Origins of the Mali Empire

Both oral traditions and references in medieval Arabic accounts suggest the demise of Ghana/Wagadu in the late 11th century was followed by a period of political fragmentation in which a constellation of competing polities strove for regional control. Two among them, the kingdoms of Sosso and Mali, would eventually come to dominate the political landscape, and war between them ensued. The main source on this conflict is the Sunjata epic. Set in the 13th century, it narrates the life of the Manding prince Sunjata Keita and his exploits leading to the foundation of the Mali Empire. The epic begins with Sunjata’s ancestry and describes how, after a troubled childhood and a prolonged exile, Sunjata returned to his homeland to unite all the Mande clans against the powerful Sumanguru Kante, king of Sosso. A long war followed until Sumanguru’s final defeat at the battle of Krina.¹ Following this victory, Sunjata reputedly set up the legal, political, and ideological framework of Mali’s imperial structure. In some versions of the epic this framework was then consolidated in a “charter of rights” signed at the plain of Karakun Fuga, but upon scrutiny this appears to have been a relatively modern addition, absent in earlier performances.²

Although as with any epic, the Sunjata story contains many elements of mythical or symbolic nature, we also know that Sunjata Keita was indeed a historical figure, as confirmed by his appearance in 14th-century texts by Ibn Khaldun and Ibn Battuta.³ Whether the accomplishments described in the epic were indeed undertaken by a single man in such a short time span or whether oral accounts conflate in a single figure and period, the results of a longer and more complex process is difficult to ascertain.

Mali’s Expansion and Apogee

The 13th and 14th centuries were times of expansion for the newly created empire, reaching its greatest territorial extent in the early 14th century, after the annexation of the regions of Walata, Gao, Timbuktu, the Gambia, and the Senegal Valley. These new territories and vassal kingdoms were added to the zones already controlled by Mali from the end of the Sosso war (i.e. the Middle and Upper Niger) and most of the western Sahel, thus establishing one of the most extensive polities in the history of West Africa (see Figure 1).

Click to view larger

Figure 1. The Empire of Mali at its greatest territorial extent (Map drawn by the author).

According to the 14th-century Arabic historian Ibn Khaldun, following Sunjata’s death, three of his sons—Wali, Wati, and Khalifa—successively rose to power. Khalifa was eventually deposed and replaced by Abu-Bakr, who in turn was killed and succeeded by a freed slave named Sakura.⁴ After the killing of Sakura upon his return from Mecca, kingship returned to Sunjata’s line with the coronation of Mansa Qu and later his son Mansa Muhammad. The fate of the latter has been the subject of some debate, as according to his successor, Muhammad wanted to discover what laid at the end of the ocean and so he sailed into the Atlantic at the head of a large fleet, never to be seen again.⁵ While this episode has sometimes been used to argue for a pre-Columbian African arrival in the Americas, a successful Atlantic crossing would have been unlikely given the riverine and coastal nature of their naval technology, the lack of prior experience in long-distance oceanic expeditions, and the fact that nobody returned.⁶

Despite Muhammad’s grand plans, it was his successor Mansa Musa who put Mali firmly on the international stage with his famous pilgrimage to Mecca and visit to Cairo in 1324–1325. Mansa Musa may not have been the first emperor of Mali to undertake the haj,j but his pilgrimage was the first one to be recorded with such a profusion of detail and diversity of sources.⁷ This was partly due to the grandiosity of his delegation, but it was also the result of the flourishing of Mamluk historiography in Cairo at the time.⁸ Despite the diversity of the accounts, one element is common to all: Mansa Musa’s opulence. His delegation is said to have included hundreds or even possibly thousands of people and to have brought so much gold that it devalued the price of gold in Cairo for over a decade.⁹ It is thus no wonder that Mansa Musa’s 25-year reign is often described as Mali’s “golden age,” but it is important to remember that while it did take place during Mali’s apogee, it probably was not the wealthiest or most powerful, just the best recorded.¹⁰

Click to view larger

Figure 2. Detail of a 14th-century Catalan atlas by Abraham Cresques showing the emperor of Mali holding a large gold nugget in front of an Arab trader.

