Show Summary Details

Page of

Printed from Oxford Research Encyclopedias, African History. Under the terms of the licence agreement, an individual user may print out a single article for personal use (for details see Privacy Policy and Legal Notice).

date: 20 January 2021

Political Complexity North and South of the Zambezi Riverfree

  • Maria SchoemanMaria SchoemanDepartment of Archaeology, University of the Witwatersrand


Complexity flourished in several regions north and south of the Zambezi during the second millennium ce. Four of these localities are explored here: the Greater Shashe-Limpopo region, the Zimbabwean plateau, the Upemba depression–Katanga area, and the Maravi state. Politics in all four regions were fluid, and processes of fusion and fission sporadically reconfigured these socio-political landscapes. Fluidity also manifested in the relocation of political centers as political loyalties were realigned, economic networks shifted, and, in the more recent period, colonial expansion affected southern Africa. Political fluidity, however, does not mean population discontinuities. Settlement and material culture distribution patterns in the Greater Shashe-Limpopo region suggest that networking between communities was vital to the development of complexity in the region. This included interaction between first peoples and newcomers, which has also been noted in the Mutapa, Khami, and Venda polities. There also was substantial continuity in local Upemba populations throughout the occupation sequence, irrespective of the configuration of political control.

Political power was closely entwined with economic wealth in these regions. Economic commodities, however, varied. In the Greater Shashe-Limpopo region and on the Zimbabwean plateau, cattle were important objects in the accumulation of wealth. In these economies cattle had intrinsic value but were also used to facilitate other forms of wealth generation, such as trade and mining. Cattle keeping played a less significant role north of the Zambezi. In the Upemba basin, salt and metal production furnished important trade goods, but trade in these products did not drive the development of complexity. In contrast, the expansion of Maravi political power was entangled with Indian Ocean trade networks.


Thirty years after Connah noted the disjuncture between recorded archaeological sites and the number of historically known urban and/or state centers in Africa, several key historically known centers in southern Africa still have not been located and/or excavated.1 These include the Mutapa, Luba, and Maravi capitals.2 A pursuit of these nodes, however, does not mean the perpetuation of capital-centric archaeology, which has largely focused on the development of hierarchies and unequal access to wealth rather than complexity. As discussed in this article, an increasing amount of contemporary research on states in southern Africa is formulated in terms of complexity, such as that advocated in the seminal McIntosh volume Beyond Chiefdoms: Pathways to Complexity in Africa (1999). Complexity-rooted approaches recognize that there are hierarchical and heterarchical forms of complexity and that these are flexible. They also transcend older ethno-state models and recognize that the populations of complex societies could be heterogeneous.3 Since the McIntosh volume, a growing body of Central and Southern Africanist scholarship has been challenging these traditional approaches. This includes Denbow and de Maret’s engagement with the complicated interface between heterarchy and hierarchy in complex systems as well as the identification of social and cultural heterogeneity in states.4

Grappling with heterogeneity has been a challenge, for not only is a theoretical shift required, scholars also have to navigate bodies of information produced in the context of colonial and nationalist intellectual paradigms. Central to these were tribalist understandings of the past, which manifested in approaches to “culture” that have been extensively critiqued by anthropologists such as Archie Mafeje and Adam Kuper as well as the archaeologist Martin Hall.5 These approaches conceptually delinked people of different origins living in the same area and entangled contemporary or historic ethnic identities with those in pre-colonial contexts.6 Consequently, contemporary or historic ethnic/linguistic identities were uncritically projected into the past.7 These invented tribal identities were frequently used to link vast areas, with very distinct archaeological assemblages, as has been done in the “Zimbabwe culture” model.8 Continuities in archaeological material culture would be read as indicators of endurance in ethno-political units.9 Ceramic style has been entwined with this approach, and it is frequently interpreted as representing linguistic or cultural groups. Accordingly, ceramics were used to trace migrations as well as in situ developments through time.10

In order to avoid the ethnic trap, this article does not discuss ceramic style as the manifestation of identity per se but rather uses the concept of communities of practice.11 This allows material culture to be viewed as manifestations of learning networks and interaction and thus transcends approaches that treat identities as static.12 Focusing on conceptual communities in complex political systems and states helps to decenter past tribalist approaches to complex societies.13

The organization of this article is guided by geographic location and temporality. The discussion starts with the development of complexity in the south, from where it shifts north and then east, while developments in each region are traced through time. This organization does not imply that the regions were not interconnected or at times affected by the same or similar processes.

The Greater Shashe-Limpopo Region

In the second half of the first millennium, the people in the greater Shashe and Limpopo river region started to change the political and economic configuration of their societies. This eventually led to the development of socio-political and economic complexity in this region. These developments centered on two areas: southeastern Botswana, with Toutswemogala as a focal point, and the Shashe Limpopo Confluence Area (SLCA), centered on Schroda, K2, and Mapungubwe.

Map of the study area and location of the sites mentioned in the text.

Source: Drawn by the Author.

Southeastern Botswana

Complexity, represented in extensive and complex networks and settlement hierarchies, manifested in southeastern Botswana from the 8th to the 14th centuries. The settlement hierarchies started to become visible after the economic and ideological importance of herding increased in the region.14 Toutswe, occupied from the 8th century, became the pinnacle of a threefold hierarchy that developed in southeastern Botswana.15 The elite component of the Toutswe center is located on a flat-topped hill, but the site as a whole extends over 7 ha. The second tier in the settlement hierarchy was occupied by sites such as Thatswane, Taukome, and Bosutswe, where the midden areas, which represent the residential zones, occupy 1 ha on average. Smaller single-household homesteads built around cattle enclosures formed the first tier.

Archaeological evidence for the presence of hunter-gatherers co-occurs with farmer material culture at first and second tier sites. Material culture associated with farmers has also been found at hunter-gatherer shelters and open-air sites in the region, suggesting two-way networks of interaction between farmers and hunter-gatherers.16


The earliest material representation of the development of complexity in the SLCA can be found at Schroda, which housed 300 to 500 people in a 15 ha area. It is the largest known site in the SLCA area between the 10th and 11th centuries. Unlike Toutswe, Schroda was not located on a hilltop.17 All other contemporaneous SLCA sites were either farmer family homesteads or hunter-gatherer shelters.18

At the start of the second millennium, a regional political realignment manifested itself in the shift of the capital from Schroda to K2. It was located less than 5 km from Schroda, at the base of Bambandyanalo hill.19 Schroda was not immediately abandoned, but K2, with approximately 1,000 residents, was substantially larger than Schroda. Under the leadership of the K2 kings, the SLCA polity expanded further. This is visible in population size, which is reflected in the increase in the number of homesteads in the Limpopo valley.20

Ongoing economic and political complexities and stratification eventually led to the shift of the center of power to a new capital at the start of the 13th century. The new political center was located at Mapungubwe hill, a kilometer away from K2. Mapungubwe started as a smaller village at the base of the hill during the K2 period. At its zenith, the village at the base of Mapungubwe was substantially larger than that at K2.21 Huffman estimated that 5,000 people lived at in the village at the base of Mapungubwe hill, with a further 15,000 residing in homesteads dotting the Limpopo valley.22 At this point the ruling elite moved onto the hilltop. Mapungubwe hill was abandoned at the end of the 13th century, but research in the larger valley indicates that the region was occupied well into the 14th century.23 Excavations at the hilltop site of Mapela, 80 km northwest of the SLCA, also yielded 14th-century dates for the most recent layers associated with Mapungubwe pottery.24

Unlike their K2 period predecessors, the Mapungubwe period elite did not live in the village. Instead, their residences and burial area were located on top of the hill. It has been argued that this residential pattern is the first manifestation of the development of class distinction in southern Africa.25 Hierarchies are also visible in the gold objects found in the burials on Mapungubwe hill. Gold objects have not yet been found in any of the other sites in the SLCA. Elite Mapungubwe period burials at K2, which probably was perceived as an ancestral site, provide further evidence for the use of rare material culture to signal status—for example, a juvenile burial contained a large multiple wound beads, and a similar bead was found on Mapungubwe Hill near the elite burial.26

Traditionally the shifting capitals were interpreted as the result of ethnic displacement.27 A critical analysis of the archaeology and ceramics found a more complicated pattern. The temporal and spatial distribution of Zhizo ceramics, which were found at Toutswe and Schroda, suggests that at the end of the first millennium a community of practice stretched from southeastern Botswana through the SLCA into southwestern Zimbabwe. People who formed part of this community of practice, however, also interacted with those in another community of practice centered in southwestern Zimbabwe that was represented in Leopard’s Kopje ceramics. Consequently, Zhizo ceramics contained Leopard’s Kopje elements.28

The geographic extent of these communities of interaction shifted after the start of the second millennium, when the SLCA capital shifts to K2. At this time the Leopard’s Kopje community of practice expanded into the SLCA. This did not replace the Zhizo community of practice, and people in the SLCA were part of two overlapping communities of practice from the 11th to the 12th centuries ce. Social and learning networks, however, appear to shift during the K2 period, and Zhizo ceramics in the SLCA incorporated a substantial amount of Leopard’s Kopje stylistic elements. This suggests a gradual shift of people in the region toward ceramic knowledge acquisition in the Leopard’s Kopje community of practice and thus a realignment of key social networks toward southeastern Zimbabwe.

The appearance of a new ceramics style, known as Leopard’s Kopje B or Mapungubwe, at the start of the 13th century indicates that state formation processes created a new community of practice, with the SLCA at its core.29 This community of practice included people in the greater SLCA and extended deep into southwestern Zimbabwe to the north. The geographical reach of the Zhizo community of practice shrunk to southeastern Botswana.30 These farmer networks intersected with contemporaneous hunter-gatherer networks, which stretched to northern Botswana to the Soutpansberg south of the SLCA.31

Cattle keeping was central to the economy and accumulation of wealth in southeastern Botswana.32 Denbow argued that the economic importance of herding increased in Botswana in the second half of the first millennium, and this eventually manifested in the unequal distribution of wealth and the associated development of settlement hierarchies in the regions of Botswana that can accommodate agropastoralism.33 In the SLCA the centrality of cattle might have shifted over time, as other forms of economic specialization developed.34

One of these was the development of specialized farming systems. The first was the extensification of flood plain farming, which suggests that farmers were optimizing sorghum and millet yields in order to feed a fast-growing population. Cattle management, including the development of cattle transhumance, speaks to the same concerns. Grazing cattle far away from the central valley, however, points to not only effective herd management strategies in the relatively dry region but also some form of political hegemony over the zone between the Limpopo valley and the Soutpansberg and deep into Botswana where cattle were grazed.35 Social networks with people in southwestern Botswana, and thus Toutswe, also might have assisted in the development and continuation of these cattle management strategies.

