SOURCE: "The Regiments of the Imperial Army: Notes on Al-Jahiz's Epistle to Al-Fath B. Khaqan," in The Shaping of Abbasid Rule, Princeton University Press, 1980, pp. 116-38.

[In the following essay, Lassner discusses al-Jahiz's On the Virtues (Manaqib) of the Turk and what it reveals about early 'Abbasid armies.]

Al-Ma'mun, al Mu'tasim, and an officer whose name is not mentioned, disagreed as to the bravest among the officers [quwwad], troops [jund], and clients [mawali]. Al-Ma'mun maintained that there were none braver than the non-Arabs among the people of Khurasan [ajam ahl Khurasan, that is, the Transoxanians who brought him to power]. Al-Mu'tasim, in turn, favored the Turks [that were the backbone of his support], but the officers held out for the abna '. They were the ones that shackled the Turks [that is, lead them to Islam], just as their forefathers led the ['Abbasid] revolution [dawlah]. They fought the Commander of the Faithful [during the civil war] but now pay him obeisance, and it is through them that his rule is secured.

Based on Ibn Tayfur (Cairo), 80.

The professional army of the 'Abbasid state came to include a variety of ethnic groups from diverse regional backgrounds and social strata. The first century of the dynasty began with an army composed of vestiges of somewhat independent Arab tribal forces and concluded with the formation of a slave corps recruited primarily from among Turkish and Iranian peoples taken prisoner in Transoxiana. Although this military organization must be regarded as among the most salient and important institutions of the government, the precise identification of various fighting constituencies and their specific functions remains problematic. Much of the confusion is no doubt derived from the haphazard presentation of the medieval chroniclers. However, in contrast with the untidy state of Arabic historiography, a work of belles letters by a well-known literary figure provides a concise and cogent statement about the general composition of the early 'Abbasid armies.

The important text, On the Virtues [Manaqib] of the Turk, is an interesting essay (risalah) by the ninth century theologian and litterateur al-Jahiz.¹ In what may be succinctly described as an ornate prose style, the qualities of the recently recruited Turkish fighting units are compared to four other groups in the army of the Caliphate: the Arabs, clients (mawali), abna', and Khurasanis. The treatise seems to call for the full acceptance of the Turks and to imply that the reconciliation of all these elements would promote an integrated and more stable society.

From the outset, al-Jahiz performs rather intricate intellectual gymnastics to demonstrate that the distinctions between the Turks and other elements of the Caliph's army are more apparent than real, but it is strongly suggested that diversity and unity are not necessarily antithetical conditions. Geography, which determines cultural values, not only divides societies, but unites them as well, for, the author contends, diverse groups inhabiting the same, or contiguous, regions may come to share common characteristics. It is not possible to tell the difference between Basran and Kufan (Iraq), Meccan and Medinese (the Hijaz), Jabalite and Khurasani (Iran), or Jazarite and Syrian. Moreover, there were Arabs and bedouins who settled in Khurasan, and "one cannot distinguish between the man who settled in Farghana [a district of Transoxiana] and the indigenous inhabitant of that country."²

This last and rather casual insertion concerning the Arab of Khurasan belies a second significant formulation. The conventional wisdom of medieval Arabic geographers not only divided the world into various inhabited climes, but it also described the ethnic composition and social behavior characteristic of each. What is implied in al-Jahiz's remarks is that the inherent cultural traits of a particular region can be transferred to a settler population. The Arab transplanted in Khurasan thereby becomes a native of that region. In this fashion, the abna' can be considered as both Khurasanis and Iraqis, and the Turks, who originate from the far reaches of eastern Khurasan, can be legitimately counted as equal to the indigenous inhabitants of the western districts. A common bond is thus created for the entire army by linking its diverse elements to Khurasan, the breeding ground of 'Abbasid revolution. Given a single geographical setting, the points of agreement in 'Abbasid military society could come to predominate over the points of difference and thus serve as a basis for unity.

The legitimacy of the Turks is also reflected by their client relationship with the ruling house, a status they shared with others among the military. The author, who was himself a client of the Banu Kinanah, observes that, from ancient times, the Arab tribes, while scrupulously guarding their genealogies, acquired clients as a matter of course. As clientage bestowed a recognizable status on both parties of the association, dissimilar groups may be found sharing a common bond. It was general knowledge that the Caliph al-Mu'tasim first acquired large numbers of Turkish slaves in order to fashion them into a regular regiment of the imperial army. By virtue of their personal ties to the ruling family, certain Turks became clients of'Abd Manaf and Hashim and were therefore linked to the Arabs in general and to Qurayshite nobility in particular.³ Nevertheless, if there is any doubt that clientage and geographical contiguity could effect the assimilation of the Turks within 'Abbasid society, the author provides additional evidence to argue his case.

Al-Jahiz also seeks to strengthen the ties between the Turks and the Caliphate (that is, the Arabs) by claiming an ancient blood relationship. It is stated that the Commanders of the Faithful are descended from Isma'il, the son of Ibrahim. The example of Isma'il demonstrates the extent to which integration can be achieved despite one's ethnic origins. The son of Ibrahim was counted among the Arabs, although he was born of foreigners, for God adapted his uvula to the correct pronunciation of Arabic without instruction or practice and then bestowed great eloquence upon him without formal educa tion or training. He thereby removed those features foreign to Isma'il and transplanted him into the Arab nation as an equal to those born of Arab stock. This fact would indicate that the most noble of contemporary Arabs are themselves the descendants of a client, or, if one were to insist on putting it more delicately, they stem from a distant ancestor who was adopted by the Arabs through a creative act of God's will.⁴ Be that as it may, in addition to Ishaq and Isma'il, the respective sons of Sarah the (north) Syrian (Suryani) and Hajar the Copt, Ibrahim had also six children by Qatura the Arab woman. Four of these offspring were fortuitously situated in Khurasan, where, tradition has it, they became the progenitors of the Turks. Therefore, should an Arab boast of his noble descent, the Turk can always offer as a rejoinder, "But Ibrahim is my [grandjfather and Isma'il is my uncle."⁵

Both in his preliminary remarks and in the main body of his presentation, which details the relative merits of each group, the author relentlessly pursues the same theme. He constantly reaffirms the mutual ties that exist between all the units and, hence, the legitimacy of the Turks within the 'Abbasid military and within society at large.

There is, to begin with, the critical issue of whether the Manaquib can be considered a historical document that actually sheds light on the structure of the 'Abbasid military. There is, in addition, the related question of the author's motivation in preparing the essay. Although the treatise is often cited, it remains elusive, for, lacking a systematic explication of the text, either by historians or literary scholars, no clear judgment of the historicity of the author's remarks can be made. The present chapter can best be described as a series of preliminary observa tions. The intention is to identify some of the many his torical allusions to groups and events found throughout the treatise and to illustrate how they serve al-Jahiz's overall strategy for legitimizing the presence of the Turks within the imperial army.

Turkish officials and their patrons within government circles would, no doubt, have favored such a line, for the introduction of Turkish contingents in large numbers, beginning with the Caliphate of al-Mu'tasim, created great tensions within the army and society at large. Given the highly charged political atmosphere, one is obliged to ask if the picture of the 'Abbasid regiments drawn by the author and cited by modern historians is an accurate reflection of contemporary military organization, or if it represents the idealized creation of a highly inventive literary mind. Even if al-Jahiz's need for literary license, to say nothing of diplomatic discretion, overwhelmed his sense of historical accuracy, there are at least vestiges of an historical truth that can be distinguished here.