(Photo courtesy of Bibliothèque Nationale de France).

If Mansa Musa’s pilgrimage firmly placed Mali’s wealth on the map (sometimes quite literally; see Figure 2), it was during the reign of his brother Suleyman that we have the first direct account of Mali’s royal court, thanks to the visit of the traveler Ibn Battuta in 1352–1353.¹¹ Ibn Battuta spent seven months in Mali, during which he witnessed the court’s operation, protocol, and rituals. He reported, for instance, how the mansa’s pavilion was decorated with golden and silver arches, and drums and trumpets announced when he was ready to hold council. He described how before every council 300 slaves carrying bows, short lances, and shields; horses, rams, lancers, and bowmen entered the room and stood around the mansa in formation. Then singers and musicians performed while the cavalry commanders entered on horse carrying bows and quivers, preceded by their followers armed with lances, while the rest of the population sat outside.¹²

Political and Territorial Organization

Although traditionally referred to as an empire, Mali’s structure and organization does not appear to abide by the traditional definition of the territorial state, with its implications of territorial sovereignty, centralized government, specialized administration, and monopoly over the legitimate use of force. Instead, it was composed of different “lands” or “vassal kingdoms” that retained considerable autonomy, with control becoming more nominal and less real as the distance from the core increased, and no assumption of ethnic, cultural, or political homogeneity.

At the center of the structure was the emperor or mansa and his court. The operation of the court is one of the elements we have the most information about, thanks to Ibn Battuta’s account. The most important official in court was the griot or jeli: spokesman, master of ceremonies, counselor, princes’ tutor, and director of court musicians. The farariya (commanders of the cavalry) and the mansa’s personal guard represented the military in court. Also important were the qasa (the mansa’s first wife); the santigi, or finance minister, master of the treasury and guardian of royal granaries and valuable deposits; and the kangoro-sigi or viceroy. Finally, slaves and Muslim officials were also valued and influential, often becoming royal confidants and advisers. Although the writing skills and northern connections of the latter were highly valued, administrative procedures remained mainly verbal.¹³

According to the Arab historian Al Umari, at its maximum extension Mali comprised fourteen districts and their towns, villages, and countryside, and its provinces were kingdoms in themselves.¹⁴ Ibn Khatir, on the other hand, claims twenty-four kings were under the control of Mali’s emperor, and in any case the number of vassal territories would have changed over time. As for the nature of the empire, although certainly more than an alliance of independent chiefdoms, the degree of centralized control would have varied for each province.¹⁵ Three types of government existed: autonomous provinces supervised by a local representative of the mansa, with local dynasties largely retaining their autonomy (applied to allies and kingdoms that had not offered resistance to conquest); provinces directly administered by a faren or farba (centrally appointed governor in charge of justice, security, and taxes), for initially hostile regions (e.g., Sosso) or regions of key economic importance (e.g., Walata); and the Malinké heartland, directly controlled by the mansa or a kinsman/ally of his.¹⁶ Where a centrally controlled provincial government existed, it reproduced the central court on a smaller scale, as shown by Ibn Battuta’s description of Walata’s farba audience.¹⁷

According to oral traditions, the representatives of the mansa met once a year at the royal court. All problems and projects were put forward and debated, and the mansa indicated the measures to be taken. After the council, the governors were not summoned until the following year, unless they disobeyed the mansa’s guidelines.¹⁸ An example of one such summoning was observed by Ibn Battuta during his stay in Mali, when the mansa’s representative in Walata was accused of not repaying a debt.¹⁹

Although unprecedented in scale, the imperial structure was rooted in, and integrated with, traditional forms and spheres of authority. Beyond the court, pre-imperial power structures appear to have remained in place. The basic unit in precolonial Mande society—as preserved by oral traditions and documented in early colonial accounts—is the lu (extended family), controlled by the fa (family head), who represents the link with the ancestors and administers property and the relations with other lu. Several lu form a dugu (village), governed by the notables of the village (the fa of the most important lu) assisted by the dugu-tigi (village head), who is the fa of the first lu believed to have settled the place, whose ancestors first established the relation with the spirits of the land. A series of dugu can form a confederation or kafu, controlled by a particular lineage and ruled by a council with a kafu-tigi at its head.²⁰ Although it is difficult to establish exactly how far back in time these structures go, the appearance of some of the terms in medieval Arab accounts suggests that at least some were indeed present by the early empire, and others may even predate it.