There also is some archaeological and spatial evidence for agricultural specialization in southeastern Botswana. Based on their research at Lose, which was continuously occupied from 900 to 1460 ce, Kiyaga-Mulindwa and Widgren argued that settlement in southeastern Botswana was more complex than previously portrayed and that the region also housed several large “agro-towns,” rather than only Denbow’s scattered homesteads with a few densely occupied nodes.36

Glass trade beads occur in all SLCA archaeological assemblages. Beads appear to become common during the K2 period and might no longer have had prestige commodity value per se. This might have resulted in an attempt to reconfigure their social/economic value. The ubiquitous small glass trade beads were melted by K2 period SLCA citizens to make larger composite beads, commonly known as K2 garden rollers.37

Some of networks of interaction or trade are visible in the small number of garden rollers found in Botswana and Zimbabwe.38 Huffman argued that long distance trade that linked the region to the Indian Ocean became a key driver of state formation.39 Denbow, on the other hand, linked the development of complexity to cattle and social networks.40 Clearly, these arguments are not incommensurate: one focuses on the introduction of external factors into the local system, while the other argues that the potential for the development of complexity resides within the system. Irrespective of the main driver, the transformation processes of the late first millennium started to reconfigure the social, economic, and political makeup of the region. Southeastern Botswanan and SLCA farmers and hunter-gatherers formed part of a network in which goods circulated and which by the late first millennium connected the interior of southern Africa to the Indian Ocean.41

At the time this network articulated with southern Africa through the earliest known trade port on the Mozambican coast, Chibuene, which was established in the 7th century.42 The earliest evidence for trade in the interior is found in the 7th- and 8th-century Chibuene series beads at Nqoma in western Botswana. The wide distribution of the 8th- to mid-10th-century Zhizo bead series indicates the rapid expansion of trade in the next century.43 The end of Chibuene as a port coincides with the arrival of a new type of bead in the SLCA. This K2 bead series, as well as the later Mapungubwe series, entered southern Africa through an (as yet) unidentified port.44

Hunting generated export goods, such as ivory and skins that could be traded for beads and other goods.45 The pattern at Mwenezi, in the Mateke hills in southwestern Zimbabwe, points to elements of local specialization in this regional system. The Mwenezi farming community occupation starts between the 8th and 9th centuries and continues into the Zimbabwe phase. Unlike other farming community sites, the faunal assemblage includes very few cattle remains. Instead a substantial amount of zebra remains was recovered. The amount of zebra remains exceeded normal dietary requirements, and Manyanga et al. suggested that, similar to ivory and leopard skin at Schroda, zebra skins were traded.46

The Zimbabwe Plateau

Great Zimbabwe

After the decline of the polities in southeastern Botswana and the SLCA, the Great Zimbabwe state was the dominant entity south of the Zambezi. At its zenith it controlled much of the Zimbabwean plateau.47 Whether this system represented hegemony, however, has been contested. Huffman favored a system of centralized control in which the political elite used a network of regional stone-walled sites, known as zimbabwe, to manage affairs of the state.48 Beach, however, argued that it is very likely that the political processes of the Great Zimbabwe state would have been as fluid and contested as those in historically known Shona contexts.49 Garlake favored economic complexity, arguing that Great Zimbabwe was the product of a stratified society and that the site of Great Zimbabwe sat at the heart of an economy based on processing resources.50 This argument was further supported by the location of zimbabwe on the edge of the plateau, which he suggested formed part of a cattle management system. In this, cattle-raising communities at the plateau’s edge paid tribute to the center.51

The center of this economic and political system was located at the stone-walled site of Great Zimbabwe, and a substantial amount of the research on the Great Zimbabwe state has focused on the site of Great Zimbabwe. Unfortunately, early excavations and subsequent “conservation” attempts have resulted in substantial deposit loss, which has hampered the secure dating of the site development, which in turn has led to a substantial amount of debate.52

This debate centers on the start of stone walling at Great Zimbabwe, which is interpreted as an indicator of statehood, and the date of abandonment. Interpretations of the occupation sequence are based on ceramics style, architecture, dated trade goods, and radiocarbon dates.53 The earliest occupation evidence (Period I) dates to the 6th to the 8th centuries. There is no evidence for stone walling at this time. A hiatus precedes Period II, during which dhaka (prepared clay) house floors appear, which Huffman and Vogel dated to the mid-12th to early 13th centuries.54 Period III signifies the zenith of Great Zimbabwe as a state capital and is characterized by type P stone walling—“Zimbabwe cement” houses. This was dated to the 13th century by Huffman and Vogel,55 but Chirikure et al. places the transition from Period II to Period III in the 12th century.56 Huffman and Vogel suggested a mid-15th-century date for the end of the occupation, whereas Collet et al. and Chirikure et al. argued for a mid-16th-century date.57 Irrespective of the date, the terminal occupation period signaled a gradual decline in the power of the Great Zimbabwe capital rather than an instant abandonment.58

The elaborate stone-walled architecture of the Great Zimbabwe elite residences, which would have required the investment of a substantial amount of labor, stood in contrast with the simpler family homesteads of the majority of citizens.59 This material contrast represented elements of this system such as centralized control of political power and unequal control over resources and is at the core of what has been termed the Zimbabwe culture settlement pattern.60 These structures invoked and reinforced power. This manifestation of power, combined with the system of rule, aided in the establishment of political and cultural hegemony.61

Wealth generated by this successful state would have further ensured social stability. This wealth was generated by the entwined elements of cattle, trade, and precious metals.62 While this meant that cattle formed an important part of the trade system,63 they also had independent value because they formed a substantial part of meat consumption at Zimbabwe sites and dominated the animal economy.64 Due to their crucial role in local economies as well as the state economy, effective systems of management were developed, including transhumance.65

As in the rest of southern Africa, Great Zimbabwe ceramics have been treated as indicators of identity. Consequently, Great Zimbabwe has been described as a Shona site.66 This consensus is supported by substantial stylistic continuities into the historic period from Zimbabwe Class 3 pottery, which marked Period III.67 This is distinct from Period I Gokomere/Ziwa/Zhizo pottery but related to Period II Gumanye pottery.68

The distinct ceramic signature is significant because Mapungubwe ceramics are distinct from Great Zimbabwe ceramics, which also developed out of Leopard’s Kopje.69 This suggests that while people in these polities might have common origins, they were no longer part of the same community of practice and thus shaped local distinct cultural trajectories. This supports the argument that Mapungubwe and Great Zimbabwe were competing polities, not successors as suggested in earlier literature.70 Recent reanalyses of radiocarbon dates from Great Zimbabwe also support the parallel polities model. These analyses suggest that the start of the occupation of the Hill Complex dates to the 12th and 13th centuries, while the beginning of construction of the Great Enclosure dates to the 13th and 14th centuries.71 This chronology means that Great Zimbabwe started to develop during the second half of the K2 occupation.

Similar to the broader Shashe-Limpopo region, the Great Zimbabwe economy included cattle and crop farming, metal mining, and trade. Various forms of economic specializations, such as gold mining, existed in the Zimbabwean state, but Pwiti argued that cattle, metal, and trade goods form part of an integrated trade system in which cattle were used to reward gold miners and producers. This gold was then traded for imported goods.72

Other key export goods included ivory, skins, and a range of metal types such as gold, copper, and tin. Trade goods reaching the site included imported glass beads and ceramics.73 The imported beads were intimately associated with elite centers and thus were a possible status marker.74 The earliest datable trade goods at Great Zimbabwe are Chinese Song dynasty ware, which dates from the 10th to the 13th centuries, as well as K2 series trade beads and garden rollers, which date to the 11th to the 12th centuries.75 Blue and white Ming dynasty porcelain found by Collett et al. indicated that Great Zimbabwe remained part of the Indian trade system until the 16th century.76

Most of these trade goods reached Great Zimbabwe through the Indian Ocean trade network, in which the relationship between Great Zimbabwe and Kilwa played an important role.77 This international trade articulated with other networks that operated at different scales. Longer-distance trade linked Zimbabwe to other regions in the southern African interior and moved objects, such as iron gongs and copper, between areas of production and areas of consumption or use.78 As discussed above, local networks within the state ensured the effective management of livestock.79

The weakening of the Great Zimbabwe capital created opportunities for the development of stronger political centers in areas in the hinterland of the state, as well as in areas previously outside of its direct control. In Zimbabwe two centers dominated the political landscape from the 16th century onwards: the Mutapa state Zambezi valley and Khami in southwestern Zimbabwe. Initially these centers shared the Zimbabwe plateau with Great Zimbabwe, which was in decline by the second half of the 15th century.80 Later, other smaller economic-political centers developed in the broader region. These included the Venda state in southern Zimbabwe and northern South Africa and the Nyanga complex in eastern Zimbabwe.


Small chiefdoms were established in farming communities in northern Zimbabwe and in the Zambezi valley by the 12th century. The occupation of these overlapped with the 15th-century shift of Great Zimbabwe state sites into northern Zimbabwe.81 The Mutapa state was led by one of the dynasties that moved into the region. The residences of these leaders were large settlements associated with stone walling.82 These represented the hierarchical organization of the Mutapa state, in which dynasties originating out of the Great Zimbabwe state ruled a society that included descendants of earlier farming communities.83 This new state eventually controlled most of the northern Zimbabwean plateau, as well as parts of the Zambezi valley, from the 16th to the late 18th centuries. The historian Beach argued that the Mutapa state did not develop out of Great Zimbabwe but instead was the result of local developments in northern Zimbabwe.84 Pwiti suggested that economic and ideological changes in the region led to the rise of the Mutapa state and that while the leadership originated at Great Zimbabwe, they ruled over a population who lived in the area prior to their arrival. Based on archaeological data, Pikirayi contended that the Mutapa state was the direct successor to Great Zimbabwe.85

The political processes in the Mutapa state were complex. Mudenge suggested internal contestations in the Mutapa state during the 15th century, and Abraham contended that there was conflict between the Mutapa and Rozvi in the 16th century.86 The fame of Mutapa gold also placed this state in conflict with the Portuguese.87

The economy in the Mutapa region appears to have been configured similar to that of Great Zimbabwe, with cattle, agriculture, gold, and trade continuing being key components. The arrival of the new dynasties from Great Zimbabwe, however, clearly stimulated economic development, and after the 15th century the number of cattle herds in northern Zimbabwe increased.88 Excavations in Zambezia have yielded a substantial amount of imported goods.89 Similar to Great Zimbabwe, the Mutapa chiefs used trade to enhance their power. This was in part achieved through the distribution of exotic goods as well as cattle. This enabled chiefs to mobilize labor successfully in order to perform a range of activities in the Mutapa area. This included gold mining. There also is evidence for craft specialization.90

Pwiti, however, recognized complex patterns of interaction between the rulers and the ruled in the ceramic record. The dominant pre-15th-century ceramics style in the region was Musengezi pottery, which had been made by potters in the small chiefdoms in the region from the 12th century onwards. In the 15th century Great Zimbabwe ceramics appeared in the region. This has been found predominantly at elite sites, whereas commoners continue to use Musengezi-style ceramics.

According to historical records the shifting Mutapa state capitals, as well as the Massapa trading market, were located close to Mount Fura in northern Zimbabwe.91 In this region there was significant shift in the architecture of the ruling elite in the 16th century—stone walling is abandoned in favor of organic materials. Archaeological research on one of these towns at Barada near Mount Fura in northern Zimbabwe yielded evidence of this new approach to space and settlement, including not using stone walling in the configuration of this kilometer-long town. Excavators also found 16th-century trade goods in association with local pottery. Pikirayi suggested that this new approach to space combined with the mixed material culture assemblage speaks to the development of dual Afro-Portuguese and Zimbabwean identities.92 The hybridity that Pikirayi identified is echoed in the historical record, and Newitt argued that a hybrid of Portuguese African culture developed in the Zambezi as a result of interaction.93


Not all stone walling in northeastern Zimbabwe was an expression of the Zimbabwe cultural complex. A notable example is the extensive stone-walled agricultural complex around Nyanga, which was occupied between the 16th and 19th centuries ce, and thus contemporaneous with the Mutapa state. This complex comprises thousands of hectares of terracing, associated hydraulic works, and settlement structures. Unlike the Great Zimbabwe stone walling, the Nyanga stone-walled structures were not coursed.94 The Nyanga architecture also does not speak to the hierarchies embodied in the construction of the zimbabwes.

Soper argued that Nyanga was a relatively isolated community. This is in part supported by the distinct form the ceramic traditions associated with Nyanga, which is different from that of the Great Zimbabwe and Mutapa states. They, however, were not completely isolated, and archaeologists have excavated evidence for trade with the coastal networks, and Nyanga pottery has been found at sites associated with other Late Farming Communities.95


The counterpart of the Mutapa state in southwestern Zimbabwe was the Torwa-Changamire-Rozvi state, where an increasingly powerful elite also competed with Great Zimbabwe in decline. The site of Khami was the largest in the region, suggesting that it was the state capital.96 The development of this distinctive archaeological complex has been linked through oral histories to the Torwa dynasty. They were unseated by the Changamire-Rozvi-Rozvi, who built their capital at Danamombe in the late 17th century. The Rozvi ruled southwestern Zimbabwe until the early 19th century, when the Ndebele arrived. The Rozvi at Khami and Danamombe were at the center of a much larger Rozvi confederacy.97

Some Khami ceramics have been found in the upper levels of Great Zimbabwe. Based on this, Robinson argued that the group that occupied Khami had also lived at Great Zimbabwe.98 Huffman argued that the Khami state was one of the Great Zimbabwe successor states and that there were substantial ideological continuities between Great Zimbabwe and Khami. Consequently, he argued that Khami followed the Zimbabwe pattern, in which leaders live in elaborate stone-walled sites, which are surrounded by commoner housing built with organic material.99 Pikirayi and Chirikure, however, favored interaction between the two centers rather than occupation of Great Zimbabwe by people associated with the Khami state.100

Robinson contended that he found a homogeneous, very distinctive pottery style in the occupation remains at Khami.101 Machiridza, however, has challenged the perceived homogeneity of the Khami ceramic assemblage and instead argued that Khami were diverse and complex. This stands in contrast with the ceramics from later Danamombe, which was simpler and less diverse.102 The uniformity of the style in the later part of the sequence suggests the development of a single community of practice in the region. This community of practice, however, developed from a number of overlapping communities of practice, including communities developing out of the earlier Mapungubwe state.103

The development of a unique elite architectural style also speaks to a community that is intentionally signaling difference from the other polities in the region as well as from the elite at Great Zimbabwe.104 These buildings also point to the ability of the Khami elite to extract a substantial amount of surplus labor.

