Koumbi Saleh: Archaeology of the Ghana Empire's Capital

Koumbi Saleh is a large ruined town in south-eastern Mauritania that most historians identify as the merchant quarter of the capital of the medieval Ghana Empire. Occupied from roughly the 9th to the 14th centuries, it was a stone-built Muslim trading centre at the southern end of the trans-Saharan caravan routes. Its excavation offered the first physical evidence for a West African state previously known almost only from Arabic texts.

Background: searching for Ghana’s capital

The Ghana Empire, known to its people as Wagadu, flourished between roughly the 8th and 11th centuries in the Sahel north of the Senegal and Niger rivers. It grew rich by taxing gold coming from the south and salt and goods from the Sahara, as explained in the article on the Ghana Empire.

The most detailed description of its capital comes from the Andalusian geographer al-Bakri, who wrote in 1068 that it consisted of two towns about six miles apart. One was a Muslim town with twelve mosques; the other, called al-Ghaba, was the king’s town with his palace and sacred groves. His account is discussed in Al-Bakri’s Account of Ghana.

Al-Bakri gave no coordinates, so locating the capital became a puzzle for later scholars. The 17th-century Timbuktu chronicle known as the Tarikh al-Fattash named the ancient capital as “Koumbi”, and local people in the Hodh region knew of extensive ruins by that name.

Excavations

In 1914 the French administrator Albert Bonnel de Mézières visited the ruins and carried out the first, small-scale digging. Much larger excavations were led by Paul Thomassey and Raymond Mauny between about 1949 and 1951, and a Franco-Mauritanian team under Jean Devisse, Serge Robert and Denise Robert-Chaleix returned in the 1970s and early 1980s.

The excavators found a dense town of stone houses built from flat slabs of local schist, some with upper storeys, narrow streets and decorative triangular niches in the walls. At its centre stood a large congregational mosque, rebuilt several times. Around the town lay extensive cemeteries with stone tombs and a few gravestones inscribed in Arabic.

Finds included North African glazed pottery, glass weights of the kind used to weigh gold, iron tools, beads and spindle whorls. Radiocarbon dates and pottery styles suggest occupation from around the 9th or 10th century until the 14th or 15th century, with the most prosperous phase in the 10th to 13th centuries.

What the ruins reveal

The architecture and objects fit a town of Muslim merchants, many of them probably North African or Saharan Berber traders, living alongside local Soninke people. This matches al-Bakri’s description of a Muslim town where foreign traders, jurists and scholars lived and where the king recruited literate officials.

Recent studies of satellite images and surveys of the cemeteries, notably by the archaeologist Chloé Capel, have shown that the burial grounds around Koumbi Saleh were remarkably large, implying a substantial population over several centuries.

The identification debate

Koumbi Saleh is widely described as Ghana’s capital, and the identification is accepted in standard works such as Nehemia Levtzion’s Ancient Ghana and Mali. Several problems remain, though.

• No inscription or document found at the site names it as Ghana’s capital.
• The royal town of al-Ghaba, with its palace, groves and tombs of kings, has not been convincingly identified nearby.
• Earlier layers from before the arrival of Muslim traders are poorly known, so the site’s beginnings are unclear.
• Ghana’s rulers may have moved their capital over time, so the town al-Bakri described in 1068 may not have been the only one.

Some scholars therefore prefer to call Koumbi Saleh “a” capital or the most likely candidate rather than a certain identification.

Decline

The town’s later history is linked to the decline of Ghana after the 11th century, the disputed role of the Almoravids, explored in Almoravids and Ghana, and the rise of the Mali Empire further south. Trade routes shifted east toward Walata and Timbuktu, and Koumbi Saleh appears to have been abandoned by about the 15th century.

Life in the merchant town

The layout of Koumbi Saleh suggests a crowded, prosperous community. Houses were often packed together along narrow lanes, with storerooms that may have held trade goods such as salt, copper and cloth waiting to be exchanged for gold. Water came from wells, and the surrounding land, somewhat wetter in the Middle Ages than today, supported farming and grazing.

Timeline

• c. 9th-10th century: Earliest well-dated occupation of the stone town.
• 1068: Al-Bakri completes his description of Ghana’s two-part capital.
• c. 1076: Almoravid attack on Ghana’s capital reported in later Arabic sources; its scale is disputed.
• c. 13th-14th century: Trade shifts toward Walata and the Mali Empire.
• c. 15th century: The town is largely abandoned.
• 1914: Albert Bonnel de Mézières investigates the ruins.
• c. 1949-1951: Excavations by Paul Thomassey and Raymond Mauny.
• 1970s-1980s: Excavations by Jean Devisse and colleagues.

Why it matters

Koumbi Saleh turned the Ghana Empire from a name in Arabic chronicles into a place that could be measured and dated. Even with its identification debated, it remains the most important archaeological window into West Africa’s first great trading state and the Muslim merchant communities that linked it to the trans-Saharan trade.