Although he achieved fame as the greatest prose writer of his time, al-Jahiz's literary interest in horses, slaves, misers, and various ladies of pleasure might cause one to suspect the acuity of his observations on so technical a subject as the 'Abbasid army. Nevertheless, the author's credentials for observing the government and military were impeccable. He was an eyewitness to events at the twin capitals, Baghdad and Samarra, and he served for a brief time as court tutor to the young children of the Caliph al-Mutawakkil (d. A.H. 247). Although not really a courtier, his views were nevertheless sought by high ranking officials, and the Manaqib is in fact addressed to al-Fath b. Khaqan, the Turkish wazir of the Caliph. The author therefore had a practical as well as scholarly acquaintance with the development of the imperial army. He may have chosen to bend history in order to suit his purposes as well as those of his patrons, but his text is nevertheless rich in allusions to historical events and personalities.⁶

Be that as it may, one critical approach might argue that al-Jahiz's facile portrayal of the military melting pot should be met with considerable skepticism. Political expediency is often the mother of rather dubious genealogies. For example, the realities of assorted alliances led the Arabs to claim descent from fictitious ancestors, thereby promoting the division of their society into tribal units of north and south Arabian origin. The author's analysis is similarly not above suspicion. The Turkish wazir of the Caliph, to whom the essay is addressed, was more than a public functionary. He was widely recognized as a great patron of the literary establishment and had artistic pretensions of his own (although with only thirteen surviving lines of verse, it is not quite possible to establish his position as a man of letters). In addition to al-Jahiz, he was familiar with the poet al-Buhturi and the polyhistor al-Tha'alibi.⁷ It is therefore possible to regard al-Jahiz, who was an occasional figure at court, as still another litterateur attempting to curry favor with an important official of state by giving literary expression to that which was certain to please his patron. The author's complimentary remarks concerning the superior martial skills of the Turks and, more particularly, his effort to integrate them into an 'Abbasid society resentful of their position can, then, be regarded as self-serving political fictions. Seen in this light, the Manaqib may be the creation of a cynical and extremely clever man, and, like other works of this genre, it would therefore be interesting only for reasons of language and not for historical content. This, however, is not the case.

The central message of the text may also reflect the ambiguities of the author's self-identification. Al-Jahiz, who was a mawla, was also part black, and the search for compatibility between disparate elements of society is one that finds ambivalent expression elsewhere in his literary endeavors. With a writer of al-Jahiz's skill and subtle sense of humor, the problem of establishing his real attitude toward race and color is likely to be difficult. One has the impression that the interests of individuals on the lower end of client-patron relationships were best served by their masking their intentions. This was true even for those of marginal background who gave every indication of being fully adapted to the society of their patrons. Moreover, it is not always clear, in these delicate relationships, whether the author is himself emotionally prepared to declare in favor of a specific reality. Al-Jahiz can thus devote a treatise to praise of the blacks, while he elegantly pokes fun at them. Who is to say what the author's true intentions are—perhaps not even the author himself.

Nevertheless, one is struck by his remarks concerning Isma'il. His interest in the rejected son of Ibrahim may be more than a literary device and tempts the reader, whether intentionally or otherwise, to draw obvious comparisons. With or without God's intervention, al-Jahiz, like Isma'il before him, was living proof that the literate client could overcome his ethnic origins by excelling in the language of his acquired patrons. Seen from this perspective, The Virtues of the Turk is not only an artistic creation and means of entry into courtly circles, it is a reflection of the author's search as well for an integrated society and his personal niche within it. Such a view suggests that the composition of the army described by al-Jahiz may have existed largely in his imagination and is, in reality, a metaphor for a much wider polity, namely, the sum of Islamic society.⁸

There may be a good deal of truth to all of this, but it should not lead us to dismiss the Manaqib out of hand. When writing on historical themes, medieval Arabic litterateurs tended to embellish rather than invent, and the embellishment was often fashioned from raw data of historical interest. The methodology of the modern historian is directed toward distinguishing the residue of historical truth from the literary shaping of the medieval author—a task easily proclaimed, but often difficult to accomplish.

Although the basic thematic lines are clear enough, the text of the Manaqib does not readily lend itself to a thorough explication. The catalogue of virtues that is ascribed to each of the five regiments is the key to the author's design. However, the allusions to historic movements and personalities are presented with particular care, so that, given an imprecise knowledge of early 'Abbasid history, the fine detail of the work is at times complex and elusive. Moreover, the specific setting is difficult to fix. The author begins his lengthy section on the Turks by indicating that it was originally composed as a letter to the Caliph al-Mu'tasim, but, for reasons not specified, it was never delivered.⁹ On the other hand, the introduction to the work in toto indicates that it was addressed to al-Fath b. Khaqan, the Turkish wazir of al-Mu'tasim's son, al-Mutawakkil.¹⁰ Elements of the Manaqib could have been composed, therefore, over a stretch of time encompassing almost three decades. Because this was, by and large, a turbulent period characterized by many changes, there is a serious problem of chronology that must be resolved. A close reading of the text seems to suggest that the basic lines of the Manaqib could have been formulated as early as the reign of al-Mu'tasim and that the text indeed reflects even earlier conditions. For, with the exception of the Turks, who are the subject of the treatise, it is not al-Mutawakkil's or even al-Mu'tasim's Samarra-based army that is being described, but that of the Baghdad Caliphate of an earlier time.

The formative years of the dynasty at Baghdad saw the evolution of a professional fighting force based almost exclusively on the revolutionary armies from Khurasan. Recent "revisionist" histories have stressed the mixed character of the troops that brought the 'Abbasids to power. Contrary to earlier assumptions of a grass roots uprising among the indigenous Iranians, the backbone of the 'Abbasid military was the Arab tribal army stationed in the region. These Arab units identified themselves according to their tribal origins and initially suffered from the damaging effects of xenophobia. There is, however, reason to believe that they may have also felt a sense of regional loyalty and identified themselves not only as Arabs, but also as Khurasanis.¹¹

Another component of this army were the descendants of the old Arab settlers in Khurasan. Transplanted into the villages of the region, they too retained their tribal identification, but intermarriage and assimilation with the indigenous Iranian population had eroded tribal sensibilities. These Arab settlers may have lost their facility with their native language. In any event, they spoke Persian among themselves.¹² When the formal structure of the 'Abbasid army was first established during the revolution, the settlers received their service pay according to a military roll (diwan) arranged by village rather than by tribal affiliation.¹³ They probably retained some martial skills, but could not be relied upon as first-rate soldiers, for, unlike the Arab tribal army, they had long ceased to function as cohesive military units in the employ of the state. Nevertheless, the old settlers produced a disproportionately high number of field commanders and political agents (naqib)—a fact no doubt related to their early and sustained ideological commitment to the 'Abbasid house.¹⁴

The last, and surely least significant, element of the early 'Abbasid army were the local mawali. Although their military contributions were no doubt negligible, various clients, nevertheless, did assume extremely important positions within the political apparatus and in the financial administration of the army.¹⁵ From almost the outset of the open revolution, all of these groups, the tribal army, the villagers, and the mawali, were subsumed within the single military force known as the ahl Khurasan. As a result, the 'Abbasid army, despite its varied Arab and Iranian components, had a distinctive Khurasani identity.¹⁶

This military force is clearly distinguished from other regional armies, tribal auxiliaries, irregular troops, and local militia. In addition to the customary distribution of bonuses before and after battle, it received subsistence pay according to a carefully regulated schedule, a measure that presumably reinforced its sense of ideological commitment.¹⁷ Moreover, the professional army had specific ties to the Caliph and his newly established capital city. Having swept into Iraq with the advance of the revolution, the Khurasanis were eventually concentrated at the recently constructed administrative center in Baghdad. Their resettlement in a new regional environment no doubt strengthened their common ties based on old geographic associations and gave rise in time to new feelings of group unity. In an ambitious and unprecedented scheme of social engineering, the general area of al-Harbiyah, the northern suburb of Baghdad, was reserved for the exclusive settlement of the 'Abbasid, that is, Khurasani, military.¹⁸ From their cantonments at al-Harbiyah, the army was sent on distant campaigns to serve as the main battle contingent of the state. However, even in these circumstances, it was sufficiently large to allow for a standing reserve at the capital, so that at no time was the Caliph's domain without the services of his trusted Khurasanis. There is reason to believe that, in time, this army, or, at least, significant elements within it, was transformed into a still more cohesive grouping, the enigmatic contingent known as the abna'. The composition of this last regiment will be discussed later.

