The Long Emancipation in Eastern Africa

Rinaldo Walcott's (Reference Walcott2021:105) The Long Emancipation analyzes the gap between legislated emancipation and Black people's true freedom “yet to come.” He argues that legislation to end slavery in the United States and the British Caribbean did not truly ensure formerly enslaved people's freedom; it instead held them “in captive relationship to their very recent past” (Walcott Reference Walcott2021:4) and in continuing subordinate labor and social positions lacking true bodily autonomy. The persistence of “a brutal outside for blackness and Black people globally” (Walcott Reference Walcott2021:9) signaled the continuation of what Walcott calls the Long Emancipation, through which Black people have yet to achieve actual freedom, even today. The Long Emancipation concept that Walcott developed in a North American and Caribbean context closely aligns with the protracted timeline for emancipation in Eastern Africa, in which even legal abolition was achieved piecemeal over decades.

The slave trade has a time-deep history in Eastern Africa, with some Greek documents describing the export of captives from the Swahili Coast as early as 2,000 years ago (Beachey Reference Beachey1976:3). Through both Portuguese and Omani colonial regimes, the trade in enslaved captives continued to flourish. More captives began to be retained rather than exported overseas in the nineteenth century in order to provide labor for a nascent coastal plantation economy under the control of Omani colonists and indigenous Swahili elites (Cooper Reference Cooper1997:46). Those captives who were enslaved tended to come from areas farther inland, away from coastal plantations, because this reduced the risk of them attempting to abscond back to their homelands. One major origin point for captives was the Lake Nyasa region, its shores abutting present-day southwestern Tanzania, along with Malawi and northern Mozambique. English explorer, missionary, and abolitionist David Livingstone (Livingstone and Livingstone Reference Livingstone and Livingstone1865:413) characterized this area as the “fountain-head” of the nineteenth-century Eastern African slave trade in his travel writing. On the shores of the lake in the 1860s, Livingstone reported the devastation that slaving had wrought: abandoned villages, roving fugitive groups, and many dead.

In the decades following Livingstone's travels, legal abolition came haltingly, piece by piece, to the region. German East Africa was established in 1891 in what we now know as mainland Tanzania following the infamous 1884 Berlin Conference, at which envoys from Europe, the United States, and the Ottoman Empire negotiated the colonial division of Africa. Even so, Omani colonialists still retained control of much of the Swahili Coast in both German East Africa and British Kenya. Under British pressure, the Omani sultan abolished slavery on Zanzibar and its sister island Pemba in 1897. In 1907, a British colonial ordinance outlawed slavery in Kenya. However, in German East Africa, slavery did not legally end until 1922, when the region was ceded to British colonial control following Germany's World War I loss (Marshall and Kiriama Reference Marshall, Kiriama, Wynne-Jones and LaViolette2018:566). Even this legal timeline, however, fails to represent the true complexity of the region's emancipation history.

Let us consider German East Africa (that is, mainland Tanzania) specifically. At the beginning of the twentieth century, the German government estimated that 15% of the colony's population, about 400,000 individuals, were enslaved (Deutsch Reference Deutsch, Lane and MacDonald2011:344). In southwestern Tanzania's Ruvuma region, where Kikole is located, there was no sustained German colonial presence at all until 1897 due to its interior location, some 600 km from the coast. After colonial contact was established, the local slave trade began to wane even without the enactment of formal legal prohibitions from the colonial government (Schmidt Reference Schmidt2010a:40). At the same time, although a German colonial decree in 1897 had prohibited the taking of captives in war, German documents describing the Ruvuma region in the early twentieth century make clear how much slavery there persisted, even into the Maji Maji War (Schmidt Reference Schmidt2010a:50). In 1904, children born to enslaved parents after 1905 were legally declared free in German East Africa; nonetheless, in practice, these children's social station closely resembled that of their parents. Given that they were not entitled to free upkeep from their parents’ enslavers, they often were put to work to offset the cost of their own food and housing (Sunseri Reference Sunseri1993:490). In 1901, German colonialists set up a ransoming system by which enslaved workers could purchase their freedom over time through work contracts with European planters (Sunseri Reference Sunseri1993:481). However, workers in these schemes often emancipated themselves without official colonial sanction; they broke their contracts, ran away, and refused to work (Deutsch Reference Deutsch2006:210; Sunseri Reference Sunseri1993:503). Historian Jan-Georg Deutsch (Reference Deutsch, Lane and MacDonald2011:344) has argued that self-emancipation was a major influence in the decades before legal abolition in German East Africa: “The archival record shows that slaves, by their own initiative, brought the ‘evil institution’ effectively to an end”—some by frequently running away (including en masse) and others by renegotiating the terms of their dependency. Indeed, the end of slavery in Tanzania was negotiated and fought for on the ground over decades rather than being achieved in any single moment by legal decree.

