. 3
( 12)


existence in competitive and often violent societies where, as later pre-colonial
evidence shows, kinship groups falling below a minimum size were simply
absorbed by more fertile rivals in a process of natural selection. ˜A race is as
fragile as a newborn child™, said a Congolese proverb. Capture of people was a
major aim of warfare. Fertility of women was a major subject of art. Care of
70 africans: the history of a continent

the pregnant and newborn was a central concern of medicine and ritual. This
African obsession with reproduction later surprised anthropologists familiar
with regions where nature was more benign.
There are no data suf¬ciently reliable to permit estimates of birth- or death-
rates at this time, although both were probably high. Average life expectancy at
birth was probably less than twenty-¬ve years (its level in the second-century
Roman Empire) and possibly less than twenty. Educated guesses have sug-
gested that population may have grown by an average of two or three per
thousand per year over the long term, although even that would have been
rapid by the standards of Ancient Egypt and other traditional societies.6 Judg-
ing from modern parallels, up to one-third of babies may have died in the ¬rst
year of life and an unusually large proportion during the next four years, for
western Africa™s malarious climate, widespread lack of animal milk (owing to
trypanosomiasis), and medical practices were especially pernicious to small
children. One Muslim leader in late eighteenth-century Hausaland fathered
forty-two children of whom only ¬fteen reached puberty, as did only thirteen
of his eldest son™s thirty-three male children.7 Among the Anyi of modern Cote ˆ
d™Ivoire, whose society took shape in the eighteenth century, only a woman™s
fourth dead child had the right to a funeral. The vulnerability of children proba-
bly explains why birthrates were not even higher. The slender evidence suggests
that most western African women married at least as soon as they could bear
children. Yoruba women freed from slave ships in the early nineteenth century,
for example, had on average borne their ¬rst child at about twenty, probably
soon after becoming fecund. Yet both the earliest colonial evidence and sub-
sequent estimates by demographers suggest that women may have averaged
little more than six births during their reproductive lifespans, many fewer than
was theoretically possible. Arti¬cial contraception is unlikely to have been the
reason, for western Africans made little use of herbs for this purpose, and then
probably ineffectively. Rather, the main constraint on fertility was probably
the spacing of pregnancies, as was still the case in the twentieth century. The
chief mechanism was probably prolonged and frequent breastfeeding, which
inhibited conception and was especially necessary where only human milk was
available. A visitor to the Gold Coast reported in 1785 that breastfeeding might
last four years. A doctor travelling in Borno in 1870 suggested an average of two
years. Breastfeeding was often supplemented by taboos against intercourse so
long as a woman had a totally dependent infant. A perceptive European trader
reported the normative rule on the River Gambia during the 1730s, although
adding his own scepticism:

No marry™d Women, after they are brought to Bed, lie with their Husbands
till three Years are expired, if the Child lives so long, at which Time they wean
their Children, and go to Bed to their Husbands. They say that if a Woman
Colonisation in western Africa 71

lies with her Husband during the Time she has a Child sucking at her Breast,
it spoils the Child™s Milk, and makes it liable to a great many Distempers.
Nevertheless, I believe, not one Woman in twenty stays till they wean their
Children before they lie with a Man; and indeed I have very often seen Women
much censur™d, and judged to be false to their Husbands Bed, upon Account
only of their sucking Child being ill.8

Practice no doubt varied, but birth intervals of three or four years were widely
reported in the early colonial period. The object was presumably not to limit
children but to maximise them by ensuring that they and their mothers sur-
vived, for modern evidence shows high mortality among children born either
before or after a short birth interval. Not only did long birth intervals limit
pregnancies, but they prevented rapid recuperation of a population decimated
by a catastrophe. In western Africa, the price for any population growth was
that it could be only slow growth.

political development in the savanna
In the West African savanna, underpopulation was the chief obstacle to state
formation. While sparse populations could not supply the surplus to support
ruling classes, denser populations had little incentive to do so when empty land
enabled them to evade political authority. The lack of evidence of a differenti-
ated ruling elite in the Niger Valley during the ¬rst millennium ad suggests that
social complexity did not require state organisation. In the second millennium,
similarly, many of the largest population concentrations remained entirely
stateless, jealously defending their freedom as colonists, regulating their affairs
by negotiation and the threat of retaliation, clustering together to resist preda-
tory neighbouring states. This pattern existed especially in Voltaic-speaking
regions (notably modern northern Ghana) and among skilled highland cul-
tivators. Whatever authority existed in these regions often belonged to the
descendants of pioneer settlers from whom late-comers ˜begged bush™. Among
the Serer of modern Senegal, for example, such ˜masters of ¬re™ were the only
political authorities until the fourteenth century. Their counterpart among
Mande-speakers, the largest group in the western savanna, was the fama, who
was both a master of the land and the political chief of a kafu, a group of villages
forming a miniature state. ˜In the middle of the forest™, wrote a nineteenth-
century traveller, ˜are immense clearings several kilometres in diameter. In the
centre are grouped seven, eight, ten, often ¬fteen villages, individually forti-
¬ed. This sort of confederation has its chosen chief who takes the title of Fama.
The chief™s village gives its name to the group.™9 The kafu was the enduring
political community of the savanna, the building-block with which larger but
more ephemeral polities were constructed. In this it had parallels throughout
72 africans: the history of a continent

the continent and in the micro-states of predynastic Egypt, the nadu of South
India, or the subimperial communities of pre-Columban America. The kafu
embodied the pervasive localism of African politics. Kings and conquerors
seeking to transcend it might root their states in concentrations of population
and wealth, the most enticing in the savanna being in the Niger Valley. They
might also rely on slave labour, long-distance trade, and sheer military force.
Invariably, however, their authority diminished with distance from the capital,
fading into a stateless penumbra where, as a later traveller put it, ˜the inhab-
itants hardly know whose subjects they are™. Underpopulation also set other
constraints on political consolidation. The polygynous marriage patterns of
colonising societies gave rulers swarms of sons to demand of¬ces, contest the
succession, and fragment the state if they could not rule it, especially where no
religious institutions provided the safety-valves for surplus sons available in
Europe and Asia. The powerful kinship groups needed to clear and defend new
land gave society a strength that the state could seldom tame. The mingling
of mobile colonists bred populations heterogeneous in customs and loyalties.
˜Power is like holding an egg in the hand™, said an Akan proverb from modern
Ghana. ˜If you hold it too tightly it breaks, and if you hold it too loosely, it
drops.™ State-building in the savanna, then as now, was a search for devices to
counteract localism and segmentation.
These dynamics can be seen best in the history of Mali, the dominant state
in the western savanna from the thirteenth to the ¬fteenth century. It began as
a kafu and then a cluster of kafus on the upper Niger, as a bard reminded its
founder, Sunjata Keita, before the battle in c. 1235 that made him king. ˜From
being village chiefs the Keitas have become tribal chiefs and then kings™, the
bard declared. ˜Cut the trees, transform the forests into ¬elds, for then only
will you become a true king.™10 At the kingdom™s core, villages of craftsmen and
other specialists clustered densely. Beyond them was the fertile agriculture of
the Niger Valley, and beyond that territories sprawled from the Atlantic to the
desert and the forest, with governors and garrisons of conquered provinces
interspersed with semi-independent vassals. In part this was a product of
Mande expansion that long predated Sunjata. In part it was stimulated by
his triumph. The ¬rst Mande-speakers to disperse widely may well have been
hunters. Behind them went a more permanent migration of traders, craftsmen,
and agriculturalists who penetrated southeastwards to the Akan gold¬elds of
modern Ghana or sought kola nuts in the forests to the southwest, where the
Vai and Dan peoples of modern Liberia, the Gouro of Cote d™Ivoire, and the
Kono and Kpelle of Guinea were all Mande-speaking groups. A third phase
of expansion was more violent, for the creation of the Mali kingdom and
the decline of rainfall allowed its horsemen to penetrate southwards and west-
wards, establishing Mande-controlled chiefdoms along the Gambia and among
the Serer during the fourteenth century. Such was Mali™s prestige that even the
Colonisation in western Africa 73

non-Mande-speaking rulers of Gonja in modern Ghana claimed descent from
Malian cavalrymen sent to control the gold trade.
Yet Mali suffered the weaknesses of a savanna polity. Its polygynous royal
family was divided between Sunjata™s descendants and his younger brother™s.
Gao, probably conquered around 1300 and the key to the fertile eastern
provinces in the middle Niger Valley, was lost again about a century later.
In 1433“4 Tuareg nomads from the neighbouring desert took Timbuktu. Jenne
appears to have regained independence from Mali at that time. Meanwhile the
western half of the empire was in¬ltrated by Fulbe cattlemen, Niger-Congo
speakers who emerged as a specialised pastoral group on the upper Senegal
and began to drift eastwards early in the second millennium. At ¬rst they
acknowledged Mali™s authority, but as it weakened and Fulbe numbers grew,
the pastoralists created a pagan state in Futa Toro (the old Takrur) at the end
of the ¬fteenth century. Mali was by then disintegrating. First its successor
on the middle Niger, Songhay, sacked the capital in 1545“6. Then a disastrous
attempt to reconquer the middle Niger in 1599 lost Mali the Bambuk gold¬eld.
Its few surviving provinces seceded. During the 1630s, Mande-speaking and
largely pagan Bambara cultivators destroyed the capital. Its bards and courtiers
retreated to Kaba, where chiefs had once sworn allegiance to Sunjata. The kafu
once more dominated the upper Niger.
Further to the west, early in the second millennium, Serer and Wolof peoples
colonised southwards from the Senegal Valley into Senegambia, perhaps in
response to desiccation and the growing power of Islam. The ¬rst new state
recorded here seems to have emerged by the twelfth century in Waalo on
the lower Senegal, where cultivators could utilise seasonal ¬‚ooding. During
the next two centuries, power shifted inland to the dry savanna region of Jolof,
perhaps responding to Mali™s commercial prosperity. By the late fourteenth
century, Jolof had renounced any loose allegiance to Mali, but its own authority
over Wolof to the south and west was slight and ¬‚uctuating, for, as in Mali,
the underlying political units were local chiefdoms headed by descendants
of pioneer colonists, military noblemen dominating commoners and slaves.
European trade may have assisted those in the coastal kingdom of Kajoor to
defeat Jolof during the mid-sixteenth century, but a more important reason
for Jolof™s disintegration into four successor kingdoms was probably the new
Fulbe state in Futa Toro, which blocked its access to the interior.
Mali™s successor to the east, Songhay, was probably created by Nilo-Saharan-
speakers and extended nearly two thousand kilometres along the Niger Valley.
From the twelfth century, its capital was at Gao. Subjected to Mali™s overlord-
ship during the fourteenth century, it recovered independence under a military
dynasty whose power peaked between 1464 and 1492 under Sonni Ali Ber. Its
settlements of slave cultivators probably made the Niger Valley more produc-
tive during the relatively favourable rainfall of the sixteenth century than at any
74 africans: the history of a continent

