The Dhiban Excavation and Development Project, the DEDP hereafter, investigates how a Middle Islamic community managed the economic and political pressures of Mamluk imperial rule and expanding “global” trade in a resource scarce, semi-arid environment. Tall Dhiban is the largest settlement on the Dhiban Plateau, a narrow slice of west-central Jordan confined by the Wadi al-Walla, the Wadi al-Mujib, the Jordan Valley and the Arabian Desert (Fig. 1). The site is positioned 64 km south of ‘Amman on the so-called King’s Highway, which connected the site with important Mamluk towns like Hisban, ‘Amman, and Karak. Dhiban receives between 250 and 400 millimeters of annual precipitation, making sustainable rain-fed agricultural just possible. A topographic survey has determined that the entire site is just over 12 hectares in area and 41 meters high on the north side, and made up of at least three major terraces (Fig. 2). The extent of the Middle Islamic settlement, however, appears limited to the central portion of the site, running east-west in an elongated and irregular elliptic shape. At present Middle Islamic Dhiban is estimated to have been approximately 5 hectares in area on the tall proper, with an additional “suburb” of some 1-2 hectares on the southern ridge next to the modern town.
Fig. 1. Tall Dhiban and contemporary Dhiban, looking south
(Photo Jamie Porter)
Fig. 2. Topographic map of Dhiban mainly emphasizing Middle and Late Islamic architecture. The gray squares represent excavation units
(Image by Andrew Wilson)
From the thirteenth through fifteenth centuries CE Dhiban was a substantial village in the al-Balqa region of the Mamluk administrative district of Damascus (Mamlakat Dimashq), but very near to the boundary (Wadi al-Mujib) with the district of al-Karak (Mamlakat al-Karak). Hence, Dhiban is likely to have been of some strategic importance, given the political competition and fluctuating administrative boundaries that characterized Mamluk rule in Jordan. Archival work by Bethany Walker indicates that the lands of Dhiban constituted an iqta‘ bestowed in 659 AH/1261 CE by Sultan Baybars on al-‘Aziz, the son of al-Mughith, an Ayyubid prince. Walker also notes that Dhiban had a mosque beside which was built a shrine, where two Mamluk amirs (Ibrahim ibn Manjak and his brother) were buried in the late fourteenth century CE. In other words, while not a major administrative center like Karak or Hisban, Dhiban was a prosperous town capable of providing both agricultural income and a desirable burial place for high ranking individuals during the thirteenth and fourteenth centuries CE.
Some indication of these higher-level connections may be seen in a large Middle Islamic building recently excavated on the north-east side of the tall by Jordan’s Department of Antiquities (Fig. 3). Built in the “piecemeal” style of the citadel (qusur) of Hisban, this irregular structure reused old building stones, columns and column bases extensively. The size and layout of the building suggests a non-domestic use, although, as the finds are still under study, it is difficult to say what this function might have been.
Fig. 3. Plan of the Middle Islamic excavated barrel vaulted building on Dhibans summit
(Photo by author)
The DEDP’s excavations on Dhiban’s summit has revealed a complex of interconnected rooms and courtyards that share many features with Middle Islamic vernacular architecture excavated in earlier campaigns at Dhiban as well as at Khirbat Faris on the Karak Plateau to the south and Hisban to the north. In particular, the core of this unit is made up of barrel-vaulted rooms with walls over one meter thick supporting the arches. These thick walls are actually constituted by two walls arching in opposite directions abutting each other for support with a rubble fill in-between. As at Khirbat Faris, the barrel vaulted rooms are trapezoid with the doorway on the widest end. However, at Dhiban excavated rooms are larger than the average given for Khirbat Faris (ca. 12 meters2) measuring ca. 20 meters2. Attached to these barrel-vaulted rooms are several communicating rooms that were roofed by sprung arches, although they do not have all the features of a classic “transverse arch house,” such as grain bins built in-between the springers. The stone masonry in the rooms with sprung arches is quite different from that in the barrel-vaulted rooms, and may represent a Middle Islamic reuse of earlier structures, as is the case elsewhere at Dhiban as well as at Khirbat Faris.
To simplify what is a very complex stratigraphic record, at least two post-construction phases that predate the final abandonment of these structures have been identified. The uppermost, Phase 2a, consists of ephemeral hearths and installations with no prepared floors, as well as piecemeal wall repairs and reconfigurations. This indicates that the buildings on the summit were gradually, rather than suddenly, abandoned, with a marked reduction in the intensity and stability of settlement in the final phase of room use. A similar pattern of abandonment in stages has been documented for the large Middle Islamic building excavated by the Department of Antiquities.
The earlier Phase 2b is marked by well-prepared floors and associated tabuns and bins. A barrel-vaulted room has yielded the most securely stratified evidence for this phase. In Phase 2b, the room was subdivided by low walls and installations constructed of large blocks laid somewhat haphazardly directly on top of a well-prepared surface. High-resolution techniques were used to excavate this surface; including point proveniencing of all surface finds, water-flotation of all floor sediments, and the collection of micromorphological, soil chemistry and phytolith samples. This mass of data is only just beginning to be analyzed, but already several items have been found such as bracelets and imported pottery that seem at odds with the initial interpretation of this building as a stable. Brief exploration beneath the Phase 2b surfaces has shown that at least one earlier Middle Islamic phase is to be found beneath Phase 2b.