The substantial number of glass beads recovered from Khami period sites show that the Torwa-Rozvi state participated in long-distance trade with the Indian Ocean coast. Mudenge, however, argued that cattle, not trade, formed the core of the Rozvi economy.105

The Khami state appears to have experienced political fission and fusion processes similar to the Great Zimbabwe state, and in the early 17th century a group separated from the Rozvi and moved into the Soutpansberg in northern South Africa, where they settled at Dzata and became known as the Singo. The links between the Singo and the Rozvi were evoked in the names of Singo leaders, such as Lozvi, Mambo, Dyambeu, and Thoho-ya-Ndou, that were early Rozvi titles.106

The Singo royals united the region politically and intensified trade with the coast. This region eventually developed into the Venda state. The Singo, however, did not create Venda identity. Instead, this was shaped through intense interaction between people who remained behind after the fall of the Mapungubwe state, Sotho-speaking communities that expanded into the region soon after the fall of the Mapungubwe state, and communities aligned to the Khami state, which expanded into the region in the mid-15th century. Loubser argued that this new identity was represented materially in Letaba ceramics.107

Dzata-type stone-walled sites were constructed at regional centers south of the Soutpansberg. The Dzata-type sites modified the Zimbabwe architectural vocabulary and in turn were later transformed by the political fragmentation processes in the Singo state. These processes led to the formation of two political centers, each associated with a related but distinct stone-walled settlement pattern.108

The Upemba Depression and Katanga

The development of complexity north of the Zambezi River, similar to the processes south of the Zambezi, started in the later part of the first millennium. The well-known 19th-century Luba state, for example, was the result of local processes in the Upemba depression that spanned more than a millennium. This sequence, however, is characterized by substantial continuity.109

These processes started to become visible in the Ancient Kisalian period, which dates to the 8th to 9th centuries. At this time population density started to increase relative to the lower densities in the preceding Kamilambian. A few ceremonial axes and an anvil, symbols of leadership in the later Luba period, have been found in burials, suggesting the development of some form of political leadership.110 During the 10th- to 12th-century period, known as the Classic Kisalian, population density increased further and remained high for the rest of the archaeological sequence. The geographic scope of the occupation also expanded to the plateau southwest of the depression. Grave goods diversified, and some graves included abundant amounts of pottery, as well as ivory, copper, and iron ornaments, but the division between the small group of wealthy individuals and others was not sharp. Grave goods also suggest that status was inherited.111 Iron, copper, ivory, bone, and pottery vessels were produced by specialized craftspeople.112

The socio-economic success of the political system in the Upemba region translated into significant population growth, accompanied by high infant mortality, which is suggested by an increase in the frequency of graves dating to the Classic Kisalian and Kambambian.113 This pattern is not dissimilar from that manifest at K2.114

The 13th-century transition from the Classic Kisalian to the Kambambian A was associated with substantial changes in grave goods. Only one symbol of power has been found in graves from this period, but, paradoxically, a greater contrast between the grave goods of the wealthy and the general population developed in the Kambambian A. Overall, metal grave goods declined, but copper goods became more common. These included copper crosses, which initially were a special-purpose currency, but in the later Kambambian B they became an all-purpose currency. As part of the process of becoming an all-purpose currency, the crosses became more uniform, but the demand also resulted in a decrease in size. Local variations in pottery and burials developed at this time, suggesting the development of local communities of practice.115 While population density remained high in the Kambambian B, which started in the 16th century, the number of pots and copper crosses in the graves that have been excavated decreased. Phiri argued that the Luba merged with their Lunda neighbors in the 15th century, which would place the merger at the start of the Kambambian B.116

The decreasing amount of grave goods continued in the Modern Luba period, dating from the 17th century onwards, when pots and metal were no longer used as grave goods. The only type of material culture found in these graves was glass beads. These beads would have been obtained through trade with the coast, which the Luba elite controlled.117 Luba period graves were avoided by de Maret.118 Furthermore, the Kabongo region, where the Luba kingdom’s center was located, has not been excavated, and thus an archaeological understanding of the consequences of their rise to power remains underdeveloped. Similar to other sub-Saharan African states, ideas about the Luba state have been entangled in ideas that reconfigured Luba political hegemony into cultural uniformity. However, it has been recognized that the Luba state was not a homogeneous ethnolinguistic empire. Instead, it comprised numerous territories, and people had a range of identities, while there is also evidence for population continuities.119 This diversity might have originated in the Kambambian A.120

Historians have linked the 17th- and 18th-century centralization processes in the Luba and Lunda kingdoms to the trade of copper, iron, and salt in which the rulers, without significant external contact, expanded their control over mining and production processes.121 This is clearly at odds with the archaeological record, which suggests that the political complexity that underlay the Luba and Lunda states developed in the Upemba region over a substantial period of time, starting in the Ancient Kisalian (700–900 ce), when pottery became more elaborate and complex. This pottery had developed into uniquely high-quality ware by the start of the Classic Kisalian (900–1200 ce), which might signal specialist production.122

Luba symbols of authority, such as iron anvils and axes, were already present in Ancient Kisalian graves. Similarly, copper crosses are present in the archaeological record from the Kambambian A, and these also are symbols of political authority and wealth.123 There was, however, a clear shift away from including socio-political status marking material culture in burials in Kambambian burials.124 This is interesting because the Kambambian period overlaps with the period in which the Luba state is historically known to have existed, and distinct socio-political hierarchies have been proposed for the Luba kingdom, which were made material in objects similar to those found in the Upemba sequence.125 Consequently, it is possible that political status objects were not present because graves from the Luba heartland did not form part of the sample.

The archaeological record suggests early origins for marked social stratification and the development of forms of political leadership, as well as continuity in the material representation of leadership.126 For de Maret this leadership, however, is rooted in social complexity, not only unequal access to goods but the model of a highly stratified society put forward for Mapungubwe and Great Zimbabwe.127 De Maret draws on Guyer and Belinga to argue that wealth-in-people rather than wealth-in-things might have played a role in the shaping of complexity in the Upemba basin.128 Wealth-in-people allowed leaders to mobilize people, skills, and knowledge to navigate various contexts. In order to achieve this, leaders unified people through the use ritual power and social networks.129 The role of ritual in Upemba leadership is visible in the inclusion of regalia in burials, such as the inclusion of an anvil in a 9th-century Early Kisalian grave, and the presence of inclusion of ceremonial axes, which at times are found in association with anvils.130

Copper, in the form of ornaments, first appears during the Ancient Kisalian and was acquired through trade with the Copperbelt. This trade expanded during the Classic Kisalian.131 Cowrie shells, found in Classic Kisalian period sites, suggest trade with the Indian Ocean coast. This trade further expanded during the Kambambian A, and by the Kambambian B period the region was firmly part of the long-distance coastal trade networks.132 The networks that carried this trade, however, developed much earlier. Local exchange or trade in the Upemba depression might have started in the Ancient Kisalian, with large curved knives and spearheads, which are typical for the period, functioning as an early currency.133

Copper became a key trade metal in regional as well as long-distance trade. The earliest evidence for copper production in the broader region dates to the 4th century, but it remained scarce in the Upemba depression until the Classic Kisalian period. Its relative scarcity suggests that it was a valuable resource, possibly restricted to elite use. Copper started to become more common during the Classic Kisalian period and is abundant in the Kambambian period. The classic copper crosses first appeared in the archaeological record at this time; these were used as used as currency and to signify wealth and authority.134 The abundance of copper in the Upemba depression during the Kambambian A points to an increase in trade with the Copperbelt region to the south.135

These networks, however, extend much further to the south, and copper originating from the region has been found as far afield as Zambia, Zaire, and Great Zimbabwe.136 During the historic Luba period these trade networks extended to the Indian Ocean and Atlantic Ocean coasts, and the Luba rulers used the trade to further enhance their status and power.137

By the 19th century the Luba state was the largest in Central Africa. As noted above, the archaeology of the historic Luba state is not as well understood as the preceding Upemba sequence, but historical sources do assist in the construction of a more comprehensive understanding. Reefe, using oral traditions, argued that the rulers of the vast Luba empire used a system of rule that was based on consensus building through ideology, reward, and institutions that integrated society. This system articulated with shared religious beliefs in which ancestors were central and in which the royal center was sacred. The political-religious authority of the Luba kings was represented in the royal regalia, which included axes.138

The empire was fully established before international trade and colonial expansion impacted the region. Historical sources suggest that the state united a network of smaller lineage-based states, which were led by sacral kings. Violence was not a favored expansion method.139 These smaller states became client states and paid tribute to the center. Local villagers also paid tribute, which included the bulk of food consumed by the Luba rulers and their families.140 This system, however, was vulnerable to challenge. In the 18th century, for example, small salt-producing states united in the Kanyok kingdom and refused to pay tribute. This threatened the Luba center’s economic interests. Luba attempts to reassert control over the Kanyok area through warfare failed.141 It is likely that similar fission and fissure processes characterized the political landscape for much of the Upemba sequence.

While the rise of the Luba state was not linked to international economic factors, its fall was. In the mid-19th century external economic-political actors started to intrude into central Africa. Their presence reconfigured the economic and political landscape, which resulted in redirection of tribute, revolt, succession struggles, and armed warfare. The extent and scale of these conflicts ripped the empire apart.142


The Maravi complex east of the Upemba and Katanga regions controlled large territories north of the Zambezi. These included areas that today form part of Zambia, Malawi, and Mozambique. Some indications of the size of the Maravi state can be gained from Portuguese records; for example, in 1608 Muzura supplied 4,000 warriors to help the Portuguese in their Mwene-Mutapa campaign. This sizable army points to a substantial population with strong allegiances to the center.

Unfortunately, the substantial disjunction between historical data and archaeological data relating to the Maravi kingdom/state has limited research insights. This disjunction stems in part from the limited amount of oral history and archaeological research that has been conducted in the Maravi area.143 Consequently, understandings of this state remain limited.

Phiri suggested that the Maravi state was the product of socio-political interaction between proto-Chewa and Maravi immigrants. Citing historical linguistics, he argued that the expansion of the state started in the mid-16th century and that they are recorded as having been important power brokers in 17th-century northern Zambezia, where their economic prosperity was due in part to their trade in ivory with the Portuguese.144

Oral traditions and material culture also suggest some form of relationship with the Luba-Lunda kingdom in Katanga. For example, white initiation paintings found in the Luba state area appear to be similar to Chewa female initiation art.145 This suggests that at some point in the past the regions formed part of a similar, or the same, initiation community of practice. Furthermore, Mgomezulu argued that Early Mawudzu ware in Zambia and Malawi are identical, which suggests some form of early interaction throughout the region, rather than later processes of diffusion that brought the Maravi leaders into contact with proto-Chewa communities.146

From an archaeological perspective, there is broad consensus that the Maravi polity was linked to Chewa speakers. The arrival of Chewa speakers is traced to the 12th or 13th century.147 Mgomezulu suggested that the arrival of Chewa speakers is visible in the archaeological record through the appearance of Mawudzu ware, which he indicated dated from the 12th to the 18th centuries. This period would coincide with the rise, flourishing, and decline of the Maravi Empire.148

Mawudzu ware has been found in south central, southern lakeshore, and southeast Malawi, in areas where at least three different ceramic wares (Kapeni, Namaso, and Longwe) previously signaled the presence of three different communities of interaction. These were not bounded, and the movement of ceramics on the landscape suggests clear interaction. These networks of interaction also stretched further afield, and sites in southeastern Malawi where Kapeni ware occurs have also yielded Zambian Longwe ceramics.149 This 14th- to 15th-century replacement of earlier wares suggests a shift from three communities of interaction to one. This development was relatively slow, and there was a substantial period of overlap, with, for example, Kapeni ware, which first appears in the 10th century, being present until the 14th century.150 In the archaeological record the shift from the Maravi state into chiefdoms is evident in the displacement of Mawudzu ware by Nkudzi ware in the 18th century.151 This signaled the development of a new community of practice.