later period. The state also exploited peasant farming and the trade of Jenne and
Timbuktu. Backed by a small standing army, probably mostly slaves, the regime
administered the Niger Valley directly through appointed royal kinsmen, leav-
ing indigenous tributary rulers to govern outlying provinces. The structure
was created partly by Sonni Ali and partly by a former provincial governor,
Askiya Muhammad Ture, who usurped the throne with Muslim support after
Sonni Ali™s death in 1492. Like Mali™s kings, however, the Askiyas never estab-
lished a stable rule of succession. Repeated con¬‚ict among the proliferating
royal family and military nobility left the state divided when competition for
the gold trade led Sultan Ahmad al-Mansur of Morocco to march twenty-¬ve
hundred newly armed musketeers and ¬fteen hundred cavalrymen across the
desert in a daring assault on Songhay. At Tondibi on 12 March 1591, they routed
an army alleged to include ten thousand to twenty thousand horsemen, but
resistance and disease prevented them from subduing a Songhay successor state
in the southeastern marches of the old empire. Instead the Moroccan troops
withdrew to Timbuktu, lost much allegiance to Marrakesh, and degenerated
into a brutal local tyranny. Between 1651 and 1750, Timbuktu had 128 mili-
tary rulers. Although Islam expanded and deepened during this period and
successor states took shape, it was in general a time of economic and political
decay. The valley population declined as agricultural settlements dispersed.
Famine and epidemic became increasingly common. Fulbe pastoralists in¬l-
trated from the west and Tuareg from the north, besieging Gao in c. 1680 and
penetrating south of the Niger by 1720. Bambara cultivators looted Jenne and
established their kafu microstates amid the ruins of the empire. Trans-Saharan
trade continued to shift eastwards into the central savanna.
Here the dominant power from perhaps the sixth century had been Kanem,
a largely pastoral state north of Lake Chad, speaking a Nilo-Saharan language,
specialising in the northward export of slaves, and ruled from about 1075 by
the Saifawa dynasty. During the fourteenth century, internal dissension and
perhaps declining rainfall caused the Saifawa to move their headquarters to
Borno on the plains southwest of Lake Chad. This region had greater agricul-
tural potential and the state lost its pastoral character, but slaves were even
more easily available among southern agricultural peoples and remained the
chief export. Borno, even more than Songhay, was dominated by an aris-
tocracy of mounted warriors who drew tribute from allotted agricultural
communities, distinguished themselves from commoners by dress and pro-
nunciation, and gloried in warfare. Given this ethos and the fact that any king™s
son was eligible for the throne, succession war remained endemic during the
¬fteenth century. Thereafter the state was somewhat stabilised. During the six-
teenth century, it conquered many surrounding agricultural peoples, provok-
ing the Mandara and other groups to organise states in self-defence. Mai Idris
Aloma (1571 “1603), Borno™s most famous warrior king, prosecuted these wars
Colonisation in western Africa 75

relentlessly. Borno prospered during the favourable rainfall of the early seven-
teenth century, administering its central territory through royal slaves and its
outlying provinces through military vassals. Its endurance as a state for over a
thousand years, like Ethiopia™s, owed much to its sense of cultural superiority
as the guardian of a world religion amidst stateless peoples and indigenous
Their retreat southwestwards from Kanem brought Borno™s rulers into closer
contact with the plains to the west, which are today occupied by the Hausa peo-
ple of Northern Nigeria. Hausa origins are a mystery. They speak a relatively
homogeneous Afroasiatic language whose closest af¬nities are to the east in
modern Chad, but many scholars think that the ancestral Hausa must have
retreated southwards from Saharan desiccation. Dalla Hill, in modern Kano
City, was certainly an ironworking site in the seventh century ad, but it is uncer-
tain whether its inhabitants were Hausa or Niger-Congo-speakers subsequently
absorbed. Traditions recorded in the seventeenth-century Kano chronicle and
by modern researchers suggest that in the early second millennium ad Hausa-
land was divided into many microstates, often clustering around ironworking
centres or the granite outcrops sacred to nature spirits. Although Islam may
have reached the region from Kanem at an earlier date, the Kano chronicle
emphasises the arrival of traders during the mid-fourteenth century, possibly
from Songhay, and certainly some impulse must then have drawn Hausaland
into savanna and desert trade, for during the ¬fteenth century new trading
polities emerged not only there but in Agades to the north and Yorubaland
to the south. By the late sixteenth century, European merchants from Ragusa
(Dubrovnik) had lived in Kano and ranked it with Fes and Cairo as one of
Africa™s three major cities. ˜Many white gentlemen live there, who have betaken
themselves there from Cairo many years ago™, they reported. ˜They have a way
of life such that many of them possess horses in their own stables and are served
like lords by numerous slaves.™11
This stress on slaves and horses suggests that Hausaland™s transformation
in the ¬fteenth and sixteenth centuries was not merely commercial but politi-
cal. The sarauta (of¬cial title) system celebrated in the Kano chronicle involved
the uni¬cation of microstates into kingdoms, the building of walled capital
towns like Kano and Katsina, the appointment of titled administrators (often
on Bornoan models), the import of more powerful war-horses, systematic
slave-raiding among Niger-Congo-speakers to the south, recurrent warfare
among the new kingdoms, the adoption of Islam by the ruling class along-
side indigenous religious practices, and urban domination of the countryside.
Muhammad Korau (c. 1444“94), founder and builder of Katsina, personi¬ed
the new order, but behind it lay profound demographic and social changes:
an in¬‚ux of people from many directions, cultural mingling, a growth of ter-
ritoriality as against kinship, economic specialisation and differentiation seen
76 africans: the history of a continent

in urbanisation and the proliferation of occupations, and probably intensi¬ed
agriculture in manured ˜close-settled zones™ surrounding walled cities. More
clearly than the long-distance trade of Songhay or Borno, Hausaland™s com-
merce was rooted in the agriculture, craft production, and local exchange
of a population dense enough to escape some of the centrifugal tendencies
of colonising societies. This, however, lay largely in the future. Sixteenth-
century Hausaland was still racked by warfare as its new kingdoms jostled for
Behind these political changes in the savanna lay military innovations. Until
perhaps the thirteenth century, infantry dominated West African battle¬elds.
Free bowmen were the core of Mali™s army, while warfare among stateless
peoples often resembled a tournament with few casualties. Horses reached the
savanna from the north during or even before the ¬rst millennium ad, but
they apparently either lost size in a less favourable environment or were small
ponies that gave their owners an advantage in mobility rather than combat,
especially because they were ridden without saddles, stirrups, or bits. These are
the horses depicted in the magni¬cent terracotta statuettes excavated from the
Niger Valley. Larger breeds of war-horses with the necessary harness probably
reached West Africa during the thirteenth century. The model may have been
the Mamluk cavalry of Egypt, for their ¬rst use in the savanna is attributed
to Mai Dunama Dibalemi (c. 1210“48) of Kanem, the state most in touch
with Egypt. Mali adopted the new techniques by the 1330s. The Kano chronicle
attributed them to Sarki Yaji (c. 1349“85). Wolof states possessed a few horses
by the 1450s and Songhay had an important cavalry force by the time of Sonni
Ali (1464“92), whose power may have rested on it. The innovation then spread
southwards. The Yoruba state of Oyo, for example, probably adopted cavalry
during the sixteenth century. Although gradual desiccation made this feasible,
horses in regions further south were vulnerable to tropical disease and became
mainly status symbols, often buried with their owners. Yet war-horses conferred
status everywhere, for their cost “ between nine and fourteen slaves on the
Senegambian coast in the 1450s “ made their owners a relatively exclusive
class new to West Africa. Their horsemanship was often dashing and ruthless.
Their swords and thrusting-spears bred the cavalryman™s contempt for missile
weapons and their users:

Our army pursued, killing and wounding, with swords and spears and whips,
till they were tired of it. The enemy™s cavalry spurred their horses, and left
the infantry behind like a worn-out sandal abandoned and thrown away, and
there was no means of safety for those on foot save the providence of God, or
recovery from a wound after crouching in the darkness.12

Horsemen cultivated codes of jealous and sel¬sh honour, expressed in self-
glori¬cation “ ˜Superior men are ignorant of humility™, the legend of Sunjata
Colonisation in western Africa 77

explained “ in extravagant display, and especially in arms, whether on the bat-
tle¬eld or in single combat, as in a famous sixteenth-century incident when
two Mossi princes fought for the throne before their men. Commoners, too,
had their codes of honour centring on courage, endurance of pain “ inculcated
especially in initiation ceremonies “ and capacity to ful¬l the roles of adults
and parents. But the horsemen™s code widened social divisions and fostered the
violent and harshly exploitative states that were replacing the more agricul-
tural and egalitarian savanna societies of the ¬rst millennium. The equestrian
ethos probably also explained these states™ failure to adopt ¬rearms. Songhay™s
warriors threw captured Moroccan muskets into the Niger. Only Borno used
¬rearms, under Idris Aloma and perhaps his predecessors, but it entrusted
them to slaves or Ottoman mercenaries and abandoned them again in the sev-
enteenth century. Perhaps, like their counterparts in France, the chivalry of
Borno came to see ¬rearms as the grave of honour.
Cavalry warfare probably increased dependence on slavery, both because
slaves were needed to pay for imported horses and because cavalrymen could
more easily capture slaves. One guess is that the Saharan routes (including that
from Darfur in the eastern savanna) carried between four thousand and seven
thousand slaves northwards each year at this time. North African traders said
that many died from want and thirst on the two-thousand-kilometre desert
crossing and that survivors were worth between ¬ve and eight times as much
in Tripoli as in Borno. Slaves also became more numerous within savanna
societies and their duties changed. Hitherto, as in other Islamic lands, most
slaves had probably been women employed as domestic servants, concubines,
and plural wives. All rulers of Songhay save one were sons of concubines,
while fourteenth-century Mali imported not only black slaves from the south
but white slave women from the eastern Mediterranean. Male slaves worked
as servants, porters, labourers, craftsmen, miners (especially in the Saharan
salt mines), and soldiers. Some cavalrymen were slave retainers riding their
masters™ horses. Borno matched slaves in wrestling contests like gladiators.
Savanna states also used slaves as administrators, for their status was thought
to check their ambition and guarantee their loyalty, although a freed slave may
have usurped the throne of Mali in 1357. One feature of Hausaland™s sarauta
system was wider reliance on slave of¬cials. The Kano chronicle records that
Sarki Muhammad Rumfa (c. 1463“99) ˜began the custom of giving to eunuchs
the of¬ces of state™; by the 1770s slaves held nine of Kano™s forty-two top of¬ces.
But the main innovation was the growing employment of slaves as agricultural
labourers, which was rare in the Islamic world but became characteristic of
the West African savanna and was a sympton of its underpopulation, which
made free labour scarce and dif¬cult to exploit. Kanem may have used slaves
to colonise new land as early as the eleventh century. Mali required servile
agriculturalists to set aside a plot and deliver its produce to the authorities.
78 africans: the history of a continent

