Project
al-Ṯurayyā Project

Project Directors
Maxim Romanov, Universität Hamburg
Masoumeh Seydi, Aga Khan University, London

Project URL
https://althurayya.github.io/

Project Reviewer
Osama Eshera, University of Maryland

Project Overview

Maxim Romanov

The al-Ṯurayyā project currently includes a gazetteer (al-Ṯurayyā Gazetteer, or al-Thurayyā Gazetteer) and a geospatial model of the early Islamic world. Both parts of the project are still under development. The gazetteer currently includes over 2,000 toponyms and almost as many route sections representing data georeferenced by Maxim Romanov from Georgette Cornu’s Atlas du monde arabo-islamique à l'époque classique: IXe-Xe siècles (Leiden: Brill, 1983), which the author developed on the basis of the Bibliotheca Geographorum Arabicorum (BGA), a series of critical editions of classic geographical texts written by several of the most famous Arab geographers (produced by M.J. de Goeje between 1870 and 1894 and published by Brill). The al-Ṯurayyā project inherits all the possible issues with the data from Cornu‘s Atlas. The current iteration of the project serves primarily as an interactive reference that allows users to check locations of settlements, get basic information about them (both technical and descriptive — the latter, in Arabic, comes from a medieval gazetteer Rawḍ al-miʿṭār of al-Ḥimyarī, matched computationally), and to model shortest routes within the network as well flood networks. The data of the gazetteer (CSV/JSON) is available on GitHub.

There are two distinct parts of the project: 1) data creation, which was largely limited to the careful digitization of the information from the Atlas and 2) technical development, building on components of ORBIS: The Stanford Geospatial Network Model of the Roman World. Specifically, we were trying to find the best way of visualizing the world of Islam experimentally without falling into the trap of modern, border-based maps. We found an efficient way of displaying major medieval provinces through color-coded networks of settlements classified as belonging to specific provinces.

The main motivation behind the project was to create a gazetteer of the medieval Islamic world that would facilitate computational analysis of premodern Arabic texts, primarily chronicles and biographical collections. In a way, it can be considered a gazetteer supplement to the Open Islamicate Texts Initiative (OpenITI) project, whose goal is to develop a unified corpus of Islamicate texts for computational research in Islamicate Studies. Together with Sarah Savant (Aga Khan University, London) and Matthew Thomas Miller (University of Maryland, College Park), Romanov is a co-principal investigator of the OpenITI project.

The core team of the al-Ṯurayyā project is Romanov (data and concept) and Masoumeh Seydi (technical development). However, it must be noted that the very first version was developed as a result of collaboration between Romanov (data and concept) and Cameron Jackson (concept and technical development), who developed the Interactive Map of Classical Islamic World as a part of her BA honors thesis at Tufts University.

The project, which has been undertaken without funding, was primarily developed for addressing internal research problems. The development was also rather ad hoc, with features being implemented when we had time, inspiration, or needed to solve some practical research questions. The project was a sort of sandbox for Masoumeh Seydi‘s work on her PhD thesis at the University of Leipzig on modeling geographical information from premodern texts. The project was never really meant to be public, but very soon after its publication it attracted attention of scholars in the field of Arabic and Islamic studies (Islamicate Studies), who keep finding it useful in both their teaching and research as an interactive geographical reference.

Project Review

Osama Eshera

The literary and scientific corpus of the premodern Islamic world includes a vast and varied body of historical geographical data that is well-suited to be studied through the aid of digital methods. Yet this domain has been largely overlooked by researchers in digital humanities. In this light, al-Ṯurayyā Project, led by Masoumeh Seydi and Maxim Romanov, is a welcome and pioneering effort.

Al-Ṯurayyā provides a digital rendering of the twenty maps of Georgette Cornu’s 1983–85 Atlas du monde arabo-islamique à l’époque Classique. The project’s data comprise approximately 2,000 toponyms and almost as many “route sections” (i.e., edges connecting nodes in the toponym network), all drawn from Cornu’s work. Some toponyms were subsequently matched with (possible) corresponding entries in a primary source, namely, Ibn ʿAbd al-Munʿim al-Ḥimyarī’s (fl. 15th cent.) al-Rawḍ al-miʿṭār fī ḫabar al-aqṭār (The Fragrant Garden on the Story of the Regions).

