East African Archaeology: Foragers, Potters, Smiths, and Traders

6. Early Ironworking Communities on the East African Coast: Excavations at Kivinja, Tanzania

6 Early Ironworking Communities on the East African Coast: Excavations at Kivinja, Tanzania Felix Chami This chapter discusses excavations at Kivinja on the central Tanzanian coast and their implications for understanding Early Iron Age inhabitants of the East African coast.This site is the first Early Iron Working (EIW) site to be found on the East African coast and one of the earliest to document extra-African trade relationships. Evidence of trade at this site strongly challenges long-held assumptions that non-African colonists founded complex cities of the Late Iron Age. Rather, these cities developed from indigenous origins. The excavations at the site of Kivinja on the Tanzanian coast are significant for East African Iron Age archaeology since this site, along with two others—the Limbo site (Chami 1992) and the Kwale Island site (Chami and Msemwa 1997a, 1997b)—provides the first evidence of the Early Iron Age cultures that existed from AD 1 to 500 on the East African coast. Prior to the discovery and excavation of these sites, only Late Iron Working (LIW) sites with standing ruins were known.All three of these sites contain beveled and fluted pottery, which is associated with farming peoples. Until 1999, Kivinja and other sites contained the earliest-known remains of foreign trade goods, including glass, beads, pottery, and Roman beads dated between the last century BC and the 5th century AD (Chami 1999). Since then several Late Stone Age/Neolithic sites have been found on the island of Zanzibar and Mafia, which have trade goods dating back to the 8th century BC (see Chami 2002b).These numerous EIW sites on the islands, littoral , and hinterland of the central Tanzanian coast paint a picture of a culture that shared its pottery style with inland groups but was also in trade contact with the Roman and the Near Eastern worlds.These sites confirm the coastal city of Rhapta, mentioned by the Romans as a trading partner (Casson 1989). The evidence of very early trade contacts on the part of indigenous Africans with hinterland origins, or at least connections, has profound implications.To understand them, we must place the finds in the context of the history of East African coastal archaeology. Archaeological Background The archaeology of the East African coast began with the work of Kirkman (1964) at Gedi and of Chittick (1962) at Mafia and Kilwa. Their interpretations of these sites vindicated long-held speculations that a Persian and Arabian empire had existed on the east coast of Africa from the 9th to the 15th centuries AD,when Islamic groups were historically known visitors to the EastAfrican coast.Chittick (1975b) argues that 9th-centuryArab immigrants founded the coastal towns and established an Islamic culture there. Chittick was the doyen of East African coastal archaeology until his death in the early 1980s. In his many archaeological works, especially that on Manda (Chittick 1984), he argued firmly that Africans played no part in coastal civilization .Arab immigrants, whose cultural and economic relationships were with the Middle East rather than with Africa, built the Swahili towns. Black Africans were hinterland peoples who arrived on the coast from the Congo forest much later in time (Oliver 1965). Soon, however, this theory became difficult to reconcile with accumulating data, in particular with EIW coastal sites, such as Usambara (Soper 1967) and Upper Tana (Phillipson 1979), that contained pottery styles similar to those of the interior. These sites yielded EIW and Triangular Incised Ware (TIW) pottery, which is associated with interior early farming villages of East Africa. Because these sites did not fit with the non-African origin of the coastal town builders, they were interpreted as belonging to a cultural group distinct from the coastal town inhabitants. Challenges to this interpretation arose when, in the early 1980s, archaeologists consistently observed,in the lowest levels of excavations at the sites like Unguja Ukuu, EIW pottery similar to that of the hinterland (Chami 1994a). Linguists in particular (Nurse and Spear 1985) corroborated these challenges, pointing out that Swahili speakers of the coastal littoral were related to Early Bantu speakers of the coast. In contradiction, archaeologists East African Archaeology: Foragers, Potters, Smiths, and Traders 88 suggested that Cushitic pastoralists, not Arabs, had migrated from southern Somalia to found early Swahili coastal sites (see Horton 1987;Abungu 1989). Cushites were also thought to have established trade contact with GrecoRomans (Horton 1990; Sutton 1990). In the early 1990s, Chami (1994a...