The ¬rst evidence of royal domains worked by slaves comes from Wolof coun-
try in the 1450s. Hausa rulers and of¬cials may have begun to employ slave
cultivators at the same period, but the institution was most developed in
Songhay, where Askiya Dawud (1549“83) is said to have had plantations, each
with twenty to a hundred slaves and overseers, at some twenty points along
the middle Niger Valley, chie¬‚y producing rice. Most slaves came from non-
Islamic, often stateless peoples to the south whom cavalrymen raided each dry
season. A song said to honour an eleventh-century ruler of Kanem celebrated his
The best you took (and sent home) as the ¬rst fruits of battle:
The children crying on their mothers you snatched away from
their mothers:
You took the slave wife from a slave, and set them in lands far
removed from one another.13

One of his sixteenth-century successors was described as setting out each year
to acquire the slaves to pay the foreign creditors awaiting his return and living at
his expense. A chronicle lauded Idris Aloma for slaughtering ˜all the fully-grown
male captives from among the pagans™, adding: ˜As for the women and children,
they merely became booty.™14 That seventeenth-century Kano appointed a new
of¬cial to supervise its slaves indicates how the expanded institution both
demanded and supported greater state power.

political development in the western forest
In the West African forest and its neighbouring grasslands, state formation was
slower than in the savanna, states were smaller, and many societies remained
stateless when Europeans ¬rst described them. Stateless societies themselves
were diverse. Segmentary lineage societies, where order rested only on the
threat of retaliation, existed mainly among herding peoples and were there-
fore rare in this region, the chief example being the Tiv of the Benue Valley,
whose history is little known. More common was the autonomous pioneering
village, headed either by a Big Man whose personal qualities attracted kins-
men and clients, as was often the case in Cameroun forest regions, or by the
senior descendant of the pioneering colonist, as in many forest areas further
west. Such coastal peoples as the Jola of modern Senegal remained stateless
by relying on hereditary ritual experts as mediators, just as others secured
an indispensable minimum of arbitration from the rulers of neighbouring
states whose authority they otherwise rejected, as among the stateless peoples
bordering Benin. Perhaps the most common religious institutions maintain-
ing cohesion within stateless communities were secret societies, notably the
Poro and Sande initiation societies for men and women whose importance
Colonisation in western Africa 79

in the forests of Guinea and Sierra Leone was attested by early Portuguese
visitors. These institutions were not mutually exclusive. The most numerous
stateless people in Africa belonged to the language group later known as Igbo,
in the southeast of modern Nigeria. Despite relatively dense population and
considerable trade, Igbo remained resolutely stateless, utilising almost all the
mechanisms mentioned. One of their ritual leaders was probably the notable
buried at Igbo-Ukwu during the ninth century. Western Igbo lived under the
shadow of Benin, while those of the north relied on age sets and systems of
titles through which men advanced with years, wealth, and in¬‚uence. Yet Igbo
observed also the law of retaliation and the concern for personal honour it
The political distance between a ritual leader or village Big Man and a terri-
torial chief was narrow and it is easy to imagine how forest peoples and their
neighbours created the microstates that probably emerged during the late ¬rst
millennium ad, initially in the country of the modern Yoruba, Edo, Nupe, and
Jukun peoples who straddle the forest-savanna edge south of Hausaland. The
earliest microstate yet identi¬ed by archaeologists was Ife, just within the forest
edge. Much uncertainty surrounds its origins, but there were small settlements
in the area by the ninth or tenth century and indications of urbanisation, houses
with potsherd pavements, and sculpture in terracotta during the eleventh or
twelfth century. The town stood on a small gold¬eld and was well situated for
trade and interaction with both the savanna and the coast, but its remains show
little evidence of such contact and suggest instead an agricultural economy that
contributed to a regional trading system by manufacturing glass beads. In this
capacity, Ife was the capital of an important kingdom from perhaps the twelfth
to the ¬fteenth century. Its fame rests on its magni¬cent sculpture in terracotta
and brass, mainly depicting people rather than the natural objects represented
at Igbo-Ukwu. The terracottas were made ¬rst. Many were probably offerings at
shrines and realistically depicted a spectrum of human conditions from kings
and courtiers to the diseased and the executed. In the fourteenth and ¬fteenth
centuries, the terracotta tradition was transferred to brass. Fewer than thirty
brasses are known. Made by a lost-wax casting process in a style of idealised
naturalism, most represent kings at the height of their powers and possess a
serene majesty unsurpassed in human art. For reasons we do not know, Ife™s
brassworkers achieved an appreciation of human worth that was to survive in
a more popular form in the life-af¬rming humanism of Yoruba woodcarving,
long after the rise of other polities had isolated Ife from its sources of brass and
power, reducing it to merely ritual primacy.
Ife™s earliest known successor was the Edo kingdom of Benin, the only other
important forest state of the period. Here the evidence that the kingdom grew
from earlier villages and microstates is especially clear from the ten thousand
kilometres of earth boundaries built by their founders in the early second
80 africans: the history of a continent

millennium. Benin City, on their western edge, may have originated as a reli-
gious centre, but it was transformed in the ¬fteenth and sixteenth centuries by
warrior kings who claimed Ife origin and introduced Yoruba innovations. The
¬rst and greatest was Ewuare, who is said to have conquered 201 towns and
villages, subjecting the surrounding microstates, resettling their populations,
and converting the city into the capital of a kingdom about 120 kilometres
across. He supposedly constructed the palace and the city defences. He con-
verted the government into a patrimonial bureaucracy by appointing freemen
as military and administrative chiefs supplanting the heads of lineage groups.
He or his successors probably established the high degree of state involvement in
foreign trade that the Portuguese found when they arrived in 1486. The regime
patronised the brassworkers who cast Benin™s famous royal heads and other
magni¬cent sculptures, combining European metal with lost-wax techniques
said to have come from Ife, although modern experts disagree on this point.
Benin™s art was a court art, practised by hereditary craftsmen within the palace
walls, divided by a vast technical gulf from popular culture. When the ¬rst
Europeans arrived, Benin was the major state of the West African forest and
deeply impressed them by its wealth and sophistication. During the seventeenth
century, however, the military and administrative chiefs came to overshadow
the king, reducing him to a secluded ritual ¬gure, ¬ghting among themselves,
and temporarily depopulating the city.
By the ¬fteenth century, several other Yoruba kingdoms coexisted with Ife,
each with its walled capital, secluded king claiming Ife origins, city chiefs head-
ing powerful coresident descent groups, and outlying villages. Trade was proba-
bly important in several of these political aggregations, especially perhaps trade
with itinerant merchants from Songhay, for the Yoruba language still retains
many Songhay loan-words for Islamic, commercial, and equestrian matters. Of
the new kingdoms, Ijebu Ode probably took shape by 1400 and was described
as a ˜very large city™ a century later, while ¬fteenth-century Owo was an artis-
tic centre to rival Ife and Benin. The arrival of the war-horse from the north
was another political stimulus. Hitherto, forest peoples had held the initia-
tive in this region. Brass sculptures from fourteenth-century Ife had passed
northwards to the savanna kingdom of Nupe, while Tsoede, traditional creator
of a new Nupe dynasty early in the sixteenth century, had an Edo-speaking
mother. Shortly thereafter, however, both Nupe and Bariba armies from the
north invaded Yorubaland, probably with cavalry, attacking especially Oyo,
the most northerly Yoruba kingdom, located in the savanna. Oyo responded
by adopting cavalry warfare and emerged during the seventeenth century as
the most powerful Yoruba state. Similar processes may have bred Allada and
Whydah, the ¬rst kingdoms among the Aja-speaking peoples (Ewe and Fon)
who occupied a savanna corridor to the coast known as the Dahomey Gap.
Both kingdoms probably existed in the ¬fteenth century.
Colonisation in western Africa 81

Further to the northwest, in the savanna regions of modern Ghana and
Burkina, war-horses enabled small groups of cavalrymen to create a series
of states among the indigenous Voltaic-speaking peoples, beginning with the
Mamprussi and Dagomba kingdoms in the late fourteenth or ¬fteenth cen-
tury and continuing with the Mossi kingdoms of Wagadugu and Yatenga. The
rulers™ origins are uncertain, but they were probably aliens, for they claimed
only political power and left control of land in indigenous hands, satisfying
themselves with tribute. In this their behaviour differed markedly from the
political changes taking place in the forest among the Akan-speaking peoples
of modern Ghana. Human settlement existed here from late in the ¬rst mil-
lennium bc, but it expanded rapidly during the second millennium ad, when
the region began to produce West Africa™s most impressive forest states. The
key to this transformation was gold. The date of its ¬rst exploitation is uncer-
tain. Begho, the trading settlement linking the Akan northwards to Jenne and
Mali, was settled from the eleventh century and Bono Manso, capital of the
¬rst Akan state, in the thirteenth, but both became substantial in the ¬fteenth
and sixteenth centuries. Gold supplied the resources to buy slaves to clear the
forest, whose conquest permanently shaped Akan culture. The pioneers were
archetypal Big Men, abirempon, whose followings of matrilineal clansmen and
incorporated slaves formed the nuclei of forest states, giving them an enduring
entrepreneurial character and an inegalitarianism that struck the Europeans
who traded with them at the coast from the late ¬fteenth century.