It is evident, from the catalogue of virtues ascribed to each group, that al-Jahiz's army is not the imperial force that existed when the Manaqib was submitted to al-Fath b. Khaqan, but this earlier army of the Baghdad-based Caliphate. There is the exception of the Turks, and a lingering influence can be detected for the abna ', but the Khurasanis described in the Manaqib were apparently those subsumed within the abna' and bear little relationship to the Transoxanians who served al-Mutawakkil and his predecessors at Samarra. Moreover, the Arab tribal units, which retained their identity, are said to have been removed from the military roll in the time of al-Mu'tasim (probably in the central provinces of the empire).¹⁹ They would hardly be mentioned, therefore, in any discussion of his son's Samarra army. Finally, although individual mawali were prominently involved in military and political service, it is very unlikely that they would have been considered a distinct contingent of the professional forces after the early stages of the open revolution. This is borne out by the text of the Manaqib, for, in enumerating the qualities of the clients, the author has nothing to say of their military activities, but draws attention, instead, to their trustworthiness in serving the political interests of the regime.²⁰

The detailed proof of these assertions is derived from an analysis of the virtues credited to each group. Moreover, an examination of this particular material suggests also the basic strategy employed by al-Jahiz in order to legitimize the Turks. It is more subtle than a contrived argument based on geography, clientage, and alleged blood ties. The Turks are seen as the last in a line of Khurasanis who served the 'Abbasid family—an unbroken connection that goes back to the revolution that brought the regime to power. There is apparently no point in describing the virtues of the Turks in relation to the other contingents in the army of al-Mutawakkil. The non-Arab Khurasanis from Transoxiana, the Faraghinah, the Ushrusaniyah, the Shashis, the Magharibah, and the Shakiriyah possessed credentials that were no better than those of the Turks.

Al-Jahiz is seeking an earlier model, when the imperial army represented the integrated fighting force of a relatively stable and unified empire. The legitimization of the Turks is, in a sense, the call for a return to the earlier age. In this respect, it mirrors the messianic propaganda of the 'Abbasid revolution, which proclaimed a spiritual return to the era of the Prophet Muhammad and the birth hour of Islam. The pagan Turks, who were brought in shackles during the campaigns in eastern Khurasan, were thus accorded a status equal to that of the old 'Abbasid revolutionaries and, by extension, to that of the earliest Muslims as well.

If Khurasan is the geographical linchpin of al-Jahiz's armies, the revolution is the predominant historical event that links them together. Al-Jahiz begins his list of virtues with the case for the Khurasanis and immediately invokes the memories of the 'Abbasid revolution.²¹ These are not the Transoxanians, who, for al-Mu'tasim's brother, the pretender al-Ma'mun, laid siege to and conquered Baghdad amid great devastation. These are the early Khurasanis who brought the regime to power. The Khurasani boasts: "We are [revolutionary] agents [naqib] and the sons [abna'] of agents, nobles [najib] and the sons of the nobles.…The twelve agents are from us as are the seventy nobles. We are the men of the moat and the sons [abna'] of the men of the moat.…We are the men [ashab] of this revolution [dawlah] and this ['Abbasid] propaganda [da'wah], the root of this tree, from which blows this wind. We are the men of the [revolutionary] black standards…who destroy the cities of tyrants and take away rule [mulk] from the hands of the oppressors, the Umayyads."²²

There then follows a list of major military campaigns, beginning with the fortification of the villages in Khurasan, which signified the declaration of the open revolution, and ending with an allusion to the siege of Wasit (nahnu…ashab Ibn Hubayrah), the Umayyad capital of Iraq and the last stronghold offering armed resistance. The Khurasani proclaims, "Ours is the old and the new, the beginning and the end of the 'Abbasid revolution."²³ It was no accident that the Khurasanis came to play this role. The author cites a variant of the fa mous tradition in which Muhammad b. 'Ali, the father of the revolution, considered where to send his propagandists (du'at). Various regions were considered but were found lacking. To be sure, there were a few 'Abbasid partisans in al-Basrah, but the city and its environs were clearly pro-'Uthmanid. The loyalty of the Syrians, on the other hand, was with the family of Abu Sufyan and the Marwanids. Finally, there was al-Jazirah, but the inhabitants there displayed distinct Kharijite sympathies. Only Khurasan was adequate to 'Abbasid needs. The Khurasani therefore boasts, "We are the best contingent for the best imam. We vindicated his [original] opinion, confirmed [the wisdom of] his idea, and proved the accuracy of his insight."²⁴

There are allusions throughout the discourse to the messianic propaganda of the revolution. This was a radical line that was tempered fairly early by the 'Abbasid Caliphs and thus suggests, once again, that al-Jahiz's description is not contemporary with his own era. When the Khurasani says, "Ours is the old and the new, the beginning and the end," he speaks not only of the political revolution, but of heralding a new age. This was to be a period that marked a return to the pristine society of early Islam. The events that brought the 'Abbasids to power did not merely signify the exchange of one dynasty for another, but the return to the ethos of an earlier age when moral authority was vested with the true believers, whose interests were now properly represented by the 'Abbasid family.

Al-Jahiz indicates that there are, in reality, two groups known as the helpers (al-ansar ansarari). He points out, through his Khurasani interlocutor, that the Aws and Khazraj supported the Prophet in the early days (fi awwal al-zaman) and the people of Khurasan (ahl Khurasan) will support his inheritors on the last day (fi akhir al-zaman).,²⁵ When the Khurasani claims, "The twelve agents are from us," he is referring, as well, to the twelve nuqaba' from the Aws and Khazraj. These were agents established by the Prophet Muhammad at al-Madinah in order to facilitate his acceptance in that community, an act that paved the way for the future success of his mission. Similarly, in the time of Muhammad b. 'Ali, twelve agents of the 'Abbasids fanned out into the various regions of Khurasan to lay the groundwork for the uprising that was to bring the latter-day Muhammad to power.²⁶ When the Khurasani boasts, "We are the people of the moat and the sons of the people of the moat," he refers to two events. The first is rooted in the age of the Prophet, when Salman al-Farisi, the original Islamic hero of Iranian origin, saved the faithful by suggesting that Muhammad dig a moat or trench (khandaq) around the oasis of al-Madinah. He thereby befuddled the opposing Meccan cavalry, who had never seen such a strategy before. This was a critical moment in the series of great victories that led to the capitulation of Mecca itself. In the same fashion, the Arab settlers and mawali later fortified their villages in Khurasan by digging a moat around the defen sive perimeter. This act, together with the unfurling of the messianic black banners, signified the beginning of revolutionary warfare and heralded the new era about to come.²⁷ It is no wonder that the Khurasani can proudly say, "We were thus nurtured by our ancestors and we thus nurture our sons (abna')."²⁸ Al-Jahiz completes the case for the Khurasanis with a description of their cultural background, fighting qualities, and armor. These are not people to be taken lightly. Even if the men of Tibet and Zabaj, the cavalry of India, and the horsemen of Byzantium were to attack the Khurasanis in concert, they would be forced to throw down their arms and flee. Is it any wonder that pregnant women give birth prematurely upon hearing the battle cry of Khurasan?²⁹

If these are the virtues of the Khurasanis, who are the Arabs and what are their claims? Al-Jahiz's description of the Arabs is tersely presented. There is no indication how pregnant women reacted in their presence, but it is clear that their claim similarly invokes Khurasan and the 'Abbasid revolution. What is more, it is the very same claim as that of the Khurasanis. The Arab boasts, "Who are most of the agents (naqib), if not, in essence, Arabs (min samim al-'arab)."¹⁰ The phrase, "in essence, Arabs," is carefully chosen, for the author indicates that, as a group, the Arabs have great respect for genealogy. As it happens, the Arabs listed by al-Jahiz among the nuqaba ' may not all have been of pure blood, for they were descended from the old settlers who took local women while partially assimilating to native life in the villages of Khurasan. Listed among them are some of the most prominent names of the revolution: Qahtabah b. Shabib al'-Ta'i, the legendary commander of the army until his mysterious and untimely death, Malik b. al-Haytham al-Khuza'i, the director of the security forces, and Sulayman b. Kathir, an early political operative of importance.³¹

