Cross-cultural After-Life of Classical Sites: including Arabic names in the Greco-Roman atlas
With thanks to the Pelagios Commons team and especially the expert and generous SIG chairs, we’re happy to announce the CALCS Project has been funded with a small development grant. The aim of this project—which serves as a pilot for a much larger investigation into the afterlives of sites we think of as classical—is to add information about mediaeval Arabic and Ottoman, and modern Arabic and Turkish, names to sites in Pleiades.
The Pleiades Gazetteer, probably the most useful authority of any kind of the Ancient and Late Antique Linked Web, as Pelagios collaborators do not need reminding, is based on the Barrington Atlas of the Greek and Roman World, the most definitive print atlas of classical antiquity. Although constantly under improvement, Pleiades is already as close to a comprehensive list of known Greco-Roman places and names as we have ever had. The majority of names in the gazetteer are those in use in Anglo-Saxon classical scholarship: either the classical names that were in the atlas (Colonia Claudia Ara Agrippinensium; αἱ Ἀθῆναι), or the English/Italian rendering of a modern or other variant place name (Naples, Cirene).
This misses a huge part of the picture, obviously. Shortly after the end of what we call the Classical Period, and before the Latin Middle Ages or Renaissance, the Islamic world took control of huge swathes of the former Empire, and held parts of it for many centuries; Iberia and Sicily were Arabic-speaking for hundred of years, and most of the Balkans were part of the Turkish-speaking Ottoman Empire for centuries after the fall of Byzantium. Today, almost half of the former Roman Empire is still made up of countries whose first language are a form of Arabic. Many sources, from mediaeval maps and manuscripts, through Renaissance scholarship, to modern references in academic and popular works in North Africa and the Near East, will be inaccessible to the sort of computational study that Pleiades and Pelagios enable if we do not take into account the Arabic and Ottoman names for sites such as Alicante (أليكانته), Messina (مسينة), Thessaloniki (سلانیك) and Leptis Magna (لَبْدَة). More importantly, this one-sided recording of historical names runs the risk of (inadvertently) perpetuating the myth of European monoculture, the idea that there is an uninterrupted and pure line—politically, geographically, linguistically, genetically—from the grandeur of antiquity to the enlightenment of modern Europe, to which no one but white, Christian, Indo-European speaking people contributed. The inclusion of data from Arabic documents (in Pleiades) and the maps themselves (in Recogito) also helps to highlight the contribution to modern cartography (including some startlingly topological maps) from the Arabic tradition.
With this pilot project we hope to show that this deficiency can be addressed through several parallel approaches. We’re focussing on a few small but promising regions in the first instance: Sicily, Cyprus and Cyrenaica, although especially for the more automated parts of the process we may catch places from other parts of the Mediterranean world (especially Arabic Spain and Ottoman Greece) in the overflow.
Firstly we are tracking down early maps and other manuscripts written in Arabic, such as al-Idrisi’s Nuzhat al-mushtāq fi’khtirāq al-āfāq (the so-called Tabula Rogeriana commissioned by King Roger II of Sicily) and the anonymous Kitāb Gharā’ib al-funūn wa-mulaḥ al-ʿuyūn (known as the Book of Curiosities), held by institutions such as the British Library and the Bibliothèque nationale de France. From these we shall record the Mediaeval Arabic names of sites that have presence in Pleiades due to their classical origins, with full attestation of their historical source. Where possible, high resolution, open licensed scans of such maps will also be made available through the Recogito platform. At the same time, we shall take advantage of modern atlases and tourist guidebooks in Arabic to record the modern Arabic names for ancient sites in the same way (although maps in such resources are unlikely to be open licensed in a way that would allow us to use them in Recogito).
Secondly, we shall attempt to harvest a bunch of low-hanging fruit, potentially at rather large scales. For example, there are many places in the Geonames.org gazetteer that have been identified with Pleiades ids by the Edinburgh Geoparser and are mapped in the “Pleiades+” dataset (now maintained by Ryan Baumann at Duke); many such places will have a name in Arabic recorded in the Geonames data. Likewise, if we can identify Wikipedia pages corresponding to Pleiades place records (some 6,000 of these are identified in Pleiades itself; others may be extracted by searching for categories relating to ancient sites, Roman settlements, etc., and hand-matching them), the associated Wikidata records will often provide the Arabic or Turkish titles from the relevant Wikipedia versions—itself a more reliable indicator of language usage than the etymology or language section at the top of an English-language entry. Other projects that collect large numbers of (ancient and modern) names of ancient sites—such as al-Thurayya (Maxim Romanov’s digital gazetteer of the Classical Islamic World), the Name Changes of Settlements in Greece database from Pandektis, and Jeremy Johns’s Arabic Documents of Norman Sicily project—can be used to generate lists of Arabic and Ottoman/Turkish name variants. All of these lists will be hand-verified by the project team and our collaborators before being entered into Pleiades, to prevent the circular introduction of self-ratifying errors based on linked data projects rashly borrowing and extrapolating from one another.
Thirdly, we shall encourage the annotation and georesolution of placenames from Arabic maps and manuscripts in Recogito, primarily through setting students at three DH workshops we are teaching this Fall (at least one of which—Sunoikisis Digital Classics—involving Arabic-speaking students) to help with the transcription of the Tabula Rogeriana (in Arabic and Latin-script Turkish versions).
We look forward to sharing our results with you in future updates, and at the Linked Pasts conference in December.
(Institute of Classical Studies, University of London)