Ghana Empire (Wagadu)

← Encyclopedia

The first wealth of Wagadu was kept underground, in two senses. There was the gold, washed in secret from the river gravels of the far south and carried north under the guard of men who would not name the country it came from. And there was the serpent, Bida, coiled in a cavern or a well at the heart of the land, to whom the founding compact bound the realm: a measure of rain and a measure of gold each year, against a single life given over in season. The Soninke who held this country between the desert and the great rivers called it Wagadu — the land of the wago, the place of herds and of those who command. It was the geographers of the Arabic-writing world, looking south across the sand toward a power they knew only by report, who fixed on the title borne by its sovereign and made that title the name of everything he ruled. The king was the ghana. To Córdoba and Kairouan, the kingdom became Ghana.

The approximate extent of the Ghana Empire (Wagadu) in the western Sahel, between the Senegal and Niger rivers — Amitchell125, via Wikimedia Commons (CC BY-SA 4.0)

That displacement of a man’s title onto a whole civilization is the right place to begin, because it marks the seam where two records of Wagadu meet and never quite fuse. From within, this was the kingdom of the serpent and the herds, ruled through the line of the king’s sister, its dead laid in groves no commoner could enter. From without, in the careful prose of merchants’ informants, it was a great market at the southern shore of the Sahara, the place where the gold of the world came up out of the ground. The empire is legible only when both records are read at once — and it is the second, the written one, that has given the first its modern name, a name that would much later cross a continent to settle, in 1957, on a republic on the Atlantic forest coast that this empire never touched.

The people and the land between two rivers

The Soninke were a Mande-speaking people of the western Sahel, the broad belt of savanna that runs east to west below the desert, watered at its edges by the Senegal River to the southwest and the Niger to the southeast. Their country was a threshold. To the north lay the desert and its salt; to the south, beyond the savanna, the forest fringe and the goldfields. A polity that could sit astride that threshold and tax what crossed it would grow rich on the difference between what the desert lacked and what the south held in surplus. Wagadu was that polity, and the Soninke its core.

Their own account of how the land came to be settled runs through a single ancestor. Dinga, in the oral tradition, crossed the Sahara out of the east and came into the country that would be Wagadu, and there he made the pact with Bida, the great python who already held the ground. The kingdom that issued from that compact was governed by a sacred kingship whose succession did not pass from father to son. It passed to the son of the king’s sister — a matrilineal line, on the reasoning that a man may carry doubt about his own son but can hold none about his sister’s child. The throne was a thing transmitted through women though held by men, and around it stood the apparatus of a divine monarchy: the grove, the tombs, the specialists of the cult.

By the standard reckoning the empire took shape across the eighth and ninth centuries CE, consolidating older Soninke settlement into a single tributary order. Its weight was felt far enough north that the Arabic geographers began to register it, and through their pens — they who never came south themselves but questioned those who did — the kingdom enters the written world.

Koumbi Saleh, the double city

The capital that the sources describe, and that the spade has partly recovered at the site now called Koumbi Saleh on the desert’s southern rim, was not one city but two, set some six miles apart on the plain and, by the eleventh century, all but joined by the habitations strung between them. One town belonged to the merchants and the faith they carried. It was built of stone, laid out in streets, and held a dozen mosques — one of them the congregational mosque where the community gathered for the Friday prayer — and the salaried Muslims who served the king as interpreters, treasurers, and scribes lived among its walls. This was the Islam of the trans-Saharan trade, arrived not by conquest but by caravan.

The other town was the king’s, and the Arabic record names it al-Ghaba — “the grove.” Here stood the palace within its enclosure, the domed pavilions where the king gave audience and heard the grievances of his subjects, and a single mosque kept for the Muslim officials who came on the king’s business. But the heart of al-Ghaba was older than any mosque. It held the sacred grove of trees in which the religious specialists of the Soninke lived and kept the cult’s sacred things, and within it lay the tombs of the kings, in a thicket that no one might enter. The double city was an exact map of the kingdom’s order: the new religion of the market lodged a few miles off from the old religion of the throne, each with its quarter, the two bound by the road of commerce that ran between them.

Part of the western necropolis at Koumbi Saleh, the density of its funerary structures visible from the air — Chloé Capel, via Wikimedia Commons (CC BY-SA 4.0)

Excavation has put flesh on al-Bakri’s report. The recovered town — dug across the twentieth century, from soundings in 1914 through the campaigns of mid-century and after — shows stone houses, a great mosque, and a litter of imported Islamic coins, ceramics, and glass that ties this place to North Africa, the Mediterranean, and the lands beyond. The archaeologist Raymond Mauny reckoned its population at fifteen to twenty thousand, which would make it the largest town of its day anywhere in West Africa. A city of stone in the sand, fed by a trade in metal and salt, governed from a grove of priests.