A description of their military exploits is the same as that of the Khurasanis. They also defeated Ibn Hubayrah and killed Ibn Dubarah and Nubatah b. Hanzalah. There is, perhaps, one difference: whereas a Khurasani is credited with killing the last Umayyad Caliph, it was an Arab who spread the good news. If there are any other subtle differences between the so-called Arabs and Khurasanis of al-Jahiz, it is not clear from a reading of the text.³² There were, to be sure, Arab tribal armies in the early years of the 'Abbasid regime, but these represented regional forces and were not counted as part of the imperial army at Baghdad. The Khurasani, in trumpeting his own praise, has a few unkind words for these Arab tribesmen, "We are not like the army of Syria who attack women and violate all that is sacred" a not so veiled reference to the rape and pillage that marked a Syrian campaign against the holy cities in Umavyad times. When al-Ma'mun declared against the central authorities, with a Transoxanian army at his back, an effort was made to enlist the Arab tribes of Syria in support of the Baghdad regime. However, they could not overcome their susceptibility to 'asabiyah, and proved a greater danger than a help. An incident at the initial muster involving a stolen horse gave rise to a debilitating tribal conflict within the greater conflict.³³

The smallest section of the Manaqib is reserved for the mawali.³⁴ This comes as no surprise for, as mentioned earlier, their role as a distinct division of the army was limited. There were units under the name of mawali that fought in campaigns subsequent to the revolution, but these were relatively small and the occasions on which they served were extremely rare. It is also true that, in certain instances, the Turks are³⁵ called mawali; however those references are to the client-patron relationship that existed between them and the Caliphate, and not to a specific element of 'Abbasid military organization.³⁶

According to al-Jahiz, the forte of the client is personal service to his patron. His value is measured here, not in terms of military prowess, but in terms of those qualities that are highly valued in government servants: good advice, trustworthiness, patience, the ability to keep secrets, and so forth. The references to specific personalities in the text clearly indicate, once again, a link between Khurasan, the revolution, and the unsullied early era of "primitive" Islam. The Prophet himself appointed his adopted son and client Zayd b. Harithah to be the commander of the Muslim forces at Mu'tah and governor of every region that he subdued. He showed the same preference to Zayd's son Usamah by appointing him over the chiefs of the muhajirun and the great men of the ansar?¹ Al-Jahiz stresses the continuation of this tradition in 'Abbasid times. The mawla proclaims, "Our service [to our patron] is like that of sons [abna '] to fathers, and fathers to grandfathers."³⁸ Emulating the example of the Prophet in an earlier age, the 'Abbasid family treated its clients with trust and generosity. Even the black (client) was not despised because of his color.…The Caliphs entrusted (the political education of) their older children to clients and assigned to them special ceremonial functions, preferring clients (at times) to members of the ruling family (wa dhalika bihadrah min al-'umumah wa bani al-a 'mam wa al-ikhwah)³⁹

The ties of the clients to Khurasan and the revolution are reflected in the boast of the mawla. He invokes the names of Abu Muslim al-Khurasani and Abu Salamah, the two most prominent political operatives in the revolutionary leadership outside the ruling family.⁴⁰ He states that the mawali are found as well among the chiefs of the revolutionary agents (nuqaba') and cites several by name. The mawla can therefore assert that he shares the virtues⁴¹(manaqib) of the clients of the 'Abbasid propaganda (da 'wah) and those of the Khurasanis (wa nahnu minhum wa ilayhim wa min anfusihim). This is something that no Muslim, no true believer can deny. The Arab may stress genealogy, but the mawla can also claim (through his clientage) a genealogy that is both correct for the Arab, and worth boasting about by the non-Arab. He has earned this right by sharing the pride of the Arab, the bravery of the Khurasani, and the excellence of the banawi (sing, of abna').⁴¹ In short, the client, despite some distinctive features, is a man of no particular group, because he is a man of all groups and, indeed, all times. As such, he has partaken in the glories of early Islam, as well as those of the revolution and the 'Abbasid triumph that followed.

Among the early contingents of the army, only the enigmatic abna ' remain to be discussed.⁴³ Who, indeed, are the abna ' and what is their excellence that is shared by the client? The conventional wisdom concerning this group is confined essentially to several paragraphs in the new edition of the Encyclopedia of Islam (EP). The relevant passages indicate two pre-Islamic designations and one that is contemporary with the early 'Abbasids. The latter falls under the rubric abna ' al-dawla (sons of the revolution or dynasty). It is indicated that this term "applied in the early centuries of the 'Abbasid Caliphate to members of the 'Abbasid house and by extension to the Khurasani and other mawali who entered its service and became adoptive members of it. They survived as a privileged group until the 3rd/9th century, after which they were eclipsed by the growing power of Turkish and other troops." There is no indication what is meant by "members of the 'Abbasid house" nor how they are to be distinguished from the mawali who "became adoptive members of it." Similarly, although there is the suggestion of a military presence for the abna ' in relation to the Turks and others, there is no clear picture of their place within the wider framework of the 'Abbasid army.

The brief bibliography appended to the last entry includes the Manaqib (listed by its other title, the Fada'il al-Atrak) and two secondary sources.⁴⁴ A careful reading of the references fails to bear out the claim in EP regarding membership and adoption into the ruling house. Only the Manaqib speaks of a distinct social unit called the abna', and al-Jahiz links them, as a group, to service in the army. One cannot deny that an individual banawi or client may have become an adoptive member of the 'Abbasid house (whatever that implies); however, in the sources that are cited, there is not the slightest hint of any specific group bearing the name abna ' outside the military context.

In addition to the bibliography cited in EP, the author also relied on the "cryptic remarks of L. Massignon's Salman Pak. "⁴⁵ A subsequent investigation of the pertinent material in that work failed to uncover anything that resembled an institution to which one can affix the name abna '. Massignon does write of adoption. He attempts to show that Salman al-Farisi, the client of the Prophet, and Abu Muslim al-Khurasani, the client of the 'Abbasids, were both accepted into the ruling house (ahl al-bayt) according to the same formula (wa anta minna ahl al-bayt).⁴⁶ However, the evidence presented by Massignon speaks to that end and bears no relationship to the subject under review. The one reference to abna ' apparently cited by him is found in a late medieval source, where it is reported that Abu Ja'far al-Mansur established "the sons of Fars as leading men of their ['Abbasid] dynasty wa ja'ala abna' Fars rijalat dawlatihim)." Specifically mentioned are two well-known client families, the Barmakids and the Banu Nawbakht.⁴⁷ Even the most casual reading reveals that abna ' Fars rijalat dawlatihim cannot be confused with a military institution known as the abna' al-dawlah. There is a tendency among other modern scholars to use various expressions containing the words abna', da 'wah, and dawlah loosely, and even interchangeably. Such a practice only adds to the confusion surrounding the identity of this particular contingent.⁴⁸

The text of al-Jahiz is a convenient point of departure from which to sort out the facts. The term abna ' literally means "sons." One is impressed by the frequent repetition of references to family in the Manaqib and, in particular, to fathers and sons. The Khurasani proclaims, "We are agents and sons of agents… we are the men of the moat and the sons of the men of the moat.…Ours is the old and the new, the beginning and the end." The mawla serves his patron as a son serves his father, and a father, his father. Zayd b. Harithah, the client of the Prophet, is his adopted son. Isma'il, the son of Ibrahim, is the "father" of the Caliphs. The sons of Qatura bt. Maftun are the progenitors of the Turks. The point in these and every other example of this sort is to stress the inherent sense of unity and continuity in 'Abbasid society and the debt that it owes to the legacy of the past.

