Regional study: trans-Saharan trade

RALPH AUSTEN

The last few centuries of the era covered by this volume mark the entry of tropical Africa into the mainstream of world history. Africa had played a major role in human prehistory as the site of the earliest Homo sapiens, but the great urban civilizations that later arose in Asia and Europe had contact - and even that very late - only with the northern and eastern coasts of the continent.

To understand how such a change came about and what its consequences would be, it is necessary to go back at least to 1200 bce and consider both the degree to which the Maghrib - that portion of northern Africa to the west (Arabic gharb) of Egypt - had been integrated into the Mediterranean world and yet remained separated from the Bilad es-Sudan (Arabic: Land of the blacks) to the south. The prime factor here is the geography of the Sahara and the adaptation of local populations to its changing conditions.

Geography and prehistory of the Sahara

The Sahara is the world's largest desert (its name in Arabic means “desert”), covering some 3.5 million square miles of sand (only 25 percent) along with gravel, rocky plains, and plateaus. It stretches from the Nile Valley in the east to

A number of colleagues more specialized in this area than myself have very generously shared their work with me, and I especially want to express my thanks to Walter Kaegi and Andrew Wilson.

the Atlantic Ocean in the west, and its arid landscape is only occasionally interrupted by oases with permanent water and palm trees or more temporary vegetation on mountains and other outcroppings. Across the Sahara’s northern sahel (Arabic: shore) lie lands with a climate, soil characteristics, and rainfall patterns which allow for the same kind of agriculture (classically centered around wheat fields and olive groves) as the rest of the Mediterranean region.

The lands of the Central and Western Sudan (as distinguished from the Eastern/Nilotic Sudan) to the south of the Sahara are savannahs (grassfields). Rainfall here is less plentiful and reliable than in the Mediterranean but more heavily concentrated, a condition that, in combination with high tempera­tures, tends to leach nutrients from the ground, leaving only thin top soils. Mediterranean crops cannot, therefore, be cultivated in the Sudan, and local agriculture is based upon indigenous grains such as millet and sorghum or, in zones with optimal irrigation, rice. Even when transport barriers were over­come, these foods were neither abundant nor (in the case of millet and sorghum) valuable enough to be exported like North African wheat and olives. (See Map 24.1.)

The conditions just described pertain to the Sahara and its surrounding lands in historical times, but it is important - especially for their human geography - to remember that they did not always prevail. From about 9000 to 3000 BCE, this region of Africa experienced a Holocene, or wet phase, with abundant rainfall and large lakes. The latter part of this era coincides with the Neolithic or late Stone Age in global human development, when commu­nities began to move from hunting and gathering to settlement in fixed locations with greater control over food resources. Saharan populations during this period were larger than in more recent times and engaged in fishing, animal herding, and possibly some agriculture. From 3000 onward the Sahara became dry again, reaching its present form around 300 ce.

As indicated by the Arabic term sudan (plural of aswad, black), the peoples living south of the Sahara in historical times are mainly dark-skinned with hair texture and facial features as well as languages (Niger-Congo) that distinguish them from Mediterranean populations. We do not have any information on the languages of Neolithic Saharan communities, but evidence from skeletal remains and rock-paintings suggests that they were of mixed race. To this day some black peoples continue to inhabit the Sahara, but the

Downloaded from https:Zwww.cambridge.org/core. Cornell University, on 07 Jan 2017 at 22:38:59, subject to the Cambridge Core term available at https:/www.cambridge.org/core/terms. http://dx.doi.org/10.1017/CBO9781139059251.026

Map 24.i Trans-Saharan trade routes

majority of the population here and in the rest of North Africa are lighter­skinned with Mediterranean body-types and speak (or once spoke) dialects of a Berber¹ language (or language family) that belongs to the same eastern Semitic Afro-Asiatic grouping as Arabic, Hebrew, and ancient Egyptian. It is not clear from where and when the Berbers first entered the Maghrib (probably Egypt around 3000 bce), but by 1200 bce they had replaced whatever early Stone Age peoples previously occupied the region and also began to dominate the Sahara.^{[889] [890]}

Foreign colonization of ancient North Africa

Urbanization and centralized state-building came late to this entire portion of Africa, in comparison with neighboring Mediterranean regions. In fact, the earliest known cities in the Maghrib were foreign colonies, beginning with Carthage (near contemporary Tunis), founded around 814 bce by Phoenicians from present-day Lebanon.

The major attraction of Africa for these foreign ventures lay in the fertile regions near the coast, which produced significant quantities of wheat, olive oil, and wine for export. The urban population of Roman Africa may have amounted to as much as 5 million people consisting of both Italian immi­grants and acculturated Berbers, although existing evidence does not allow us to measure their origins with any precision. In the immediate rural hinter­lands, the majority of the population appears to have remained Berber although their languages (at least as expressed in inscriptions) included Latin, Punic (a Western Mediterranean form of Phoenician), and various versions of Berber.^[891]

The example or perhaps just the stimulus of foreign colonization inspired Berber groups to form their own states in the coastal regions of North Africa. The large “Numidian Kingdoms,” extending westward from present-day Tunisia, played a major role in the third- and second-century bce Punic Wars between Rome and Carthage. After the fall of Carthage, however, a combination of internal instability and Roman imperial ambition ended the independence of the Berber polities and drew the Romans further along the North African coast and southwards towards the Sahara.