The locations of the Maravi centers have been the subject of much dispute. This includes the location of the capital of the 17th-century Maravi leader, Muzuru. Based on Chewa Kalonga oral traditions, Alpers situated the capital at the southwestern shore of Lake Malawi.152 Rangley, using a range of historical sources, including Gaspar Bocarro’s travelogue, argued, on the other hand, that the capital was located on the banks of the Wamkurumadzi River, which would mean that the capital was associated with the Mang’anja Paramount Kaphwiti.153 Schoffeleers favors the latter scenario and suggests that Muzura was not a member of the Kalonga family.154 There is also substantial debate regarding the date for Muzura’s capital.

Juwayeyi’s excavations of Mankhamba, which formed the core of Kalonga’s state in the southern Lake Malawi area, have started to integrate historical and archaeological data. This site was occupied by pre-Maravi people from the early 13th century. The inclusion of stamping in the decoration of the Mawudzu pottery excavated at the site suggests either origins or networks of interaction to the west. The Chewa-speaking Maravi joined the pre-Maravi communities by the mid-15th century, and under their leadership Mankhamba developed into a prosperous political and economic center, which traded with the coast. The wealth in trade goods, size of the site, and evidence of ritual and political power suggest some form of political and economic consolidation under the leadership of the Kalonga. This, however, was short lived, and the absence of Nkudzi ware at the site suggests that the site was abandoned in the 18th century.155

Trade was an important part of the Maravian economy, and the abundance of imported items at Mankhamba, such as Chinese porcelain manufactured in the 16th century and Khami series glass beads, suggests that this trade was well established by the 16th century.156 Linden suggested that Maravi contact with the Indian Ocean trade network started in the 13th century, through the Lake Tanganyika-Malawi corridor.157 The main trade export commodity appears to have been ivory. Glass beads recovered in Mngomezulu’s excavations also show that sites in southern Malawi formed part of the Indian Ocean trade network. The associated Mawundzu ware dates from the 12th to the 18th centuries. Some of the bead descriptions (e.g., Indian red on green) point to European origins, but whether Swahili trade is also represented is unclear. A link with Kilwa is possibly supported by 10th- to 14th-century Kapeni ware, found in south central Malawi, which is said to be similar to pottery found at Kilwa.158 The continued presence of Kapeni ware, in parallel with Mawundzu ware, until the 14th century could indicate that the people on upper lands and the southern lake shores in the south continued to trade with the coast during the rise of the Maravi complex.

Langworthy argued that Portuguese contact with the region only started after 1600, with one of the earliest mentions being traded with chief Bundy (Undi?) in 1614.159 Until then contact appears to have been with the Swahili coast. Trade in ivory and possibly gold between the Swahili coast and the Maravi increased in the 1500s when Portuguese presence at Sofala blocked Swahili access.160

Langworthy argued that Portuguese interest in trade further declined after the Kalonga increased the price of ivory by restricting it to him and two other chiefs at the end of the 17th century.161 The Yao’s increased dominance in the ivory trade in the early 1700s would have further shifted the focus on the Portuguese.162 The pattern in Undi’s kingdom was different. Initially the center controlled the ivory trade. This started to unravel when, at the end of the 17th century, the Portuguese were blocked from the Mutapa state area; however, they turned their full attention to Undi’s area. The Portuguese established settlements to access the ivory trade, and by the mid-18th century a Portuguese gold mine, which relied on slave labor, was established.163 The overall amount of Portuguese trade with the Maravi region, however, was lower than with the Mutapa south of the Zambezi River.164

The Maravi state had broken into chiefdoms by the mid-18th century. The lack of political hegemony also was expressed in challenges to paramountcies by local chiefs. Gold discovered in Undi’s region also resulted in conflict with the Portuguese.165 This shift is reflected in the archaeological record.166

Continuities, Differences, and Complexity

The development of complexity in the four regions discussed here speaks to regional as well as distinctly local dynamics. Developments on the Indian Ocean coast impacted all regions, directly or indirectly. The most direct impact was the effects of Portuguese colonialism on the Maravi and Mutapa areas. Earlier, Indian Ocean trade was important in the elite status signaling, as well as labor-mobilization processes, of leaders in the SLCA, Great Zimbabwe, Mutapa, and Maravi areas.

Indian Ocean trade, however, was not the sole driver in the development of complexity in states north and south of the Zambezi, and the role trade goods played differed between regions. Cattle and trade were important in the southernmost regions discussed here and have been linked to the rise of complexity in southeastern Botswana, the SLCA, and Torwa -Changamire-Rozvi. Cattle and trade were also important in the Great Zimbabwe and Mutapa states, but in these two cases gold played an equally important role, and Pwiti suggested that cattle and gold were fundamentally entwined. Metal production and trade also played a role in the development of complexity in the Upemba basin and Luba area, but there wealth-in-people was of central importance. The rise and fall of the Maravi state has been directly linked with trade, but it is distinctly possible that systematic archaeological research would point to more complex processes. These might include political contestations and fluidity similar to the Zimbabwean processes discussed by Beach.167

At the core of the development of complexity in eastern Botswana and the SLCA was societal diversity. In both these areas this included hunter-gatherers and farmers.168 An additional element of diversity in the SLCA stemmed from the different regional affiliations of farmers, which resulted in two distinct but contemporaneous and complementary networks of practice.169 Societal diversity among farming communities characterized the Kambambian B period in the Upemba depression as well as the later Luba state. In contrast with the SLCA, Mutapa, and Torwa-Changamire-Rozvi areas, some of the diversity in the Upemba basin arose during the development of complexity in the region. The social configuration of the Great Zimbabwe society, however, appears to be unlike that in the Greater Shashe-Limpopo region and in the Katanga states. The rise of Great Zimbabwe appears to be linked to a displacement of earlier inhabitants or “cultures,” evidenced in pottery, from the area.170 This might, however, be a reflection of the ideological framework in which past research in the Great Zimbabwe state was framed.

Substantial evidence also exists for continuity in local populations in complex societies irrespective of changes in ruling elites. This is visible in ceramic assemblages from the SLCA,171 Mutapa, and Khami and Danamombe, as well as in human remains from the Upemba depression. In the SLCA and Torwa-Changamire-Rozvi states the assemblages provide substantial evidence for the development of forms of hybridity.172 This societal diversity in the states north and south of the Zambezi is a profound challenge to earlier understandings of state formation and instead speaks to recent understandings rooted in complexity.

Discussion of the Literature

One of the consequences of tribal understandings of the past is the fragmentary interpretation of the development of complexity in the SLCA. Conventional approaches focus on Mapungubwe or, at times, K2 and Mapungubwe.173 Often Schroda, with its 500 residents, is excluded from the narrative. Instead the role of Mapungubwe in the “Zimbabwe culture” is the dominant narrative regarding identity of the SLCA.174 It is likely that a substantial part of the exclusion of Schroda stems from the Schroda Zhizo pottery not forming part of the “Zimbabwe culture” tradition. This ceramic difference is read as a difference in “cultural” affiliation, and since states were perceived as ethnically homogeneous, Schroda is conceptually excluded the Mapungubwe narrative.175 These tribalist approaches are also entangled with historic and modern national agendas, which are in part the result of colonial borders that intersect the Greater Shashe-Limpopo Region. Research in Botswana, South Africa, and Zimbabwe is aligned with national research agendas, such as the African renaissance in South Africa, and emphasis on Shona history in Zimbabwe.176

These cultural entanglements were further perpetuated through research practices. Most past research projects focused on specific sites or areas rather than regional systems. Denbow’s research stands in sharp contrast to this culture and location-biased model. His research explored horizontal complexity beyond cultural or mode of production categories, thereby pioneering the investigation of regional scale complexity in southern Africa.177 Recent research in the SLCA has drawn heavily on Denbow’s earlier insights.178

Research on the Zimbabwe plateau states also has been entangled in identity matters from the start. Initially these focused on race, with the first excavators claiming exotic origins.179 These bizarre claims were soon shown to be false, and it was shown that the sites were occupied by Africans.180 More recently, understandings of culture have obscured possible complexity. In this culture-driven nation-state model, some archaeologists have argued for continuity in the “Zimbabwe culture” for almost a millennium in the vast region encompassing modern Zimbabwe, eastern Botswana, and northern South Africa.181 Consequently, it has been argued that Mapungubwe, Great Zimbabwe, Mutapa, and Torwa-Changamire-Rozvi (and for some the Venda state) share a common ideology and cultural pattern as part of the “Zimbabwe culture.”182 A notable exception to this approach is the analysis of Mutapa ceramic style by Pwiti.183 This is one of the most theoretically informed approaches in Zimbabwe state archaeology. Instead of viewing ceramics as a passive indicator of culture, he linked ceramic style to economic and associated ideological change.

Claims of continuity have been made in spite of very distinct material culture, architectural, and locational differences between these states.184 These differences are important because there is clear evidence for hybridity in the formation of the Mapungubwe185 and Torwa-Changamire-Rozvi states. These local interaction–based processes would have reshaped the “cultures” of these states. Recent reanalysis of radiocarbon-based research also challenged this unitary culture model and suggests that Mapungubwe and Great Zimbabwe were parallel developments.186 Furthermore, the regional and historical contexts affecting these states were fundamentally different.187

Archaeological research in the Upemba basin has explored the full farming community occupation sequence. The excavation of graves formed a key part of the excavation strategy, and over 200 graves have been excavated. There is remarkable continuity in research focus and the recognition of complexity through time.188 The archaeological long-term insights stand in contrast with historical research, which locates the development of the Luba state in more recent historical processes.189 Memory-based work, however, has raised the potential for constructive engagement between archaeology and historical sources.190

Much of 20th-century historical research into the development of Maravi complexity had two entangled concerns: Maravi origins and the location of key capitals. Viewing the origins of Malawian communities and their leaders as the result of in-migration was common. Two models were prominent. In the first Maravi rulers imposed themselves on earlier proto-Chewa communities that migrated to the area from central Africa at the beginning of the Late Iron Age, and in the other they were part of these migrant communities.191 The location of the Maravi leader Muzura’s capital, and by implication possible ethnic or lineage affiliation, was of equal concern.192

Older understandings rooted in ideas of homogeneity of state societies informed research on the identity of the Maravi.193 These concerns regarding affiliation also influenced the identification of the Maravi capitals.194 Similarly, ethnic affiliation has informed research into the origins of the ancestral Malawian ceramics that form part of the Luangwa complex, which develops into local traditions associated with contemporary communities, such as the Chewa.195

Primary Sources

Archaeologists and historians have pursued the development and transformations of complexity north and south of the Zambezi. But while their sources speak to the same places, archaeological and historical sources often inform on very different contexts and scales of analysis. At times this makes it challenging to reconcile the information garnered through the methodologies of the two disciplines.

Understandings in both disciplines are derived from a range of sources, both using oral and archival sources in grappling with the more recent past. Staring in the south, Nicholas J. Van Warmelo played a leading role in the collection of Venda oral traditions. Some of these accounts were published in the Department of Native Affairs Ethnographic Papers. Many more remain unpublished and can be accessed in the South African National Archive in Pretoria.196 To the north, Donald P. Abrahams collected oral histories (ca. 1950 to 1971) about the Mutapa and Rozvi dynasties.197 His published notes have been crucial sources for other historians and archaeologists.198 Much of his work remains unpublished.199 Reefe’s The Rainbow and the Kings on the Luba state is based on his own field work, which made a very extensive collection of oral sources.200 Fieldwork-based collection of new oral accounts has been a key source of writing on the Maravi state.201 Some of the unpublished interviews have been lodged in the National Archive of Malawi.202

The earliest known written documents relating to east and southern Africa are the chronicles of Islamic traveler Al-Mas’udi (843–957 ce) and Buzurg ibn Shahriyar, who compiled a book titled ‘Aja’ib al-Hind (Marvels of India) based on stories collected from sailors between 930 and 947 ce. Translations by Greville S. P. Freeman-Grenville made these texts more widely available.203 The writing of the Portuguese priest João dos Santos, who lived in southeast Africa in the later part of the 16th century, is a key source regarding the history of the Mutapa and other Zambezi states. This text, as well as other key texts, was published in Theal’s Records of South-eastern Africa.204 The earliest written account on the Luba dates to the mid-18th century, when the Portuguese-speaking Manuel Leitao visited the area. Other visitors followed over the next centuries, including English and Swahili traders who wrote about their travels. These sources are discussed in The Rainbow and the Kings.205

Archaeological data provide evidence for the earlier part of the sequence during which complexity developed north and south of the Zambezi and processes where written records do not exist. It can also be used to test the veracity of oral and written historical sources. At the core of archaeological interpretations are the study of the spatial and architectural configuration of archaeological sites, analysis of excavated material culture, and the study of human remains.

