bards. This would mean that the early Arabic poetry of which we possess records, transmitted as written in medieval books, of course, represents only specific iterations of poems that varied on each occasion. The application of the “Oral Formulaic” hypothesis to early Arabic poetry (most concertedly made by Zwettler 1978) has, however, essentially been invalidated by the research of Gregor Schoeler (2006, pp. 87–110), who has documented reports of the great effort that early Arabic poets took to prune and edit their own poetry over long periods, months or more, until they were finished. Explicit statements by early medieval students of this poetry show that it was not spontaneous and formulaic. Arabic poetry was certainly orally performed, and there are reports of spontaneous versification over the centuries (reported because it was remarkable), but long poems represented a specially honed craft of deliberate, thoughtful, sometimes time‐consuming composition.

  The numerous variations in the recorded versions of individual Arabic poems have several sources. One is surely just the error and whimsy of devoted reciters (rāwīs) and copyists who, unfamiliar with ancient or dialectal words and expressions, replaced them mistakenly or deliberately with words that made more sense to them and their contemporary audience. Another is the sort of variation that one encounters in song lyrics today, too, whereby the words may be changed when either remembered imperfectly, so as to make sense, or deliberately, to make a point. There is no single source of variation in the lyrics, then, of these popular poems and songs but, in short, they seem to be due primarily just to the lively long‐lived human appreciation and reiteration they enjoyed, as with all ancient textual traditions preserved by copying. Sometimes modern scholars romantically attribute almost supernatural memories to premodern peoples, but a moment’s reflection reminds us that, even today, individuals may know the lyrics to hundreds of popular songs and musicians may be able to recall countless tunes, even if sometimes imperfectly. In view of this, the survival of many genuine ancient Arabic songs in writing, with considerable variations, is not strange.

  Another matter of debate concerns the nature of the Arabic language in which the poetry is composed. A typical view is that the language of the early Arabic poetry is a “supra‐tribal koine,” more or less artificial, employed by nobody as an everyday language and used only for poetry and special, solemn occasions. In a strong, widely held version of this theory, the ancient Semitic nominal case endings used in the poetry were confined to it and to non‐poetic solemn utterances. This view retrojects the modern Arabic diglossia (the concurrent use of colloquial varieties for ordinary communication and a universal literary standard for formal and written communication) onto antiquity without regard for the social circumstances and system of education that generate modern Arabic diglossia. Contrariwise, other scholars, seemingly currently a minority, hold that the Arabic of the poets represents a normal language of general use up to the time of the Arab conquests. Its use in poetry was certainly in a poetic register, but is not endowed with special grammatical systems of inflection used only in poetry. In this minority view, the Arabic poets intended their songs to be appreciated by a normal, untrained audience. The stance one adopts on the Arabic language of the poetry may make a difference in how one interprets it and uses it for history.

  Numerous sources indicate that the Arabic language of everyday life underwent drastic changes following the Arabic conquest. The reports come from the earliest Arabic linguists, from the eighth century onward, who strove to preserve the language in a classical state from the “corruption” that they perceived it to be undergoing due to social mixing with conquered peoples. The very survival of late ancient Arabic poetry, along with the standardization of an idealized pre‐Islamic Arabic language maintained among bedouin who had not mixed with colonized peoples, is therefore partly due to the attempt of those scholars to record early, “uncorrupted,” aesthetically pleasing, and especially authentic‐seeming samples of that language. This, they believed, was the language known to their prophet and used in their scripture. Without their effort to find, preserve, collect, and evaluate the corpus of old lyrics and verses available to them, we would have almost no literary products from late ancient Arabia besides the inscriptions.

  The poetry of late ancient Arabia, representing mo
stly the sixth and early seventh centuries, is remarkable in its apparent isolation from other contemporary forms of literature in the languages of neighboring peoples. It is not that Arabic poetry was produced in a world sealed off from the outside. It is rather that Arabic poetry seems to have been unknown to outsiders. It had its own forms and conventions. It refers largely to a world aware of and often oriented toward the great and powerful states of the Romans, Persians, and Ḥimyarites, but on the margins of that world. The setting of ancient Arabic poetry is rather a society of mostly pastoralist tribes in competition over resources of water, pasturage, and livestock, who were not part of large, complex states. The inhabitants of this world are portrayed as having no laws but those of revered custom, the might of arms levied by kings, warlords, or bands, and the overriding connections of kinship. Small‐scale violence appears to have been common. The poets impart melancholic wisdom about the inevitability of death and seek refuge in the glory of victory at arms and the good reputation of one’s family and tribe, transitory though they may be. They praise unstinting generosity. They long for happy days spent with friends now gone away by distant roads. Poetry by women is preserved, mostly in the form of moving laments for brothers who died violently, of whom only the report of their distant death returns. Other poems survive in the voice of the outcast or brigand, the ṣuʿlūk, glorifying a harsh but independent life.