The strong links between traditional kinship notions and the imperial structure were an important aspect of Mali’s imperial ideology. As the primary authority system, the family and clan system provided the basis for the buildup of power networks, from the local to the regional and up to the imperial level. Thus, the mansa’s office grew out of the family headman and the village head and, in the case of Mali, from a kafu controlled by the Keita clan to an extensive centralized empire. Initially supervised by elders’ councils, the mansa gradually escaped such control, expanding his territorial dominion and becoming the hereditary office found immediately before the rise of Mali and the framework of hierarchical structures on which the future empire would be based.²¹

Furthermore, the office of mansa was not just based on kinship principles but also depended on them for the deployment and maintenance of authority. In the “provinces” the state did not relate to individuals but to kinship groups—lineages and clans—and other traditional authorities.²² Besides, the local dugu-tigi (master of earth) and wula-tigi (master of the bush), who controlled the spirit of the place, had to be respected so the soil would keep producing. For example, al-Dawadari reports how during his stay in Cairo, Mansa Musa was asked why he sent tribute collectors to the gold-producing lands instead of ruling them directly. He replied that for the land to produce gold, it had to be controlled by its inhabitants, as conquest and direct rule would destroy the land’s productivity.²³

The Quest for Mali’s Capital

Of all the issues concerning the empire of Mali, the location and nature of its capital has been by far the most debated. Arabic sources commonly describe the capital—Ibn Battuta even visited it in person—but they do not provide any definitive clues regarding its name or location. The capital is referred to as BYTY by Al-Dukkali, as BNY by Ibn-Khaldun, as Mali by Ibn Battuta, and as Malal by Ibn Said and al-Idrisi.²⁴ Other names present in the literature include “Songo,” suggested in 1492 by the Portuguese de Barros, while the 17th-century Tarikh-el-Fattash claims the two successive capitals of the empire were Diâriba (sometimes interpreted as Kangaba) and Yan’ (also read as Niani, Nyang, and Dyang, depending on translation).²⁵

Regarding its location, Al-Dukkali’s description—a first-hand account of a city encircled on all sides by a river south of a great lake, with seasonal floods and in hilly and verdant country—could refer to somewhere in the vicinity of the Inland Niger Delta. Al-Idrisi located it twelve stages south of Ghana.²⁶ The city visited by Ibn-Battuta, on the other hand, was ten miles from the river, and although different interpretations have been made of his itinerary from Walata to Mali, a west bank location north of Bamako seems likely, given the trip’s duration and the fact that he does not mention crossing the Niger.²⁷

Oral traditions are not much clearer. When taken globally, the Sunjata corpus weighs heavily in favor of Mali’s 13th-century administrative center being located somewhere southwest of Bamako and northeast of Kela, but it is inconsistent regarding which bank of the Niger was involved. Narena and Farakoro are the towns most commonly associated to Sunjata’s father, whereas Dakajalan, Kangaba, and Niani are the most frequently identified with Sunjata.²⁸

Bearing in mind the diversity of the sources, it is hardly surprising that the academic debate on the matter has been complex (see Figure 3). Already in 1841, Cooley argued that the capital of ancient Mali was located near Niamina on the northern bank on the Niger, in current Mali.²⁹ This opinion was seconded thirty years later by Binger, who further specified the site of Niani Madougou, between Tougouni and Kondou, west of Niamina as the location. Delafosse, on the other hand, favored Kangaba as the ancient capital but, influenced by Binger, accepted Niani as Sunjata’s town.³⁰

Click to view larger

Figure 3. Locations of the suggested capitals/power centers of the Mali Empire (Map drawn by the author).