In order to allow for possible reanalysis, excavated archaeological material is housed in archaeological storage facilities after excavation. The diverse research focuses in the four regions resulted in dissimilar collections that could be reanalyzed. The Upemba collection, for example, is largely derived from graves, while the bulk of the Greater Shashe-Limpopo collection comprises material resulting from residential contexts. In the regions discussed here archaeological collections are housed in the national museums (Botswana, Democratic Republic of Congo, South Africa, and Zimbabwe), research universities (South Africa), and the Department of Antiquities (Malawi).

While excavation is central to archaeological methodology, published archaeological data are frequently used as primary sources, with site excavation reports being vital. In the first part of the 20th century, excavation reports were generally published in book form,206 but in the second half of the century journals became the favored site for the publication of excavation data. A substantial amount of archaeological site reports from Botswana, South Africa, and Zimbabwe since 1945 were published in the South African Archaeological Bulletin. In Zimbabwe, Zimbabwean Prehistory (Rhodesian Prehistory Until 1979): Journal of the Prehistory Society of Zimbabwe also played a key role in the publication of data. Since 1966, Azania: Journal of the British Institute in East Africa has provided another avenue for the publication of archaeological data.

Further Reading

  • Calabrese, John A. The Emergence of Social and Political Complexity in the Shashi-Limpopo Valley of Southern Africa, AD 900 to 1300. Ethnicity, Class, and Polity. Oxford: Bar International Series, 2007.
  • Childs, S. Terry, and Pierre de Maret. “Re/Constructing Luba Pasts.” In Luba Art and the Making of History. Edited by Mary N. Roberts and Allen F. Roberts, 49–59. New York: Museum of African Art, 1996.
  • Chirikure, Shadreck, Mark Pollard, Munyaradzi Manyanga, and Foreman Bandama. “A Bayesian Chronology for Great Zimbabwe: Re-threading the Sequence of a Vandalised Monument.” Antiquity 87 (2013): 854–872.
  • de Maret, Pierre. “The Power of Symbols and the Symbols of Power Through Time: Probing the Luba Past.” In Beyond Chiefdoms: Pathways to Complexity in Africa, edited by Susan McIntosh, 151–165. Cambridge, U.K.: Cambridge University Press, 1999.
  • de Maret, Pierre. “From Kinship to Kingship: An African Journey into Complexity.” Azania: Archaeological Research in Africa 47 (2012): 314–326.
  • Denbow, James. “Material Culture and the Dialectics of Identity in the Kalahari: AD 700–1700.” In Beyond Chiefdoms: Pathways to Complexity in Africa, edited by Susan McIntosh, 110–123. Cambridge, U.K.: Cambridge University Press, 1999.
  • Huffman, Thomas N. Snakes and Crocodiles: Power and Symbolism in Ancient Zimbabwe. Johannesburg: Witwatersrand University Press, 1996.
  • Huffman, Thomas N. Handbook to the Iron Age: The Archaeology of Pre-Colonial Farming Societies in Southern Africa. Pietermaritzburg: University of KwaZulu Natal Press, 2007.
  • Machiridza, Lesley, H. “Material Culture and Dialectics of Identity and Power: Towards a Historical Archaeology of the Rozvi in South-Western Zimbabwe.” Master’s thesis, University of Pretoria, 2012.
  • Mgomezulu, Gabi G. Y. “Food Production: The Beginnings in the Linthipe/Changoni Area of Dedza District, Malawi.” PhD dissertation, University of California, 1978.
  • Pikirayi, Innocent. “The Archaeological Identity of the Mutapa State.” PhD dissertation, Uppsala University, 1993.
  • Pikirayi, Innocent. The Zimbabwe Culture. Walnut Creek, CA: AltaMira, 2001.
  • Pwiti, Gilbert. “Economic Change, Ideology and the Development of Cultural Complexity in Northern Zimbabwe.” Azania 39 (2004): 265–282.


  • 1. Graham Connah, African Civilisations: An Archaeological Perspective, 2d ed. (Cambridge, U.K.: Cambridge University Press, 1987), 11; see also Timothy Insoll, The Archaeology of Islam in Sub-Saharan Africa (Cambridge, U.K.: Cambridge University Press, 2003), 3–12.

  • 2. Gabi G. Y. Mgomezulu, “Food Production: The Beginnings in the Linthipe/Changoni Area of Dedza District, Malawi,” (PhD dissertation, University of California, 1978); Gabi G. Y. Mgomezulu, “Recent Archaeological Research and Radiocarbon Dates from Eastern Africa,” Journal of African History 22 (1981): 435–456; Innocent Pikirayi, The Archaeological Identity of the Mutapa State (Uppsala, Sweden: Uppsala University, 1993); Pierre de Maret, “The Power of Symbols and the Symbols Of Power Through Time: Probing the Luba Past,” in Beyond Chiefdoms: Pathways to Complexity in Africa, ed. Susan K. McIntosh (Cambridge, U.K.: Cambridge University Press, 1999), 151–165; Pierre de Maret, “From Kinship to Kingship: An African Journey into Complexity,” Azania: Archaeological Research in Africa 47 (2012): 314–326; Yusuf M. Juwayeyi, “Archaeological Excavations at Mankhamba, Malawi: An Early Settlement Site of the Maravi,” Azania 45 (2010): 175–202.

  • 3. For more detailed discussions, see Susan K. McIntosh, ed., Beyond Chiefdoms: Pathways to Complexity in Africa (Cambridge, U.K.: Cambridge University Press, 1999); Susan K. McIntosh, “Pathways to Complexity: An African Perspective,” in Beyond Chiefdoms: Pathways to Complexity in Africa, ed. Susan K. McIntosh (Cambridge, U.K.: Cambridge University Press, 1999), 1–30; see also de Maret, “Power”; de Maret, “Kinship”; James Denbow, “Material Culture and the Dialectics of Identity in the Kalahari: AD 700–1700,” in Beyond Chiefdoms: Pathways to Complexity in Africa, ed. Susan McIntosh (Cambridge, U.K.: Cambridge University Press, 1999), 110–123; Roderick J. McIntosh, “Alternative Polities,” Emerging Trends in the Social and Behavioral Sciences: An Interdisciplinary, Searchable, and Linkable Resource, Wiley Online Library.

  • 4. John A. Calabrese, “Interregional Interaction in Southern Africa: Zhizo and Leopard’s Kopje Relations in Northern South Africa, South-Western Zimbabwe, and Eastern Botswana, AD 1000 to 1200,” African Archaeological Review, 17 (2000): 183–210; John A. Calabrese, The Emergence of Social and Political Complexity in the Shashi-Limpopo Valley of Southern Africa, AD 900 to 1300 Ethnicity, Class, and Polity (Oxford: Bar International Series, 2007); Denbow, “Material Culture”; de Maret, “Power”; de Maret, “Kinship”; Lesley H. Machiridza, “Material Culture and Dialectics of Identity and Power: Towards a Historical Archaeology of the Rozvi in South-Western Zimbabwe” (MA thesis, University of Pretoria, 2012); Gilbert Pwiti, “Economic Change, Ideology and the Development of Cultural Complexity in Northern Zimbabwe,” Azania 39 (2004): 265–282; Bronwen Van Doornum, “Tshisuku Shelter and the Shashe-Limpopo Confluence Area Hunter-Gatherer Sequence,” Southern African Humanities 19 (2007): 17–67; Bronwen Van Doornum, “Sheltered from Change: Hunter-Gatherer Occupation of Balerno Main Shelter, Shashe-Limpopo Confluence Area, South Africa,” Southern African Humanities 20 (2008): 249–284; Bronwen Van Doornum, “Balerno Shelter 3: A Later Stone Age Site in the Shashe-Limpopo Confluence Area, South Africa,” Southern African Humanities 26 (2014): 129–155.

  • 5. For ramifications of these approaches in the present, see Archie Mafeje, “Ideology of Tribalism,” Journal of Modern African Studies 4 (1971): 253–261; Adam Kuper, The Invention of Primitive Society: Transformations of an Illusion (London: Routledge, 1988); Adam Kuper, Culture: The Anthropologists’ Account (Cambridge, MA: Harvard University Press, 1999); see also Martin Hall, “Tribes, Traditions and Numbers: The American Model in Southern African Iron Age Ceramic Studies,” The South African Archaeological Bulletin 38 (1983): 51–57; Mahmood Mamdani, Citizens and Subjects: Contemporary Africa and the Legacy of Late Colonialism (Princeton, NJ: Princeton University Press, 1996).

  • 6. For critiques, see Edwin N. Wilmsen and James R. Denbow, “Paradigmatic History of San-Speaking Peoples and Current Attempts at Revision,” Current Anthropology 31 (1990): 489–524; Denbow, “Material Culture”; Hall, “Tribes.”

  • 7. For a more comprehensive discussion, see Eric Hobsbawm and Terence Ranger, eds., The Invention of Tradition (Cambridge, U.K.: Cambridge University Press, 1983).

  • 8. David N. Beach, “Cognitive Archaeology and Imaginary History at Great Zimbabwe,” Current Anthropology 39 (1998): 47–72.

  • 9. See, for example, Thomas N. Huffman, Snakes and Crocodiles: Power and Symbolism in Ancient Zimbabwe (Johannesburg: Witwatersrand University Press, 1996); Thomas N. Huffman, Handbook to the Iron Age: The Archaeology of Pre-Colonial Farming Societies in Southern Africa (Pietermaritzburg: University of KwaZulu Natal Press, 2007).

  • 10. See, for example, Timothy M. Evers, “The Recognition of Groups in the Iron Age of Southern Africa” (PhD thesis, University of the Witwatersrand, 1988); Thomas N. Huffman, “Ceramics, Settlement and Late Iron Age Migrations,” The African Archaeological Review 7(1989): 155–182; Thomas N. Huffman, Iron Age Migration: The Ceramic Sequence in Southern Zambia (Johannesburg: Witwatersrand University Press, 1989); Huffman, Handbook; for critiques of this approach, see, for example, Hall, “Tribes”; Pwiti “Economic”; Stephanie Wynne-Jones and Sarah K. Croucher, “Theorizing Identity in African Archaeology,” Journal of Social Archaeology 7 (2007): 283–285.

  • 11. Suzanne L. Eckert, Kari L. Schleher, and William D. James, “Communities of Identity, Communities of Practice: Understanding, Santa Fe Black-on-White Pottery in the Espanola Basin of New Mexico,” Journal of Archaeological Science 63 (2015): 1–12; Oliver Gosselain, “Thoughts and Adjustments in the Potter’s Backyard,” in Breaking the Mould: Challenging the Past Through Pottery, ed. Ina Berg (Oxford: BAR, 2008), 67–79; Oliver J. T. Harris, “(Re)assembling Communities,” Journal of Archaeological Method and Theory 21(2014): 76–97.

  • 12. For example, McIntosh, “Alternative”; Ann B. Stahl, “Archaeological Insights into Aesthetic Communities of Practice in the Western Volta Basin,” African Arts Autumn 46 (2013): 54–67.

  • 13. See, for example, Timothy R. Pauketat, Chiefdoms and Other Archaeological Delusions (Lanham, MD: AltaMira, 2007); Timothy R. Pauketat, “The Grounds for Agency in Southwestern Archaeology,” in The Social Construction of Communities: Agency, Structure, and Identity in the Prehispanic Southwest, ed. Mark D. Varien and James M. Potter (Walnut Creek, CA: AltaMira, 2008), 233–249; Steven A. Wernke, “Negotiating Community and Landscape in the Peruvian Andes: A Transconquest View,” American Anthropologist 109 (2007): 130–152.

  • 14. James Denbow, “The Toutswe Tradition: A Study in Socioeconomic Change,” in Settlement in Botswana, ed. Renee Hitchcock and Mary Smith (Johannesburg: Heinemann, 1982), 73–86; James Denbow, “Cows and Kings: A Spatial and Economic Analysis of a Hierarchical Early Iron Age Settlement System in Eastern Botswana,” in Frontiers: Southern African Archaeology Today, ed. Martin Hall, Graham Avery, Mary Avery, Monica Wilson, and Tony Humphreys (Oxford: BAR, 1984), 24–39; James Denbow, “A New Look at the Later Prehistory of the Kalahari,” The Journal of African History 27 (1986), 3–28; James Denbow, “Congo to Kalahari: Data and Hypotheses About the Political Economy of the Western Stream of the Early Iron Age,” African Archaeological Review 8 (1990):139–176.

