For a number of reasons, the size of the Arabic textual tradition still remains a mystery to historians. The proliferation of digital libraries of premodern Arabic texts and the development of methods of textual analysis offer, however, a new means and opportunity to obtain our first data-driven view of this tradition. The paper will begin with an evaluation of the Arabic textual tradition through the analysis of the Hadiyyat al-`arifin, a bibliographical collection from the early 20th century. Building on the famous Kashf al-zunun of Hajji Khalifa (d. 1656), it offers sufficient information on about 8,800 authors and 40,000 titles to analyze a variety of dimensions of book production in the Islamic world. While showing that the rate of overall book production across the Islamic world steadily increased over time, the data also shows that regionally book production peaks during the reign of major Islamic dynasties that flourished in different periods across the Islamic world. The second part of the paper will focus on an overview of the textual interconnectedness of premodern Arabic texts that are now available in a number of digital collections (al-Jami' al-Kabir, Shamela, ShiaOnlineLibrary). A comparison of their contents with the bibliographical data from the Hadiyyat al-`arifin shows that these digitized texts---despite their impressive volume of about 7,800 unique texts (1.2 billion words) unevenly distributed across 14 centuries of Islamic history---are just a fraction of the entire tradition. Our computational analysis further shows that nearly all these texts are interconnected with each other through an extremely high volume of shared quotations, strongly suggesting that these digitized texts---despite representing a number of forms and genres---in fact form a clear sub-tradition. The final section of the paper will turn to one of the major nodes in this sub-tradition---"The History of Islam", a 50-volume biographical collection written by the Damascene scholar al-Dhahabi in the 14th century. After characterizing al-Dhahabi's extensive network of sources, the paper will examine commonalities among passages that are being quoted by al-Dhahabi and those that are being omitted by him, thus attempting to identify what may constitute the core elements of that generate this interconnectedness and create a textual tradition.

Arabic authors frequently made use of past works, cutting them into pieces and reconstituting them to address their own outlooks and concerns. Digital technology can now detect, measure, and trace these repurposings, shedding entirely new light on how books were produced and received, how the Arabic textual tradition developed over time, and what its priorities and blind spots were. This technology, arising out of sequence alignment in genetics, the removal of duplicates in web-crawls and search engines, and anti-plagiarism software, is absolutely cutting-edge. One of the most innovative uses of it will be to map networks in new and more precise ways. In particular, historians are now able to discern much more easily transmission of texts across boundaries set by geography, sect, and genre. This paper focuses on text transmission across geography. It begins with a discussion of text reuse detection methods and specifically, how, building on our bibliographic data, our algorithms and visualisations can generate evidence for the transmission of texts. With reference to the presenters' web application, I highlight issues of methodology, and outline what our data can and cannot reveal. The heart of the paper focuses on the Arabic textual networks to which Khurasan belonged in the late 10th and early 11th centuries. In cultural history, this is an important region and period for the rise of Persian literature (e.g., Ferdowsi’s Shah-nameh, completed in 1010), as well as for the development of a vibrant court-centered literary culture in Arabic (as studied by Louise Marlow and Bilal Orfali, among others). I focus especially on the context of Abū Manṣūr al-Thaʿālibī (d. 1038), a man whose career benefited from the financial support of Sāmānid, Khwārizm-Shāh, and Ghaznavid rulers, and whose copious writings in Arabic provide extensive evidence for the world view of a major man of letters. I address three questions: first, based on our text reuse data, what can we say about the likely Arabic corpus of works available to al-Thaʿālibī? Secondly, by considering this corpus and citations from it by other, contemporary authors throughout the Arabic-writing world, what can we learn about the geographic spread and network connections of Arabic literature in al-Thaʿālibī’s lifetime? Finally, by examining the contents of specific passages, can we say anything particular about literary tastes in Khurasan, versus other regions? By addressing these questions, I will bring a new perspective to literary culture at this fascinating point in cultural history.

Modern scholars of Islamic history tend to differentiate between Sunnī and Shī‘ī historical works. Leaving aside the general skepticism surrounding the entirety of early Muslim historiography, Shī‘ī texts are often dismissed as products of hagiography composed solely for polemical purposes. This attitude dates back to the seminal work of Orientalist scholars such as Julius Wellhausen and (especially) Leone Caetani. These (and other scholars) offered negative assessments of the reliability of Shī‘ī reports on a range of topics from the succession to the Prophet to the moral rectitude of the Umayyad dynasty. In most cases, any pro-‘Alid report was considered suspect and relegated to the status of polemic. My paper challenges the marginalization of Shī‘ī historical sources through a network analysis of a collection of Twelver Shī‘ī biographies (entitled Kitāb al-irshād) composed/compiled by one of the school’s seminal theologians (al-Shaykh al-Mufīd [d. 413/1022]). This relatively small work includes apparent references to a number of 3rd/9th century sources that are no longer extant but are cited by contemporaneous and near contemporaneous Sunnī sources. I am particularly interested in assessing the degree to which al-Mufīd (and other Shī‘ī authors) participated in a common and intertwined historical tradition. Were there any clear distinctions between Sunnī and Shī‘ī sources? Or were authors navigating a shared reservoir of interdependent historical reports? It may be possible to begin answering these kinds of questions through new digital tools that offer a wider and more comprehensive portrait of the historiographical tradition. By interrogating the interconnectedness of a large corpus of sources, these tools provide insight into the methods that informed the compilation of historical works. They also have the potential to shed light on the epistemological concerns that governed historical authorship. My paper attempts to harness this potential in an effort to elucidate the commonalities of Sunnī and Shī‘ī historical writing.

Although the genre of ‘nasab’ (genealogical literature) was a familiar one to the medieval Muslim historian it appears to have had very fluid boundaries. In this paper I will show how digital approaches can be combined with more traditional forms of text criticism to help us delineate this genre, and in so doing further our understanding of a type of historical writing that has been largely neglected in modern studies of Islamic historiography. In order to carry out this analysis I have compared passages taken from a number of works where they report on the same genealogical connections. It will be shown that the traditional means of doing this (i.e. searching manually through indexes by theme or name) can only ever give us a partial answer because the historians in question were drawing from many unreferenced sources. A more systematic approach is therefore needed and this is provided by text reuse detection methods which can scan for instances of common passages and their variants across an extensive digitised corpus of Arabic literature. The results are twofold. First, the combined methodologies create an extensive database list of examples showing variant ways of transmitting similar historical information and this allows us to consider the distinct contexts of transmission. It will be shown that nasab literature is well-placed to allow us to discuss genre more widely given that its long-term popularity transcended both sectarian and geographic divides; it also helps that the information preserved within it is relatively stable. The second result is that we will be able to compare the digital and traditional methodologies in light of each other. Where the traditional methodology is shown to have shortcomings this has an impact on what we think we know about other types of historical writing that have yet to be systematically searched for text re-use (e.g. chronicles and tabaqat works). Conversely, where the traditional methodology is shown to uncover connections missed by automatic text reuse detection, we will be able to consider ways to improve our digital methods.