Having visited the site of Niani Madogou in 1922 and not having found the large city he expected, Vidal proposed that Mali had had four capitals in succession: Diériba (nowadays Dieliba-Koto, in the Milo), Niani-on-Sankarani, Manikoura near Figuira, and Kangaba.³¹ Soon after, Gaillard’s initial 1923 excavations at Niani uncovered a site sizable enough to convince Delafosse to accept Vidal’s hypothesis.³² Monteil, on the other hand, argued that Mali had two capitals: Tabou (near Sigui) for the early period under the Konate dynasty and an unspecified town in the area between Kangaba and Siguiri from Sunjata onward.³³

Following the 1960 publication of D. T. Niane’s version of the Sunjata epic, in which Niani-on-Sankarani (in current Guinea) appeared as the only and permanent capital, the second set of excavations were undertaken by Filipowiak.³⁴ These uncovered several structures supposedly corresponding to the royal palace and audience hall in Ibn Battuta’s account. Nevertheless, the structures’ association with smoking pipes (not introduced until two centuries later) and a clustering of radiocarbon dates between the 6th and the 10th centuries, with a reoccupation in the 16th century, leave a 600-year gap coinciding precisely with the Mali Empire. Consequently, all available evidence indicates that Niani-on-Sankarani was not the city visited by Ibn Battuta or Mali’s capital at its apogee, although it may have been a power center during its decline.

Another possibility, suggested by the historian David Conrad, is the site of Dakajalan near the village of Kirina, in current Mali, which is the town oral traditions most frequently associate with Sunjata.³⁵ A brief archaeological reconnaissance in the site in 2014 found the site to be of great sacred importance to the current populations but yielded very limited archaeological material.³⁶ More recently, a magnetometry survey uncovered some possible evidence for walls as well as some arrowheads, but archaeological evidence for an important power center in the area still remains limited.³⁷ Further north, in the Segou region, survey and excavations at the site of Sorotomo have documented an unusually large, nucleated settlement mound (72 ha) whose foundation and abandonment coincide well with the Mali Empire’s chronology, but research at the site is still in its early stages.³⁸

Beyond the issue of the capital’s name and location, there is a parallel—and perhaps more interesting—debate regarding whether in fact “capital” is a relevant term altogether. On the one hand, it is possible that archaeologists just have not found the capital yet; after all, archaeological work in the heartland of ancient Mali has been limited and focused largely on the great commercial centers of Arabic texts. On the other hand, it is possible that the notion of a permanent seat of power and administration may be altogether inappropriate, resulting exclusively from our own assumptions and the cultural preconceptions of Arab authors on the nature of government and urbanism.

The notion, first suggested by Monteil, that Mali might have had not of a single permanent capital but several seats of power in succession has gradually gained popularity.³⁹ Whether peripatetic, as suggested by Haour, or sedentary but changing location several times during history, as argued by Monteil, Hunwick, and Conrad, several factors suggest that the power center might have been more mobile than traditionally thought.⁴⁰ In fact, except for versions adapted to Western audiences, the notion of capital is largely absent from the Sunjata epic. Instead, we find references to mansaduguw or “king’s towns,” conveying the idea that the center of authority was wherever the mansa happened to be. The same sources also clearly distinguish between the royal lineage’s ancestral residence and the king’s court, and only in relation to the first is a clear sense of place conveyed.⁴¹ On the other hand, the number of places described as mansadugu is limited, suggesting a certain length of occupation for each site. This perception is reinforced by Ibn Battuta’s description of large, permanent royal structures, not consistent with a pattern of continuously moving capitals.⁴²

Social Structure

Colonial sources often describe precolonial Mande society as characterized by a tripartite division into horonw, nyamakalaw, and slaves. While not as rigid as some of these sources might imply, Arab/European sources and linguistic evidence confirm a version of this structure does indeed go back to imperial times.⁴³Horonw (sing. horon)—sometimes confusingly translated as “nobles”—comprised the majority of the population, including farmers, warriors, traders, and clerics, as well as the ruling elites. Nyamakalaw (sing. nyamakala), on the other hand, were endogamic specialists such as blacksmiths, leatherworkers, and griots (oral historians/musicians). The status of nyamakalaw was ambiguous: as people with special capacities to perform activities considered dangerous, polluting, and connected to mystical forces, they were feared and respected but also despised and banned from political power.