Ancient Saharan and Sudanic urbanism

If the combination of indigenous and colonial North African ventures had already linked the Mediterranean and Saharan worlds by the early Christian centuries, to what degree were the contemporaneous societies of the Sudan and within the desert itself also prepared to enter global circuits of exchange? Posed in these terms, the answer must be that they were somewhat prepared to do so: both of these southern regions took major steps toward urbaniza­tion and expanded markets, but it is doubtful that these developments led to the initiation of a regular or at least large-scale and direct trans-Saharan trade. One of the problems here is that the most dynamic centers of economic and political change in the Sudan and Sahara emerged at a great distance from one another.

The major sites of Sudanic development before the Islamic era are located in the western portion of the region, from Dar Tichett in Mauretania through Gao on the Niger River Bend but especially Jenne in the Inner Niger Delta (with a population of between 10,000 and 26,000).^[893] The development of such dense (and in the cases of Jenne and Gao clearly urban) settlements between the late first millennium bce and early first millennium c e were all connected in some degree to the Sahara, first involving pastoral migrations southward due to desiccation of the desert and then also comprising exchanges for such Saharan goods as copper and salt.

None of the Western Sudanic centers of this era were organized as hier­archical and expansive states, a phenomenon that did occur within the Sahara itself but somewhat to the east in the Libyan region of Fazzan, where the Garamantian civilization arose about 1000 bce and survived to c.

The one relatively extended and complimentary contemporary descrip­tion of the Garamantes, by the Greek historian Herodotus (c. 484-425 bce), presents them as possible precursors of later trans-Saharan commerce; they used chariots to “chase the Troglodyte Ethiopians”^[895] and stood at the center of a series of geographical sites extending westward across the Sahara that might be interpreted as a desert caravan route linking Egypt with the Niger Bend.^[896] This reading of Herodotus stands on one side of a long-standing historical debate about pre-Islamic Saharan trade.^[897] Given the limited evidence for such traffic, the easiest conclusion is that it did not exist, but recent scholarship has been pushing in a more affirmative direction, revealing, at the very least, the conditions necessary for such a system of cross-desert exchange and the documentation that may reveal when it first came into place.

Ancient (trans-?) Saharan trade

The main conditions for conveying commodities through a vast and difficult expanse like the Sahara are two: first, the presence of goods with sufficient value to warrant the expense and risk of preindustrial carriage across an environment of this kind; second, the availability of a transport system that could make the cost of such goods competitive with alternative sources. For Islamic times we have numerous descriptions of trans-Saharan caravan commerce by direct participants or at least authors who were in contact with them. The absence of such accounts in the ancient era^{[898] [899]} does not prove that these voyages did not occur, but it does force us to turn to less direct evidence, mainly from archaeology.

The kinds of manufactured goods that made their way southward across the Sahara in the Islamic era (especially textiles, ceramics, glassware, and metalware) were, for the most part, available during Carthaginian and Roman/Byzantine times; thus, a key test for a trans-desert trade in this earlier period is the presence of Mediterranean commodities in Saharan and Sudanic sites that can be dated before 700 ce. Such goods are abundant in the Fazzan sites of the Garamantes state; some are also found in the fifth-century ce tomb ascribed to “Tin-Hanan” in the Hoggar mountains, a west central Sahara location that marks the end point of the desert route described by Herodotus and posited as a link to the Western Sudan.11 Excavations in the Sudan, however, have produced what one leading scholar calls “a total lack of archaeological evidence on the Niger Bend or elsewhere in the Western Sahel attesting to Roman or other pre-Arab Mediterranean pre­sence, albeit fleeting.”¹² On the other hand, much archaeology remains to be done in the Sudan, and recent analysis of copper objects and glass beads found in the rather obscure Kisii region of Burkina Faso at least suggests greater importation of northern goods during this time than was pre­viously known.1³ A further possibility is that Sudanic goods may have been purchased by Saharan peoples with their own products and then exchanged for Mediterranean commodities that never reached the Sudan. This argument makes sense for the Garamantian civilization, with the rich agricultural and artisanal production of its peak era (c. 70-300 ce), but probably not, at least on the same scale, for other times and places in the Sahara before 700 ce.

The most important items exported by the Sudan in Islamic times were gold and slaves. Most of the historical controversy about earlier trans-Saharan trade centers around gold, whose mediaeval Sudanic sources lay to the south and west of the Niger Bend. This location puts it a great distance from the Garamantes and also, if exports are tied to urban development in the Niger region (somewhat after 200 c e), far later than the trade route implied by the fifth-century bce account of Herodotus and somewhat after the flourishing era of the Garamantes. Andrew Wilson and Kevin McDonald posit the possibility of gold trade from the Niger Bend via Algeria, on the basis of Roman artifacts at Tin-Hanan and a few other neighboring places as well the juxtaposition of Kisii to the sources of ore.1⁴

The most serious argument for a Sudanic gold reaching the Mediterranean in ancient times refers to the late Roman and Byzantine eras (c. 100-680 ce) when camels were already in use for travel in the northern Sahara and a good

12 Timothy Insoll, “Islamic Archaeology and the Sahara,” in David Mattingly (ed.), The Libyan Desert: Natural Resources and Cultural Heritage (London: Society for Libyan Studies, 2006), p. 230.

