Page 76

may have been integral to the political and commercial activities of the Takrur polity (Robert-Chaleix and Sognane 1983; Robert-Chaleix 1994).

For the other two great polities of the western Sudan, Gao (known to Arab authors as Kawkaw) and Mali, our knowledge remains largely historical. The two archaeological sites known at Gao are the old town (Gao ancien), north of the existing town, and Gao Sané, several kilometers to the east on the opposite side of a channel (which likely flowed perennially in the late first millennium) leading north to the Tilemsi Valley. Numerous brick structures are visible on the surface of both sites, some of which were excavated by Colin Flight of the University of Birmingham to reveal a confusing sequence of building and razing episodes (Insoll 1996 provides a very useful summary). Tim Insoll’s archaeological work in 1994 included excavations at both Gao ancien and Gao Sané. Although very limited in scale, his excavations have provided concrete details on the appearance of trans-Saharan trade goods, including ceramics and glass, in the tenth or early eleventh century and insights into the process of Islamization. Gao apparently was a part of trade networks extending to North Africa and Spain.

In the thirteenth century, all the regions discussed above became consolidated within the hegemony of the Empire of Mali, and little archaeological information is available for Mali. The Polish research project at Niani in Guinea (Filopowiak 1979) is noteworthy for its regional perspective, even though claims that the site was the capital of Mali visited by Ibn Battuta in the fourteenth century are problematic. There is, in fact, little compelling evidence from the excavations for occupation deposits dating to this period (McIntosh and McIntosh 1984). Excavations in the Arab Quarter, for example, produced mostly material dated to the seventh to eleventh centuries a.d. and a notable lack of imported goods.

However, from that material we can get a glimpse of Mali society during the period of small villages and petty chiefdoms described by eleventh- and twelfth-century Arab authors in the area called Malal. In this regard, Filopowiak’s surface survey of fifty sites (iron-smelting sites, funerary tumuli, small villages) located within a four-square-kilometer area around Niani offers the potential for understanding an evolving settlement system. Of particular interest are small stone tumuli with rock-cut shaft-and-chamber collective graves similar to graves in the southwestern part of modern Mali, which indicate strong first-millennium cultural connections within this whole area that became the political heartland of the Empire of Mali in the thirteenth century. Unfortunately, the systematic study and description of the Niani pottery necessary to establish the chronological relationships of these hinterland sites has not yet been accomplished.

Conclusions

From a colonial, medievalist paradigm concerned mainly with recovering objects, inscriptions, and architecture that testified to outside influences at work in the empires of western Sudan, archaeology has moved since the 1970s to documenting local contexts and components of change. The opening of archaeology in the former French West Sudan to broad international collaboration has resulted in a variety of more regionally based research projects concerned with the establishment of the basic chronological sequences and site inventory data that are the foundation of all sustainable archaeological interpretation. As comparative regional data on settlement systems become available for the first time from different parts of western Sudan, important theoretical issues for archaeology as a whole emerge from the diversity of trajectories leading to the social complexity that may be detected there (McIntosh, S., 1999a, 1999b, 1999c). As a consequence, the empires of western Sudan, which have been largely marginal to mainstream archaeological theorizing about the rise of complexity until now, may in the future play a more prominent role.

Susan Keech McIntosh