  • 15. Denbow, “Toutswe”; Denbow, “Cows”; Denbow, “New Look”; Denbow, “Congo.”

  • 16. Denbow, “Toutswe”; Denbow, “Cows”; Denbow, “New Look”; Denbow, “Congo”; Wilmsen and Denbow, “Paradigmatic”; see also Tim Forssman, “A Preliminary Report on Fieldwork in the Northern Tuli Game Reserve, Northeastern Botswana,” South African Archaeological Bulletin 68 (2013): 63; Tim Forssman, “Dzombo Shelter: A Contribution to the Later Stone Age Sequence of the Greater Mapungubwe Landscape,” South African Archaeological Bulletin 69 (2013): 182.

  • 17. Edwin Hanisch, “An Archaeological Interpretation of Certain Iron Age Sites in the Limpopo/Shashe Valley” (MA thesis, University of Pretoria, 1980); Edwin Hanisch, “Schroda: A Zhizo Site in the Northern Transvaal,” in Archaeological Sites in the Northern and Eastern Transvaal, ed. Elizabeth Voight (Pretoria: Transvaal Museum, 1981), 37–53; Edwin Hanisch, “Schroda: The Archaeological Evidence,” in Sculptured in Clay: Iron Age Figurines from Schroda, Limpopo Province, South Africa, ed. Johnny Van Schalkwyk and Edwin Hanisch (Pretoria: National Cultural History Museum, 2002), 21–39; Thomas N. Huffman, “Mapungubwe and the Origins of the Zimbabwe Culture,” South African Archaeological Society Goodwin Series 10 (2000): 14–29.

  • 18. Calabrese, Emergence; Huffman, “Mapungubwe”; Van Doornum, “Tshisuku”; Van Doornum, “Balerno”; Van Doornum, “Sheltered.”

  • 19. Huffman, “Mapungubwe”; Andrie Meyer, The Archaeological Sites of Greefswald: Stratigraphy and Chronology of the Sites and a History of Investigations (Pretoria: University of Pretoria, 1998).

  • 20. Calabrese, Emergence; Huffman, “Mapungubwe.”

  • 21. Huffman, “Mapungubwe”; Meyer, Archaeological.

  • 22. Huffman, “Mapungubwe”; Thomas N. Huffman, “Social Complexity in Southern Africa,” African Archaeological Review 32 (2015): 71–91.

  • 23. Meyer, Archaeological, 300.

  • 24. Shadreck Chirikure, Munyaradzi Manyanga, A. Mark Pollard, Foreman Bandama, Godfrey Mahachi, and Innocent Pikirayi, “Zimbabwe Culture Before Mapungubwe: New Evidence from Mapela Hill, South-Western Zimbabwe,” PLoS ONE 9(2014): e111224, doi:10.1371/journal.pone.0111224; for a critique of the approach to dating used by the Mapela team, see Thomas N. Huffman, “Mapela, Mapungubwe and the Origins of States in Southern Africa,” The South African Archaeological Bulletin 70 (2015): 15–27.

  • 25. Calabrese, Emergence; Huffman, “Mapungubwe”; Huffman, Snakes.

  • 26. Marilee Wood, “Making Connections: Relationships Between International Trade and Glass Beads from the Shashe-Limpopo Area,” South African Archaeological Society Goodwin Series 10 (2000): 78–90.

  • 27. See, for example, Thomas N. Huffman, “Ceramics, Classification and Iron Age Entities,” African Studies 39 (1980): 123–174; John Schofield, “Work Done in 1934: Pottery,” in Mapungubwe: Ancient Bantu Civilization on the Limpopo, ed. Leo Fouché (Cambridge, U.K.: Cambridge University Press, 1937), 32–61.

  • 28. For more detailed discussions of Zhizo ceramics, see Calabrese, “Interregional”; Calabrese, Emergence; Denbow, “Toutswe.”

  • 29. Calabrese, “Interregional”; Calabrese, Emergence; Huffman, “Mapungubwe”; Meyer, Archaeological.

  • 30. For a more detailed discussion of the distribution of Zhizo ceramics through time, see Calabrese, Emergence.

  • 31. Simon Hall and Benjamin W. Smith, “Empowering Places: Rock Shelters and Ritual Control in Farmer-Forager Interactions in the Northern Province,” The South African Goodwin Series 8 (2000): 30–46; Wilmsen and Denbow, “Paradigmatic”; Van Doornum, “Tshisuku”; Van Doornum, “Balerno”; Van Doornum, “Sheltered.”

  • 32. Denbow, “Toutswe”; Denbow, “Cows”; Denbow, “New Look”; Denbow, “Congo”; Calabrese, Emergence.

  • 33. Denbow, “Toutswe”; Denbow, “Cows”; Denbow, “New Look”; Denbow, “Congo.”

  • 34. Calabrese, Emergence; Huffman, “Mapungubwe”; Thomas N. Huffman, “Iron Age Settlement Patterns and the Origins of Class Distinction in Southern Africa,” in Advances in World Archaeology, ed. Fred Wendorf and Angela E. Close (New York: Academic Press, 1986), 291–338.

  • 35. Jeanette, M. Smith, “Climate Change and Agropastoral Sustainability in the Shashe/Limpopo River Basin from 900 AD” (PhD dissertation, University of the Witwatersrand, 2005); Jeannette Smith, Julia Lee-Thorp, and Simon Hall, “Climate Change and Agropastoralist Settlements in the Shashe-Limpopo River Basin, Southern Africa, AD 880 to 1700,” South African Archaeological Bulletin 62 (2007): 115–125; Jeanette Smith, Julia Lee-Thorp, Steve Prevec, Simon Hall,and Andreas Späth, “Pre-colonial Herding Strategies in the Shashe-Limpopo Basin, Southern Africa, Based on Strontium Isotope Analysis of Domestic Fauna,” Journal of African Archaeology 8 (2010): 83–98.

  • 36. Cited in Calabrese, Emergence.

  • 37. Wood, “Making”; Marilee Wood, “A Glass Bead Sequence for Southern Africa from the 8th to the 16th Century AD,” Journal of African Archaeology 9 (2011): 67–84; Marilee Wood, Interconnections: Glass Beads and Trade in Southern and Eastern Africa and the Indian Ocean—7th to 16th Centuries AD (Uppsala, Sweden: Uppsala University, 2011).

  • 38. Wood, Interconnections, 41.

  • 39. Huffman, “Mapungubwe.”

  • 40. Denbow, “Toutswe”; Denbow, “Cows”; Denbow, “New Look”; Denbow, “Congo.”

  • 41. Denbow, “New Look”; Huffman, “Mapungubwe.”

  • 42. Anneli Ekblom, Changing Landscapes: An Environmental History of Chibuene, Southern Mozambique. Global Studies in Archaeology 5 (Uppsala, Sweden: Uppsala University, 2004); Paul J. J. Sinclair, “Chibuene—An Early Trading Site in Southern Mozambique,” Paideuma 28 (1982): 150–64; Paul J.J. Sinclair, Space, Time and Social Formation: A Territorial Approach to the Archaeology and Anthropology of Zimbabwe and Mozambique c 0–1700 AD (Uppsala, Sweden: Societas Archaeologica Upsaliensis, 1987); Paul J. J. Sinclair, Joåo M. F. Morais, Leonardo Adamowicz, and Ricardo T. Duarte, “A Perspective on Archaeological Research in Mozambique,” in The Archaeology of Africa; Food, Metals and Towns, ed. Thurstan Shaw, Paul Sinclair, Bassey Andah, and Alex Okpoko (New York: Routledge, 1993), 409–431

  • 43. Wood, “Making”; Wood, Interconnections; Denbow, “New Look”; Huffman, “Mapungubwe.”

  • 44. Wood, Interconnections, 30.

  • 45. Munyaradzi Manyanga, Innocent Pikirayi, and Webber, Ndoro, “Coping with Dryland Environments: Preliminary Results from Mapungubwe and Zimbabwe Phase Sites in the Mateke Hills, South-Eastern Zimbabwe,” Goodwin Series 8 (2000): 69–77; see also Hanisch “Archaeological”; Hansich, “Schroda”; Huffman, “Mapungubwe”; Tim Forssman, Bruce Page, and Jeanetta Selier, “How Important Was the Presence of Elephants as a Determinant of the Zhizo Settlement of the Greater Mapungubwe Landscape?” Journal of African Archaeology 12 (2014): 75–87.

  • 46. Manyanga et al., “Coping with Dryland Environments.”

  • 47. Huffman, Snakes; Innocent Pikirayi, The Zimbabwe Culture (Walnut Creek, CA: AltaMira, 2001).

  • 48. Huffman, Snakes; Pikirayi, Zimbabwe.

  • 49. Beach “Cognitive.”

  • 50. Peter S. Garlake, “Prehistory and Ideology in Zimbabwe,” Journal of the International African Institute: Past and Present in Zimbabwe 52(1982): 1–19.

  • 51. Peter S. Garlake, “Pastoralism and Zimbabwe,” The Journal of African History, 19 (1978): 479–493.

  • 52. For more details on the dating challenges facing archaeologists, see Roger F. H. Summers, Keith, R. Robinson and Anthony Whitty, “Zimbabwe excavations 1958,” Occasional Papers of the National Museums of Southern Rhodesia 3 (1961): 157–332; ; Thomas N. Huffman and John C. Vogel, “The Chronology of Great Zimbabwe,” South African Archaeological Bulletin 46 (1991): 61–70; Shadreck Chirikure, Mark Pollard, Munyaradzi Manyanga, and Foreman Bandama, “A Bayesian Chronology for Great Zimbabwe: Re-threading the Sequence of a Vandalised Monument,” Antiquity 87 (2013): 854–872; Shadreck Chirikure, Munyaradzi Manyanga, Innocent Pikirayi, and Mark Pollard, “New Pathways of Sociopolitical Complexity in Southern Africa,” African Archaeological Review 30 (2013): 339–366.

  • 53. For example, Chirikure et al., “Bayesian”; Huffman and Vogel, “Chronology”; Thomas N. Huffman, “Mapungubwe and Great Zimbabwe: The Origin and Spread of Social Complexity in Southern Africa,” Journal of Anthropological Archaeology 28 (2009): 37–54; Keith Robinson, “Zimbabwe Pottery,” Occasional Papers of the National Museums of Southern Rhodesia 3 (1961): 193–226:; Roger Summers, “Excavations in the Great Enclosure,” Occasional Papers of the National Museums of Southern Rhodesia 3 (1961): 236–288.

  • 54. Huffman and Vogel, “Chronology.”

  • 55. Huffman and Vogel, “Chronology.”

  • 56. Chirikure et al., “Bayesian.”

  • 57. Chirikure et al. “Bayesian”; David P. Collett, Alex E. Vines, and E. Gwilym Hughes, “The Chronology of the Valley Enclosures: Implications for the Interpretation of Great Zimbabwe,” The African Archaeological Review 10 (1992): 139–161; Huffman and Vogel, “Chronology.”

  • 58. Collett et al., “Chronology”; Chirikure et al., “Bayesian.”

  • 59. Huffman, Snakes; Innocent Pikirayi and Shadreck Chirikure, “Debating Great Zimbabwe,” Azania: Archaeological Research in Africa 46 (2011): 221–231.

  • 60. Huffman, “Iron Age”; Pikirayi, Zimbabwe.

  • 61. Innocent Pikirayi, “Stone Architecture and the Development of Power in the Zimbabwe Tradition AD 1270–1830,” Azania 48 (2013): 282–300.

  • 62. Gilbert Pwiti, “Southern Africa and the East African Coast,” in African Archaeology: A Critical Introduction, ed. Anne Stahl (Oxford: Blackwell Publishing, 2005), 378–391.

  • 63. Pwiti, “African.”

  • 64. Manyanga et al., “Coping with Dryland Environments”; Carolyn Thorp, Kings, Commoners and Cattle at Zimbabwe Tradition Sites. Museum Memoir Series No. 1. (Harare: National Museums and Monuments of Zimbabwe, 1995).

  • 65. Garlake, “Pastoralism.”

  • 66. Chirikure et al,. “Bayesian”; Chirikure et al., “New Pathways”; Huffman, Snakes; Pikirayi, Zimbabwe.