political development in the equatorial forest
The Big Man who pioneered forest clearance was also the mainspring of political
evolution in western equatorial regions. In this fragile and sparsely populated
environment, control of manpower was the key to change and several alterna-
tive paths of institutional development were possible. Many groups remained
wholly stateless until European conquest. This had obvious advantages for
mobile hunters and ¬shermen, but it could also be deliberately maintained
even where population was concentrated, by devising institutions to chan-
nel social pressures away from political consolidation. Like the northern Igbo,
for example, the Lega people in the eastern equatorial forest institutionalised
ambition through title-taking societies, which also provided the focus for their
exquisite miniature art. Elsewhere, however, the successful Big Man might
establish his village™s leadership over its neighbours, perhaps through demo-
graphic growth or trade or marriage alliances, until he or a successor emerged
as territorial chief. This happened especially among the peoples bordering the
northern edge of the equatorial forest who, around 1600, created the Nzakara
and Ngbandi chiefdoms of the modern Central African Republic. Religious
resources could also facilitate political consolidation. In the Cuanza river basin
82 africans: the history of a continent

of modern central Angola, for example, the Ndongo kingdom was created dur-
ing the ¬fteenth century by conquest, but its ruler drew ritual authority from
control of iron symbols known as ngola.
Within this equatorial region, however, large and enduring states came into
being only (although not necessarily) where exceptional wealth of resources “
especially the conjunction of forest, savanna, and waterside environments “
supported unusually dense and stable populations. Two examples stand out.
Close to the southern forest edge, the Kuba kingdom was a product of gradual
colonisation by the Mongo language group southwards into the territory of
another Bantu-speaking people, the Kete. The mixture created several small
chiefdoms, which a great ruler named Shyaam (perhaps an immigrant from
the west) and his successors uni¬ed by force into a single kingdom during
the seventeenth century. Drawing on rich environmental and cultural diver-
sity, the aristocracy of Shyaam™s dominant central chiefdom ruled outlying
areas and subject peoples through institutions of unusual complexity (includ-
ing a separate judicial organisation) and enjoyed a lifestyle distinguished by
the artistry that woodcarvers and weavers applied to everyday articles. Unmili-
tarised and remote from external in¬‚uences, the Kuba kingdom was to preserve
its counterbalanced institutions and re¬ned culture until colonial invasion. Yet
princes commonly fought one another at each succession and even in the late
nineteenth century the population numbered perhaps only 150,000 in an area
two-thirds the size of Belgium.
The Kongo kingdom on the western coast south of the Congo estuary had a
very different future, but its origins were not dissimilar. It too was located on the
southern forest edge, at a point of exchange between the specialised products
of the coast and the hilly interior. Control of the redistribution of goods from
different ecological zones was one basis of royal power for the Mwissikongo, an
immigrant royal clan who, during the fourteenth and ¬fteenth centuries, uni-
¬ed several small chiefdoms into a conquest state centred on Mbanza Kongo.
Captives taken during the conquest were enslaved and concentrated around
the capital, partly as agricultural labourers. This created a slavery more like that
of the Niger Valley than the lineage slavery predominant elsewhere in equa-
torial regions and the West African forest, where slaves were generally people
separated from their kin groups and incorporated into their master™s group
as junior members whose status rose with time. Royal kinsmen administered
Kongo™s sparse rural population as provincial governors. This administration,
together with a centralised army and tribute system and a state-controlled shell
currency, made Kongo an unusually authoritarian African kingdom but also
encouraged warfare at each succession, for which there was no clear rule. The
¬ssiparous forces of underpopulation remained strong in Kongo, despite the
devices invented to counter them.
Colonisation in western Africa 83

trade and industry
Western Africa™s nonagricultural economies were also shaped by underpopula-
tion, which impeded transport, inhibited exchange, and encouraged local self-
suf¬ciency. Narrow markets in turn restricted technological innovation, as did
relative isolation from the outside world, so that production was often labour-
intensive even where labour was scarce. With only a small surplus available,
the powerful often competed for it in ways that damaged its very production.
Only slowly and with great effort could Africans break out of this impasse by
enterprise and demographic growth.
Like most premodern peoples, western Africans preferred pack-animals to
wheeled vehicles. The Garamantes had abandoned chariots for ridden horses
and then for the camels that ¬rst made it possible to transport goods across
the Sahara. South of the desert only a relatively narrow belt was suitable for
animal-drawn vehicles before reaching tsetse-infested country, but even here
pack-animals were more ef¬cient, for they did not need roads, which sparse
population and tropical rains rendered uneconomic. The wheel simply did
not pay, as was discovered by every European from the seventeenth to the
nineteenth century who tried to introduce animal-drawn vehicles into tropical
Africa. Among pack-animals, camels were the most cost-effective. They carried
the bulk of freight north of the Niger, where they unloaded on to donkeys,
which were twice as expensive to employ and worked only in the dry season.
On reaching the tsetse belt further south, each donkey might transfer its load
to two human porters. But here, as throughout western Africa, the cheapest
transport was by river on craft ranging from the small dug-out canoes of the
shallow upper Niger to the twenty-metre trading canoes of the Niger Delta and
other broad rivers and coastal lagoons.
Transport imposed severe constraints on trade. Most was purely local
exchange of specialities, either informally, as at a river bank where ¬sh might
be bartered for vegetable food, or in a more formal market: ˜cotton, but not
in large quantities, cotton thread and cloth, vegetables, oil and millet, wooden
bowls, palm leaf mats, and all the other articles they use in their daily life™, as
a European merchant noted in the Kajoor area of modern Senegal in 1455.15
Western Africa had highly organised market systems, whether within towns
and villages, as in Igbo country, or on neutral ground between them, as in the
Kongo kingdom. Often they rotated among neighbouring villages on succes-
sive days, forming complex ˜market weeks™. An anthropologist noted in Nupe
in the 1930s that the market circle of eight or sixteen kilometres radius was also
the normal ¬eld of intermarriage and common local interest. Most market-
sellers appear to have been women, although the itinerant professional traders
attending markets were men. Many political authorities levied tolls on traders
84 africans: the history of a continent

and supervised markets; such of¬cials were part of Hausaland™s new sarauta
system. Other states managed redistribution systems. In the Kuba kingdom,
for example, the state™s demand for specialised produce was the chief stimulus
to surplus production.
Where trade relied on human porters, as was widely the case in equatorial
and forest regions, agricultural produce could travel only short distances “
the concentrated settlement around Mbanza Kongo probably indicated the
range “ and almost all commerce was local. It was often facilitated by a regional
currency like the small imported shells used in the western provinces of the
Kongo kingdom and the locally manufactured cloth employed in the east.
Animal transport allowed a wider bulk commerce “ legend ascribed Kano™s
prosperity to the donkey-borne grain trade “ but professional long-distance
trade rested chie¬‚y on high-value goods produced only in con¬ned areas.
The gold trade was the extreme case, peaking in Mali in the late fourteenth
century, shifting thereafter to the new Akan gold¬eld, and declining when
American bullion ¬‚ooded Europe. Copper retained its value, but the main
Saharan source at Azelik was abandoned in the ¬fteenth century. Two hundred
years later western Africa™s chief mineral product was salt, whose value was
probably greater than the whole trans-Saharan trade. It was not only essential
to life for those with vegetable diets but, in West Africa, it took much the place
that spices occupied in Europe. The Hausa language had over ¬fty words for
different kinds of salt. The largest and best supplies came from the Sahara, where
slaves working under appalling conditions exploited the great reserves left by
the lakes occupying the region in earlier periods. Tuareg and other nomads
carried the heavy blocks of rock-salt southwards, sometimes in caravans of
twenty thousand to thirty thousand camels stretching twenty-¬ve kilometres
across the desert. The exchange of Saharan salt for savanna gold or grain was
the core of the north-south commerce between climatic zones on which long-
distance trade centred. Its transport costs were high, natron being said to cost
sixty times as much in Gonja as in Borno, close to its source nearly two thousand
kilometres away, so that poorer West Africans had to use inadequate local
substitutes. But desert trade routes did penetrate beyond the grain, gold, slaves,
and cattle of the savanna to the kola nuts of the West African forest, which were
valued by Muslims as stimulants, aphrodisiacs, symbols of hospitality, and
astringents making even local water taste sweet. Kola nuts reached North Africa
by the thirteenth century. Mande-speaking traders used donkeys to transport
the perishable crop, whose price, in the nineteenth century, could multiply
forty times between Gonja and Borno.
As power and wealth in the savanna shifted eastwards from Ghana and Mali
to Songhay and then to Hausaland, so merchants from Morocco and modern
Algeria gave way to visitors from Tripoli and Cairo. Yet most desert trade was
conducted by desert peoples: Moors in the extreme west and Tuareg elsewhere.
Colonisation in western Africa 85

Although they showed much local diversity, both were nomadic peoples whose
warrior aristocracies claimed ˜white™ identity: Tuareg noblemen were of Berber
origin, while Moorish aristocrats were mainly Arabs who, from the fourteenth
century, subjected Berbers. Both aristocracies dominated clerical strata, pas-
toral vassals, dependent black cultivators, black slaves acting as menials in
nomadic camps, and stigmatised craftsmen. Both employed these followings
much like commercial ¬rms integrating camel transport, desert salt, pastoral
and agricultural produce, and resident urban agents. When sedentary states
on the Senegal and middle Niger disintegrated between the ¬fteenth and sev-
enteenth centuries, Moors and Tuareg extended their predominance into the
Such networks extending across environmental borders were the main form
of West African commercial organisation. The earliest comprised the Soninke
traders of Ghana, later renowned as itinerant leaders of donkey caravans. When
Mali supplanted Ghana, many Soninke formed the nucleus of a larger, mul-
tiethnic, Muslim, Mande-speaking trading class, known most commonly as
Dyula, who chie¬‚y traded gold, kola, and cloth. While many Dyula were trav-
elling merchants, others settled in foreign territory, established local ties, and
created a diaspora through which trade could pass across political and ethnic
borders from the forest to the desert and from the Atlantic to Hausaland. Fur-
ther east, itinerant merchants from Borno specialised in salt and horses, while
the kingdom also operated a state trading system across the Sahara that was
unique in the region™s otherwise free-enterprise economy. From perhaps the
¬fteenth century, however, Hausa traders emerged as competitors and created
a diaspora on Dyula lines that became the most extensive in the continent.
Hausa traders relied especially on a currency of cowrie shells, imported from
the Indian Ocean via North Africa, whose uniqueness, durability, limited sup-
ply, and utility in small transactions gave them great advantages over the local
currencies used in forest regions. When Ibn Battuta found cowries in general
use in Mali in 1352, they may already have been known in the region for four
hundred years. They penetrated most of the Dyula and Hausa trading zones
except Senegambia and Akan country. They lubricated commercial expansion,
but because they were imported and desert traders refused to accept them back
in payment for goods, they introduced some dependency into West Africa™s
external commerce.
By restricting trade, transport dif¬culties also limited the specialisation of
production. Most crafts supplied only local markets, even when techniques were
highly skilled; one twentieth-century anthropologist was to list over a hundred
kinds of baskets made by the Jola people of the Senegalese coast, and modern
cultivators have generally had at least one nonagricultural skill. Some crafts,
however, demanded special expertise. One was ironworking, which needed
esoteric knowledge and long experience, had af¬nities to magic, and produced
86 africans: the history of a continent