Dating Phases 2a and 2b with precision is difficult, given current knowledge of artifact sequences in the Middle Islamic period. In 2005 a hoard of 30 copper coins was found in association with the foundation level of a Phase 2b cobble stone surface. These thirty copper coins have the fabric and size consistent with the copper coins (fals; pl. fulus) minted in areas of Egypt and greater Syria in the Ayyubid (567-648 AH/1171-1250 CE) and Mamluk (648-923 AH/1250-1517 CE) periods. They are not well preserved and only four definitive identifications have been made; all of them Ayyubid (three from al-Malik al-Kamil Muhammad [615-635 AH/1218-1237 CE] and one from al-Malik al-'Aziz 'Uthman [589-595 AH/1193-1198 CE]. Only one of the coins seems to be Mamluk, but has not yet been identified with a known Mamluk type. Because the material culture associated with this hoard is clearly late Middle Islamic in date (i.e. Mamluk), and the coins are very heavily worn, it is assumed that most of this hoard was in circulation for more than a century before its final curation.
The favourable shape of the calibration curve during the latter part of the Middle Islamic period means that radiocarbon dates have proven more useful than numismatic dates. At present the beginning of the final phase of building use, Phase 2a, is radiocarbon dated to the first half of the fifteenth century CE with a two-sigma range of 1409-1445 CE. The calibrated dates from Phase 2b are more dispersed, but clearly predate the fifteenth century CE. This evidence makes the Phase 2a “squatters” occupation contemporary with the well-known late fourteenth and early fifteenth century fiscal and political crises within the Mamluk Empire. Because piecemeal site abandonment was already underway in the first half of the fifteenth century, it is presumed that Middle Islamic Dhiban was largely abandoned before the beginning of the sixteenth century. In this regard, it is interesting to note that Dhiban does not appear in the Ottoman tax registers (defters) of the sixteenth century CE.
As is typical of Middle Islamic sites in Jordan, green and yellow lead-glazed, hand-made geometric painted, and undecorated hand-made and wheel-made vessels dominate the ceramic assemblage. While petrographic analysis of this pottery is not yet complete, all of these wares appear to be of regional origin. More unusual is the abundance of relief-moulded lead glazed wares decorated with Quranic verses. These wares are thought to have been manufactured in the vicinity of Jerusalem, and are common at Karak and Hisban, but more rare further north in Jordan. Much less common, but still present, are examples of stonepaste wares (also called fritware/faience), generally presumed to be manufactured in the region of Damascus and in central and northern Syria (Fig. 4). Small quantities of so-called “sugar pots” have also been excavated. However, their limited abundance and secondary find contexts do not allow much to be said regarding Dhiban’s role in the sugar industry, which played such a large role in Jordan during Mamluk rule.
Fig. 4. Decorated stone-paste wares from one surface in the barrel vaulted building
(Photo by Coleen Morgan)
Overall, Dhiban fits the general pattern described by Bethany Walker of a “boom and bust” expansion of agricultural settlement in Jordan from the thirteenth through fifteenth centuries CE. Walker links this expansion and contraction of settlement with Mamluk investment in the cash-cropping of sugar cane in the Jordan Valley, as well as the use of waqf endowments to provide “tax sheltered” agricultural investments in Jordan for Mamluk elites. This period of investment came to an end during the fiscal and political crises of the late fourteenth and early fifteenth century. As Walker notes, the local impact of these large-scale trends was quite variable, with areas that had been transformed into private waqf endowments often continuing to prosper through the transition to Ottoman rule in the sixteenth century. For many regions, however, the withdrawal of Mamluk state investment led to a marked decline in settlement. Dhiban would seem to fit into this later category. The town existed prior to the thirteenth century but expanded considerably under Mamluk administration. Between the thirteenth and fifteenth centuries CE, Dhiban developed as a prosperous local centre with connections to regional and international trade networks, as well as imperial elites. Middle Islamic Dhiban also hosted some small-scale non-agricultural production. The evidence suggests that Dhiban was abandoned gradually after a period of marked decline in the first half of the fifteenth century.
Precisely how and why Middle Islamic Dhiban expanded and was abandoned are interesting questions that deserve further research. In contrast to Hisban, no evidence for earthquake damage at Dhiban has been identified, despite numerous quakes being documented in the wider region over the century between 1341 CE and 1458 CE. Dhiban’s lack of a perennial water source and its reliance on cisterns would have made it more susceptible than most towns to periods of drought. Such droughts are certainly offered as causes of agrarian decline in Mamluk documentary sources. The impact of more intensive agro-pastoral production on soil erosion and fertility are also likely to have been severe at Dhiban. However, the relevant proxy data to examine these issues are still being collected, it would be premature to credit the site’s abandonment to earthquakes, drought or soil degradation. The shifting of the regional capital of al-Balqa from Hisban further north to ‘Amman in 757 AH/ 1356 CE may have impacted Dhiban. However, as of yet, there is no evidence for dependency between the two sites. Similarly, although the withdrawal of Mamluk state investment from Jordan coincides closely with the abandonment of Dhiban, it is not yet possible to show specific evidence for such investment and divestment in the archaeological record. Regarding the initial expansion of the site there is little that can be said at this time, as the earliest Middle Islamic phases are only now coming to light. In other words, much remains to be done in terms of both excavation and analysis. The preliminary results suggest that such effort will be worthwhile and that Dhiban will soon prove a key site in understanding the historical dynamics of agrarian expansion and “collapse” during the Mamluk administration of the Levant.
Benjamin W. Porter
Near Eastern Studies Department
University of California, Berkeley
References: see www.dhiban.org for DEDP publications and updates