In interviews with Deutsch in the mid-1990s, residents of Mikindani, on the southern Tanzanian coast, reported that abolition had happened around the turn of the twentieth century (Deutsch Reference Deutsch, Lane and MacDonald2011:354). In Songea, in southwestern Tanzania, residents reported that slavery there had continued until the country won independence in 1961 (Deutsch Reference Deutsch, Lane and MacDonald2011:352). Of course, neither reported date coincides with legal abolition in 1922. Deutsch (Reference Deutsch, Lane and MacDonald2011) interprets these discrepancies by pointing out how enslaved people were increasingly able to self-emancipate in coastal locations such as Mikindani by ransoming themselves in the first decades of the twentieth century. He likewise suggests that the memories of his Songea interviewees have been shaped by local forced labor schemes that arose under indirect rule by German—and later, British—colonizers; in such schemes, European district officers relayed their labor demands to colonially appointed “chiefs” (or akida, see paragraph below) who then passed them on to local, typically marginalized people under their control. These demands disproportionately fell to ex-slaves (Deutsch Reference Deutsch, Lane and MacDonald2011:352). The protracted timeline by which freedom was achieved points to an unsettled sociopolitical landscape; here, the legal timeline of abolition is simply a backdrop to a more complex reality in which the practice and definition of slavery were negotiated and contested on the ground.

Rashid bin Masoud and the Making of Kikole

Rashid bin Masoud, an ivory and slave trader, had been active in the Ruvuma region of southwestern Tanzania since the early 1870s (Schmidt Reference Schmidt2010a:45). Ruvuma strategically abuts Lake Nyasa, a major slaving area. Though of Arab descent, bin Masoud was born in Eastern Africa, with some accounts maintaining that he originated in the inland caravan trade town Tabora (Giblin Reference Giblin, Zimba, Alpers and Isaacman2005:272), and others linking him to the Swahili Coast city of Kilwa, where he later had a residence (Schmidt Reference Schmidt, Giblin and Monson2010b:190). In any case, his birthplace was far afield of the large trading entrepots he established in Ruvuma—first around 1890 at Mang'ua and, less than a decade later, at Kikole. At Kikole, bin Masoud maintained an African following of heterogenous ethnic origin, including Ngoni, Ngindo, and Matumbi people, in addition to other Arabs (Mapunda Reference Mapunda2004:66) and Yao people from further south (Mzee Ally Hamisi Ally, personal communication 2021). Bin Masoud also kept a house on the Swahili Coast in Kilwa, where several of his “slave wives” reportedly resided (Becker Reference Becker2001:178). Keeping a coastal residence importantly supported his trade activities there. Back inland, German colonists appointed bin Masoud as akida for the Ruvuma region in 1894. The term akida originated with Omani colonists on the coast and was adopted by Germany in its indirect rule to denote a local official placed and backed by the colonial power structure (Becker Reference Becker2004:11).