goods vital to farmer, hunter, and warrior. Almost everywhere in western Africa
ironworkers were to some degree a category apart. They might be wealthy,
credited with introducing the skills of civilisation, or closely associated with
chieftainship, especially among long-settled agricultural peoples. The Marghi
to the south of Lake Chad, for example, often buried their chiefs seated on
iron stools and surrounded by charcoal. In regions with equestrian nobilities,
however, ironworkers increasingly became a stigmatised group with whom
others would not eat or marry, often alongside other male specialists such as
leatherworkers, woodworkers, and bards. Their womenfolk were frequently
potters and midwives. Stigmatised groups were often about 5 percent of the
population. They were exempt from codes of honour and normally could not
bear arms or be enslaved. Most also cultivated but were paid for their craft
products in grain.
The origins of stigmatisation are uncertain. Tradition claimed that it had
existed in ancient Ghana. The earliest documentary evidence of the practice
comes from fourteenth-century Mali, where ironworkers may have suffered for
being a focus of resistance to Islamic state-formation. Mande-speaking horse-
men may then have spread their prejudice. It did not extend to New Jenne, a free
city hostile to Islamic empires, nor beyond the western savanna and its envi-
rons. Elsewhere crafts were nevertheless often hereditary, partly because skills
were transmitted, partly because any expertise was widely regarded as lineage
property “ a principle that extended equally to hunting or the playing of par-
ticular musical instruments. Yoruba crafts were generally, but not necessarily,
hereditary in certain lineages, one of which provided a headman for all prac-
titioners of that craft within a town, making formal craft guilds unnecessary.
The nearest approximation to guilds in West Africa was probably in Nupe and
among Benin™s skilled craftsmen, who formed tight-knit groups under titled
chiefs, but these organisations depended heavily upon rulers, as in the Islamic
world. Nupe was famous for glasswork, one of many specialities that evolved as
trade expanded markets. Within Yorubaland, Ijebu Ode became renowned for
work in precious metals, Ilorin for pottery, and Oyo for leatherwork “ an associ-
ation between that craft and cavalry warfare paralleled further north in Katsina.
Specialisation was most advanced in the textile industry. In equatorial Africa,
many peoples wove raf¬a or other ¬bres, often producing cloth of great beauty.
Both cotton and its weaving appear to have been introduced by Muslim traders,
for excavations at Jenne have revealed spindle-whorls only in deposits later than
about ad 900. Although not always con¬ned to Muslims, weaving was often
monopolised by them, as in Voltaic-speaking areas, and the prestige of Islam
may explain why only Fulbe and Wolof stigmatised weavers. The women who
spun cotton used only a spindle and not a wheel, a major constraint on output
and an important technological consequence of West Africa™s semi-isolation
from the Old World. Women also wove on an inef¬cient broad loom, while
Colonisation in western Africa 87

men used a less clumsy but still relatively unproductive narrow loom, usually
the only machines West Africans knew. One estimate is that they produced
respectively only one-sixth and ¬ve-ninths of the output of an English broad
handloom. Narrow looms often produced cloth of high quality, especially in
Wolof and Mande country and in Nupe and Yorubaland where some weavers
were full-time professionals. Such regional specialities were traded widely in
West Africa, where the desert gave them natural protection, but they failed
to win important export markets in the Atlantic economy created after 1450,
probably because they were too expensive to compete, owing to reliance on
labour-intensive techniques in a labour-scarce economy.
Specialised craft production enriched West Africa™s ancient urban tradition.
Trans-Saharan trade and declining rainfall largely destroyed the urban clus-
ters without visible political authorities that had characterised the pre-Islamic
Niger Valley. Old Jenne and nearly three-quarters of its surrounding villages
had been abandoned by the ¬fteenth century, while Timbuktu alone survived
from its urban cluster as an entrepˆ for desert trade and a major centre of
Islamic education. Generally, however, West African towns possessed a mul-
tiplicity of functions. Hausaland™s major walled towns were capitals, markets,
craft centres, and refuges for surrounding cultivators. Ife was not only a religious
centre and political capital but had a renowned glass industry and doubtless
housed the agriculturalists who were the majority of Yoruba townsmen “ a
cultural preference with no apparent functional explanation, especially when
contrasted with the more dispersed settlement found in even the most densely
populated Igbo areas. Many towns merged imperceptibly into the countryside.
Mbanza Kongo was really a large cluster of villages, while an eighteenth-century
description of the capital of Loango, on the coast of modern Gabon, claimed
that ˜a missionary who was a bit nearsighted could have traversed the whole
town without seeing a single house™. It contained some ¬fteen thousand people
but was said to be as large as Amsterdam, while Benin City was six kilometres
across and Kano™s early walls enclosed seven square kilometres. Because most
large savanna towns were capitals, their citizens had little political autonomy,
although New Jenne and Timbuktu remained partial exceptions with turbu-
lent histories. Major Yoruba towns were also capitals but appear to have had
stronger political traditions, perhaps because their citizens formed military
units and there was a sharp divide between the palace and the popular quar-
ters where commoners lived in large patrilineages occupying rooms ranged
around internal courtyards, sometimes with underground water-tanks “ a res-
idential pattern that nineteenth-century Yoruba aptly described as squares.
Yoruba temples were unimposing, but savanna towns centred on palaces and
mosques built in a picturesque mud-and-timber style attributed to Abu Ishaq
al-Saheli, a North African architect employed by an early fourteenth-century
king of Mali, although in fact the technique mainly evolved locally. Around
88 africans: the history of a continent

them were ethnic and occupational quarters in permanent materials, the rich
often distinguishing themselves by multistorey houses. Beyond these were the
straw huts of the poor who clustered beneath the high walls protecting them
from the world outside.

religion and culture
To reconstruct the cultural and social patterns of largely preliterate people of
the mid-second millennium risks overreliance on later ethnographic evidence,
but some insights are possible. One is that western African thought and culture
were deeply shaped by the experience of colonising land. Distinctions between
the cultivated and the wild, civilisation and savagery, provided an intellectual
framework, as in medieval Europe and Hindu India. The distinction was not
simply between good and evil, for Africans had diverse attitudes towards nature.
Generally, the more a culture sought to dominate the natural world, the more
hostile nature seemed. For Akan peoples “ to judge from the Asante who
became dominant among them during the eighteenth century “ the experience
of clearing dense forest made civilisation appear a brittle artefact, ever at risk of
being overrun. The low, symbolic barrier separating an Asante village from the
surrounding forest epitomised their concern to maintain boundaries between
the two spheres. Benin had similar attitudes. By contrast, later anthropologists
found that Pygmies who relied on wild produce regarded the forest as innately
good, while many peoples had ambivalent and symbiotic relationships with it.
Women probably often saw the dichotomy differently from men.
For cultivators, to clear and plant the bush was to create both civilisation “
often associated with implanting seed in a woman to increase the community “
and property, which gave the ¬rst settlers™ descendants their claim to the land.
Initiation rituals, generally performed in the bush, civilised the young, remov-
ing them from the animal to the human world, just as Yoruba used the same
terms for civilisation as for the carving of lines of identity into a person™s face.
Vital human activities were con¬ned to the cultivated area, whether it be sexual
intercourse, which was often forbidden in the bush, or burial, which must take
place in the house or the ¬elds “ only victims of smallpox, leprosy, drowning,
suicide, or execution being cast into ˜bad bush™. The forest, by contrast, was
associated with sorcery and magic, whose practitioners could transform them-
selves into wild animals. For a thing of the bush to enter the civilised world was
an evil omen. Mossi described nightmares as ˜bush creeping up™.
Art frequently expressed the dichotomy. Among the Senufo of northern Cote
d™Ivoire, for example, the two major woodcarving traditions of the Poro ini-
tiation society counterposed the ancestral couple, personifying civilisation, to
magni¬cent helmet masks representing hyenas in which ˜the menace of open
jaws and jagged teeth and the explosive force of bundles of feathers, quills,
Colonisation in western Africa 89

bristling layers of skins, and other materials are icons appropriated from the
bush world as symbols of power™.16 The dangerous was also the powerful, a
force that courage and skill might appropriate for good. The experts were
hunters, who in humanised regions like Yoruba, Mande, and Hausa country
were specialised groups with associations, folklore, spheres of power, military
functions, and reputations for violence and magic. Herbalists needed a similar
mastery of nature; modern analysis has shown that many Kongo medicines
combine a forest and a cultivated plant. For unskilled and unprotected per-
sons to leave the safety of civilisation, by contrast, was to incur risk, which
especially affected the poor, who often survived by exploiting the bush, and
those obliged to seek wild produce during famine. However dangerous, the
forest was a vital resource. Africa™s most elaborate (but undated) myth, that
of the Bagre initiation society among the LoDagaa people of modern north-
western Ghana, tells of dwarf-like beings who herded wild animals, cultivated
wild plants, and taught the ¬rst human beings hunting, cultivation, smelting,
cooking, and death. Forest motifs coloured religious systems, especially the
widespread symbolism of the axe, sometimes a ground stone axehead surviv-
ing from the beginnings of colonisation. Kingship, too, shared the menace of
the wild. Benin identi¬ed its kings with the leopard, the lord of the forest,
whose skin symbolised chieftainship throughout equatorial Africa. Wild ani-
mals were major objects for the play of imagination by which human beings
conceptualised themselves and the world.
Religious ideas and practices are especially dif¬cult to reconstruct. They
too were shaped by the needs of colonising peoples, which bred both a cen-
tral concern with the fertility of women and crops and a scarcity of elaborate
religious institutions, by contrast with the more settled societies of Asia and
pre-Columban America. Lacking literacy, non-Islamic Africans lacked sacred
texts to de¬ne orthodoxy and heresy. The test of religious practices and prac-
titioners was whether they worked, especially whether they relieved human
misfortune and secured fertility, prosperity, health, and social harmony in this
world. In 1563 drought led the king of Ndongo in modern Angola to execute
eleven rainmakers. This demand for constant validation by success made many
Africans sceptical of religious claims. The myth of the Bagre association told
that although its practices were not God™s way, they at least brought earthly
In God™s country
what is there
to surpass this?17