This theme is continued by al-Jahiz in his description of the abna ', hence, the explanation of the name. He notes that the banawi, like the Turk, is also considered a Khurasani. Because the identity of the abna' has never been firmly established in any published work, the author's remark that "the banawi is a Khurasani" may not, at first, seem informative.⁴⁹ By way of explication, al-Jahiz simply asserts, "the genealogy of the sons is that of the fathers, and the deeds of the fathers and grandfathers before them are reckoned to the account of the sons (abna')"⁵⁰ Although the statement has a sermonic quality and a morally edifying tone, it is, in itself, hardly revealing. Nevertheless, when it is juxtaposed against al-Jahiz's detailed description of the qualities of these regiments and considered together with incidental information obtained from the chroniclers, the author's vague observation takes on a specific meaning.

It is possible to establish a clear relationship between the abna ' and Khurasan and, hence, a link between the abna ' and the Turks. It is also possible to show that the abna ' are designated as sons of the dynasty (dawlah) and the 'Abbasid propaganda (da 'wah), not because they have been adopted into the ruling house, but because they are, in fact, the second generation of the revolutionaries that brought the 'Abbasids to power. Of the early contingents listed by al-Jahiz, only the abna ' do not invoke the events and personalities of the revolution in establishing their claims. That is because the abna ' are the Baghdad-based regiments of the imperial army, which were fashioned from descendants of the older revolutionary forces. They are mindful, indeed, boastful, of their origins as well as of their ties to both Khurasan and the revolution, but it is the new capital, Baghdad, and a subsequent martial history in which they glory.

The banawi proclaims that the root of his lineage (asI) is Khurasan, the same region from which the 'Abbasid revolution (dawlah) and propaganda (da 'wah) burst forth to bring about a new age. This reference to a new age is both explicitly stated and clearly imbued with messianic allusions.⁵¹ However, if Khurasan is the root, then Baghdad is the branch (far ') and is therefore called the Khurasan of Iraq. It is the (new) seat of the Caliphate, and houses the remaining veterans (baqiyah) of the ('Abbasid) propaganda and the sons (abna') of the 'Abbasid faction (shi'ah).⁵² If the banawf's fate is inextricably linked to the new Khurasan, it is justifiably so. He considers himself more rooted to the dynasty than his father, and more a part of it than his grandfather. He presumes to make this claim though the Manaqib specifically mentions his ancestors (aba') as leading the revolution. The expression "more rooted" (a 'raqu) appears to be a play on words to indicate more Iraqi, that is, Baghdadi, for it is the tie to the capital that gives the banawi his distinction.⁵³

The abna ' can thus be described as the foster sons of the Caliphs and the neighbors of the wazirs. They are born in the court of their kings and under the wings of their Caliphs. They follow in their rulers' footsteps and imitate their example. The banawi declares, "We recognize only them and will not be recognized except by them.…"⁵⁴ In a sense, the passage reaffirms the ties between the abna' and the seat of the Caliphate at Baghdad, but the words surely lend themselves to a still wider interpretation, namely, the personal relationship between the banawi and his sovereign.

A tradition preserved by the historian Tabari, under the year A.H. 163, immediately comes to mind in this connection. It is, coincidentally, the earliest evidence I have encountered that may refer to the abna' as a specific group. The chronicler reports that the Caliph al-Mahdi, anxious for his son, al-Rashid, to cut his teeth in a provincial campaign, is about to send him off against the Byzantines. The young prince, however, is seen in need of some mature advice. To that end, the Caliph ordered the secretaries (kuttab) of the abna ' al-da 'wah⁵⁵ to appear before him so that he might choose his son's companion from among them. The choice ultimately fell upon Yahya, the son of Khalid b. Barmak, the old revolutionary. Yahya was given far-reaching powers and, along with his father and two close relatives, accompanied the heir apparent.

Examining this account in relation to the passage from al-Jahiz describing the banawi as an intimate of the Caliph, P. Crone attempts to identify the abna' and establish their relationship to the ruling family.⁵⁶ There can be no argument that al-Jahiz's banawi stresses his loyalty and close association with the 'Abbasid house. For some, depicting the banawi as foster brother to the Caliph may be a mere figure of speech, but for Crone, it is nothing less than evidence of an institution. She argues that there were two groups drawn from the ranks of the old Khurasani army and their descendants. The first was the ahl-al-dawlah. The second, a smaller circle, were the abna' al-dawlah or abna' al-da'wah. Thus, the concept of dawlah provided two hierarchical ranks, which were in turn subdivided, according to Crone, thereby resulting in a still smaller group belonging to the circle of the ruling family itself.

It cannot be denied that various individuals and family groups who were the descendants of the old revolutionaries maintained close contacts with the 'Abbasid house and were to be numbered as members of the ahl al-bayt in its widest sense.⁵⁷ The example of the Barmakids perhaps stands out. above all others. Indeed, al-Rashid and al-Fadl b. Yahya suckled at the same breast, and the latter entered into marriage with the Caliph's sister, 'Abbasah, however calamitous the result.⁵⁸ Nevertheless, the isolated reference by the chronicler to the "scribes of the abna' al-da'wah," even when bolstered by the rather vague passage in al-Jahiz, is hardly reason to assume the existence of three well-defined institutions: ahl aldawlah, abna ' al-dawlah, and the ahl al-bayt.

The only certain claim that can be made for the abna ', before the great civil war between the brothers, is that they are Baghdadis descended from the Khurasanis who come to Iraq with the advance of the revolution. The first specific indication of this is a report that dates from the reign of the Caliph al-Rashid, under the year A.H. 187. An informer reporting the whereabouts of the Caliph's enemy Yahya b. 'Abdallah is interrogated by none other than al-Rashid himself. When asked his identity, the informer replies, "a descendant [grandson?] of the abna ' of this dynasty (rajul min a 'qab abna ' hadhihi al-dawlah). My place of origin [asl] is Marw [the revolutionary capital of Khurasan], but my place of birth is Baghdad [Madinat al-Salam, the capital of the established dynasty]." When the Caliph asked, "Is your domicile there [that is, at Madinat al-Salam]?" the man answered affirmatively.⁵⁹

Abna, as a generic term signifying military units, does not appear with any frequency until the outbreak of the conflict between the pretender al-Ma'mun and his brother al-Amin (A.H. 195).⁶⁰ They are then described as the dominant force in the army of the deposed al-Amin and were designated by the expression, "people of Baghdad" (ahl Baghdad).⁶¹ In connection with this loyalist army, the following expressions are found: the people (ahl) of al-Harbiyah,⁶² the abna' of the suburbs (arbad),⁶³ the abna ' of the Khurasani faction,⁶⁴ and the abna ' of Khurasan who have become "Arabicized" (muwallad),⁶⁵ that is to say, born in Baghdad and raised among Arabic speakers (one recalls that the Khurasanis of Arab descent spoke Persian perhaps better than Arabic).