Bin Masoud's participation in the slave trade is well established by both oral historical and written colonial accounts. A German officer who visited Songea in 1897 noted that

Local oral histories we collected, as well as those recorded by previous scholars (e.g., Giblin Reference Giblin, Zimba, Alpers and Isaacman2005:272; Mapunda Reference Mapunda2004:65), are likewise unambiguous on the question of bin Masoud's involvement in the trade of enslaved captives. In our research, Kikole's descendants and outsiders were equally frank on this point. Although David Livingstone's description of the turbulence and devastation on Lake Nyasa's shores in the 1860s points to warfare and raiding (e.g., Livingstone and Livingstone Reference Livingstone and Livingstone1865:413), there were multiple other entry points to enslavement in the region. For example, many captives were likely pawned by their families to local political leaders or wealthy relatives in exchange for food during famines, later being sold to caravans such as bin Masoud's (Mapunda Reference Mapunda2004:65). Bin Masoud was an equally prominent player in procurement and trade of ivory, with a reported 400 elephant hunters under his command in the region in 1899 (Gissibl Reference Gissibl2019:93). At the turn of the twentieth century, he was one of the most politically and socially powerful men residing in the Ruvuma region (Redmond Reference Redmond1972:231).

In 1905, bin Masoud was offered sacred water (maji in Swahili) by emissaries from Ngoni leader Nkosi Mputa Gama; drinking it would have bound him to the nascent anti-colonial Maji Maji uprising (Schmidt Reference Schmidt2010a:46). However, he refused the water and instead allied himself and his followers at Kikole with German colonialists, including providing soldiers for German-led attacks (Schmidt Reference Schmidt, Giblin and Monson2010b:202). He and his men also captured and delivered anti-colonial Ngoni leaders to the Germans for execution, including Mputa Gama in early 1906 (Schmidt Reference Schmidt2010a:47, 49). In September 1905, Kikole twice repulsed attacks by followers of Mputa Gama in retribution for bin Masoud's alliance with German colonialists. The war ended in 1907, its conclusion hastened by a brutal famine, which the Germans orchestrated by burning grain stores and crops. In the peacetime that followed, the colonial government rewarded bin Masoud for his loyalty by granting him special permission and arms to hunt elephants (Gissibl Reference Gissibl2019:131), turning a blind eye to complaints against him for kidnapping and assault (Schmidt Reference Schmidt, Giblin and Monson2010b:212), tolerating his continuing practice of enslavement (Becker Reference Becker2021:91), and discharging a series of debts, together amounting to more than 15,000 German East African rupees (Schmidt Reference Schmidt2010a:48)—the equivalent of about £1,000 at the time (Pallaver Reference Pallaver2023) and approximately £100,000 or $125,000 in 2024 currency.

Bin Masoud left Kikole sometime following the close of the Maji Maji War, although there is some discrepancy over the timing of his exodus in oral historical (e.g., Mapunda Reference Mapunda2004:72) and archival accounts (e.g., Schmidt Reference Schmidt2010a:48). Oral histories generally peg bin Masoud's departure to an earlier date, just a year or two following the close of the Maji Maji War, whereas some written colonial records suggest bin Masoud spent time in the Ruvuma region into the late 1910s. Whatever the case, Kikole was occupied several decades following bin Masoud's departure, persisting until the 1970s, when Tanzania's villagization policies forced residents to move to Lyangweni, a village 4 km southwest, in order to consolidate the local rural population. Several elders we interviewed had been born at Kikole, but all but a few had relocated to Lyangweni as young children, so they had limited memories of the settlement. The social memory of Kikole is, however, expressed in other ways, specifically through the persistence of Muslim practice in a minority of Lyangweni's occupants as an “island” of Islam in the “ocean” of Christianity suffusing the town and surrounding region (Mapunda Reference Mapunda2004:66).