Such pragmatism made indigenous religions eclectic: ideas and practices from
any source were acceptable if they worked, with little regard for mutual con-
sistency but much tolerance, not because Africans were simple-minded or
90 africans: the history of a continent

unre¬‚ective “ their myths and symbolism refute that “ but because they had
no reason to be systematic, unless challenged by a systematic imported creed.
Religions were therefore mutable, probably the most swiftly changing aspect
of African culture. This was why they struck Muslim and Christian observers
as so diverse, fragmented, and incoherent, especially in the absence of written
The religions of the equatorial region are the most accessible in western Africa
because Bantu-speaking peoples preserved a measure of religious homogeneity
evident in their languages. These show that they shared ideas of a creator
spirit, ancestral and nature spirits, charms, ritual experts, and witches. From
this common basis, each society had evolved distinctive ideas and practices.
The late ¬fteenth-century Kongo people, for example, appear to have had a
vague notion of a ˜highest or ultimate power™, nzambi mpungu, but the main
spiritual powers active among them were ancestral and nature spirits. Each
matrilineage communicated with its ancestors through public rituals at their
graves. Agricultural fertility, a communal concern, was the sphere of nature
spirits, who were served by ˜chiefs of the land™ and could also communicate
through men and women by possessing them. Both could also be localised in
objects known as nkisi. Yet this is a vast simpli¬cation. As an indication of the
complexity and mutability of religions, the eighteenth-century Kuba kingdom
probably venerated three distinct creator spirits: Mboom (the spirit recognised
by Mongo immigrants from the north), Ngaan (recognised by the indigenous
Kete people), and Ncyeem apoong (Nzambi mpungu, perhaps brought from
the Kongo region by the founding King Shyaam).
A similar range of spiritual powers generally underlay the even more complex
religious ideas and practices of West Africa. The Bagre myth stated the grounds
for religious action:

Earth shrine,
say we should perform.
Failure of childbearing,
the scorpion™s sting,
failure of farms,
caused the elder one
to sleep badly,
so he seized ten cowries
and hurried off
to see a diviner.18
Colonisation in western Africa 91

Misfortune might be collective, for Africans widely believed that natural dis-
aster arose from social and moral disorder. Al-Bakri mentioned cattle sacri¬ce
in eleventh-century Ghana to bring rain, one of few purposes for which West
Africans directly approached the Creator. Or misfortune might be individual:
Benin™s most widely venerated divinity was Olokun, provider of children and
wealth, benefactor especially of women. A diviner determined the spiritual
power to be propitiated or invoked. He might direct the supplicant to an earth
shrine, especially in the Voltaic-speaking regions where the Bagre myth orig-
inated. Served by a descendant of a pioneer colonist, an earth shrine claimed
power over all people within its territory. By contrast, ancestors “ the subjects
of sub-Saharan Africa™s oldest surviving ¬gure-carvings by the Tellem of Mali “
concerned themselves only with their descendants, blessing or punishing them
for observing or neglecting custom. Ancestors were widely believed to sur-
vive in a shadowy world that reproduced earthly conditions, like the Field of
Reeds, so that kings of Ghana were buried with their ornaments, food, and
retainers in a manner akin to that unearthed at Igbo-Ukwu. Ancestral cults
¬‚ourished where patrilineal descent dominated social structure. Where village
organisation was strong, religion might focus on a secret association like Poro,
which drew on the power of ancestral or nature spirits to cure disease, restore
fertility, and counter witchcraft. Larger societies might create more complex
institutions and pantheons. The Yoruba and related peoples, for example, ven-
erated innumerable divinities (orisa). Some were known only in one town or
region; others, as in ancient Egypt, had gained recognition throughout the
culture area. Individual Yoruba generally served one orisa, either by inheri-
tance or by the divinity™s choice as revealed by a diviner. An orisa™s devotees
might form a local cult group with a temple, images, priests, collective rituals,
and a role in the town™s colourful cycle of festivals. Yoruba also practised a
divination system, ifa, in which a skilled professional identi¬ed which among
hundreds of memorised verses were relevant to an enquirer™s situation. The
cult existed by the seventeenth century “ a divination board of that period
survives “ and looked to Ife as its headquarters, thereby af¬rming the city™s
ritual primacy. Oyo, its rival, countered by extending throughout its domin-
ions a cult of Sango, supposedly an early King of Oyo identi¬ed with an even
earlier thunder god. Sango communicated by possessing his devotees, a means
of access to divinity that probably grew increasingly common in West Africa.
When the new urban-centred Katsina kingdom was created in Hausaland and
its rulers adopted Islam, for example, spirit possession (bori) became a popular
cult of af¬‚iction for those marginalised within the new order. Its pantheon
expanded beyond the old nature spirits to embrace personi¬cations of the new
forces that caused misfortune. Much the same probably happened to Songhay™s
indigenous religion.
92 africans: the history of a continent

As possession cults suggest, western Africans with limited power to control
their environment or experience generally looked outside themselves for expla-
nations of misfortune and means of relief. Many blamed witches, especially
where misfortune attacked the fertility of women, the survival of children, and
the multiplication of the community. Some peoples distinguished witchcraft,
as an inborn and perhaps hereditary psychic power for evil, from sorcery, the
manipulation of material substances for evil purposes, but the distinction was
not universal. Sixteenth-century Kongo, for example, used the same word to
describe both witches and sorcerers, dividing them into at least three categories:
those so born, those possessed by evil spirits, and those harnessing spiritual
power for malign purposes. No full account of witch beliefs appears to survive
from that period, but those recorded in modern times commonly associated
witches with the bush and the inversion of normality:

those who kill without war,
those who are blessed at midnight,
those who gnaw ¬‚esh from the belly,
those who eat liver and heart.19

These beliefs ¬‚ourished in concentrated settlements where interpersonal ten-
sions were high and institutions to resolve them were weak. Those suspected
of witchcraft were commonly relatives or neighbours who might gain by the
misfortune, especially women whose age, childlessness, deformity, misery, or
ill-nature suggested jealousy. Suspects might be tested by a poison ordeal, for
which the original Bantu language contained three separate words, or by other
techniques. Several peoples had associations of young men, frequently masked,
to identify and execute witches, often with great cruelty. The Egungun society
among the Yoruba included ˜the hanger of witches™ among its masks. An alter-
native penalty was expulsion; nineteenth-century Christian missionaries found
small communities of such women on the fringes of Igbo towns. Persecution
of supposed witches was a cruel feature of pre-colonial Africa, as of many other
The nature of indigenous religions makes it easier to understand responses
to Islam. By the fourteenth century, the Moors and Tuareg of the desert had
accepted the new faith and the Kounta clerical group among the Moors was
af¬liating to the Qadiriyya order, the most important brotherhood engaged in
deepening Islamic commitment. In Senegambia, an early centre of sub-Saharan
Islam, ¬fteenth-century Portuguese merchants described Islamic courts with
non-Islamic subjects. Borno, another early centre, was more deeply committed;
between 1574 and 1728 at least twelve of its rulers passed through Cairo on
Pilgrimage and there are indications of more extensive Islamisation in the
countryside, although indigenous religious practices survived even at court. In
the Niger Valley, Mansa Musa proclaimed Mali™s Islamic commitment by an
Colonisation in western Africa 93

extravagant Pilgrimage in 1324“5, one of at least sixteen pilgrim caravans from
West Africa to reach Egypt during the next two centuries. The Sonni dynasty
who created the Songhay empire, by contrast, retained strong ties to the indige-
nous religion; their overthrow in 1493 by a coalition of military of¬cers and
Muslim clerics headed by Askiya Muhammad was West Africa™s ¬rst Islamic
coup d™´ tat, although the new regime continued to patronise the indigenous
religion of its rural subjects. In Hausaland the later ¬fteenth-century rulers
of Kano, Katsina, and Zaria were all reputed to be keen Muslims, but during
the next century there was much tension with subjects who held to indige-
nous beliefs and some evidence of retreat from Islam at court, so that even
in seventeenth-century Katsina, a reputed centre of learning, enthronement
rites remained pagan and the palace was a stronghold of the old spirit cult.
Islamic expansion to the south was even more tentative. A Nupe ruler is said
to have been deposed during the eighteenth century when he tried to make
Islam the of¬cial faith. Muslims from Songhay probably reached Yorubaland
by the ¬fteenth century and had a profound intellectual impact, demonstrated
by the existence of many abstract Songhay words in the Yoruba language.20 Yet
clerics in late eighteenth-century Hausaland still regarded Yoruba as legally
enslaveable because they were pagan. Neighbouring peoples were openly
resistant, notably the Jukun kingdom in the Benue Valley and the Bariba of
Borgu, who, in the twentieth century, still beat drums at the ¬rst moon of
Ramadhan to show de¬ance. More common was the situation in Mossi states,
where Mande-speaking Muslim traders were tolerated as a quietist community,
winning some sympathy among chiefs but apparently no converts among rural
people loyal to ancestral cults and earth shrines.
Muslim teachers, like their Christian counterparts in Ethiopia, regarded
African religions as works of the Devil. In Kano, for example, they too cut
down sacred trees, whence ˜came forth strange devils that no one can describe™,
and built a mosque on the site.21 As in Ethiopia, however, African responses
were dominated by the eclectic and pragmatic traditions of indigenous reli-
gions. Eager to utilise any power that worked, West Africans often saw Muslims
initially as powerful magicians. Gonja™s chronicle claims that its king was con-
verted by observing Islam™s superior war magic. The Poro society adopted an
Islamic cleric as an of¬cer to provide magic against its enemies. Islamic amulets
were especially valued, while a document from Kanem exempts the descendants
of a twelfth- or thirteenth-century imam from civic duties in return for a special
obligation to pray. Islamic medicine was also attractive, as was the prestige to
be gained by attaching one™s genealogy to Islam™s sacred history, a characteris-
tic response of peoples partially integrated into the Old World. Kings of Mali
took the Prophet™s black muezzin, Bilal, as their ancestor, rulers from Kanem
to Yorubaland claimed Middle Eastern origin, and even the resolutely anti-
Islamic Bariba chiefs acquired an ancestor driven from Mecca for rejecting
94 africans: the history of a continent