The precise origin of nyamakalaw is unclear: most oral traditions attribute their formal establishment to Sunjata, but some nyamakalaw trace their ancestry back to Wagadu.⁴⁴ Griots certainly existed by the 14th century when Ibn Battuta visited the court of Mali, as corroborated by the Portuguese Valentim Fernandes a century later.⁴⁵ Linguistic evidence indicates some of the words used to refer to these specialists appeared among the Malinké no later than the 13th century.⁴⁶ The current regional variability in a clan’s status, however, with cases in which the same clan is considered horon in one zone and nyamakala in another, suggests that the process by which these classifications were established may have been longer and more diverse than implied by oral sources.⁴⁷ In any case, the creation of a social category of endogamic specialists by Mali’s elites would have been a fundamental tool in the articulation and control of traditional spheres of power. Through the control of groups especially suited to harness mystical energy, elites benefited from occult-sanctioned persuasion without being polluted by it.⁴⁸

As for slaves, their status and process of enslavement greatly varied, from prisoners of war to members of other groups seized during raids, youngsters pawned by their families, and condemned criminals. Except for chattel slaves sold into the trans-Saharan, Indian Ocean, and later the transatlantic trade, slaves were always attached to a family and very often adopted the family name. They could also rise to positions of significant military and political importance, and the head of slaves was one of the most powerful figures in court.⁴⁹ We even know of cases of freed slaves who became emperors, like Sakura in the 14th century, as described by Ibn Khaldun.⁵⁰

The View from the Periphery: Kaabu

Although our knowledge of Mali’s provincial organization is generally limited, there is one exception: Kaabu. Located in the Senegambia, Kaabu was Mali’s westernmost province and main link with the Atlantic world. It is also the part of the empire where Mali’s structures survived the longest, for after Mali’s collapse, Kaabu continued to flourish as an independent kingdom until its demise in the 19th century. Consequently, for both geographical and chronological reasons, European literature on Kaabu is far more abundant and richer than for any other part of the empire, giving us a unique glimpse into the diversity of Mali’s provincial structures.

According to the Sunjata epic, Kaabu was conquered and annexed to the Empire by Tiramakan Traoré, one of Sunjata’s generals. Tiramakan was then given control of the region and, together with his descendants, created a political and administrative structure combining Mali’s principles with local idiosyncrasies. While we have no proof of Tiramakhan’s existence or deeds, European travelers’ descriptions of Kaabu’s political traditions are consistent with this hybrid political structure.

In some aspects, Kaabu reproduced Mali’s organization and protocols to the detail. For instance, in 1623, the British traveler Richard Jobson described how in an audience with a Gambian king, the guest kneeled and sprinkled dust upon his head two or three times, a protocol that had been described in almost the same exact terms for the imperial court of Mali two centuries earlier by Al-Umari and later by Ibn Battuta.⁵¹

In terms of social structure, Kaabu followed Mali’s tripartite division into horonw, nyamakalaw, and slaves, but it also had its own peculiarities—most notably, the existence of two aristocratic classes, nyanthio and koring, unknown in the rest of the Manding world. The nyanthio were at the top of the social hierarchy and the group to which all of Kaabu’s rulers belonged. Identified by the patronyms Sane and Mane, nyanthio status was transmitted only through the maternal line. The nyanthio looked down on both agriculture and trade, as war and hunting were the only occupation worth their status.⁵² Below the nyanthio were the koring. Koring status could be inherited patrilineally, or from a nyanthio father and a non-nyanthio mother. They had a greater variety of patronyms than the nyanthio and could not become rulers of Kaabu but were allowed to rule over certain territories. Like the nyanthio, they did not farm and reveled in military exploits and hunting.⁵³