13 Thomas R. Fenn, DavidJohn Killick, John Chelsey, Sonja Magnavita, and Joaquin Ruiz, “Contacts between West Africa and Roman North Africa: Archaeometallurgical Results from Kisii, Northeastern Burkina Faso,” in Sonja Magnavita, Lassina Kote, Peter Breunig, and Oumarou A. Ide (eds.), Crossroads/Carrefour Sahel: Cultural and Technological Developments in First Millennium BC/ad West Africa (Frankfurt am Main: Africa Magna Verlag, 2009), pp. 119-46; Kevin C. MacDonald, “A View from the South: Sub-Saharan Evidence for Contacts between North Africa, Mauritania and the Niger, 1000 bc - ad 700,” in Dowler and Galvin, Money, Trade and Trade Routes, pp. 72-82, notes that this site lay along a potential gold trade route, but see below.

14 Wilson, “Saharan Trade,” and MacDonald, “A View from the South,” pp. 72-82.

RALP H AUSTBN

deal of gold was being minted in the region.^[900] Closer examination of Byzantine gold sources, the weights and measures used in later Islamic gold trade, and, especially, analyses of trace elements in Byzantine and early Islamic North African coinage do not, however, provide any evidence of a trans-Saharan gold trade before the mid-eighth century ce.^[901] We can be pretty sure that such trade was well under way by the mid-ninth century, on the basis of gold coin molds dating to this time excavated at Tadmekka in the southern Sahara, a development that suggests a level of commerce that must have had an earlier history.^[902] The only literary documentation for a prior date (or any trans-Saharan gold trade before the tenth century) is the report of an Arab “expedition to Sous [southwestern Morocco] and the country of Sudan” sometime after 734 which “got as much gold as he [the leader] wanted.”^[903] However, this narrative was recorded more than a century after the event it describes and is contradicted by subsequent accounts of the same raid.1^[904]

Black African slaves were also found in the ancient Mediterranean world but not - despite a very high general demand for servile labor - in large numbers.^[905] Moreover, those slaves observed on the North African coast and either remaining there or brought on to other destinations probably did not, for the most part, arrive via trans-Saharan trade. A fourth-century ce Syrian

merchant's account of “the Whole World and its Peoples” notes that “the country of Mauretania” (roughly western Algeria and Morocco) “carries on trade in clothing and slaves” but then goes on to describe the lands south of “Africa” (coastal Tunisia and Libya) as “a great desert country” with nothing in it but “a perverse race of barbarians who are called Mazices or Ethiopians.”^[906] This text can be read in two ways: it indicates either little trans-Saharan slave trade (the slaves from Mauretania, on the basis of both geography and independent literary evidence^[907], were Berbers and the Libyan desert populations are of no commercial interest) or the opposite (perhaps sub-Saharan slaves did pass through Mauretania and the knowledge of “Ethiopians” in or around the Fazzan suggests that they regularly came from there to the Mediterranean).

Most of the scant documentation we do have for trade in black Africans comes from Egypt, in part because the survival of papyrus in the dry climate there gives us better records than from other parts of the ancient world. However, there are also some inscriptions telling of the Garamantes deliver­ing such captives to sites in Tunisia as well as the frontier post of Golaia (now Bu Njem) in Libya.^[908] This information is consistent with both Herodotus' account and a later Roman account of Garamantes chariot raids against “Ethiopians.” It seems clear that the extensive imports of Roman goods by the Garamantes had to be paid for with something. Earlier historians assumed this export was “carbuncles,” a semi-precious red carnelian stone found in the northern Fazzan and the only commodity from this region given much attention in Mediterranean texts.^[909] However, these references date to a time prior to the fullest development of the Garamantian civilization, and Wilson makes a good case for the probability that during this time, the Garamantes (whose skeletal remains indicate a significant black African population) would have needed large numbers of slaves to maintain their irrigation system. If such traffic existed, based on the described chariot raids and/or peaceful exchange, the Garamantes could also have sold thousands of such captives to the north each year, a number that would not have made a great demographic impact upon even the servile population of the Roman Empire.^[910] Based upon the number of Roman artifacts imported into the Fazzan, such trade continued well into the period of Garamantian economic decline after 300 ce, but this diminution may explain its apparent invisibility in the fourth-century Syrian merchant's report.

In his account of the Sahara, Herodotus says nothing about gold and very little on slaves but gives a great deal of attention to salt. The Garamantes and all the people around them reside on “hillocks of salt” and in the western desert “There is a mine of salt on it every ten days' journey” and men dwell in houses “all built of the blocks of salt.”^[911] In later times such block salt was highly valued in the Sudan, where it met an especially acute need for supplements to a low-saline cereal diet in a high-temperature region. The Sudan had other sources of salt via locally produced vegetable ash and sea salt from the Atlantic Ocean, but neither of these provided the concentrated salinity as well as transportability of Saharan blocks.