Muhammad™s teaching. Islam offered West Africans clearer notions of the
Creator and how to approach him, powerful visions of heaven and hell, a sense
of purpose and destiny, and a cosmology claiming the authority of divine rev-
elation. In Songhay these were adopted even by those who otherwise remained
faithful to indigenous cults. Such eclecticism was the normal West African
response, adding Allah to the pantheon or synthesising Him with the creator
spirit, borrowing notions of angels or the devil, or adopting a Prophet-like
¬gure who had revealed divine knowledge to men. The result was a spectrum
of beliefs that rulers patronised in the interests of harmony. Just as Ibn Battuta
had watched the King of Mali celebrate the end of Ramadhan in the morning
and listen to bards in bird-head masks sing his dynastic praises in the after-
noon, so a ruler of Jenne is said to have built a mosque divided into two sections
for Muslims and pagans. Cattle were annually sacri¬ced to the Koran at the
Katsina court until late in the eighteenth century. Its ruler was praised in an
Arabic poem of 1659 as both commander of the faithful and elephant slayer.22
Muslims disagreed about how to respond to eclecticism. In sixteenth-century
Timbuktu, for example, rigorous clerics followed the Algerian teacher al-
Maghili and denounced such unbelief, while accommodators relied upon the
Egyptian scholar al-Suyuti and saw unbelief as ignorance to be eradicated grad-
ually by pious example. In practice, Muslims bravely opposed certain indige-
nous rites, especially retainer sacri¬ce at the funerals of great men, which had
virtually disappeared from areas under Islamic in¬‚uence by the nineteenth
century. Muslims apparently abolished the Songhay custom of female circum-
cision in Timbuktu and they generally treated all non-Islamic spirits and magic
as evil, even when non-Muslims thought them benevolent. Yet in other respects,
many Muslims were eclectic, especially on the southern frontier of Islam, where
many tales described Dyula pilgrims whose descendants sacri¬ced to copies of
the Koran purchased at Mecca.
Veneration of the book emphasises that literacy was a crucial difference
between Islam and indigenous religions. Among Arabic loan-words in the
Songhay language, that for ink coexisted with others for religion, paradise,
amulet, and pro¬t. Underpopulated societies, dif¬cult to tax or control, had
not supported the bureaucracies that had invented literacy in Sumer and Egypt.
African oral cultures were consequently exceptionally rich. Oratory, debate,
poetry, and conversation were sophisticated arts at the Kuba court. Imagination
had free play when uncon¬ned by written texts. Memory was cultivated by
diviners who expounded ifa verses, remembrancers who carried the messages
or treasured the traditions of kingdoms, and traders whose powers of recall
struck a European merchant as ˜beyond what is easy to imagine™. A few West
African groups invented systems of written signs, notably the Nsibidi script
used by the Ekpe secret society in the southeast of modern Nigeria, but their
antiquity is uncertain and they lacked the ¬‚exibility of true literacy. Stereotyped
Colonisation in western Africa 95

messages could be transmitted by drum calls or clusters of natural objects.
Hausa weavers ˜signed™ their cloth by working into it a tiny pattern. Mothers in
Jenne identi¬ed their children by distinctive cicatrisation. Such devices suggest
that many West Africans could pro¬t from Islamic literacy. Yet, as in ancient
Egypt, the impact of writing was shaped by the interests of its transmitters and
the social context of its recipients. Mid-sixteenth-century Timbuktu was said
to have between 150 and 180 Koran schools, but most concentrated on teaching
the memorisation of Arabic texts. Relatively few West Africans became literate;
none ceased to be oral. Mali had secretaries to conduct foreign correspondence,
but its internal administration used only word of mouth. Much writing served
such purposes as making amulets. Yet for a tiny minority in savanna regions,
literacy did become more than an adjunct to oral communication, especially
perhaps when Arabic script was used to write African languages, possibly ¬rst
for Hausa during the ¬fteenth or sixteenth century. In the long term, literate
orthodoxy would gain over oral eclecticism. But the victory was not yet.
Islam also adapted itself to indigenous social relationships. The harsh strug-
gle with nature gave western Africans a primary concern with prosperity and
harmony in this world, an ideal embodied in the image of the Big Man wealthy
in grain stores, cattle, gold, and above all people to provide labour, power,
and security. Ifa verses marked out an ideal career of wealth, wives, children,
titles, and long life. Pursuit of prosperity bred competition and enterprise,
but its attainment could be signalled only by lavish display. Noblemen on the
seventeenth-century Gold Coast observed each year several festival days at
which they dispensed food, drink, and goods to all who came; the result, so
a European observed, was to prevent accumulation. Arziki, fortune in both
senses in the Hausa language, was easily lost where nature was so hostile and
death so close. Social mobility therefore went together with frank acceptance of
inequality, even “ perhaps especially “ among stateless and ostensibly egalitarian
peoples for whom all status had to be achieved through competition. The poor,
in this world of ample land, were those who lacked both the labour to work it “
because they were old, handicapped, sick, young, or burdened with children “
and the labour of others (especially kinsmen) to support them. Charity beyond
the family was mostly informal largess, although political authorities on the
Gold Coast provided sheltered employment for the blind and handicapped,
while Benin™s rulers displayed characteristic panache:

The king being very charitable, as well as his subjects, has peculiar of¬cers
about him, whose chief employment is, on certain days, to carry a great
quantity of provisions, ready dressed, which the king sends into the town for
the use of the poor. Those men make a sort of procession, marching two and
two with those provisions in great order, preceded by the head of¬cer, with a
long white staff in his hand.23
96 africans: the history of a continent

Muslims accepted these patterns, neglecting the institutionalised endowment
(waqf) that was the normal basis of Islamic charity elsewhere and concentrat-
ing instead on personal almsgiving, especially perhaps to beggars who were
very numerous in savanna states when evidence ¬rst becomes available in the
nineteenth century. Islam had ascetic ideals distinct from those of indigenous
religions, but the paragon in the chronicles of Timbuktu was the cleric who was
both ascetic and generous, giving away to the poor what he received as alms. It
was a satisfying blend of Islamic and indigenous values.

the family
Concern to build up populations and colonise land gave western Africa dis-
tinctive and enduring family structures. Where land was almost a free good,
there was no need to hold it within the family through monogamous or endog-
amous marriage. Rather, the need was for wives and children to appropriate
and cultivate it, in addition to providing the social support on which a man™s
standing rested. The result was intense competition for women, great inequal-
ity in access to them, obsession with male potency and female fertility, little
anxiety about chastity, and extreme tension between male generations. The
struggle for wives bred several forms of marriage, ranging from abduction
by youthful daring to formal payment of bridewealth by which the husband™s
family compensated the bride™s for the loss of her fertility and labour. Although
bridewealth prevented powerful men from completely monopolising women,
it probably enabled them to practise high levels of polygyny; one estimate is
that two-thirds of rural wives in nineteenth-century Yorubaland may have been
in polygynous marriages. This assisted their husbands to respect the long post-
partum taboo on sexual intercourse, but it also required most women to marry
very young, probably at or before puberty, and demanded that all but the rich-
est men should marry very late, probably in their thirties, although there were
exceptions to this in monogamous societies. In the Kuba kingdom, where for
unknown reasons polygyny ceased (except for chiefs) at some time between
the seventeenth and nineteenth centuries, bridewealth fell sharply and men
of less than twenty began to marry. Elsewhere, however, up to half the adult
men may have been unmarried at any time, while the proportion without a
sexually available wife would have been higher. The result, despite brutal pun-
ishments, must have been extensive extramarital sex, which encouraged both
blatant machismo and the relaxed and hedonistic attitudes that horri¬ed Ibn
Battuta. In western Africa, in¬delity probably endangered a woman less than
Women™s share of agricultural labour varied widely and inexplicably, from
preponderant among the Tio of modern Zaire to only a minority among the
Yoruba. Commonly, however, heavy clearing work fell on men, tedious planting
Colonisation in western Africa 97

and weeding on women, and peak activities like grain-harvesting on both. Hus-
band and wife seldom held property in common, so that women retained more
economic autonomy than was normal in agricultural societies, especially in
the West African forest and southern savanna where women predominated in
small-scale trade. Yet women, of course, were not a homogeneous category. In
Borno, for example, the high status of senior royal women, who shared in
political activity and controlled important territories, contrasted with the sub-
missiveness expected of peasant women, usually so much younger than their
husbands, and the drudgery no doubt exacted from the women who were most
of West Africa™s slaves. The Mossi kingdom of Yatenga had two palace gates,
one for freemen and the other for captives and women. Yet women had some
protection in the vigorous competition for wives and the relative ease of divorce
in many societies.
In most regions, the ideal of social organisation was the large complex house-
hold headed by a Big Man surrounded by his wives, married and unmarried
sons, younger brothers, poor relations, dependents, and swarming children.
Households of this kind, with perhaps ten to forty members, were the key
colonising groups in equatorial Africa around which villages and eventually
chiefdoms formed, as also among Malinke and Hausa in West Africa. An urban
notable™s household on the seventeenth-century Gold Coast might contain
over 150 people, while a Kongo nobleman™s might number several hundreds.
Where extensive agriculture was the norm, large households were probably
more ef¬cient as producers and guarantors of economic security, in addition
to physical protection in a violent world. When households were ¬rst enumer-
ated in these areas early in the twentieth century, their average size was quite
large: 10 to 15 people in 5 Hausa villages surveyed in 1909, 10.4 people in 61
Tallensi households in northern Ghana in 1934. Yet while some Hausa house-
holds contained up to ¬fty members, most must have been quite small, which
would correlate with evidence from other continents that the large complex
household was more the ideal than the norm. Data for the Kongo kingdom in
the seventeenth and early eighteenth centuries suggest an average cultivator™s
household of ¬ve or six people, much the same as in ancient Egypt.25 Such
households, of course, were linked to others by kinship, which is probably
what ˜extended family™ generally meant. The contrast between the Big Man™s
household and the cultivator™s was probably crucial to social organisation in
this family-dominated world.
Competition for wives in polygynous societies made con¬‚ict between male
generations one of the most dynamic and enduring forces in African history,
whereas the ample availability of land minimised other forms of social con¬‚ict “
the exact reverse of the situation in India, where both sexes married young, land
was often scarce, and con¬‚ict centred on social strati¬cation rather than age.
At its simplest, in sub-Saharan Africa father and son might contest the same
98 africans: the history of a continent

