In this series of blog posts, we have argued for the imperative to rethink writerly culture in ways that allow for a more meaningful exploration of al-Tabari’s (d. 310/923) sources. We have also advanced a more general argument about the nature of writing in the period in which al-Tabari lived. The question that we should be asking is not whether a particular piece of writing constituted a published, redacted book but rather what forms writing took, what evidence we have of these forms in the works that survive, and how ideas about writing itself might have contributed to the evolution of the tradition – including the loss and oblivion of many of its layers.

We have observed that al-Tabari rarely refers to titled works as his sources. Interestingly, Yaqut also fails to mention titles for al-Tabari’s sources in his discussion of his famed predecessor’s sources. Instead, he refers to ‘the book’, kitab, of so-and-so. For example, Yaqut shares a report noting that al-Tabari failed to use the kitab of Muhammad b. al-Saʾib al-Kalbi or those of Muqatil b. Sulayman or Muhammad b. ʿUmar al-Waqidi when composing his tafsir (Yaqut’s source makes the point that al-Tabari did use these books in writing up his taʿrikh). Yaqut uses the same general terminology – the kitab of so-and-so – when describing the sources that al-Tabari did use when he wrote about taʾrikh, siyar and akhbar.1

Al-Tabari’s own writings were recognised early on as important works in their own right. His students copied them, and their titles are recorded in the early biographical works. Yaqut reproduces a comment he says he found in a volume of al-Tabari’s Tafsir in the handwriting of al-Tabari’s student Abu Muhammad ʿAbd Allah b. Aḥmad b. Jaʿfar al-Farghani (d. 362/972–3):

Dear ʿAli b. ʿImran and Ibrahim b. Muhammad, I have given you an ijaza (ajaztu la-ka) for the following books which I heard from Abu Jaʿfar al-Tabari (God have mercy on him!): the book on tafsir, called Jamiʿ al-bayan ʿan taʾwil ay al-Qurʾan and the book, Taʾrikh al-rusul wa’l-anbiyāʾ wa-l-muluk wa-l-khulafaʾ, both of which I heard (samiʿtu); for al-Qaṭʿayn min al-Kitab, which I did not hear, but which I obtained through an ijaza; the book treating the dating of transmitters (rijal), called Dhayl al-mudhayyal 2

Al-Farghani is remembered as a major transmitter of al-Tabari’s works. He wrote a continuation of the Taʾrikh, and his son Abu Mansur Ahmad b. ʿAbd Allah in turn wrote a continuation of the continuation.3 In the comment quoted by Yaqut, he tells two people – perhaps his own students – that he has received al-Tabari’s two major works in person through dictation. The al-Qatʿayn min al-Kitab was transmitted through a mechanism that granted him permission to transmit the work without having heard it read aloud.

The term al-qaṭʿayn is a curious one. Yaqut subsequently quotes Abu al-Hasan b. al-Mughallis who describes the contents of the Taʾrikh, and says the term al-qatʿayn refers to the qatʿ of the Umayyads and the qatʿ of the ʿAbbasids.4 Ibn al-Nadim (d. 385/995) also refers to qatʿan, when describing the Taʾrikh. He says that al-Tabari added the qatʿan to the Taʾrikh as well as other contents that ran up to the year 302.5 This may well suggest an original conception of the Taʾrikh that ended prior to the Umayyad period.

For al-Tabari, and also his immediate readers, the Taʾrikh therefore might have been viewed not as a continuous presentation of history, but rather as discrete segments added together and comprising it. Readers would know to read the Taʾrikh differently depending on where in the narrative they landed. Segments on the Prophet’s life would be read differently than those on the Umayyads not just because of the topic, but also because of the manner of assembly.

What authorized breaking books up? A significant factor would have been knowledge of their assembly, as was the case with the Taʾrikh - both of the individual reports and also of the larger segments (e.g., the qatʿayn/qatʿan).

Citing al-Tabari

We should not assume that al-Tabari’s words were conveyed onwards in the form of published works, in the sense described by Schoeler. Nor should we assume that a person who passed on information from one of al-Tabari’s books necessarily had access to the whole work; we know that later authors did not, and even early users of his oeuvre may not have had the complete texts at their disposal. In the long term, the manuscript evidence points even more strongly to the partial transmission of al-Tabari’s works. For example, there is a not a single complete manuscript witness of the Taʾrikh. Our text reuse data shows, furthermore, just how extensively the works were excerpted and reused at an early date and for centuries afterwards.