Kaabu’s territorial organization—both as part of Mali and later as an independent kingdom—was that of a confederation of diverse territories, linked by fluctuating ties of subordination and collaboration. Three of these territories (Jimara, Pathiana, and Sama) were considered nyanthio provinces, with the right to provide rulers for the whole Kaabu, which they did following a system of rotation. Another important difference of Mali’s rulership was the existence of female rulers, which, although not frequent, were far from rare. For instance, a ruler list from Kankelefa—one of Kaabu’s territories—includes three female rulers out of a list of ten, and a similar account from another province, Niumi, features twelve queens.⁵⁴

While Kaabu is not necessarily representative of other territories within the Mali Empire, it does illustrate the richness and diversity of political organizations within a common tradition that characterized Mali’s imperial organization.

Economy and Trade

Military Power and the Army

Military power, and more specifically horsemanship and iron weaponry, are recurrent themes in the material culture and oral traditions of Mali. Whether this reflects their physical importance in the maintenance and expansion of the territory or its symbolic significance as prestige markers is difficult to ascertain. In Mali a permanent army was institutionalized, with large garrisons stationed in sensitive frontiers and important cities, including Walata, Gao, and Timbuktu.⁶⁴ Al Umari describes a contingent of 100,000 soldiers, both infantry and cavalry, with the latter constituting a tenth of the total number.⁶⁵ While the reliability of such figures is doubtful, Mali’s military force was undoubtedly substantial, as shown by the fact that north African princes approached Mansa Musa during his trip to Cairo to request his assistance in their campaigns.⁶⁶ Mansa Musa himself claimed during his trip that he had conquered “by his sword and armies” twenty-four cities with their surrounding estates.⁶⁷ Furthermore, Ibn Battuta describes Mali as an exceptionally safe territory, where “neither traveller there nor dweller has nothing to fear from thief or usurper.”⁶⁸

Religion and Belief

Mali’s Decline

By the time of Ibn Battuta’s visit to the court of Mali in 1352 in the times of Mansa Sulayman, increasing instability was already undermining the central power, resulting in the successive independence of the vassal states of Gao and Méma in the early 15th century.⁸⁰ This process of internal division was further encouraged by the rise of the Songhay Empire in the northeast and its gradual conquest of most of Mali’s eastern territories during the 16th century. Subsequent defeats at the hands of Moroccan troops in the north and the loss of the Bambuk goldfields in the southeast further reduced Mali’s territory and power. Although no longer an empire, Mali did survive as a much smaller polity into the 17th century, possibly even longer.⁸¹ So did its legacy: Kaabu, its Atlantic province, survived as an independent kingdom until the late 19th century, and Mali’s structures and political traditions deeply shaped its successors, including the Songhay Empire itself.

Discussion of the Literature

For its historical importance, the available literature on the Mali Empire is limited. The main and most comprehensive syntheses remain Nehemia Levtzion’s 1973 Ancient Ghana and Mali and Monteil’s 1929 “Les Empires du Mali,” both of which present important problems in terms of source identification and criticism.⁸² Two separate syntheses exist for Kaabu, Mali’s westernmost province: D. T. Niane’s Histoire des Mandingues de l’Ouest and Carlos Lopes’s Kaabunké: espaço, território e poder na Guiné-Bissau.⁸³ In recent decades, historians have devoted substantial energy to the recording, publication, and analysis of oral traditions and medieval Arabic accounts and epigraphic evidence, as discussed in the Primary Sources section. As a result, a parallel discussion has emerged critically evaluating the role of different sources, particularly with regard to the historical value of oral traditions and the limitations and biases of Arab sources.⁸⁴

Walking the line between the ethnographic and the historical, anthropologists have exponentially expanded our understanding of Mali’s past and present social structures. These have mostly focused on nyamakala or endogamic specialists, and in particular on blacksmiths and griots.⁸⁵ Archaeological debates, on the other hand, have tended to focus on two topics: trade and its role in the emergence and operation of West African states and the location and nature of Mali’s capitals/political centers.⁸⁶

Primary Sources