A local historian's debt to al- Tabari: The case of al-Azdi's Ta'rikh al-Mawsil.

• Author: Robinson, Chase F.
• Position: Muhammad ibn Jarir al-Tabari, Yazid ibn Muhammad al-Azdi - Critical essay

In 1969 Paul Forand published in this journal the results of his reading of Yazid ibn Muhammad al-Azdi's Ta'rikh al-Mawsil, which was at this point only available to him on a film of the Chester Beatty manuscript. (1) Although chiefly interested in reconstructing some of the administrative history of the city during the late Umayyad and early Abbasid period, Forand concluded with some historiographic comparisons, principally between al-Azdi and his more celebrated Mosuli successor 'IZZ al-Din Ibn al-Athir. He also had something to say about al-Azdi and his contemporary, al-Tabari. Noting that al-Azdi's isnads often begin with the same authorities cited by al-Tabari in his Ta'rikh al-rusulwa'l-muluk, only to diverge from them later on, he proposed that "neither was copying from the written work of the other. The significance of this, for the study of early Muslim historiography, has yet to be studied in detail." (2) In fact, aside from A. Habiba's comments in the introduction to his edition of the Ta' rikh, which appeared shortly after Forand's article was published, (3) nothing of detail has been said of al-Azdi and his work, including his handling of isnads and his sources more generally. For all that the work has been praised (4) and also been put to use in several studies on early Islam, (5) it has not been properly studied. In this respect it is scarcely unique. Much of a general nature has recently been said about the emergence of Islamic historiography, (6) but a great deal of detailed work remains to be done. What follows is a modest installment.

It is regrettable that the relationship between al-Azdi and al-Tabari, which was first recognised by Forand, has been left unstudied, and this for at least three reasons. First, we shall shortly see that the Ta'rikh al-Mawsil was written in the 320s or very early 330s, and that al-Azdi was familiar with al-Tabari's work; it is thus an extremely early witness to the reception of al-Tabari's text--indeed much earlier than the sources that are customarily pressed into service to improve our understanding of the Ta'rikh al-rusul wa'l-muluk, e.g., Miskawayh, Ibn Asakir, Ibn al-Athir, and Ibn Khallikan. (7) Second, since al-Azdi was writing in the decades following al-Tabari, his Ta'rikh can say something about the reception of al-Tabari's Ta'rikh among those who immediately followed the great master. That al-Tabari's history was immensely significant we can all agree; but as to precisely how he became so significant there is no clear consensus. (8) Third--and returning to Forand's insight--al-Azdi fre-quently drew on the same authorities tapped by al-Tabari, but whose works are for the most part now lost, such as Abu Ma'shar (170/786), Abu Mikhnaf (157/774), al-Haytham ibn 'AdI (207/822), al-Mada'ini (around 228/843), and 'Umar ibn Shabba (262/878). The Ta'rikh al-Mawsil can thus throw some light on an early stage of the historical tradition, which al-Tabari's great work, among others, largely eclipsed. In what follows I therefore return to the problem first identified by Forand and since ignored, intending only to lay the foundation for more detailed work. More specifically, I intend to date the history; to suggest that its relation to al-Tabari's is more complex than Forand thought; to illustrate some of the problems that a comparison of al-Azdi and al-Tabari raises; and finally to point out some of the problems it can help us solve.

Two more introductory comments are in order; the first concerns al-Azdi's work in general, and the second the limits of our evidence. First, in terms of provenance, one can identify two categories of material in al-Azdi's Ta'rikh: local Mosuli material, and what I shall call "imperial" material. Generally speaking, the local material predominates in the later sections of the history, particularly as the tribal 'asabiyat between the Azd and Hamdan accelerated in the final quarter of the second century. (9) In terms of form, this material is heterogeneous in the extreme: lists (of qadis and governors); documents (letters of appointment and iqta's;wasiyas); genealogical material; rijal material; and, finally, narratives of local history, generally outfitted with isnads composed of local authorities. By contrast, the "imperial" material, which predominates in the Umayyad period, is considerably more homogeneous. Although here too there are some letters and obituaries of non-Mosulis, most of it consists of isnad-equipped akhbar, the majority of which are which are credited to the great second-and third-century compilers on whom al-Tabari also relied. This contrast between local knowledge of local affairs and what I have called the "imperial" tradition is one that al-Azdi himself seems to acknowledge, (10) and can be explained easily enough. As a high-born qadi of the city, al-Azdi had direct access to a range of local material, at least some of which might fairly be called archival; (11) but as an otherwise obscure qadi in a modest provincial city, his knowledge of most non-Mosuli history was necessarily imported and very much bookish. It is with the bookish al-Azdi that I shall be concerned here.