Archaeology at Kikole

Kikole, approximately 10 ha in size, has stayed unoccupied since its inhabitants’ forced exodus in the 1970s. Nonetheless, material, architectural, and ecological indications of its prior occupation remain abundant. These include fired brick, unfired brick, and wattle-and-daub house ruins in addition to dozens of coastal trees (mango, date palm, and guava) that are not native to the region. These non-native trees and Kikole's structural remains closely coincide spatially, (Figure 2) and we can infer the trees were originally cultivated by bin Masoud and his followers (see also Doerr Reference Doerr1998). Coastal trees also persist at the ruins of bin Masoud's first trading post in the area, Mang'ua. Indeed, Elzear Ebner (Reference Ebner1987:122), a missionary historian, credited “Arabs” in the broader region around Kikole with introducing not only mango trees but also rice and sugarcane cultivation. The presence of such trees is a pattern also seen at other interior entrepots built by Arab traders in Eastern Africa (e.g., Becker Reference Becker2004:13; Mapunda Reference Mapunda, Giblin and Monson2010:233–234). Although surface artifact scatters are less visible due to a bushy and dense vegetative ground cover, multiple seasons of excavation have yielded thousands of artifacts, including local pottery, imported pottery, glass, beads, and metal objects.

Figure 2. Location of houses and non-native trees, Kikole. (Map by Beth Wilkerson, GIS Center, DePauw University.) (Color online)

Bertram Mapunda pioneered the archaeological investigation of Kikole in 2002, when he conducted a reconnaissance survey of the settlement and excavated a single 1 × 1 m test pit inside one of the houses with wattle-and-daub surface ruins (Biginagwa and Mapunda Reference Biginagwa, Mapunda, Wynne-Jones and LaViolette2018:549; Mapunda Reference Mapunda2004). Six additional units (1 × 1 m, 1 × 2 m, and 2 × 2 m) were excavated in 2014, including in collapsed houses, on middens, and in yard areas (Katto Reference Katto2016:39). The most extensive excavations of the site happened in 2021 during a University of Dar es Salaam (UDSM) field school, where 25 excavation units were dug to sterile soil (Figure 3). These were primarily but not exclusively 1 × 1 m units located inside prior structures—both collapsed earthen structures, indicated by surface mounding, and structures with standing wattle-and-daub or brick ruins. These excavations yielded local pottery (n = 2,255), imported pottery (n = 79), beads (n = 210), glass (n = 159), nails (n = 20), wire (n = 11), slag, animal bone, archaeobotanical material, and smaller amounts of a variety of miscellaneous objects, including buttons and five colonial German coins dating between 1904 and 1916 (Table 1). Field school students also participated in a larger regional surface survey of the surrounding area, in which locally made decorated pottery (n = 154) was recovered and analyzed as a comparative dataset.

Figure 3. Location of houses, excavation trenches, water wells, and termite mound watch towers, Kikole. (Map by Beth Wilkerson, GIS Center, DePauw University.) (Color online)

Table 1. Summary of Artifact Finds Recovered at Kikole during the 2021 University of Dar es Salaam Field School.

The 2021 field school was run jointly by Thomas J. Biginagwa and Sinyati Robinson Mark, both of the UDSM, with Lydia Wilson Marshall joining as a third volunteer instructor and supervisor for the 56 undergraduate participants; two individuals with BA degrees from UDSM, Baraka Edward and Javern Sabas, also contributed to the field school as secondary supervisors. The archaeological component of the 2021 field project was enhanced by a concurrent collection of oral histories conducted by Baraka Edward and Lydia Marshall. The most informative interviewee by far was 88-year-old Mzee Ally Hamisi Ally, who had lived at Kikole into early adulthood and therefore had a much clearer memory of the settlement than most other (younger) interviewees.

Kikole's material and architectural remains, in conjunction with local oral histories, indicate a diverse range of inhabitants, all of whom maintained a relationship of dependency with bin Masoud but also varied in the degree and nature of that dependence. Our analysis will focus on evidence of Kikole residents’ negotiation of and participation in a predatory landscape, the connections of the settlement to broader economic and social networks, and the social status and identity of Kikole's inhabitants.