We have no direct evidence of Saharan salt trade in the pre-Islamic era, and historians of this industry are reluctant to project it back beyond the time of the earliest written records.^[912] But Herodotus' account implies a prominence and level of production that is usually associated with commercialization, and the development of agriculture in the contemporaneous Sudan adds to the plausibility of such a market system. If these exchanges existed, the Sudanic exports would probably not have been gold or any large number of slaves but rather agricultural goods and dried freshwater fish, for which there was (and is) a demand in the desert. Whether or not, as Wilson argues, such shorter-range commercial networks indirectly connected the Sudan to the Mediterranean in ancient times, they did lay the groundwork for subsequent trans-Saharan trade and would also continue into the Islamic era, “providing the infrastructure for much trans-regional commerce.”^[913]

While the pre-Islamic inhabitants of the Sahara produced no usable written records, they did leave a legacy of rock art that provides, among other

Figure 24.ι Rock painting of a warrior, probably of the Garamantes tribe, with team of horses, southern Algeria (Robert Estall Photo Agency / Alamy)

information, good indications of the animal transport available during var­ious eras.^[914] The chronology of this art and the worlds it depicts is not easy to establish, but scholars seem agreed that it falls into three periods: a bovidian, corresponding to the Neolithic era and slightly beyond (c. 4000­1000 bce) in which cattle are the main domestic animals and used (possibly along with donkeys) for limited transport; a horse era corresponding to Mediterranean colonization of the Maghrib and Berber domination of the Sahara (c. 1000 bce - 100 ce) and a camel era immediately preceding the beginning of known trans-Saharan trade (beginning c. 100 bce).

Horses and chariots are not only shown on numerous Saharan frescoes but also described in accounts of the Garamantes by Herodotus as well as later Roman sources (see Fig. 24.ι). However, apart from the capture of slaves, this information tells us little about the economic history of the region since chariots are shown, corresponding to their use throughout the Mediterranean, in contexts of warfare, racing, and heroic display rather than commercial transport. The Sahara, as it became increasingly dry, was not very hospitable to wheeled vehicles or even unharnessed horses, who could not survive long journeys across the desert. Merchant caravans con­necting the Maghrib and the Sudan would thus await the arrival and adapta­tion of the camel.

A form of wild camel is native to North Africa, and there is evidence to suggest that such animals never disappeared from the Sahara.^[915] However, they do not appear in local rock art or reports of domestic use until shortly before the Christian era. It is quite probable that camels were reintroduced into the region at this time from farther east or, at the very least, their usage for transport was adopted from practices first developed in Arabia and passed through Egypt.^[916] However this chronology is understood, there remains a gap between the presence of domesticated camels, capable of carrying goods over long distances under the limited food and water conditions of the desert, and the clear initiation of direct and sustained trans-Saharan trade. The explanation for this delay is no longer technolo­gical but political and cultural.

The disorders of North African late antiquity

Widespread camel usage in North Africa was established by the early cen­turies C e yet was soon followed (for no related reason) by a decline in Roman rule as well as general social order throughout the region. For much of the third century C e, the entire Roman Empire was in crisis, from which it was rescued only by the harsh reforms of the Emperor Diocletian (284-305), who initiated a reduction of garrisons from the southern frontiers of the Maghrib, particularly in the Moroccan and Libyan zones that would become the main avenues of trans-Saharan trade.^[917] On the other side of this border, the Garamantian regime also fell into gradual decay from the early 300s, marked by a diminishing supply of water from the foggara irrigation system and the fortification of agricultural settlements. These conditions are linked to raiding by Berber tribes against both coastal cities (recorded in Roman texts) and, apparently, the urban settlements of the Fazzan.^[918]

The reign of Diocletian was followed by that of Constantine, who sought a new basis of unity in the Roman Empire through the embrace of Christianity as its official religion. The impact of this faith upon North Africa was ambiguous. On the one hand, it produced the globally most significant figure of that region's spiritual and intellectual history, St. Augustine of Hippo (354-430 ce). Yet, along with his inspiring Confessions and The City of God, Augustine also wrote lengthy polemics against what he called “the Donatists," a group of dissenting clergy probably representing the majority of local Christians who identified the bishops now recognized by the emperor as heirs of those who had collaborated in Diocletian's earlier persecutions. The “Catholic-Donatist" split divided North African Christians throughout the last period of Roman rule and involved very high levels of violence, including the mobilization of “Circumcellions," itinerant rural laborers who often sought martyrdom in suicidal attacks against both the established church and wealthy landowners.^[919]

The climax of this weakness in the Roman regime was the invasion of coastal North Africa in 429 ce by the Vandals, a Germanic people who ruled the Maghrib for about a century.^[920] Despite their official designation as “barbarians" and the English connotation of their name (based on their sack of Rome in 455), the Vandals did not wreak deliberate destruction on their African territorial base. Instead, they developed a modus vivendi with the Roman landholding elite and maintained exports from the major coastal cities. Nonetheless, the Vandals represented yet another Christian sect (Arians) and were (like their late Roman predecessors and Byzantine successors) unable to establish any effective control over the North African hinterland.

In 533 ce Roman authority was restored to the Maghrib, although this time under the Byzantine (Eastern) Empire based in Constantinople. The Byzantine authorities, like their Vandal predecessors, “were not well-suited to control inland North African transportation and travel or military

Figure 24.2 Camels carrying salt into the village of Araouane, Mali. Araouane is the main stopping point for camel caravans working the salt trade between Taodenni and Timbuktu (© George Steinmetz/Corbis)

movements.”^[921] The two centuries of their rule witnessed the continuation of religious divisions and Berber uprisings, along with general economic decline, dissension in the military ranks and “seemingly insatiable demands for revenue from the central Byzantine imperial authorities.”^[922] The establish­ment of sustained and direct links between Mediterranean and Sudanic Africa would thus await yet another external colonization, that of Muslim Arabs in the seventh century c e.

