Academic journal article
By Insoll, Timmothy
Antiquity , Vol. 72, No. 276
Surprisingly little archaeological research has been conducted within the well-known city of Timbuktu, which is listed as a World Heritage site on account of its historical significance. In spite of surveys of the standing monuments (Pefontan 1922; Mauny 1952), and of parts of the surrounding region (McIntosh & McIntosh 1986), the city remains largely untouched. In contrast the history of Timbuktu is well known (see for example Herbert 1980; Saad 1983). To rectify this imbalance a preliminary survey was conducted within the city limits of Timbuktu in November 1996. The aims of this were twofold; to look at both the standing monuments, about which a fair amount is known already, but also to delimit possible areas for future archaeological excavation. Thus all the quarters of the old city were covered on foot, as were most of the outer suburbs and cemeteries which ring the city [ILLUSTRATION FOR FIGURE 1 OMITTED].
Unfortunately, the survey was not as simple as originally envisaged. Primary factors were modern occupation and shifting sands obscuring archaeological deposits. 'Peaks' or 'islands' of archaeological deposits appeared though the sand deposits and areas of modern housing, providing snapshots into the settlement history of Timbuktu. The issue of contemporary occupation overlying and obscuring archaeological deposits is obviously not unique to Timbuktu, but the degree to which 'ensablement', or the build-up of wind-blown sand, occurs in Timbuktu is dramatic. Within the Sankore mosque, for example, the recent accumulation of sand was attested by the literally disappearing doorways [ILLUSTRATION FOR FIGURE 2 OMITTED]. Aerial photographs and satellite images would not be much help in overcoming this problem, except at the outer edges of the city, where they have proved their worth in the survey of the now un-occupied hinterland of Timbuktu (McIntosh & McIntosh 1986). The problem of archaeological visibility evident at Timbuktu is not such a factor at comparable urban centres on the Niger Bend. At Gao, for example, where two seasons of excavation and survey have thus far been completed, the 'medieval' settlement is clearly discernible, and it has proved possible to begin to delimit the different quarters of the town (Insoll 1995; 1996; 1997).
Delimiting the precise areas of primary occupation in Timbuktu based upon survey evidence alone has therefore proved impossible because of the problems already outlined. However, we have some information on early settlement gleaned from the historical sources. According to local tradition and an indigenous chronicle written in the 17th century, the Tarikh al-Sudan, Timbuktu was founded c. AD 1100 as a seasonal pastoral Tuareg/Masufa camp (Es-Sadi 1900: 20-21). No data was found which can either confirm or deny this, but a variety of evidence was recorded which has suggested possible areas of early occupation which might prove informative once excavation commences later this year.
Two primary areas of archaeological material were recorded; the first in the northeast of the city around the Sankore mosque (locally called Casbah Marocaine), one of the 'islands of archaeology' of approximately 80 m length x 40 m width/[ILLUSTRATION FOR FIGURE 1 OMITTED]). The second was west of the Djinguereber mosque, where an extensive area of mounds was found covered with a mixture of modern rubbish and earlier archaeological material (locally called 'Azalai') [ILLUSTRATION FOR FIGURE 1 OMITTED]. Dating of these two areas is at present far from certain, but a sherd of southern Chinese Celadon dated to the late 11th-early 12th centuries (Rosemary Scott pers. comm.; Insoll 1998), was found at Azalai (on a mound given the identifying code 'B') [ILLUSTRATION FOR FIGURE 3 OMITTED]. The mix of material scattered across the surface of the so-called Casbah Marocaine precludes any serious attempt at dating settlement in this area. Indication of 'medieval' occupation was suggested by a fragment of a glass bracelet which was collected from the surface of the site. …