A Predatory Landscape

Kikole's historical involvement in the slave trade is today acknowledged by both descendants of and neighbors to the settlement. It is also well established in the historical documentary record. Our excavations additionally yielded strongly suggestive material evidence, the most striking of which is a single iron shackle with an attached spike (Figure 4), which was located in a unit placed in the yard of one of the larger brick houses, reportedly owned by Mohamed Bushiri, an Arab trader and follower of bin Masoud. Wooden “slave sticks,” which attached around the neck, were a more common form of restraint for enslaved captives in Eastern Africa throughout most of the slave trade (e.g., Livingstone and Livingstone Reference Livingstone and Livingstone1865:137, 378, 537, 587); the late date of Kikole's occupation may account for the use of iron constraints at the settlement, given that metals were becoming less scarce and less costly (Edward A. Alpers, personal communication 2021). Certainly, bin Masoud's later capture and delivery of anti-colonial Ngoni leaders to the Germans during the Maji Maji War points to the possibility that these shackles may have been used in contexts beyond slaving. Since bin Masoud's participation in the slave trade is not in doubt, what is more important here is that the shackles indicate forced constraint of outsiders to the settlement and the potential for violence against them.

Figure 4. Partial iron shackle, excavated in a yard unit outside of standing house ruins at Kikole, 2021, pictured with a 10 cm scale. (Photo by Lydia Wilson Marshall.) (Color online)

Additional material remains demonstrate Kikole residents’ contribution to and negotiation of a predatory landscape. Archaeologically recovered weaponry and ammunition include bullets (n = 2), a bullet casing (n = 1), iron knives (n = 3), and a Ngoni-style iron spearpoint (n = 1). An understanding of the world as a place rife with danger is also reflected in more diverse materials such as keys (used to secure objects or people), a sharpening stone (to care for knives), and a half-moon folded copper alloy bell (to signal the movement of people or animals). This bell also exactly resembles some bells from Kikole that were ethnographically collected and are now on display in the Maji Maji War Memorial Museum in nearby Songea, Tanzania.

If we expand our analysis to include the spatial organization and regional placement of Kikole, other indications of latent violence come into focus. Bin Masoud set up his headquarters and entrepot to enable the procurement of ivory and captives. Kikole's relative proximity to Lake Nyasa, where slaving was deeply entrenched, is no accident. The settlement also directly abuts a knoll that rises about 100 m above ground (Figure 5); this elevation could have given scouts a clear view and advance notice of any approaching foes (Mapunda Reference Mapunda2004:64–65). Though long since decayed, a palisade reportedly ringed the settlement, at least during the Maji Maji War, and funneled those approaching from outside to two controlled entry points (Schmidt Reference Schmidt, Giblin and Monson2010b:203). Site survey also revealed the presence of four massive termite mounds near the settlement edges that oral histories suggest were adapted as watchtowers. Missionary observers actually reported that Kikole residents built watchtowers when anticipating attack during the Maji Maji War (Ebner Reference Ebner1987:141); however, our archaeological and oral historical research makes the adaptation of a natural feature—termite mounds—more likely. Kikole residents also dug three wells inside their settlement (Schmidt Reference Schmidt, Giblin and Monson2010b:203). These wells ensured that residents had access to fresh water without having to leave the protection of the palisade. The danger felt in the larger landscape is clear in their investment in multiple wells despite the very close proximity of the Mwangaza River. This investment signals Kikole residents’ wariness of outsiders and fear of attack. This wariness was certainly warranted, as shown by two fierce assaults by Ngoni forces in September 1905 against Kikole during the anti-colonial Maji Maji War. Investment in wells elsewhere in Africa has been likewise interpreted as a protective measure against slave raids (e.g., Monroe Reference Monroe2011:406).