This does not imply that the abna ' were the only element in the ahl Baghdad, but it appears clear, from a review of the conflict, that they were the dominant force in the Caliph's contingents—certainly they were the professional backbone of the army.⁶⁶ This was a role for which they were well suited, because they now represented the standing army of the dynasty, which was originally recruited in Khurasan during the revolution and was then transplanted to the military cantonments of al-Harbiyah, the northern suburb of the capital. The positioning of the army in what were intended to be exclusive military colonies to the north of the city administrative center explains some of the expressions connecting the abna' with the suburbs.⁶⁷

The initial successes of al-Ma'mun's Transoxanian army portended serious changes in 'Abbasid military society, particularly in the distribution of payments ('ata'). Baghdadis no doubt suspected that a victory for the pretender would enhance the position of the new Khurasani forces from Transoxiana that supported him.⁶⁸ It is no small wonder that the abna' fought with great tenacity during the early siege of Baghdad and continued to resist a change in status long after. This is strikingly reflected in al-Jahiz's description of the military skills of the banawi.⁶⁹ He is an expert at close quarter combat. When weapons are exhausted, the banawi grabs the neck of his foe. He knows how to stab with the knife and fend off the dagger. When surrounded, he and his comrades have the ability to respond and are therefore called the sons of difficult, that is, close combat in confined places (abna' al-mada'iq). The banawi boasts, "We know how to fight at the entrances to protective moats [khandaq] and at the bridge heads [a reference to the many masonry structures, qantarah, that bridged the canals of the western city, giving access to various quarters].⁷⁰ Bloody death confronts [those who oppose] us at the breaches in the protective wall [wa nahnu al-mawt al-ahmar 'ind abwab al-nuqab] and exhaustion in the narrow lanes [aziqqah].…We are masters of night fighting and kill openly in the markets and the roads.…We fight in the water as well as on land [a possible reference to the flotillas of al-Ma'mun that bombarded the city from the Tigris river]."

Other images of the banawi during the siege of Baghdad reappear in al-Jahiz's description of the qualities of the Turks. Unlike the Turks, who are cavalry, the abna' are infantry, whose lances are for protecting the entrances to moats and narrowly confined places (mada'iq). Their skills are suited to fighting in side streets (sikak) and trenches (?sujuri). These are all vivid references to the grim battle for Baghdad, when the defensive lines of al-Amin's forces were established along the natural barriers protecting the approaches to the Round City. The administrative complex was the last bastion of the beleaguered Caliph and his weary defenders.⁷¹

After the fall of Baghdad, the abna ' remained a highly volatile element with considerable military skill, so that the banawi of al-Jahiz claims, "All of Baghdad is ours. It is quiescent when we are quiescent, it is in turmoil when we are in turmoil."⁷² Although prudence dictated that the new authorities continue to pay their salaries (with all that this implied for the financial condition of the government), the abna' were not fully committed to the current regime. It would be some years before al-Ma'mun would set foot in Baghdad. His successor, al-Mu'tasim, was forced to abandon the city altogether and built a second capital sixty miles to the north at Samarra.⁷³ There he garrisoned a new imperial army with Faraghinah, Khurasanis and Turks from Transoxiana, Magharibah from the west, and an enigmatic elite corps known as the Shakiriyah. This army, however, never attained the sense of unity of the old Baghdad regiments, and the history of the Caliphate at Samarra, following al-Mu'tasim, was one of increasing turmoil.

By describing the Baghdad army, al-Jahiz recalls an earlier age, before the cumulative effects of the great civil war, and before the move to Samarra, for the imperial army that served the early Baghdad Caliphs represented, by and large, the integrated fighting force of a relatively stable and unified empire. This early 'Abbasid state made for a very attractive model in contrast to the turbulent reign of al-Mutawakkil, which was to end with the Caliph's assassination, the murder of al-Fath b. Khaqan, and, shortly thereafter, the eruption of still another civil conflict between elements favoring Baghdad and the partisans of the new capital. Had 'Abbasid society chosen to emulate this earlier model, it might have been spared the political chaos reflected in contemporary events. The Turks, because of their central political and military role, were the key element to any reconciliation within society. Although devoid of the messianic overtones that characterized early 'Abbasid propaganda, al-Jahiz's appeal for the acceptance of the Turks and the creation of a united society also called for a new era, one whose values were to be rooted in the early years of the 'Abbasid regime. One recalls that the 'Abbasids themselves originally established their legitimacy on the legacy of a still earlier time, the unadulterated Islamic age contemporary with the birth of Islam. The author's vision of a new society was, however, not destined to be realized.

Notes

¹ Al-Jahiz, Risalah ila al-Fath b. Khaqan fi manaqib al-Turk wa 'ammatjund al-khilafah, ed. A. M. Harun, 1:1-86; ed. G. van Vloten, 1-56. Harun's edition differs slightly from that of G. van Vloten and is cited throughout the text. See also the article of C. Pellat in Encyclopedia of Islam, 2nd edition, s.v. al-Djahiz; and his Le Milieu basrien et la formation de Gahiz. See particularly F. Gabrieli, "La 'Risala' di al-Gahiz sui Turchi," 477-83.

² Jahiz, 9-11, 63-64.

³ Ibid., 12-14.

⁴ Ibid., 31, 33.

⁵ Ibid., 74-75. The reference here is to the Qeturah genealogy in Genesis, 25:1-4. The account of al-Jahiz is probably filtered through traditions that are ultimately based on rabbinic material. In the Midrash, Qeturah and Hajar are sometimes regarded as the same. In any event, the sons born to Qeturah-Hajar were all idolaters and were thus sent by Abraham as far away as possible to the east. This may serve to explain the later connection with Khurasan, a land originally inhabited by unbelievers. For a list of the rabbinic sources dealing with Qeturah, see the notes of L. Ginsburg, The Legends of the Jews, 5:264-66.

⁶ The main medieval biographies are the Khatib (Cairo), 12:212-22; Yaqut, Irshad al-arib, 6:56-80. For a list of recent Arabic biographies, see Encyclopedia of Islam, 2nd edition, s.v. al-Djahiz, to be supplemented by A. H. Balba', Al-nathr al-fanni wa athar al-Jahizfihi (Cairo, 1969); M. 'A. Khafaji, Abu 'Uthman al-Jahiz (Cairo, 1965). Concerning the period in which the Manaqib was composed, see C. Pellat, "Gahiz à Bagdad et à Samarra," 48-67. See also C. Brockelmaan, Gesehiehte order Arahischen Litterafur, S, l:241ff.

⁷ See O. Pinto, "Al-Fath b. Haqan favorito di al-Mutawakkil," 133-49.

⁸ Concerning the problems of race and color in Islam, there is B. Lewis, Race and Color in Islam, esp. 15-18, on al-Jahiz. Lewis made use of K. Vollors, "Uber Rassenfarben in der arabischen Literatur," 1:84-95; and frequently quotes "an excellent German doctoral thesis" by G. Rotter, Die Stellung des Negers in der islamischarabischen Gesellschaft bis zum XVI Jahrhundert (Bonn, 1967). I have not seen Rotter's work.

⁹ Jahiz, 36.

¹⁰ Ibid., 1.

¹¹ Concerning the Arab tribes of Khurasan serving in the 'Abbasid army, see Sharon, '"Aliyat," 138-64, esp. 155ff.

¹² Wellhausen, trans. Weir, 493-94; Sharon, '"Aliyat," 30ff. As regards Arabs speaking Persian, see Tabari, 3/1:16, 50, 65. Note also Ibrahim al-Imam's command to Abu Muslim to "kill every speaker of Arabic [in Khurasan]." See Tabari, 3/1:25. The statement certainly does not mean, however, that every Arab in Khurasan was to be killed (not even those who spoke pure Arabic). Sharon, '"Aliyat," 32, 156-59, deals with this and similar traditions (Tabari, 2/3:1974; Dinawari, 358) and concludes that they actually signify a policy towards those Arab tribesmen who would not join forces with Abu Muslim. Regarding lughat Khurasan, see Sharon, 32: Jahiz, K. al-bukhala', 1:99.

¹³ See Sharon, '"Aliyat," 177ff.

¹⁴ The organization of an 'Abbasid protogovemment in Khurasan is treated in detail by Sharon, '"Aliyat," 176-85. A full list of the 'Abbasid operatives in Khurasan during the clandestine phase of the revolution is preserved in the Akhbar, 215-223. It supplements the partial list in Tabari, 2/3:1358. The 'Abbasid agents are almost exclusively from among the old settlers. The twelve highest ranking agents (naqib) were all from the early Arab settlements around Marw. Forty of the remaining fifty-eight were also from Marw, and there were, to be sure, old settlements elsewhere. The command structure of the army was largely made up of people on these lists, for example, Qahtabah b. Shabib and his sons al-Hasan and Humayd, Khazim b. Khuzaymah, Muqatil b. Hakim al-'Akki, al-Fadl b. Sulayman al-Tusi, 'Uthman b. Nahik and his relatives, and al-Musayyab b. Zuhayr. Concerning the early organization of the da 'wah in Khurasan, see Sharon, '"Aliyat," 105-117.