The Arab conquest

What the new Muslim regime eventually provided for this larger region was a stable political order extending throughout the Maghrib, an established practice of long-distance camel caravan commerce, and a religion that would bind rather than divide the communities on both sides of the Sahara (see Fig. 24.2). These conditions did not come into place until well after the initial Arab invasions of North Africa and even then in a somewhat paradox­ical manner.

The Muslim conquest of North Africa was a prolonged, often disorderly affair extending from 643 ce until the first decade of the following century and concentrating, as had earlier colorizations, on the Mediterranean coast. The forces of Islam took control of Egypt in 642 and within the next year had established a foothold in eastern Libya. For the next twenty­eight years, however, their ventures further west consisted mainly of raids designed both to defend Egypt against Byzantine counterattacks and to exact tribute from North African populations. Only with the foundation of the new city of Qayrawan (caravan) in the near interior of Tunisia between 670 and 675 did Arab forces create a base in the heartland of the Maghrib. It took them until 698 to drive the Byzantines definitively from the major coastal cities and a few more years to establish control over northern Morocco and move from there, in 711, into the neighboring European region of Spain.

Among the raids which accompanied or immediately followed this westward march were at least two expeditions into the Sahara. We know very little about the c. 734 move into southern Morocco and (possibly) its gold trade, but a venture of 666-67 into and beyond the Fazzan by the legendary Uqba ibn Nafi has been recorded in some detail. In geographical extent, Uqba's raid did not represent a greater penetration of the desert than several known Roman efforts in conjunction with the Garamantes. The formulaic account, by the ninth-century Egyptian chronicler Ibn ‘Abd al-Hakam, of Uqba's encounters with a series of Saharan rulers, also reads more like a literary trope than a plausible statement of what actually occurred:

He seized the king and cut off his finger. The king asked him: “Why have you done this to me?” Uqba answered: “As a lesson to you, for when you look at your finger, you won't make war on the Arabs.” Then he imposed on him a tribute of 360 slaves.^[923]

Yet however fanciful these procedures and numbers may be, they reflect a new relationship between Mediterranean conquerors and the African hinter­land, one that would ultimately produce enduring and wide-ranging trans­Saharan links, but only through the initial creation of even greater disorder in the Maghrib. A major basis for both these conditions was the renewal and intensification by the Arabs of a Roman policy that had largely been aban­doned by the Vandals and Byzantines: the forcible recruitment of large numbers of Berbers as slaves for both local use and export.^[924]

The inland Berbers responded to these impositions as they had to previous invasions, by forming their own large-scale states. As a result of Berber counterattacks, as well as political instability in the Caliphate at home, Uqba's base of Qayrawan was abandoned several times between 670 and 700. Uqba himself died around 683 in a battle against Kusayla, a Berber leader who maintained his own empire in Tunisia and eastern Algeria for several years. In about 688 Kusayla was, in turn, killed by a newly arrived Arab force. But in 790s the invaders again gave way, this time to a Berber queen known mainly by her Arabic sobriquet, “al-Kahina” (the diviner), whose looser regime was only destroyed in 703.^[925]

Kusayla and al-Kahina are justifiably remembered in North African histor­iography as bearers of “the torch of resistance” against alien domination.^[926] Nonetheless, as with earlier Berber kingdoms, both drew upon the culture of coastal colonizers for their political practices and hybrid identity. Kusayla was even an ally of the Byzantine regime, still based at his time in Carthage, and al-Kahina, who flourished after Carthage had fallen to the Arabs, nonetheless appears to have been more or less Christian. The future of the North African interior was, however, manifested by the army that defeated Kusayla and consisted of both Arabs and Berbers. From this point on, resistance to the external forces as well as the capacity to reach across the Sahara would draw upon dissident versions of Islam.

Kharajite Muslims and the beginnings of trans-Saharan trade

Once the conquest of North Africa had been completed, Islam gradually became the overwhelmingly dominant religion of the Maghrib. Christianity survived for another 400 years but as an increasingly marginal faith which disappeared almost completely with the persecutions of the twelfth-century Almohades regime.^[927] The major echo of Christian North African history in the early career of its Muslim successor was the prevalence of dissident sects. In the case of Maghribi Islam, the major threat to Sunni orthodoxy during this period came not from Shia, followers of an alternative succession to the heritage of the Prophet Muhammad, but rather through sects of the Kharajite (secessionist) movement, which rejected any line of hereditary (and necessa­rily Arab) caliphs or imams.