Citations of al-Tabari point to a fragmentary reception of his writings, where subsequent generations passed on his works without direct reference to his works.

We have just begun work on a further data set that documents the afterlife of al-Tabari’s writings using manuscripts, text reuse data and citations of his works and his name. Within our data, we can already see that the haddathani/haddathana formula was used to refer to subsequent persons’ connections back to him in a way that suggests his words passed into notebooks much like his teachers’ words had done. Like al-Tabari, these people also seem to have copied his material into notes that were then recycled into other works, some of which have come down to us. Although more work needs to be done to understand the usages of the haddathani/haddathana formula in specific later works and contexts, it is noteworthy that it continued at all. Its persistence seems consistent with what scholars have described as the ongoing prestige of aurality.6

To locate citations of al-Tabari, we created a regular expression that searched our database of machine-readable files of books written in the roughly 260-year period between al-Tabari’s death in 310/923 and the death of Ibn ʿAsākir in 571/1176. Our search targeted the following variant forms of Abu Jaʿfar Muhammad b. Jarir al-Tabari’s name combined with haddathani/haddathana and its variants (a method similar to that which produced our data set on al-Tabari’s three books).

محمد بن جرير

أبي جعفر الطبري

أبو جعفر الطبري


أبا جعفر الطبري

أبي جعفر محمد بن جرير الطبري

أبو جعفر محمد بن جرير الطبري

أبا جعفر محمد بن جرير الطبري

This search did not catch all occurrences of al-Tabari’s name in isnads written after his lifetime, because we excluded certain surface forms (such as Abu Jaʿfar and Ibn Jarir) that would have produced too many false hits. Even with the more restrictive formula, we found 315 instances of al-Tabari’s being cited in an isnad across sixty-six works.

As an example, consider this long isnad from Ibn ʿAsakir, which appears in his Taʾrikh madinat Dimashq. Ibn ʿAsakir invokes al-Tabari’s name within his biography of al-Walid II b. Yazid (r. 125-6/743-4) in the context of a report regarding the so-called third fitna, during which Yazid b. al-Walid (r. approximately six months in 126/744) seized the caliphate by overthrowing his cousin al-Walid II:

I [Ibn ʿAsakir] read aloud to (qaraʾtu ʿala) Abu al-Wafaʾ Huffāz b. al-Hasan b. al-Husayn from ʿAbd al-ʿAziz b. Ahmad, who said that ʿAbd al-Wahhab al-Maydani told us that Abu Sulayman b. Zabr said that ʿAbd Allah b. Ahmad b. Jaʿfar told him that Muhammad b. Jarir al-Tabari told him:

Ahmad b. Zuhayr told us that ʿAli b. Muhammad told him, from ʿAmr b. Marwan al-Kalbi, that Yaʿqub b. Ibrahim b. al-Walid told him that when Yazid b. al-Walid went out (kharaja, i.e. rebelled), a mawla belonging to al-Walid went out on a horse …7

Both the isnad in the second paragraph and the report can be traced to al-Tabari’s Taʾrikh,8 with some precisely reproduced phrasing (as well as some differences, which do not concern us here).9 But the Taʾrikh itself is not mentioned. This is typical of Ibn ʿAsakir’s citations: even when referring to people who wrote books, he very often cites the people, not the books.

In both Ibn ʿAsakir’s text and the Taʾrikh itself, al-Tabari’s direct informant is named as Ahmad b. Zuhayr (d. 279/892), who, according to Gilliot, wrote an histoire (taʾrikh) of the caliphal period and attended the lessons of Musʿab b. ʿAbd Allah al-Zubayri (d. 233/848 or 236/851) on the genealogy of the Arabs.10 Ibn al-Nadim attributes four works to Ahmad b. Zuhayr (Kitab al-Taʾrikh, Kitab al-Mutayyamin, Kitab al-Iʿrab and Kitab Akhbar al-shuʿaraʾ).