¹⁵ Among the twelve ranking agents (naqib) were the following clients: Abu al-Najm, 'Imran b. Isma'il, Abu 'Ali Shibl b. Tihman, Abu Hamzah 'Amr b. A'yan. Of the fifty-eight remaining nuqaba' of lesser rank, four are listed as mawali. Khalid b. Barmak, who is listed as a replacement (nazir) for a naqib, was a mawla, though he is not listed as such. It is possible that other figures on the lists could similarly have been clients, though there is no certain indication to that effect. Note that Khalid b. Barmak was responsible for the distribution of booty and later became the director of the military payroll. Various clients are also listed among the propagandists (du'at.)

¹⁶ Not all the Khurasanis sided with the revolution. Some forces remained loyal to the encumbent regime, and with grave consequences. During the siege at Nihawand, the Syrian troops surrendered with a guarantee of safety. Their Khurasani compatriots were, however, slaughtered. See Tabari, 3/1:6-8. Regarding the Khurasani identity of the army, note the juxtaposition of the terms ahl Khurasan and ahl al-Sham in the exhortations to the troops before the first critical battle against Nubatah b. Hanzalah and the main Umayyad contingents. See Tabari, 2/3:2004-2005. Qahtabah's main address, which plays on Iranian sentiments, may be, however, the invention of later times.

¹⁷ See Chapter IV, Section D, for a discussion of Ibn al-Muqaffa's Risalah fi al-sahabah.

¹⁸ The topography of military settlement at Baghdad is dealt with in detail in Chapter VIII.

¹⁹ Tabari, 3/1:1142; Ibn Tayfur (Cairo), 144-45; Kindi, 193, Maqrizi, 63. Sharon '"Aliyat," 283-87, stresses that al-Ma'mun had built up an army of non-Arabs from Transoxiana that served as the backbone of his regime vis-à-vis the earlier Khurasanis. The latter, largely of Arab stock, had been resettled in Iraq. The policy of removing the Arabs from the diwan was, however, not directed towards the Iraqis, who would not have been registered as Arabs in any case. The Arabs spoken of here are the tribal armies from Syria. Concerning 'Abbasid attitudes to this last group, see Chapter IV, Section D. Note that al-Mu'tasim also removed the Arabs from the diwan in Egypt while introducing the Turks in large numbers to the 'Abbasid army.

²⁰ Jahiz, 23-25.

²¹ Ibid., 14-23.

²² Ibid., 14-16.

²³ Ibid., 17-18.

²⁴ Ibid., 16-17. Note Chapter VI, which deals with the subsequent 'Abbasid decision to establish Iraq as the central province of the dynasty.

²⁵ Ibid., 15.

²⁶Ibid., 14. The relationship between the early agents and those of Muhammad b. 'Ali is not coincidental, but part of a well-conceived ideological campaign in which the 'Abbasids portrayed themselves as the heirs of the Prophet Muhammad. This is explicitly stated by the author of the Akhbar, 215ff., in his description of the organizational structure of the revolutionary underground Al-Jahiz's reference (p. 14) to seventy nobles (najib) is probably a reference to the group of Aws and Khazraj that met with Muhammad at al-'Aqabah. From this group of about seventy, the Prophet chose the twelve agents (naqib) to spread his message. The twelve ranking 'Abbasid agents were similarly chosen from a wider group of seventy. See Akhbar, 216. Note that the 'Abbasids also employed a group of seventy known as the du 'at (Akhbar, 221-22). To my knowledge, only Sharon, '"Aliyat," 105-107 and Crone, 132, have recognized the full connection between al-Jahiz's description of the Khurasanis, the Abbasid revolution, and the pristine era of early Islam.

Tabari, 2/3:1952ff. Note that al-Jahiz's Khurasani (p. 15) boasts "Our dress is known [that is, black] and we are the people of the black flags…(wa libasuna ma'rufwa nahnu ashab al-rayat al-sud).

²⁸ Jahiz, 15 (ghadhana bidhalika aba'una wa ghadawna bihi abna'and).

²⁹ Ibid., 20.

³⁰ Ibid., 21.

³¹ Ibid., 22. Others listed include: Abu Dawud Khalid b. Ibrahim, Abu 'Amr Lahiz b. Qurayz, Abu 'Utaybah Musa b. Ka'b, Abu Sahl al-Qasim b. Mujashi' al-Muzani. All of the individuals mentioned are to be found on the list of nuqaba'm Tabari, 2/3:1358, and the Akhbar, 215ff. (see also Text, ns. 13, 14, 15). Malik b. Tawwaf al-Muzani is listed by al-Jahiz, as one who functioned as a naqib though not one of them (yajri majri al-nuqaba' wa lam yadkhul fihim). He is nevertheless also on the list of the Akhbar, p. 218, that denotes the fifty-eight nuqaba' directly below the original twelve. Al-Jahiz also indicates that others from the second category were similarly Arabs but he does not list them by name. Almost the entire leadership of the revolutionary apparatus in Khurasan consisted of Arabs (see n. 14), as did the central element of the revolutionary army.

³² Jahiz, 23, also includes Musa b. Ka'b, who conquered al-Sind, and Muhammad b. al-Ash'ath, who conquered Ifriqiyah, among the Arabs. Both were Khurasanis. Both campaigns were conducted with Khurasani troops.

³³ For 'Abbasid attitudes towards the Syrians, see Chapter IV, Section D.

³⁴ Jahiz, 23-25.

³⁵ See Chapter IV, Section C.

³⁶ Tabari, 3/2:1516, 1539; Fragmenta Historicorum Arabicorum, 577; Mas'udi, Muruj (Beirut), 4:60, 77,90-91.

³⁷ Jahiz, 24.

³⁸ Ibid., 23.

³⁹ Ibid., 23-24. The specific indication of the paternal uncles and their offspring (al-'umumah wa bani ala'mam) may be a very subtle reference to the conversation between 'Abd al-Samad b. 'Ali and his great nephew al-Mahdi. See Tabari, 3/1:534, whose text is the point of departure for chapter IV. The relationship between the ruling family, the Caliph, the mawali, and the army is the central problem of early 'Abbasid political organization and is thus reflected throughout this study.

⁴⁰ For some reason, the passage concerning Abu Muslim and Abu Salamah is missing from Harun's edition of the text. It is cited here on the basis of G. van Vloten's Tria Opuscula, 14. One should remember the mawali were also well represented in the revolutionary apparatus in Iraq. See Sharon, '"Aliyat," 75-84.

⁴¹ Jahiz, 24-25. Mentioned are the ranking agents (ru'us al-nuqaba'): Abu Mansur, mawla of the Khuza'ah, Abu al-Hakam 'Isa b. A'yan, and Abu al-Najm 'Imran B. Isma'il. All three are listed among the ranking nuqaba by Tabari, 2/3:1358; and the Akhbar, 216.

⁴² Jahiz, 25.

⁴³ Ibid., 25-28.

⁴⁴ The secondary sources are Wellhausen, 556ff. and A. Mez, Renaissance of Islam, 155 ff.

⁴⁵ This information was kindly supplied to me by the author, Professor B. Lewis. The work in question is a short publication of the Sociētē des Ētudes Iraniennes et de l'Art Persan (7) entitled Salman Pak et les prēmices spirituelles de l'Islam Iranien.

⁴⁶ Ibid., 18.

⁴⁷ Maqrizi, 72 (according to Massignon, p. 50). Note Section E below, where the Barmakid, Yahya b. Khalid, is identified with a group called the abna' al-dawlah. The reference is, however, vague and certainly does not point to the military institution spoken of by al-Jahiz.