North Africa along with Spain was originally conquered in the name of the Sunni Umayyad Caliphate, based in Damascus, with authority over both regions delegated to governors based in Qayrawan. In the course of the eighth century, this structure came apart because of local Berber rebellions and instability at the center of the Caliphate, which shifted in 750 to Baghdad under a new (and initially less clearly orthodox) dynasty, the Abbasids. The first definitive break came in 750 when the Spanish Umayyads refused to recognize the Abbasids and formed their own caliphate based in Cordoba. The Maghrib experienced more prolonged and vacillating efforts to assert caliphal rule, although, by the end of the century, it had broken into separate Muslim states that, except for a brief Almohades effort at reunification, characterizes the region's political geography up to the present. The religious inspiration for Berber rebellions against representatives of the Caliph and the basis of at least two of the regional successor states was Kharajism.^[928]

This dissident movement originated at the centers of the early Islamic world in western Arabia and Iraq but achieved political power only in peripheral areas such as North Africa, where emissaries of this sect are said to have arrived around 719. The first political manifestation of Maghribi Kharajism occurred between 739 and 742 when a massive Berber revolt broke out, first in the north of Morocco but subsequently as a threat to the Arab governing base of Qayrawan. A second such outbreak in the latter 750s, when the Abbasids were still consolidating their authority, represents the peak of Kharajite power, in which the dissidents occupied both Qayrawan and coastal Tripoli. In 762 caliphal forces regained control over the territories as far west as central Algeria. The Arab family that came to dominate the governorship here, the Aghlabids, remained nominally loyal to the Abbasids but by 800 had been recognized as hereditary rulers of what amounted to an autonomous state that lasted until 909. In northern Morocco another Arab dynasty, the Idrissids, claiming descent from the Prophet Muhammad via his nephew Ali, took independent power in 788 and survived until 959. But the interior of western Algeria and the south of Morocco came under the rule of two Kharajite regimes that also pioneered trans-Saharan trade.

The largest of these states, the Rustamid imamate (778-909), based at Tahert in western Algeria, belonged to the Ibadiyya, a moderate version of Kharajism that managed to coexist peacefully with Sunni Muslims both in Iraq and, once the Abbasids had retaken Qayrawan, in North Africa. The Rustamids violated Kharajite doctrine by granting leadership to the descen­dants of their founder, Abd al-Rahman ibn Rustam, a scholar of Persian merchant origin whose initial status derived from his training at the center of Ibadi learning in Basra, Iraq. At the base of Tahert state was a confedera­tion of Berber tribes and “the resulting Ibadi-Berber fusion incorporated all the elements necessary for a long-distance trade network.”^[929]

The other Kharajite state of this era, built around Sijilmasa in Morocco (757-977), was located immediately on the northern desert edge and adhered to the Sufri sect. Because the Sufris did not, like the Ibadis, produce many surviving writings of their own, we know them only from descriptions by others. They played a major role in the first wave (739-61) of Berber-Kharajite rebellions in North Africa and continued to sponsor uprisings against the Idrissid regime in Morocco. However, even if less “moderate” than the Ibadis, the rulers of Sijilmasa (eventually ruled by their own hereditary Midrari dynasty) appear to have retained peaceful relations with neighboring North African states, an attitude that may be attributed in both Kharajite cases to their increasing role in trans-Saharan trade.

In later and better-documented periods, the western Sahara trade routes passing through the region around Sijilmasa specialized in gold, a logical consequence of its greater proximity to the sources of this precious metal in the Sudan. There is some evidence suggesting gold trade in this region even before the Sufris settled here.^[930] It is very likely that the Banu Midrar Sufris initiated or at least significantly expanded this commerce which was clearly under way via Sijilmasa by the early ninth century, but we have no detailed records about such commerce before this time.⁴⁶

Again, as in Roman times, there is more information about early Muslim Saharan trade in slaves than in gold. The excessive demands for Berber slaves and even treatment of Islamic troops as if they were private property lay behind the Kharajite rebellions of the mid-eighth century and were even recognized as illegitimate by the first Abbasid governor of the Maghrib, since the Berbers were now Muslim. Farther south in the Sahara, particularly around the still relatively populated Fazzan, lay two new potential sources of slaves: still unconverted Berbers and the much more numerous peoples of the Sudan.

A phrase attributed to the second Imam of Tahert, Abd al-Wahab (784­823), claims that the success of his Ibadi religion was based upon “the Nafusa sword” and “Mazata wealth.” The Nafusa, a Berber community from the highlands of the northern Libyan coastal region, provided the Rustamid state with its critical military and administrative support. But the key to trans­Saharan trade was the adherence to Ibadism of the Mazata, herders of donkeys, goats, and camels from Waddan, close to the desert edge. In yet another paradox, the Mazata, who were heirs to the conflicts between Berber tribes and the Garamantes, reinvented and expanded that regime's cross­desert trade. Around the same time as Rustamids established their capital at Tahert, the Mazata took over Jarma, the capital of a now completely collapsed Garamantian regime, and then established a new trading entrepot farther south in the Fazzan at Zawila. From here they extended their commercial network to the farthest extent of Uqba's earlier expedition, Kawar in the eastern Sahara, and by at least the early ninth century were in regular contact with the Saharan-Sudanic kingdom of Kanem. This region of the Sudan produced no gold, so the principal commodity passing northward through its Ibadi commercial system was again slaves.

Early Arabic Sources, pp. 62 and 66; for more such indications of early commercial activity, see Tadeusz Lewicki, “Les origines de l'Islam dans les tribus berberes du Sahara occidental: Miisa ibn Nusayr et'Ubayd Allah ibn al-Habhab,” Studia Islamica, 32 (1970): 203-14; evidence from Tadmekka (Nixon et al., “New Light,” and Nixon, “Excavating Essouk-Tadmakka”) points to a more eastern gold route.