Figure 5. Kikole in relation to the adjacent knoll and rivers. (Map by Beth Wilkerson, GIS Center, DePauw University.) (Color online)

However, it is important to remember that bin Masoud and his followers not only experienced a predatory landscape but also were key contributors to it. After the two attacks on Kikole, which were both repelled, bin Masoud persisted in his strong material support of German efforts to quash the anti-colonial uprising. Most notable is that Masoud's men captured the prominent Ngoni anti-colonial leader, Mputa Gama, in January 1906 (Mapunda Reference Mapunda, Giblin and Monson2010:238; Schmidt Reference Schmidt2010a:49). Mputa Gama had in fact offered bin Masoud the magic water that he refused; drinking it would have allied him with Mputa Gama and other anti-colonialists. Mputa Gama also later organized the two attacks against Kikole. He died with approximately 100 other Ngoni elders and leaders who were hanged at the regional German colonial headquarters at Songea in the spring of 1906. Existing grievances and tensions—including the Ngoni's recent arrival to the region and their own history of raiding—complicated relationships between different ethnic and linguistic groups in the area, and even among different Ngoni communities, during the Maji Maji War. It is nonetheless clear that Kikole's neighbors had cause to fear bin Masoud and his followers given their raiding, warring, and slaving activities. During our 2021 field project, one local Ngoni elder with no familial connection to the settlement declined to sit for an oral historical interview about Kikole, saying simply that bad people (watu wabaya in Swahili) had lived there.

Connections and Coherence

Certainly, bin Masoud and his retainers never sought complete cultural or social integration in the Ruvuma region; the persistence of their outsider status was buoyed by a continued adherence to the cultural norms of the Swahili Coast. Although located some 600 km inland from the Indian Ocean, Kikole is in many ways a distinctly coastal place, with clear indications of Swahili cultural practice. In cultivating coastal trees, in building rectilinear houses with baraza (verandas), in using coastal architectural materials such as large round-headed wrought iron nails, and in building a mosque, bin Masoud carefully crafted Kikole as a coastal place deep in the interior.

Date palm, mangoes, and guava are not native to the region and yet persist at Kikole to this day (Figure 6). The planting of coastal trees at Kikole aligns with colonial strategies observed elsewhere in the world to tame, control, and remake foreign landscapes (Cagnato et al. Reference Cagnato, Fritz and Dawdy2015; Gray and Sheikh Reference Gray and Sheikh2021). At Kikole, the message of these trees seems clear enough: this is a coastal place. Partially standing walls of brick and wattle-and-daub make clear how ubiquitous rectilinear structures were at Kikole (Figure 7); however, we know that curvilinear houses were much more common in all of the Tanzanian interior at the turn of the twentieth century. Rectilinear houses were widely adopted in southern Tanzania only between 1910 and 1920 (McKim Reference McKim1985:68–70). As German missionary influence was still nascent at the time of Kikole's founding in 1899 (notwithstanding the use of fired and unfired brick), it is reasonable to infer adherence to Swahili house style from the evidence for structure shape.

Figure 6. University of Dar es Salaam undergraduate field school students measuring the diameter of an extant mango tree to estimate its age, 2021. (Photo by Javern Sabas.) (Color online)

Figure 7. Overhead drone view of trader Bushiri's house and yard, showing its rectilinear orientation, 2021. (Photo by Thomas John Biginagwa.) (Color online)

This interpretation is additionally supported by archaeological evidence for the use of heavy round-headed wrought-iron nails at some structures. These partial and complete large nails (n = 8) ranged up to 13 cm in length and 57 g in weight, with head diameters typically more than 2 cm. Such nails are clearly distinct from their smaller utilitarian counterparts, 12 of which we also recovered; some of these smaller nails did not register even 1 g in weight when weighed. In Eastern Africa, large wrought-iron nails like those recovered at Kikole strongly suggest the construction of carved wooden doors. On the coast, such “Swahili” doors incorporated larger iron nails and stakes from the Portuguese colonial period in the fifteenth century onward (Aldrick Reference Aldrick1988:95–98, 202). The type of door indicated by the presence of these wrought-iron nails (Figure 8) suggests not only coastal cultural affiliation but also architectural investment as a claim to an elite coastal status; as British explorer Richard Burton (Reference Burton1856:205) explained succinctly of the coast, “The door is a valuable and venerable article in this part of the world” (quoted in Aldrick Reference Aldrick1988:218).