  • 67. Huffman, Handbook.

  • 68. Huffman and Vogel, “Chronology”; Pikirayi and Chirikure, “Debating”; Summers, “Excavations.”

  • 69. Pikirayi, Zimbabwe, 124.

  • 70. For more information on the competing polities model, see, for example, Denbow, “New Look”; Catrien Van Waarden, “The Late Iron Age,” in Ditswa Mmung: The Archaeology of Botswana, ed. Paul Lane, Andrew Reid, and Alinah Segobye (Gaborone: Botswana Society, 1998), 115–160; for the successor model see, for example, Huffman, Snakes.

  • 71. Chirikure et al., “Bayesian”; Chirikure et al., “New Pathways.”

  • 72. Pwiti, “African,” 386.

  • 73. Gertrude Caton-Thompson, The Zimbabwe Culture: Ruins and Reactions (Oxford: Clarendon Press, 1931); Collett et al., “Chronology”; Wood, Interconnections.

  • 74. Chirikure et al., “Bayesian.”

  • 75. Chirikure et al., “Bayesian”; Collett et al., “Chronology”; Wood, Interconnections.

  • 76. Pikirayi, Zimbabwe, 241; Pwiti, “African”; Foreman Bandama, Abigail J. Moffett, Thomas P. Thondhlana, and Shadreck Chirikure, “The Production, Distribution and Consumption of Metals and Alloys at Great Zimbabwe,” Archaeometry 58 (2016): 164–181.

  • 77. Pikirayi, Zimbabwe.

  • 78. Pikirayi, Zimbabwe, 242; Lorraine M. Swan, Minerals and Managers: Production Contexts as Evidence for Social Organization in Zimbabwean Prehistory (Uppsala, Sweden: Uppsala University, 2008).

  • 79. Garlake, “Pastoralism.”

  • 80. Chirikure et al., “Bayesian”; Pikirayi, Zimbabwe; Pikirayi and Chirikure, “Debating.”

  • 81. Pwiti, “Economic.”

  • 82. Pikirayi, Archaeological; Pwiti, “Economic.”

  • 83. Pwiti, “Economic.”

  • 84. David N. Beach, “The Mutapa Dynasty: A Comparison of Documentary and Traditional Evidence,” History in Africa, 3 (1976): 1–17.

  • 85. Pikirayi, Archaeological; also see Pikirayi and Chirikure, “Debating.”

  • 86. Donald P. Abraham, “Ethnohistory of the Empire of Mutapa: Problems and Methods,” in The Historian in Tropical Africa, ed. Jan Vansina, R. Mauny, and V. Thomas (London: Oxford University Press, 1964), 104–211; Stanislaus I. G. Mudenge, “The Role of Foreign Trade in the Rozvi Empire: A Reappraisal,” The Journal of African History 15 (1974): 373–391.

  • 87. Pikirayi, Zimbabwe; Pikirayi, Archaeological; Pwiti, “Economic”; Newitt, “Portuguese”; Newitt, Portuguese.

  • 88. Pwiti, “Economic.”

  • 89. Pikirayi, Archaeological; Pwiti, “Economic.”

  • 90. Pwiti, “Economic.”

  • 91. Pikirayi, Archaeological.

  • 92. Pikirayi, Zimbabwe.

  • 93. Malyn D.D. Newitt, “The Portuguese on the Zambezi: An Historical Interpretation of the Prazo system,” The Journal of African History, 10(1969): 67–85; Malyn D. D. Newitt, Portuguese Settlement on the Zambesi: Exploration, Land Tenure and Colonial Rule in East Africa (New York: Africana, 1973).

  • 94. Roger Summers, Inyanga: Prehistoric Settlements in Southern Rhodesia (Cambridge, U.K.: Cambridge University Press, 1958); Robert Soper, “The Nyanga Terrace Complex of Eastern Zimbabwe: New Investigations,” Azania 31 (1996): 1–35; Robert Soper, Nyanga: Ancient Fields, Settlements and Agricultural History in Zimbabwe (Nairobi. British Institute in Eastern Africa, 2002); Robert Soper, The Terrace Builders of Nyanga (Harare, Zimbabwe: Weaver Press, 2006); Munyaradzi Manyanga and Plan Shenjere, “The Archaeology of the Northern Nyanga Lowlands and the Unfolding Farming Community Sequence in Northeastern Zimbabwe,” The South African Archaeological Bulletin 67 (2012): 244–255.

  • 95. Soper, “Nyanga terrace”; Manyanga and Shenjere, “Archaeology.”

  • 96. Ivan M. Murambiwa, “Survey of Archaeological Sites Within a 50km Radius of Khami,” in Urban Origins in Eastern Africa: Proceedings of the 1990 Workshop, Harare and Great Zimbabwe, ed. Paul J. J. Sinclair and Pwiti Gilbert (Stockholm: The Central Board of National Antiquities, 1990), 93–94.

  • 97. David N. Beach, The Shona and Zimbabwe 1900–1850 (Gweru, Zimbabwe: Mambo Press, 1980); G. C. Mazarire, “Changing Landscape and Oral Memory in South-Central Zimbabwe: Towards a Historical Geography of Chishanga, c.1850–1990,” Journal of Southern African Studies 29 (2003): 701–715; Keith R. Robinson, Khami Ruins: A Report on Excavations in Southern Rhodesia (Cambridge, U.K.: Cambridge University Press, 1959); Keith R. Robinson, “The Archaeology of the Rozwi,” in The Zambezian Past, ed. Eric Stokes and Richard Brown (Manchester, U.K.: Manchester University Press, 1966), 3–27.

  • 98. Keith R. Robinson, Khami Ruins (Cambridge, U.K.: Cambridge University Press, 1959).

  • 99. Huffman, Snakes.

  • 100. Pikirayi and Chirikure, “Debating.”

  • 101. Robinson, Khami; Robinson, “Archaeology.”

  • 102. Machiridza, “Material Culture.”

  • 103. Pikirayi and Chirikure, “Debating.”

  • 104. Pikirayi and Chirikure, “Debating.”

  • 105. Stanislaus I. G. Mudenge, “Eighteenth-Century Portuguese Settlements on the Zambezi and the Dating of Rhodesian Ruins: Some Reflections on the Problems of Reference Dating,” The International Journal of African Historical Studies, 10 (1977): 384–393; Stanislaus I. G. Mudenge, A Political History of Munhumutapa c 1400–1902 (Harare: Zimbabwe Publishing House, 1988).

  • 106. Johannes H. N. Loubser, “Archaeology and Early Venda History,” Goodwin Series 6 (1989): 54–56; Johannes H. N. Loubser, The Ethnoarchaeology of Venda-Speakers in Southern Africa (Bloemfontein: Nasionale Museum, 1991); Beach, Shona; see also accounts in Nicholas J. Van Warmelo, ed., The Copper Miners of Musina and the Early History of the Soutpansberg. Ethnological Publications (Pretoria: Government Printer), 8: 51–70.

  • 107. Loubser, “Archaeology”; Loubser, Ethnoarchaeology.

  • 108. Loubser, “Archaeology”; Loubser, Ethnoarchaeology.

  • 109. S. Terry Childs and Pierre de Maret, “Re/Constructing Luba Pasts,” in Luba Art and the Making of History, ed. Mary N. Roberts and Allen F. Roberts (New York: Museum of African Art, 1996), 49–59; Pierre de Maret, “Sanga: New Excavations, More Data, and Some Related Problems,” Journal of African History 28 (1977): 321–337; Pierre de Maret, “Luba Roots: The First Complete Iron Age Sequence in Zaire,” Current Anthropology 20 (1979): 233–235; Pierre de Maret, “The Iron Age in the West and the South,” in The Archaeology of Central Africa, ed. Francis L. Van Noten (Graz, Austria: Akademische Druck, U. Verlagsanstalt, 1982), 77–95; Pierre de Maret, “Recent Archaeological Research and Dates from Central Africa,” Journal of African History 26 (1985): 129–148; Pierre de Maret, “The Ngovo Group: An Industry with Polished Stone Tools and Pottery in Lower Zaire,” African Archaeological Review 4 (1986): 103–133; de Maret, “Power”; de Maret, “Kinship.”

  • 110. Childs and de Maret, “Re/Constructing”; de Maret, “Luba”; de Maret, “Recent”; de Maret, “Power”; de Maret, “Kinship.”

  • 111. Childs and de Maret, “Re/Constructing”; de Maret, “Luba”; de Maret, “Recent”; de Maret, “Power”; de Maret, “Kinship.”

  • 112. S. Terry Childs, “Transformations: Iron and Copper Production in Central Africa,” in Recent Trends in Archeo- Metallurgical Research, ed. Petar D. Glumac (Philadelphia: MASCA, 1991), 33–46.; de Maret, “Luba Roots”; de Maret, “Power.”

  • 113. de Maret,“Iron Age.”

  • 114. Maryna Steyn and Maciej Hennenberg, “The Health Status of the People from the Iron Age Sites at K2 and Mapungubwe (South Africa),” Rivista di Antropologia 73 (1995): 133–143.

  • 115. de Maret, “Iron Age”; de Maret, “Power.”

  • 116. D. D. Phiri, History of Malawi: From Earliest Times to the Year 1915 (Blantyre: Christian Literature Association in Malawi, 2004).

  • 117. de Maret, “Power”; Thomas Q. Reefe, The Rainbow and the Kings: A History of the Luba Empire to 1891 (Berkeley: University of California, 1981).

  • 118. Childs and de Maret, “Re/Constructing.”

  • 119. Nonhlanhla Dlamini, “The Early Inhabitants of the Upemba Depression, The Democratic Republic of Congo: A Biological Review of the Cultural Continuity” (PhD thesis, University of Cape Town, 2014); Mary N. Roberts and Allen F. Roberts, eds., Memory: Luba Art and the Making of History (New York: The Museum of African Arts, 1996).

  • 120. de Maret, “Power”; Roberts and Roberts, Memory.

  • 121. Reefe, Rainbow; Miles Larmer, “At the Crossroads: Mining and Political Change on the Katangese-Zambian Copperbelt,” Oxford Handbooks Online:

  • 122. de Maret, “Luba Roots.”

  • 123. de Maret, “Power.”

  • 124. de Maret, “Power”; Childs and de Maret, “Re/Constructing.”

  • 125. Reefe, Rainbow.

  • 126. de Maret, “Luba Roots.”

  • 127. See for example Garlake,“Prehistory”; Huffman, “Mapungubwe.”

  • 128. Jane Guyer and S. E. Belinga, “Wealth in People as Wealth in Knowledge: Accumulation and Composition in Equatorial Africa,” Journal of African History 36 (1995): 91–120.

  • 129. de Maret, “Kinship,” 315.

  • 130. de Maret, “Recent”; de Maret, “Kinship”; also see de Maret, “Power.”

  • 131. de Maret, “Power.”

  • 132. de Maret, “Power”; Childs and de Maret, “Re/Constructing.”

  • 133. de Maret, “Power.”

  • 134. de Maret, “Iron Age.”

  • 135. de Maret, “Luba Roots.”

  • 136. Michael Bisson, “Copper Currency in Central Africa: The Archaeological Record,” World Archaeology 6.3 (1975): 276–292; Lorraine M. Swan, Minerals and Managers: Production Contexts as Evidence for Social Organization in Zimbabwean Prehistory (Uppsala, Sweden: Uppsala University, 2008); Childs and de Maret, “Re/Constructing.”

  • 137. Reefe, Rainbow, 97.

  • 138. Reefe, Rainbow; Pierre De Maret and Alexandre Livingstone Smith, “Who’s Who? The Case of the Luba 7,” in Ethnic Ambiguity and the African Past: Materiality, History, and the Shaping of Cultural Identities, ed. François G Richard and Kevin C MacDonald (London: Routledge, 2016), 192–216.

  • 139. Reefe, Rainbow.

  • 140. Reefe, Rainbow, 6.

  • 141. Reefe, Rainbow.

  • 142. Reefe, Rainbow.

  • 143. K. M. Phiri, “Northern Zambezia,” The Society of Malawi Journal 32 (1979): 6–22; J. Matthew Schoffeleers, “The Zimba and the Lundu State in the Late Sixteenth and Early Seventeenth Centuries,” The Journal of African History, 28 (1987): 337–355.

  • 144. Phiri, “Northern.”