In this process, books were rendered doubly invisible. First, al-Tabari copied material from Ibn Zuhayr’s writings (which he obtained directly from the author) into his own Taʾrikh. Then, what al-Tabari wrote was converted back into note format for several generations (as witnessed in the isnad). Finally, the material was absorbed into Ibn ʿAsakir’s Taʾrikh madinat Dimashq, with the identities of its previous book-form carriers – the writings of Aḥmad b. Zuhayr and al-Tabari – erased. In Ibn ʿAsakir’s work, these other writings have become invisible as books. Each of the people involved in this and similar chains of writing has gone to the trouble of creating a work. But it is as if each person afterwards has reserves credit for authoring a book only to himself.

Concluding Remarks

The data for these blog posts was collected over several weeks during the Covid-19 pandemic. It was an extraordinary time to be buried in spreadsheets, doing this sort of intensive work. Perhaps it would have seemed less interesting in a different time. We did not set out to write eight posts, only two. But as we immersed ourselves in the data, we came to see the potential for gathering more data to address different aspects of the questions outlined at the start. In the months to come, we hope to publish our work in a peer-reviewed journal with additional reflections, data and corrections.

The method we are developing can be used with any text that is well structured (including having regularly marked paragraphs) and whose author cites sources according to a regular pattern. The Transmission Chains table can be filtered and sorted to address questions we have not raised (though keeping in mind the problems outlined in Post 2). The Name Normalisation table can be reused and expanded to, for example, study citation patterns in works written earlier or later. Scholars often work on a single book by a well-known author. The OpenITI corpus provides access to many less-known works, and digital methods applied to the entirety of an author’s surviving oeuvre can often shed more light on the author’s working methods than can close scrutiny of a single work.

Our data provided us with a complex new perspective on how al-Tabari worked, but we want to emphasise how much can also be learned by far more simple explorations. We encourage you to download the corpus through Zenodo and, as a matter of routine, to search individual works and the corpus broadly. Doing so will reveal hitherto unseen patterns, including in the structures of various works and in the frequencies of citation partners (as in our case). There is now a vast number of texts at your fingertips that you can not only read and search but also annotate and return to again and again on your own computer. We will be adding more book files to the OpenITI corpus in the future. If you do not see something that you would like to see in the corpus, please be in touch: if you can get hold of the work in question in a PDF file of reasonable quality, you can now create a machine-readable file of it through our Optical Character Recognition pipeline.

Finally, Digital Humanities is a field built on teamwork. We worked intensively together, even if separated by geography (Sarah in the United Kingdom and Masoumeh in Germany). We relied on the contributions, feedback and questions of a team of KITABis, whose input has been invaluable, and on those of reviewers of the project, especially Letizia Osti and Konrad Hirschler, who read these posts and improved them with their comments. We welcome your thoughts, too.

  1. Yaqut, Muʿjam al-ʿudabaʾ, OpenITI, Shamela0009788-ara1.mARkdown, ms. 1812. 

  2. The quotation goes on to list other works of al-Tabari. Yaqut, Muʿjam al-ʿudabaʾ, OpenITI, ms. 1801. 

  3. F. Rosenthal, ‘Al-Farg̲h̲ānī’, in Encyclopaedia of Islam, 2nd ed., []{.ul}

  4. Wa-innama kharaja dhalika ila al-nas ʿala sabil al-ijaza ila sanna arbaʿ wa-tisʿin wa-miʾatayn. Yaqut, Muʿjam al-ʿudabaʾ, OpenITI, ms. 1815. 

  5. Ibn al-Nadim, al-Fihrist, OpenITI, 0385.Fihrist.Shia003355-ara1.mARkdown, ms. 285, search term القطعان. On the term qatʿan/qatʿayn, see Savant, Cultural History of the Arabic Book (forthcoming). 

  6. See, for example, Letizia Osti’s forthcoming monograph on al-Suli. 

  7. Ibn ʿAsakir, Taʾrikh madinat Dimashq, OpenITI, JK000916-ara1, ms. 25095. 

  8. Al-Tabari, Taʾrikh al-rusul wa-l-muluk, OpenITI, Shamela0009783BK1--ara1.completed, ms. 3267. 

  9. Al-Tabari, Taʾrikh, OpenITI, ms. 1132. 

  10. Claude Gilliot, ‘La formation intellectuelle de Tabari (224/5–310/839–923)’, Journal asiatique 276 (1988), 203–44, at 220–1.