Figure 8. Wooden house door, showing the decorative use of wrought-iron nails in a Swahili Coast style, Mikindani, Tanzania, 2022. (Photo by Thomas John Biginagwa.) (Color online)

Archaeological remnants and oral historical accounts also point to the existence of baraza, a type of exterior veranda with a bench on which elite coastal men would congregate (Croucher Reference Croucher2015:128–129) at two houses in Kikole, including that of bin Masoud himself. The ruins of a mosque also still stand at Kikole, and many descendants of residents remain Muslim today. The reach of Islam beyond the Swahili Coast at the turn of the twentieth century was limited to only a few interior trade cities, such as Tabora. Muslim practice again suggests coastal influence. Bin Masoud maintained a second residence at Kilwa, on the coast, and these cultural connections demonstrate his continued adherence to coastal cultural norms even when physically in the interior. Bin Masoud, in fact, named Kikole after a coastal area of the same name near Kilwa in his efforts to remake the interior as the coast.

Trade goods at Kikole further demonstrate connections to global economic networks. In our 2021 excavations, we recovered 210 trade beads, including 16 of the famous drawn compound red-on-white variety sometimes called the rose bead, or Cornaline d'aleppo. Most common were drawn white glass beads (n = 81) followed by drawn blue glass beads (n = 46). Glass shards were overwhelmingly container glass; most pieces recovered were various shades of green (n = 86) or colorless, with or without an aqua tint (n = 56). Imported pottery (n = 79) was rare compared to the 2,255 local pottery pieces collected; it comprised only slightly more than 3% of the site's total combined pottery assemblage. The imported collection, though small, is revealing of the settlement's economic integration into the broader world; it included significant amounts of refined earthenware (n = 30), predominantly whiteware, and also coarse earthenware (n = 30), including Indian redware. We recovered smaller amounts of porcelain (n = 11) and stoneware (n = 8). Decorations ranged from simple red or black slip on the coarse earthenware to hand-painted floral polychrome and cut-sponge decorations on whiteware and flow blue on the single recovered sherd of pearlware. The imported beads, glass, and ceramics recovered are primarily of European origin, and they indicate the broad reach of global trade networks, given that these objects would have come through ships docked at the coast and then snaked their way into the interior on the backs of porters, whether hired or enslaved. The glass beads in particular represent some of the most important goods that flowed through the caravan trade, along with cloth, ivory elephant tusks, and enslaved captives (Karklins Reference Karklins1992:56; Mann Reference Mann2000:44). Their presence reminds us that Kikole residents did not just negotiate a predatory landscape; they also contributed to it.

Dependence by Degree

In considering the social status of Kikole's inhabitants, this section aims to illustrate the diversity of residents by comparing artifacts and architectural remains at three distinct structures at the site. According to the German colonial archive (and historians who have relied on that archive), bin Masoud's followers were simply his slaves. For example, based on colonialist and missionary accounts, historian Heike Schmidt (Reference Schmidt, Giblin and Monson2010b:207) estimates that bin Masoud's “slave army” boasted 1,400 individuals during the Maji Maji War. However, if one listens to descendants today, their accounts of their forebearers differ. Certainly, they affirm, Kikole inhabitants were not paid by bin Masoud except in kind with food or clothes for their labor in elephant hunting (for ivory), slaving, or warfare. But, as Mzee Ally Hamisi Ally rhetorically asked us, would you pay your children? The implication of this question is clear: the bonds tying followers to bin Masoud were familial in character. They were connections borne of love rather than bondage. In the remainder of this article, we call on archaeological evidence to intervene in this debate by comparing the houses of three followers of bin Masoud to demonstrate distinction in their statuses and caution against any blanket representation of their identity. Material and architectural remains together suggest that any single depiction of bin Masoud's followers may be inadequate.