  • 145. Roberts and Roberts, Memory, 126; Leslie F. Zubieta, “The Rock Art of Chinamwali: Material Culture and Girls’ Initiation in South-Central Africa” (PhD dissertation, University of the Witwatersrand, 2009); for a more comprehensive discussion of this painting tradition, see Benjamin W. Smith, “Rock Art in South-Central Africa: A Study Based on the Pictographs of Dedza District, Malawi and Kasama District Zambia” (PhD dissertation, University of Cambridge, 1995); Benjamin W. Smith, Zambia’s Ancient Rock Art: the Paintings of Kasama (Livingstone: The National Heritage Conservation Commission of Zambia, 1997); Benjamin W. Smith, “The Tale of the Chameleon and the Platypus: Limited and Likely Choices in Making Pictures,” in The Archaeology of Rock Art, ed. Chris Chippindale and Paul Taçon (Cambridge, U.K.: Cambridge University Press, 1998), 213–228; Benjamin W. Smith, “Forbidden Images: Rock Paintings and the Nyau Secret Society of Central Malawi and Eastern Zambia,” African Archaeological Review 18 (2001): 187–211.

  • 146. Mgomezulu, “Food”; Mgomezulu, “Recent.”

  • 147. For example, Mgomezulu, “Food”; Zubieta, “The Rock Art.”

  • 148. Mgomezulu, “Food”; Mgomezulu, “Recent.”

  • 149. Yusuf M. Juwayeyi, “The Later Prehistory of Southern Malawi: A Contribution to the Study of Technology and Economy During the Later Stone Age and Iron Age Periods” (PhD dissertation, University of California, 1981); Yusuf M. Juwayeyi, “The Distribution of Longwe Pottery Sites in Southern Malawi,” Nyame Akuma 27 (1986): 23–25; Yusuf M. Juwayeyi, “Iron Age Settlement and Subsistence Patterns in Southern Malawi,” in The Archaeology of Africa: Food, Metals and Towns, ed. Thurstan Shaw, Paul Sinclair, Bassey Andah, and Alex Okpoko (London: Routledge, 1993), 391–398; Yusuf M. Juwayeyi, “Wealth and Influence in Southern Malawi During the Proto-Historic Period: The Evidence from Archaeology, Oral Traditions and History,” The South African Bulletin 68 (2008): 102–115.

  • 150. Mgomezulu, “Food,” 216; Keith R. Robinson, Iron Age Occupation, North and East of the Mulanje Plateau, Malawi. Department of Antiquities Publication 17 (Zomba: Malawi Government Printers, 1977).

  • 151. Mngomozulu, “Food,” 330.

  • 152. Alpers, Ivory, 49.

  • 153. W. H. J. Rangley cited in Zimba Schofeleers and M. D. D. Newitt, “The Early History of the Maravi,” Journal of African History 23 (1982): 145–162.

  • 154. Shoffeleers, “Early”; J. Matthew Schoffeleers, “Towards the Identification of a Proto-Chewa Culture: A Preliminary Contribution,” Journal of Social Science 2 (1973): 47–60; J. M. Schoffeleers, River of Blood: The Genesis of a Martyr Cult in Southern Malawi, c. A.D. 1600 (Madison: The University of Wisconsin, 1992).

  • 155. Juwayeyi, “Archaeological Excavations.”

  • 156. Brian Morris, “The Ivory Trade and Chiefdoms in Pre-colonial Malawi,” The Society of Malawi Journal 592 (2006): 6–23; Juwayeyi, “Archaeological Excavations.”

  • 157. Zubieta, “Rock Art,” 73.

  • 158. Mgomezulu, “Food,” 217; Keith R. Robinson, The Iron Age of Upper and Lower Shire, Malawi. Department of Antiquities Publication 13 (Zomba: Malawi Government Printers, 1973).

  • 159. Henry W. Langworthy III, “A History of the Undi’s Kingdom to 1890: Aspects of Chewa History in East Central Africa” (PhD dissertation, Boston University, 1969).

  • 160. Edward A. Alpers, Ivory and Slaves in East Central Africa (Berkeley: University of California Press, 1975); Langworthy, “History”; Edward A. Alpers, “The Mutapa and Malawi Political Systems,” in Aspects of Central African History, ed. Terence O. Ranger (London: Heinemann, 1968), 17–26.

  • 161. Langworthy, “History.”

  • 162. Alpers, “Ivory”; Alpers, “Mutapa.”

  • 163. Langworthy, “History,” 263–265, 271.

  • 164. Langworthy, “History,” 169.

  • 165. Phiri, “Northern.”

  • 166. Mgomezulu, “Food.”

  • 167. Beach, Shona.

  • 168. Denbow, “New Look”; Wilmsen and Denbow, “Paradigmatic”; Van Doornum, “Tshisuku”; Van Doornum, “Balerno”; Van Doornum, “Sheltered.”

  • 169. Calabrese, “Interregional”; Calabrese, Emergence; Huffman, “Mapungubwe.”

  • 170. Manyanga et al., “Coping with Dryland Environments”; Chirikure et al., “Bayesian”; Pikirayi, Zimbabwe.

  • 171. Calabrese, “Interregional”; Calabrese, Emergence; Pwiti, “Economic”; Machiridza, “Material Culture”; Dlamini, “Early.”

  • 172. Calabrese, “Interregional”; Calabrese, Emergence; Machiridza, “Material Culture.”

  • 173. Chirikure et al., “Bayesian”; Chirikure et al.,“New Pathways”; Pikirayi and Chirikure, “Debating.”

  • 174. Huffman, Snakes; Chirikure et al., “Bayesian”; Chirikure et al., “New Pathways.”

  • 175. John Schofield, “Work Done in 1934: Pottery,” in Mapungubwe: Ancient Bantu Civilization on the Limpopo, ed. Leo Fouché (Cambridge, U.K.: Cambridge University Press, 1937), 32–61; Huffman, Handbook.

  • 176. See, for example, Mary Leslie and Tim Maggs, eds., African Naissance: The Limpopo Valley 1000 Years Ago. South African Archaeological Society Goodwin Series 8 (2000).

  • 177. Denbow, “Toutswe”; Denbow, “Cows”; Denbow, “New Look”; Denbow, “Congo.”

  • 178. Van Doornum, “Tshisuku”; Van Doornum, “Balerno”; Van Doornum, “Sheltered.”

  • 179. For example, Richard N. Hall and W. G. Neal, The Ancient Ruins of Rhodesia, 2d ed. (London: Methuen, 1904); Karl Mauch, E. Bernhard, F. O. Bernhard, and E. E. Burke, The Journals of Carl Mauch: His Travels in the Transvaal and Rhodesia, 1869–1872 (Salisbury: National Archives of Rhodesia, 1969); Edward A. Maund, “On Matabele and Mashona Lands,” Proceedings of the Royal Geographical Society and Monthly Record of Geography, 13 (1891): 1–21.

  • 180. D. Randall MacIver, “The Rhodesian Ruins: Their Probable Origin and Significance,” The Geographical Journal, 27 (1906): 325–336; Caton-Thompson, Zimbabwe.

  • 181. Huffman, Snakes.

  • 182. For example, Huffman, Snakes; Johannes H. N. Loubser, “Ndondondwane: The Significance of Features and Finds from a Ninth-Century Site on the Lower Thukela River, Natal,” Natal Museum Journal of Humanities 5 (1993): 109–51; Pikirayi, Zimbabwe.

  • 183. Pwiti, “Economic.”

  • 184. For a more detailed review of these differences, see Chirikure et al., “Bayesian”; Pikirayi and Chirikure, “Debating.”

  • 185. See Calabrese, “Interregional”; Calabrese, Emergence; Machiridza, “Material Culture.”

  • 186. Chirikure et al., “Bayesian”; Chirikure et al., “New Pathways.”

  • 187. Machiridza, “Material Culture”; Smith, “Climate”; Pikirayi, Zimbabwe.

  • 188. de Maret, “Power”; de Maret, “Kinship.”

  • 189. Reefe, Rainbow.

  • 190. Childs and de Maret, “Re/Constructing.”

  • 191. Langworthy, “History,” 127.

  • 192. For example, Schoffeleers, River; Alpers, Ivory; Langworthy, “History.”

  • 193. For example, Schoffeleers, River; Alpers, Ivory; Langworthy, “History”; Newitt, “Early”; Kings M. Phiri, “Chewa History in Central Malawi and the Use of Oral Tradition, 1600–1920” (PhD dissertation, University of Wisconsin, 1975).

  • 194. Langworthy, “History”; Phiri, “Northern.”

  • 195. Mgomezulu, “Food.”

  • 196. Van Warmelo, Copper.

  • 197. Donald P. Abraham, “Ethnohistory of the Empire of Mutapa: Problems and Methods,” in The Historian in Tropical Africa, ed. Jan Vansina, R. Mauny, and V. Thomas (London: Oxford University Press, 1964), 104–211.

  • 198. See, for example, Donald P. Abraham, “The Principality of Maungwe,” NADA 28 (1951): 56–83; Donald P. Abraham, “The Monomotapa Dynasty,” NADA 36 (1959): 59–84; Donald P. Abraham, “The Early Political History of the Kingdom of Mwene Mutapa (850–1589),” in Historians in Tropical Africa: Proceedings of the Leverhulme Inter-Collegiate History Conference, September1960, ed. Erik Stokes (Salisbury: University College of Rhodesia and Nyasaland, 1962), 61–91; Donald P. Abraham, “Maramuca: An Exercise in the Combined use of Portuguese Records and Traditions,” Journal of African History 2 (1961): 211–225; Donald P. Abraham, “Ethnohistory of the Empire of Mutapa: Problems and Methods,” in The Historian in Tropical Africa, ed. Jan Vansina, R. Mauny, and V. Thomas (London: Oxford University Press, 1964), 104–211; Donald P. Abraham, “The Roles of ‘Chaminuka’ and the Mhondoro-cults in Shona Political History,” in The Zambesian Past, ed. Eric Stokes and Richard Brown (Manchester, U.K.: Manchester University Press, 1966), 88; see, for example, Edward A. Alpers, “Dynasties of the Mutapa-Rozwi Complex,” The Journal of African History 11 (1970): 203–220; see also D. N. Beach, “The Mutapa Dynasty: A Comparison of Documentary and Traditional Evidence,” History in Africa 3 (1976): 1–17.

  • 199. Beach, “Mutapa.”

  • 200. See Reefe, Rainbow, 14–19, for a more detailed discussion on written sources; De Maret and Livingstone Smith, “Who’s.”

  • 201. See, for example, Langworthy, “History”; Phiri, “Northern”; Schoffeleers, “Zimba.”

  • 202. Phiri, for example, lodged 100 edited interviews on Chewa History in the National Archives of Malawi, Zomba, and the University of Malawi Library in Zomba; see Phiri, “Northern,” 20.

  • 203. See, for example, Greville S. P. Freeman-Grenville, The East African Coast—Selected Documents from the First to the Early 19th Century (Oxford: Clarendon Press, 1962); Greville S. P. Freeman-Grenville, The Book of Wonders of India by Captain Buzurg ibn Shahriyar of Ramhormuz (London: East-West Publications, 1981); see also J. Spencer Trimingham, “The Arab Geographers and the East African Coast,” in East Africa and the Orient; Cultural Synthesis in Pre-Colonial Times, ed. Neville Chittick and Robert I. Rotberg (New York: Africana Publishing Company, 1975), 115–146; J. Spencer Trimingham, “Notes on Arabic Sources of Information on East Africa,” in East Africa and the Orient; Cultural Synthesis in Pre-Colonial Times, ed. Neville Chittick and Robert I. Rotberg (New York: Africana Publishing Company, 1975), 272–283.

  • 204. J. dos Santos, Ethiopia Oriental (Lisbon: Bibliotheca de Classicos Portuguezes, 1891 [1609]), English translation in G. Mccall Theal, ed., Records of South-Eastern Africa, vol. 7 (Cape Town: C. Struik, 1964 [1901]), 291–292; G. McCall Theal, Records of South-eastern Africa Collected in Various Libraries & Archive Departments in Europe (Cape Town: Government of the Cape Colony, 1898–1903).

  • 205. For a more detailed discussion on written sources, see Reefe, Rainbow, 14–19; de Maret and Livingstone Smith, “Who’s.”

  • 206. See, for example, D. Randall MacIver, Medieval Rhodesia (London: Macmillan, 1906); Caton-Thompson, Zimbabwe; Fouché, Mapungubwe; Guy A. Gardner, Mapungubwe, vol. 2 (Pretoria: J.L. van Schaik, 1963).