The central mausoleum of the Column Tomb on the ridge above Koumbi Saleh, an elite stone burial monument of the eleventh to thirteenth centuries — via Wikimedia Commons (CC BY-SA 4.0)

The engine: gold from the south, salt from the sand

Everything in Wagadu turned on a single asymmetry. The desert had salt and no gold; the lands south of the savanna had gold and no salt. Salt the body cannot do without and the savanna could not produce; gold the wider world would take in any quantity offered. Wagadu sat on the line where the two met, and it made itself the toll-house of that exchange.

West African gold had reached the Mediterranean long before there was an empire to tax it — the mints of Carthage and, after it, Rome had drawn on the southern metal through Saharan intermediaries — but it was Wagadu that first organized the southern end of that flow into a state. The gold came up from the south, from the river fields of Bambuk and the country the Arabic writers called Wangara, panned from gravels and carried north by traders who guarded the secret of its origin so jealously that the location of the mines stayed hidden from the Muslim merchants for generations. The salt came down from the desert, cut in slabs from the Saharan pans — Taghaza and the others — and borne south by the camel caravans that set out from Sijilmasa and the Maghrebi marts. Centuries earlier the long-distance crossing of the central Sahara had been pioneered by the desert kingdom of the Garamantes, whose chariot and caravan tracks first knit the desert into a trade network; by Wagadu’s day that network’s western branch ran its gold and salt straight to the Soninke tax-gate. At Audaghost, the trade town that Wagadu brought under its control near the close of the tenth century, the desert and the savanna economies touched.

Early trans-Saharan trade routes carrying gold and salt across the desert, with Koumbi Saleh and Audaghost on the southern, Sahelian end — Aa77zz, via Wikimedia Commons (CC0)

The kingdom did not need to dig the gold or cut the salt. It needed only to stand at the chokepoint and tax what passed. On every donkey-load of salt that entered the country the king’s agents levied one gold dinar; on every load that left, two. Other goods, copper and the rest, were charged in their turn. Out of that levy, taken at the seam between two worlds that could not do without each other, the treasury of Ghana was filled.

There was a second mechanism, subtler, recorded by al-Bakri and pointing to a state that understood its own market. All the gold nuggets — the worked metal, the heavy pieces — were reserved to the king; only the gold dust was left to circulate among the people. The reasoning given is precise: were the nuggets not held back, the people would hoard gold until it lost its value. Here is a sovereign managing scarcity to defend a price, draining the bullion that would have flooded the market and keeping it in the throne’s own hands. The famous regalia of the court followed from the same monopoly — the king in silk with his bow and quiver, the pages bearing gold-mounted swords and shields, the very guard-dogs collared in gold and silver, while his subjects fell to their knees and put dust on their heads as they came before him. The display was the surplus of the monopoly made visible.

The grove and the mosque: a kingdom of two religions

Wagadu did not choose between its gods. The Soninke cult of the divine kingship — the serpent Bida, the sacred grove, the royal tombs, the specialists who tended the sacred objects — remained the religion of the throne and of the land, while Islam settled into the merchant quarter and the chancery. The king kept literate Muslims about him because a trading state needs men who can read a contract and reckon a tax, and the prestige of the wider Islamic world, then in the full flood of its golden age, traveled south with the goods. But the king himself stood within the older order. The two faiths shared a capital without merging, the grove a few miles from the mosques.

The Islam that arrived came in the legal coloring of the Maghrib, which was overwhelmingly that of the school founded by Malik ibn Anas of Medina — the Maliki rite that governed jurisprudence across North Africa and the Sahara and that the desert reform movements would carry with a hardening edge. The deepest current of that world, the inward path of Sufism, would in later centuries thread the Sahel through the trading and scholarly diasporas; in Wagadu’s day the faith was still chiefly the merchant’s creed and the chancery’s tool. Conversion among the Soninke proceeded the way trade proceeds — by contact, by intermarriage, by the slow advantage of belonging to a literate, far-connected community — not by the sword. The throne held its grove for as long as Wagadu held its primacy.

The Almoravid century and the long contraction

The eleventh century, which shows Wagadu at its height in al-Bakri’s account, also brought the pressure that older histories named as its undoing. From the western Sahara rose a militant reform movement, austere and Maliki to the bone, that swept the desert and would carry Islam by force into Morocco and Iberia. In 1054 its southern arm took Audaghost from Ghana’s control. The traditional narrative, descending through later Arabic chronicles and the region’s own oral memory, held that this movement went on to conquer the capital itself in 1076, sack Kumbi, and force the kingdom to Islam — and that the shock of it broke Wagadu’s power.