The toponym and route data are in GeoJSON format and are displayed on a modern world map using the Leaflet JavaScript library on al-Ṯurayyā’s web-interface (whose code is available publicly). The interface also facilitates use of a pathfinding tool that highlights the “shortest” and “optimal” paths between an origin and a destination, with or without specified intermediate destinations.

Inasmuch as al-Ṯurayyā is an experiment in producing a machine-actionable abstraction of historical geographical data, it is useful to scholars interested in “distant” or computational analysis. The 2017 article written by al-Ṯurayyā’s authors is similarly beneficial and deals with the related problem of parsing various kinds of geographical data from narrative, historical texts.

But for those interested in historical or geographical analysis in a “close” sense, al-Ṯurayyā will have a rather limited benefit. This is due to certain constraints in the sources and methods used in the project, which I shall discuss in turn.

Choice of Sources

Al-Ṯurayyā’s main source, Cornu’s atlas, is based on Bibliotheca Geographorum Arabicorum, a collection of nine classical Arabic geographical texts edited by M.J. de Goeje. First published in the late-19th century, de Goeje’s work was excellent for its time but hardly meets the philological and textual-critical standards of modern historical research. Moreover, while Cornu’s work may be beneficial in some cases, in many cases it is, by Cornu’s own admission, provisional, incomplete, speculative, and imprecise. Having relied exclusively on Cornu’s work for the project’s core data, al-Ṯurayyā is three degrees removed from the primary sources of geographical data (i.e., the extant medieval manuscripts) and can only be as accurate as Cornu. Indeed, al-Ṯurayyā’s data includes inconsistencies both in how toponyms are spelled and in the proposed identifications with Ḥimyarī’s al-Rawḍ al-miʿṭār.

The decision to match Cornu’s toponyms with entries in Ḥimyarī’s work is itself curious, especially since al-Ṯurayyā is described as “a gazetteer and a geospatial model of the early Islamic World” (emphasis added). Ḥimyarī, who was active in the 15th century, can hardly be classified as belonging to the “early” period of Islamic civilization. It is true, of course, that Ḥimyarī’s work encompasses many places and regions that fall within the geographical range of “early” Islam, but his is a second-hand, posterior account. Moreover, Ḥimyarī was not himself a geographer or scientist of note; to the contrary, he drew nearly all the geographical content of al-Rawḍ al-miʿṭār from earlier sources. Why should Ḥimyarī’s work be preferred when the canonical works of Islamic geographical writing are available and offer direct and more reliable data for the early period?

Methods

Al-Ṯurayyā’s documentation is minimal and does not always provide a clear picture of and sufficient justification for the project’s methods. How were the geographical coordinates and route sections determined? Cornu placed the toponyms on a map and drew route paths between some of them, but Cornu did not provide explicit values for the longitude and latitude of each toponym, or for the route paths. So, al-Ṯurayyā would have applied some inferential procedure to arrive at the numerical values that now populate its GeoJSON objects. Including information about that procedure would be helpful for users. 

Further, project creators might address how these procedures were validated. Over centuries, premodern Islamic geographers calculated and recalculated the geographical coordinates of thousands of localities. The data are well preserved, and some scholarly efforts have already gone toward organizing the massive corpus into a structured form (most notably, E.S. Kennedy and M.H. Kennedy’s 1987 Geographical Coordinates of Localities from Islamic Sources). Did al-Ṯurayyā use such historical data to validate the values inferred from Cornu’s atlas? Similarly, premodern Islamic geographers also provided first-hand data of the actual networks of their time and the travel distances between localities. Were such primary sources used to validate the route segments inferred from Cornu’s atlas? After a thorough examination of the material and software published by al-Ṯurayyā, it may seem that the data has not been validated against the primary sources. This raises the question of whether al-Ṯurayyā’s data is too far removed from the primary sources to be relied upon for “close” historical research.

Nevertheless, al-Ṯurayyā’s main contribution remains significant, namely, presenting provisional approaches for parsing geographical texts, tagging geographical entities, and aligning them across the dataset.