It must also be conceded that comparing al-Tabari and al-Azdi is not without some problems. Much more work needs to be done on the textual tradition that produced the manuscripts on which our modern al-Tabari editions are based, (12) and this includes trying to identify which branch of the tradition al-Azdi himself knew. Moreover, only the middle third of al-Azdi's Ta'rikh al-Mawsil survives, and this in a single manuscript. (13). If collating parallel traditions were not tedious enough, constraints such as these make the task of distinguishing between mere variants, on the one hand, and real divergent traditions on the other, extremely difficult. (14). Finally, since al-Azdi was writing local history in a relatively lean style, much of what concerned al-Tabari in his world history obviously had little or no appeal. If it is fair to characterize at lease some of al-Tabari's method as "cut and paste"--a view that is unlikely to survive a close comparison of his work with the surviving parts of one of his principal sources (15)--then so too can al-Azdi's; but to compare the Ta'rikh al-Mawsil with the Ta'rikh al-rusul wa'l-muluk is to conclude that al-Azdi was especially active with his scissors: the bits are smaller, and much has been left on the cutting-room floor. All of this said, the problems are not insuperable: our understanding of al-Tabari and his history improves all the time; al-Azdi's editor has managed to re-assemble a jumbled manuscript with some success; (16) and garbled names aside, the text--that is, the printed edition used in conjunction with the manuscript--can sustain the comparative scrutiny that I propose. I begin with problems of sequence and dating.

I

Although he provided no explicit evidence, Forand was correct to assert that al-Tabari, unlike his famous Mosuli abridger, Ibn al-Athir, did not copy from al-Azdi's work.' (17) Not only is the material characteristic of the Ta'rikh al-Mawsil--local information transmitted by local tradents--absent in al-Tabari, (18) but there is no reason to think that the work had even been composed when al-Tabari put down his pen in 302/915. (19) Since the first and last third of the Tarikh have been lost, dating its composition may at first seem difficult but an account attributed by al-Najashi (d. 450/1058) to a certain Salama ibn Dhaka' al-Mawsili can provide us with a safe terminus post quem of 321/933; for according to Salama (fl. late fourth/early fifth century), his teacher, Ali ibn Muhammad al-Shimshati, undertook a continuation (tammama) of al-Azdi's history, beginning in year 322/934. (20) Year 334/945, which al-Dhahabi tentatively dates as al-Azdi's death, may stand as an approximate terminus ante quem. (21) The latter date is corroborated by the work's inclusion in al-Mascudi's introductory comments to his Muruj al-dhahab, the existing version of which was completed by 336/947 at the latest. (22) Whether the work was finished in the later 320s or early 330s is currently impossible to say.

So Forand's instinct proved correct on at least one count: al-Tabari did not--indeed could not--rely on al-Azdi' s finished Ta'rikh. (There is no reason to think that al-Tabari had access to a draft or to lectures of al-Azdi, if such existed.) But what of the reverse: was Forand right in arguing for al-Azdi's independence of al-Tabari? With an edited and respectably indexed text, we are now better equipped than was Forand, and here it turns out that he was mistaken. (23) In the preserved second section of the Ta'rikh al-Mawsil--428 pages of printed text--al-Azdi refers to al-Tabari at least four times, as Abu JaTar, (24) Muhammad al-Amull, (25) and Muhammad ibn Jarir. (26) In all these cases al-Tabari is cited directly, without intermediary, and with the verb anbaani ("He reported to me"). It is impossible to know if this modest rate of explicit use was representative of the work as a whole; indeed, the sample is too small to suggest much at all. Considering that al-Azdi narrows his focus on local matters during the second half of the second century, one might speculate that al-Tabari would have appeared more frequently in the now-lost first section; but since we also know that al-Azdi had access to al-Tabari's sources for the period covered by the first part, (27) perhaps he did not.

How al-Azdi was familiar with al-Tabari's material must also be inferred. As used by al-Azdi, anbaani usually describes the direct transmission of material by contemporaries; but there are exceptions, such as when he uses the same term to introduce material attributed to al-Madaini, who had died almost a century before him. (28) Perhaps here al-Azdi's Zwischenautoritat was omitted by a negligent scribe, (29)or perhaps al-Azdi simply copied the term from an earlier, written source; in either case...