⁴⁸ Forand, "Slave," 59-60, who also follows Encyclopedia of Islam, 2nd edition, s.v.; and Crone, 131-34. Crone speaks of an 'Abbasid aristocracy based on early service to the revolution. Hence the abna' are correctly identified as the descendants of those Khurasanis who brought the 'Abbasids to power; however, her desire to establish clear distinctions between formal groupings of public figures seems to strain the evidence. There were, to be sure, abna' of varied rank and importance, but it appears that was a function of their general duties and background, and not of any specific subgroups in which they claimed association.

⁴⁹ I Lapidus, 371 correctly calls the abna' "the descendants of the supporters of the 'Abbasid revolution," without further comment or annotation. A fuller exposition is found in the unpublished dissertation of Crone, 131-34, and, more particularly, in that of Sharon, 288-90. The real key to unlocking the puzzle of the identity of the abna' was, however, provided by D. Ayalon, Reforms, 6ff. It is his unpublished paper concerning the military reforms of al-Mu'tasim that is the foundation of all subsequent studies. With his usual flair for reading a text, he isolated many traditions that make it clear that the term abna' does not simply mean sons, but alludes as well to a well-known military group in Baghdad. See below.

⁵⁰ Jahiz, 12.

⁵¹ Ibid., 25.

⁵² Ibid., 26.

⁵³ Ibid., 26.

⁵⁴ Ibid., 28.

⁵⁵ Tabari, 3/1:497-98. The apparatus of the text indicates that the term abna' is missing from MS A. Yahya al-Barmaki is also mentioned as one of the abna' dawlatihi (that is, Caliph) in Tabari, 3/2:703, sub anno 189.

⁵⁶ Crone, 132-34.

⁵⁷ Regarding this term, see Chapter IV, n. 6.

⁵⁸ The significance of this marriage is discussed in Chapter IV, Section C. Other examples of close contacts, that is, foster-brothers of the ruling house, are cited by Crone, 134.

⁵⁹ Tabari, 3/2:672.

⁶⁰ Note the brief reference to a contingent of abna' in Tabari, 3/2:732, sub anno 192. The abna' dawlatihi (that is, al-Rashid) are also mentioned in Tabari, 3/2:703, sub anno 189; however, in that context they are distinct from the army officers (min kufatihi wa ansarihi wa abna' dawlatihi wa quwwadihi).

⁶¹ See Tabari, 3/2:817, which indicates that this group was registered in the military roll and was not a makeshift force as the name may imply. The references to the ahl Baghdad during the civil war are numerous. Concerning their dominant role in the royalist army, see Ayalon, Reforms, 8-10, 36; Tabari, 3/2:831, 858, 865, 866, 898, 1001, 1006, 1013; Ya'qubi, Historiae, 2:532-33, 535, 547, 548.

⁶² Tabari, 3/2:825, 847, 848, 865, 984, 998, 1000, 1002, 1008-12; Ya'qubi, Historiae, 2:547, 548, 564. See n. 67 below.

⁶³ Tabari, 3/2:934, 935. For references to the people (ahl) of the suburbs, see Tabari, 3/2:849, 866, 867, 871, 872, 934, 936.

⁶⁴ Ayalon, Reforms, 6; Isfahani, Aghani (Beirut), 18:61; and the abna' of the Khurasani military cantonment (jund) in Aghani, 18:100. Ayalon points out that the connection to Iran is also found in the expression "the abna' are the sons of the dihqans [Iranian gentry]": Khuwarizmi, Mafatih al-'ulum, 119.

⁶⁵ Ibn Tayfur (Cairo), 80.

⁶⁶ Ayalon, Reforms, 9-10; Tabari, 3/2:824, 826ff. (a force of 20,000 abna' under the command of one of their own officers), 911-12, 922-23, 931, 932 (large force of abna').

⁶⁷ Ayalon, Reforms, 11-12, surmises that not all the suburban fighters were in fact abna'. "There must have been in Baghdad strong military and paramilitary bodies outside the descendants of the early Khurasanis.…Admittedly the sources do not make it possible quite often to draw a clear line between various fighting elements defending the 'Abbasid capital." Furthermore, "from the available data it is impossible to decide whether the abna' lived only in one quarter or were scattered all over the town." There is, however, reason to assume that the backbone of the fighting forces in al-Harbiyah were abna'. With regard to the topography of military occupation in Baghdad, see Chapter VIII.

⁶⁸ The composition of the imperial army in Khurasan on the eve of the civil war cannot be determined with certainty. In all likelihood, the Transoxanians were by far the predominant element. The provincial forces, largely Arab at the outset of the regime, were no doubt replenished by captives taken in campaigns to the east. See Tabari, 3/1:631, sub anno 169. Note also the advice of al-Fadl b. Sahl, the wazir of al-Ma'mun, that the latter resist his brother and seek alliances with the Transoxanians (Tabari, 3/2:815ff.). Concerning the problem of 'ata' see n. 73 below. See also D. Ayalon, "preliminary remarks on the Mamluk Military Institution in Islam," for some general remarks concerning Transoxiana.

⁶⁹ Jahiz, 26-28.

⁷⁰ Concerning the masonry bridges of Baghdad, see Lassner, Topography, 321, s.v. qantarah.

⁷¹ Jahiz, 53; 54. From the passages describing the military skills of the abna', Ayalon, Reforms, 37, concludes: "According to Jahiz there were amongst the abna' a sizeable body of excellent infantrymen whose weapons and tactics he describes.… I did not find any clear evidence corroborating al-Jahiz's contention.…" It is, of course, al-Jahiz's purpose to compare the abilities of the abna' with the superior fighting skills of the Turks, who specialized in cavalry combat. Nevertheless, there is no reason to suspect al-Jahiz's analysis. He is not referring to the abna ' specifically as foot soldiers, but to the tactics that were imposed upon them by the siege of Baghdad. For reference to the abna' as cavalry, see Tabari, 3/2:817, 911-12.

⁷² The reference here is to the continuing turmoil in the city after the war was over, particularly the events of A.H. 202-203, when the populace of Baghdad expelled the Caliph's governor, al-Hasan b. Sahl, and installed the counter-Calpihate of Ibrahim b. al-Mahdi. On the other side of the coin, note the remarks of Ibn Tayfur (Cairo), 80: "It was they [the abna'] who took up the fight against the Commander of the Faithful [al-Ma'mun], but then they obeyed him and the Caliphate was maintained in good stead owing to them." This last passage is taken from an apocryphal account in which the qualities of the abna' are contrasted with al-Mu'tasim's Turks and the Transoxanians ('ajam ahl Khurasan) of al-Ma'mun. The purpose of the discussion was to ascertain which of the three groups was the most proficient and courageous in battle. The disputed issue was resolved in favor of the abna'. The text is essentially the mirror image of the Manaqib encapsulated. It indicates that the abna' continued to function as a group and enjoyed support after the Turks were introduced in large numbers by the Caliph al-Mu'tasim. Note the attempts of the Caliph al-Muhtadi (A.H. 255-256) to create a counterweight to the Turks by using the abna' (Ya'qubi, Historiae, 2:604, 618-19). See also Ayalon, Reforms, 34-36.

⁷³ The antagonism of the local Baghdadis to the Turks was the major reason behind al-Mu'tasim's decision to move the Caliphate from Baghdad. This led eventually to the creation of a new capital at Samarra. When al-Musta'in later (A.H. 248) wished to free himself of the Turks at Samarra, he quite naturally opted to establish his government at Baghdad. For evidence that the abna' continued to be kept on the military rolls, see Tabari, 3/2:867, 1032-33, 1038, 1112. There are numerous references to the tremendous financial drain caused by the civil war. The shortages continued after the conflict had ended. See, for example, Tabari, 3/2:934ff; 977, 1012-13, 1014, 1016, 1028, 1030, 1038, 1065; FHA, All, 435, 438; Shabushti, 148.