46 The oldest references to gold trade as such are in al-Masudi (written 947-956), citing an earlier author, al-Fazari (c. 820); see Hopkins and Levtzion, Corpus ofEarly Arabic Sources, pp. 30 and 32; the Midraris are also the oldest North African dynasty to produce coinage based on Sudanic gold, although when exactly this occurred is not clear (Gondonneau and Guerra, “The Circulation of Precious Metals,” 587).

Postscript: forming the classic trans-Saharan world

Sometime near the end of the tenth century, a prominent Muslim scholar of Qayrawan, Ibn Abi Zayd (922-996), produced a canonical catechism of the dominant Maliki school of Sunni law in which he stated: “Trade to the territory of the enemy and to the Bilad-as-Sudan is reprehensible. The Prophet said: ‘the journey is a part of the punishment.'”^[931] This pronounce­ment indicates that at the date the present volume is supposed to termi­nate, 900 ce, the existing trading links across the Sahara were still far from uniting the two sides of the desert into a common culture. For Ibn Abi Zayd's judgment to be altered, it was necessary that an orthodox version of Islam prevail across the entire Maghrib and Sahara as well as extending at least to the rulers in the Sudan and also that caravan traffic across the desert become secure enough so that it could engage resources from all over North Africa. These conditions would not be met until the end of the following century.

The initiation and early domination of trans-Saharan trade by Kharajites was both a symptom and a cause of its bad reputation among orthodox Muslims. We can assume that the retreat of both Sufris and Ibadis from efforts to control Qayrawan and Tripoli left them with few options for prosperity other than to exploit their links to desert Berbers. For Ibadis there was also a doctrine of kitman (concealment), developed when they had to live under hostile Muslim authorities in Iraq that presumably allowed them to accept the rule of “enemy,” that is, pagan kings in the Sudan, while also justifying the purchase of such kaffirun (non-believers) as slaves.^[932]

The shift toward a Sunni trans-Saharan regime came in stages. The first, again paradoxically, was the rise, in eastern Algeria during the early 900s, of the Shia Fatimid movement. The Fatimids both destroyed the Rustamid state of Tahert in 909 and a year later took over Tunisia from the Aghlabids. The Ibadis fled from their northern base of political power but became even more committed to controlling trans-Saharan trade by shifting their center of learning and commerce to Wargla, on the desert edge, a site that gave them access to Western Sudanic gold sources via Tadmekka, as well as Fazzan routes to the slaves of the Central Sudan.^[933] The major centers of the Fazzan also remained under Ibadi rule until the end of the twelfth century.^[934] However, in Tunisia the efforts to promote Shiisim by the Fatimids and the governing dynasty they left behind when departing for Egypt in 972, the Zirids, only provoked stronger commitments to Maliki Sunnism by Qayrawan scholars such as Ibn Abi Zayd. By about 1140 the Zirids formally broke with Fatimid Egypt and its doctrines, thus ending the most extended episode of Shiism in North Africa.

Maliki orthodoxy only penetrated the western Sahara and Morocco in the mid-eleventh century with the rise of the Almoravids (Arabic: al-Murabitun) among the Sanhaja Berbers in the desert portions of this region. The founder of the movement, Abdullah ibn Yasin, was himself a Maliki scholar sent into the desert by Qayrawan authorities at the request of local leaders. Politically, the Almoravid empire (1073-1146) represents a sensational but short-lived conquest of Morocco, Spain, and even a portion of the Sudan by Saharan Berbers. In religious and cultural terms, Almoravid influence was more enduring since they appear to have destroyed the remnants of Ibadism, Shiism, and other more local Islamic sects in at least the western half of the trans-Saharan world.^[935]

There can be little doubt that the Almoravids were driven by religious fervor, but their political emergence also owed something to competition over western Sahara trading entrepots between Sanhaja Berbers and newly developing Sudanic states, particularly the Soninke Ghana empire. The eleventh-century geographer al-Bakri describes the capital of Ghana, in terms consistent with Ibn Abi Zayd's vision, as divided into two towns, one “inhabited by Muslims” (i.e. Maghribi merchants) and the other “king's town,” whose “religion is paganism and the worship of idols.” Yet al-Bakri also notes that the rulers of two lesser Western Sudan kingdoms of this era, Takrur and Gao on the middle Niger, had already converted to Islam.^[936] By the latter eleventh century, Ghana also became Muslim, possibly under the influence of the Almoravids, and may even have joined the latter in campaigns for “the eradication of early Ibadi influences from the southern Sudan and the western Sahara.”^[937]

Almoravid orthodoxy also threatened another group that played a part in early trans-Saharan trade, the Jews of North Africa. Like Kharajites, Jews were left to carry on a whole range of occupations seen as inappropriate for good Muslims, although, unlike Christians, they welcomed the shift from the rule of the Byzantines (who persecuted them) to that of the Arabs.^[938] Kharajite communities in Tahert, and at such desert-edge settlements as Wargla and Sijilmasa, proved hospitable to Jewish artisans and merchants who provided valuable metalworking skills as well as commercial links with their brethren in both the east (Tunisia and Egypt) and the northwest (Spain).^[939] The rise of the Almohades in the twelfth century dealt more serious blows to Jewish religious practice and commerce than had the Almoravids, but again, unlike Christians, Jewish communities recovered from this setback and played an even greater (and better-documented) role in later trans-Saharan commerce.^[940]

In the Central Sudan the Kanuri Kanem empire was founded, with a base in the southern desert, by sometime in the ninth century and became a major slave-trading partner of Ibadi merchants in the Fazzan.^[941] The Kanem rulers converted to Islam, in its Sunni orthodox form, by at least the eleventh century, and not long after this began to undertake pilgrimages to Mecca.^[942]

Pilgrimages from the Sudan, which always passed through North Africa, doubly confirmed the establishment of an integrated trans-Saharan world.