property with which to pay bridewealth for the same woman, but probably con-
¬‚ict was more commonly diffused through the society as a whole. It was implicit
in the stress that myth and folklore laid on respect for seniority “ in Dogon
thought, disobedience to an ancestor brought death into the world “ and in the
widespread use of initiation into adulthood as a painful and psychologically
traumatic means of imposing the authority of age. Kuba myth told that the
¬rst man instituted initiation to punish his sons for mocking his nakedness.
Linguistic evidence suggests that both circumcision of boys and their organi-
sation into age groups were already key cultural elements in equatorial regions
during the ¬rst millennium ad. Yet western African societies were not immo-
bile gerontocracies. People struggling against nature could not afford much
deference to old age, which in any case few can have achieved. ˜They do not
like to become old,™ a European wrote of the seventeenth-century Gold Coast,
˜for they are not esteemed or honoured here when they are: they are cast aside
and nowhere respected.™26 Tradition said that Ewuare of Benin favoured brass-
casters because their work made him look younger than did that of woodcarvers.
Ife™s sculptors mostly represented kings in early manhood, just before decay.
When African art or masquerade represented elderly people, it was often as
caricature. Elders exercising authority, therefore, were not necessarily toothless
greybeards, nor were the young necessarily immature, for the term generally
embraced all unmarried men. In Benin, for example, the junior male age grade
contained those between roughly 12 and 30. The potential for con¬‚ict was great,
not only over women but over young men™s labour, a crucial resource for their
parents but one that the young could convert into independence where land
was easily available. The early seventeenth-century Gold Coast already had
˜common boys™ waiting on the beach to perform unwanted services at a price.
Moreover, if elders gained power from control of wives and property, the young
gained it from violence in a world where every man had to be ready to defend
himself and his family. That was why abduction and elopement coexisted with
marriage by bridewealth. Youthful violence was licensed and encouraged, as
in the ritualised battles that Hausa youth groups fought with other villages or
urban wards, in a pattern common in western Africa. Young men might culti-
vate a distinct subculture stressing beauty, dress, ornament, virility, insolence,
and aggression, to judge both from accounts by the ¬rst European observers
and from objects as old as Nok ¬gurines showing intense concern for per-
sonal appearance and adornment. In extreme situations, these tensions might
make the young available for political upheaval, as in the warrior society of the
Imbangala in sixteenth-century Angola, based on initiation and the inversion of
normal social values, or in the proliferation at the same period of refugee Arawa
chiefdoms in modern Niger after a young ruler mobilised his contemporaries
against their elders. Alternatively, generational tension might be dramatised in
festival, as in the Do masquerade that Ibn Battuta appears to have witnessed at
Colonisation in western Africa 99

the court of Mali. Yet we know little of the history of dance, the most important
African cultural form and the one most likely to have embodied the values of
The only leisure activity with a history stretching back beyond European
contact is the board game often known in West Africa as mankala. Played in
ancient Egypt, where a stone board of c. 1500 bc has been found, the game
appears to have spread to other speakers of Afroasiatic languages and thence
throughout the continent, except its southern tip, changing its form in complex
ways. Everywhere it was seen as a test of intelligence. Legend said that Sunjata
played it for his life against his rival for power in Mali. Shyaam™s ritual statuette
shows him with the mankala board that he allegedly introduced when found-
ing the Kuba kingdom. Being African, the game was played quickly, publicly,
socially, noisily. Islam frowned on it and replaced it by the more sedate dara,
a form of draughts, while the Ethiopian nobility either played an especially
complicated variant or preferred chess. Chess was the game of a strati¬ed soci-
ety, with unequal pieces and the objective of destroying the opposing forces. In
mankala all pieces were of equal value and the aim was to capture the opposing
pieces and add them to one™s own. It was the game of a society dedicated to
building up its numbers.27

Colonising society in eastern and southern Africa

this chapter considers the regions east and south of the
equatorial forest during the thousand years between the end of the early iron
age and the outside world™s ¬rst extensive penetration in the eighteenth century.
The central themes were the same as in western Africa: colonisation of land,
control over nature, expansion of populations, and consolidation of societies.
But the circumstances were different. Because neither Muslims nor Europeans
commonly penetrated beyond the coast, few written sources for this region exist
to compare with Islamic and early European accounts of West Africa, while oral
traditions seldom extend back reliably beyond three centuries. Much therefore
remains uncertain, although archaeological research indicates the wealth of
knowledge awaiting recovery. Moreover, whereas West Africa™s lateral climatic
belts tended to separate pastoralists from cultivators, the two were interspersed
in eastern and southern Africa, where faulting and volcanic action had left dra-
matic local variations of height, rainfall, and environment. The grasslands in
which mankind had evolved now supported cattle as the chief form of human
wealth. Settlements were dispersed and often mobile, with few urban centres to
rival Jenne or Ife. Interaction between pastoralists and cultivators created many
of the region™s ¬rst states, although others grew up in the few areas with exten-
sive trade. Pastoral values shaped social organisation, culture, and ideology.
Not only men but their herds were engaged in a long and painful colonisation
of the land.

southern africa
By about ad 400, early iron age cultivators speaking Bantu languages occupied
much of eastern and southern Africa, although sparsely and unevenly. Archae-
ological evidence shows that they usually preferred well-watered areas “ forest
margins, valleys, riversides, lakeshores, and coastal plains “ suggesting that they
relied chie¬‚y on yams, sorghum, ¬shing, hunting, and small livestock rather
than millet or many cattle. In East Africa their remains have been found espe-
cially around Lake Victoria (where forest clearance was already well advanced),
on the foothills of high mountains like Mount Kenya, and close to the coast, but
not in the grasslands of western and northern Uganda, the Rift Valley, or western
Colonisation in eastern and southern Africa 101

Tanzania, which were either uninhabited or occupied by earlier populations.
Further south, in modern Central Africa, Bantu-speakers were quite widely
dispersed alongside Khoisan forager-hunters and had evolved regional pottery
styles, but here too their preferred settlements, as in the Zambezi Valley above
the Victoria Falls, were ˜clusters of small thatched wattle-and-daub huts set in
a clearing hewn out of the wooded margins of a dambo [moist depression]™.1
Watercourses and shorelines also attracted the ¬rst Bantu-speaking settlers in
modern South Africa, who generally occupied the wooded lowveld close to the
coast, eschewing the treeless grasslands of the inland highveld.
Southern Africa provides the best evidence of subsequent evolution through
the growth of pastoralism and its role, together with trade, in fostering large-
scale polities. By ad 500 Bantu-speaking groups from the coastal lowveld were
settling in valleys running up into the highveld, perhaps using the uplands
for grazing. By that date, people in the Soutpansberg of northern Transvaal
were building homesteads of circular huts around central cattle pens in which
they dug storage pits and graves, a settlement layout that became as distinctive
in much of southern Africa as the straight streets and rectangular huts that
characterised western equatorial villages. Further west, on the eastern pastoral
fringe of the Kalahari in modern Botswana, rainfall in the mid ¬rst millennium
ad was substantially higher than it is today and supported a strongly pastoral
culture, known as the Toutswe tradition, which practised the same settlement
pattern and differentiated about ad 1000 into a hierarchy of larger and smaller
settlements, implying the existence of political authorities and demonstrating
the importance of cattle as stores of wealth and means of strati¬cation.
The Toutswe tradition survived until at least the thirteenth century, when
drought and overgrazing may have depopulated the region. In modern Natal,
by contrast, early iron age pottery was supplanted quite radically between
the ninth and eleventh centuries by a new ceramic style and the smaller
settlement sites generally associated with pastoralism in this wetter environ-
ment. Whether this discontinuity was due to immigration or a local expansion
of pastoralism is uncertain, but the new pattern was still found among the
Nguni-speaking peoples of the region when Europeans ¬rst described them.
To the west of the Nguni-speakers, across the Drakensberg Mountains, the
closely related Sotho-Tswana peoples also entered the archaeological record
together with more extensive cattle-keeping. Their origin is contentious, but
during the twelfth century the Moloko pottery later associated with them
replaced earlier wares at sites in the northern and eastern Transvaal. From
this base, they colonised the highveld of the southern Transvaal and, from
the ¬fteenth century, the Orange Free State, whose treeless, drought-prone
grasslands had deterred cultivators as much as they attracted pastoralists. The
chief building material here had to be stone; remains show settlements of up
to ¬fteen hundred people composed of interlocking circles of huts clustered
102 africans: the history of a continent

7. Colonising society in eastern and southern Africa.
Colonisation in eastern and southern Africa 103

around cattle pens and linked into communities by dry-stone walls. The mul-
tiplication of these settlements during the sixteenth and seventeenth centuries
suggests population growth in favoured areas, from which Sotho and Tswana
groups dispersed in all directions, using their cattle wealth to assimilate earlier
peoples and form the small chiefdoms that became the action groups of the
highveld™s later history. Hereditary chieftainship, the homestead as social unit,
and the ideological predominance of cattle became shared cultural character-
istics of South Africa™s Bantu-speaking peoples, who had few of the stateless
societies so common elsewhere in the continent. Yet until the eighteenth cen-
tury chiefdoms were small, partly because ample land enabled the ambitious
or discontented to establish new micro-units and partly because inheritance
systems encouraged ¬ssiparation by giving almost equal status to the ¬rst son
of a chief™s ¬rst wife and that of his great wife (married after his accession). No
settlement hierarchy of the Toutswe kind, suggesting a larger political system, is
visible on the southern highveld until the nineteenth century. Similarly, among
the Xhosa, the most southerly Nguni-speaking group, all chiefs belonged to the
Tshawe royal family “ allegiance to them de¬ned Xhosa identity “ but chief-
doms multiplied in each generation as sons settled unoccupied river valleys,
retaining only loose allegiance to the senior line. If a Xhosa ruler displeased his
subjects, so their ¬rst missionary reported, they gradually emigrated until he
North of the Limpopo, many (but not all) early iron age pottery traditions
were supplanted during the centuries around ad 1000 by new styles. Some “
especially the Luangwa style, which became predominant in the north, centre,
and east of modern Zambia and the north and centre of Malawi “ probably
signify migration eastwards from the Katanga area of Congo and the Copper-
belt of Zambia. Others, notably Kalomo and later wares on the Batoka Plateau
of southern Zambia, may indicate expansion by local cattle herders, although
this is disputed. The most complex changes took place in modern Zimbabwe,
the plateau between the Limpopo and Zambezi Valleys, where a higher degree
of political organisation developed. Cultivators had inhabited this region since
about ad 200. Some probably spoke languages ancestral to those of the Shona
peoples now numerically dominant there. Late seventh-century plateau sites
contain beads imported from the Indian Ocean coast. Two centuries later,
Schroda, in the Shashi-Limpopo basin south of the plateau, reveals the ¬rst
large quantities of imported beads found in Central Africa, along with scraps
of ivory that suggest that this material was the source of its prosperity. Yet
the region™s chief nonagricultural wealth was a gold vein running along the
plateau™s highest ridge from southwest to northeast. Four goldworking sites
show signs of exploitation at the end of the ¬rst millennium ad. The earli-
est reference to gold reaching the coast is al-Masudi™s account of ad 916. Less
than a century later, at a site known as Leopard™s Kopje (Nthabazingwe) in the
104 africans: the history of a continent

sweet-grass country near modern Bulawayo that attracted successive pastoral-
ists, there is evidence of greatly increased cattle herds and a new pottery style.
These Leopard™s Kopje people probably spoke a southern variant of the Shona
language. Between the tenth and twelfth centuries, they spread widely across
the Zimbabwe plateau.2 Their expansion left much of previous local cultures
intact. Hoes and grain-bins scarcely changed. But a new political pattern did
result, because the cattle-keepers could accumulate followers and power not
only by deploying livestock but by exploiting international trade in gold.
The effects were ¬rst seen on the southern bank of the Limpopo at Mapun-


. 3
( 12)