FROM THE CARAVANSERAI TO THE ROAD: PROPOSAL FOR A PRELIMINARY RECONSTRUCTION OF THE SYRIAN ROAD NETWORKS DURING THE MIDDLE AGES CINZIA TAVERNARI AbstrAct Many of the most important roads of medieval Bilād al-Šām were dotted with caravanserais, sheltered stopovers whose construction really developed in the region during the Ayyubid and Mamluk periods. 1 These buildings can be considered as evidence of the passage of a road in a precise spot within a given period and consequently provide useful information to map out the routes which crossed the region at that time. The purpose of this article is to use data provided by wayside caravanserais to compare and integrate with available data on the routes in order to provide a more comprehensive view of the Levant communications network from the end of the 12 th century to the beginning of the 16 th century, particularly trying to highlight its evolution and changes. IntroductIon: AIms And methods Sources such as chronicles (e.g. Ibn Ṣaṣrā 1963), travel books (e.g. Ibn Ğubayr 1953; 1994; Niccolò da Poggibonsi 1968, inter alia), chancellery manuals (e.g.al- Qalqašandī 1913-19; 1923) and geographical dictionaries (e.g. Yāqūt al-Ḥamawī 1979-86) or treaties (e.g. Ibn Šaddād 1956; 1982; 1984) undoubtedly constitute the primary resource we can draw upon to identify routes in use during the Ayyubid and Mamluk periods, and studies of the Levant roads have traditionally relied on the historical data provided by these materials (see e.g. Dussaud 1927; Elisséeff 1967). Surprisingly, caravanserais only rarely seem to have been considered important elements in the reconstruction of routes, 2 though these buildings constitute the least elusive proof of a road passing by within a given period. Caravanserais represent spatial-temporal tags 1 For discussion on this point see e.g. O.R. Constable (2003: 75-76) who says that ‘Whatever earlier numbers, it is clear that the quantity of commercial funduqs and khāns increased substantially under the Ayyūbids’. 2 See the opinion of R. Hillenbrand et al. (1992: 55) who state that ‘No serious attempt has been made to coordinate the data on medieval trade in the area with the facilities available at surviving khāns, or the location of khāns with the major trade routes as known from literary sources’.