Not only did they act out one of the major pillars of the newly shared religion, but they also demonstrated that a secure system of movement across the desert was now in place. The technology of camel transport did not change much in the ensuing centuries, although caravans - formed by the seasonal joining together of various merchants around a single paid guide and his staff - grew considerably in size. The influence of Islam in both the Sahara and the Sudan also increased in width as well as depth, eventually impacting on the management and adjudication of caravan trade.^[943] However, the basis for all these developments was laid out by the environmental, social, political, and religious struggles in this portion of Africa during the Neolithic, ancient, and early Islamic eras.

Further Reading

Primary sources

Herodotus, The Histories, trans. A. D. Godley, Cambridge, ma: Harvard University Press, 1921, vol. ii, pp. 383-89.

Hopkins, J. F. P., and Nehemia Levitzon, Corpus of Early Arabic Sources for West African History, Cambridge University Press, 1981.

Secondary sources

Abun-Nasr, Jamil M., A History of the Maghrib, Cambridge University Press, 1975.

Austen, Ralph A., Trans-Saharan Africa in World History, New York: Oxford University Press, 2010.

Brett, Michael, and Elizabeth Fentress, The Berbers: The People of Africa, Oxford: Blackwell, 1996.

Bulliet, Richard W., The Camel and the Wheel, Cambridge, ma: Harvard University Press, 1975.

De Villiers, Marq, and Sheila Hirtle, Sahara: A Natural History, New York: Walker, 2002.

Dowler, Amelia, and Elizabeth R. Galvin (eds.), Money, Trade and Trade Routes in Pre- Islamic North Africa, London: The British Museum, 2011.

Encyclopedie berbere, Aix-en-Provence: EDISUD, 2008.

Encyclopedia of Islam, 2nd edn., ed. P. Bearman, Th. Bianquis, C. E. Bosworth, E. van Donzel, and W. P. Heinrichs. Leiden: Brill, 2005.

Gruen, Erich S., Rethinking the Other in Antiquity, Princeton University Press, 2010.

RALP H AUSTBN

Harper, Kyle, Slavery in the Late Roman World, ad 275-425, Cambridge University Press, 2011.

Hirschberg, H. Z., A History of the Jews in North Africa, Leiden: Brill, 1974, vol. 1

KaegiJr., Walter E., Muslim Expansion and Byzantine Collapse in North Africa, Cambridge University Press, 2010.

Law, R. C. C., “North Africa in the Hellenistic and Roman Periods, 323 bc to ad 305,” in J. D. Fage and Roland Oliver (eds.), The Cambridge History of Africa, vol. ii, Cambridge University Press, 1978, pp. 148-290.

Lydon, Ghislaine, On Trans-Saharan Trails: Islamic Law, Trade Networks and Cross-Cultural Exchange in Western Africa, New York: Cambridge University Press, 2008.

Mattingly, David (ed.), The Libyan Desert: Natural Resources and Cultural Heritage, London: Society for Libyan Studies, 2006.

McIntosh, Susan Keech, “Changing Perceptions of West Africa's Past: Archaeological Research Since 1988,” Journal of Archaeological Research 2 (1994): 165-98.

Merrills, A. H. (ed.), Vandals, Romans and Berbers: New Perspectives on Late Antique North Africa, Aldershot: Ashgate, 2004.

Muzzolini, Alfred, “Livestock in Saharan Rock Art,” in Roger M. Blench and Kevin C. MacDonald (eds.), The Origins and Development of African Livestock: Archaeology, Genetics, Linguistics, and Ethnography, London: UCL Press, 2000, pp. 87-110.

Savage, Elizabeth, A Gateway to Hell, a Gateway to Paradise: The North African Response to the Arab Conquest, Princeton, nj: Darwin Press, 1997.

Schiedel, Walter, “The Roman Slave Supply,” in Keith Bradley and Paul Cartledge (eds.), The Cambridge World History of Slavery, Cambridge University Press, vol. I, 2011, pp. 287-310.

Shaw, Brent D., Sacred Violence: African Christians and Sectarian Hatred in the Age of Augustine, Cambridge University Press, 2011.

Vik0r, Knut S., The Oasis of Salt: The History of Kawar, a Saharan Centre of Salt Production, Bergen, Norway: Centre for Middle Eastern and Islamic Studies, 1999.

Wilson, Andrew, “Saharan Trade in the Roman Period: Short-, Medium- and Long­Distance Trade Networks,” Azania: Archaeological Research in Africa 47 (2012): 409-49.