That story has been taken apart. In a two-part study, “The Conquest That Never Was: Ghana and the Almoravids, 1076,” David Conrad and Humphrey Fisher returned first to the external Arabic sources and then, in its second part, to the local oral record, and found in neither any clear evidence of a military conquest of the capital. The dramatic sack, they argued, is a back-formation — a story assembled from misread sources and later folklore, not a reported event. The spade has tended to agree: there is no destruction layer at Koumbi Saleh that answers to 1076, no burning, no rupture. What the ground shows instead is gradual — a slow abandonment, a long economic contraction spread across centuries.

The real decline was structural, and it grew out of the very thing that had made Wagadu rich. The kingdom lived by holding the chokepoint of a trade, and a chokepoint can be bypassed. New goldfields opened to the east and south; the trade routes shifted toward them; the desert edge dried and the pasture thinned; rival powers in the savanna learned the same trick of taxing the caravans. The serpent’s own people kept a different account of the loss — the pact with Bida broken when a young man slew the serpent to save the woman owed to it that year, and the land cursed with seven years of drought for the breach. Read as etiology, the legend names exactly what the economics confirm: the rain failed, the gold went elsewhere, and the compact that had bound prosperity to the throne came undone.

By the thirteenth century Wagadu was a husk of its former reach, its old vassals contending over the inheritance. The Sosso kingdom of Kaniaga, itself a former tributary, rose under the king Sumanguru Kanté and pressed into the Mande heartland. And then, around 1235, a Mande prince named Sundiata Keita broke the Sosso at the battle of Kirina, and in 1240 his forces seized and razed Kumbi itself — extinguishing the last ember of Ghana’s imperial dignity and folding what remained of Wagadu into a new and larger order.

Texts and the recovery of Wagadu

Wagadu is known through two archives that rarely overlap and must be read against each other. The first is the corpus of Arabic geography. The indispensable witness is the Andalusian al-Bakri, who in 1067–68 compiled his Book of Routes and Realms (Kitab al-Masalik wa’l-Mamalik) in Córdoba without ever crossing to Africa, drawing on earlier accounts and on merchants who had made the journey; his chapter on Ghana — the two towns, the twelve mosques, the dinar levied on salt, the king’s gold monopoly, the audience with its gold-collared dogs, the grove of the dead — remains the fullest single portrait of the empire at its height. These sources, with the later notices of al-Idrisi and others, are gathered, translated, and annotated in the standard scholarly collection assembled by Nehemia Levtzion and J. F. P. Hopkins, the Corpus of Early Arabic Sources for West African History (Cambridge, 1981) — the foundation of nearly all modern work on the subject, including Levtzion’s own Ancient Ghana and Mali (1973).

The second archive is Soninke oral tradition — the legend of Dinga and the serpent Bida, the genealogies of the founding clans, the memory of the broken pact and the drought — which preserves the kingdom’s view of itself and which the conquest debate forced historians to weigh seriously rather than dismiss as fable. The synthesis of both, with the archaeological record of Koumbi Saleh and Audaghost, anchors the West African chapters of the UNESCO General History of Africa, Volume III (eds. M. El Fasi and I. Hrbek, 1988), the collaborative reference work that set the medieval Sahel within African historiography on its own terms. For the commercial machinery that made the empire possible, the Metropolitan Museum’s survey essay on the trans-Saharan gold trade traces the gold and salt routes that converged on Wagadu’s tax-gate. Against this body of work the older racialized fantasies — the search for foreign founders behind the stone city, the refusal to credit a Sahelian people with its own urban empire — have no standing; the evidence is Soninke from the gravel of the goldfields up.

What Wagadu handed down

When Sundiata razed Kumbi he did not invent the order he built. He inherited it. The thing Wagadu had perfected was not a city or a dynasty but a method: plant the state on the seam between the salt of the desert and the gold of the south, tax the caravans at the crossing, hold the heavy metal in the sovereign’s hand to govern its worth, and let a literate Islam serve the throne without unseating the older cult that sanctified it. The Soninke themselves told its fall as a broken covenant. Wagadu prospered, the legend held, because each year a maiden was given to the great serpent Bida coiled in its well, and Bida in return drew gold-bearing rains over the land; when a young man drew his sword to save his betrothed and struck the serpent’s head from its body, the gold ceased, the wells soured, and the trade drifted south. Whatever drought and shifting caravan-roads lie behind the story, the kingdom that had first married gold to sovereignty on the desert’s edge contracted, across centuries, back into the sand it had once taxed.

→ Related: Mali Empire · Songhai Empire · Kanem Bornu Empire · Djenne Djenno · Islam · Islamic Golden Age · Sufism · Garamantes · Carthage · Malik Ibn Anas