Through survey, we identified dozens of small collapsed wattle-and-daub houses at Kikole, and we excavated several. What is notable about these structures is that although we were able to identify house floors and occasional postholes, we generally found few to no artifacts in collapsed earthen structures. Although residents likely took their belongings with them when moving away, such extreme material paucity still suggests a greater level of poverty and dependence than some other structures at the settlement. Consider, for example, a larger wattle-and-daub house belonging to Makonyola, who was reportedly an askari, or soldier, but with enough importance to be remembered by elders by name to this day. At his wattle-and-daub house, we unearthed an attached veranda with a fired brick floor (Figure 9). Fired brick (a technology adapted from German colonists and missionaries) was also used to build the house identified as belonging to Rashid bin Masoud as well as nearby houses of elite traders and allies reportedly living in Kikole. Although Makonyola's house was itself earthen, the use of brick in a veranda and the presence of the veranda itself are both significant. Verandas (often with a long stone bench running along the front of the house) were typical architectural features of elite coastal Eastern African houses and were well established by the nineteenth century (Croucher Reference Croucher2015:128–129). Verandas were also reported at Masoud's house in oral histories and archaeologically identified at the adjacent house of an elite ally and trader, Bushiri. That Makonyola's house included greater material investment through the use of brick and, specifically, the construction of a veranda (an elite coastal housing feature) suggests he maintained a greater social status than the inhabitants of the small earthen houses that we excavated. Artifacts recovered at Makonyola's house—including 11 European glass beads, a fragment of European whiteware ceramic, two shards of glass, six nails (including one large wrought-iron nail), and a 1916 German colonial coin—all point to greater material abundance as compared to the wattle-and-daub houses at which we recovered no artifacts.

Figure 9. Brick-paved veranda at Makonyola's house, Kikole, 2021. The orientation of the bricks indicates that it is a floor rather than a fallen wall. (Photo by Javern Sabas.) (Color online)

Let us now consider the remains of a third dwelling, closer to bin Masoud's house and reportedly belonging to his friend, an Arab trader named Said. This structure generated an even larger range of materials, including 13 imported pottery sherds (both Indian and European), 10 imported European glass beads, and 20 shards of glass; these were overwhelmingly container glass, suggesting the importation of mass-produced products stored inside glass bottles. We also found two nails, 10 pieces of wire, a 1904 German colonial coin, and a coat button. Most significant of all was the head of a walking stick, or in Swahili, fimbo (Figure 10). A fimbo was not generally used only to assist in walking; instead, it was a potent symbol of a “big man,” or “bwana mkubwa,” identity and consequently was strongly associated with elite masculinity on the Swahili Coast. The house represented here also showed evidence of brick building material, again suggesting greater material investment in house construction.

Figure 10. A partial wooden walking stick, or fimbo (Swahili), excavated at Said's house, Kikole, 2021. (Photo by Lydia Wilson Marshall.) (Color online)

Considering this diversity, it would be inaccurate to classify all of Kikole's inhabitants as dependent on Rashid bin Masoud in the same way. Certainly, to refer to Kikole inhabitants simply as a “slave army” is a significant oversimplification of their varying social standing and identities. But, even more, our research at Kikole calls us to question a single definition of slavery itself. Global understandings of slavery are typically rooted in Americanist historical conceptualizations of plantation agriculture and overwhelming violent domination. These depictions do not always align with well-documented slave systems elsewhere in the world (Croucher Reference Croucher2015), as the semi-independent but still enslaved mafundi on the Swahili Coast show. Indeed, they are also far from representing the diversity of enslaved people's experiences in the Americas, some of whom (like Kikole inhabitants) were armed and served as soldiers. Archaeologists of slavery and other relationships of dependence would do well to grapple with this diversity. Kikole shows us that there was more than one way of being dependent. There was also certainly more than one way of being enslaved. When we ignore this reality, we flatten enslaved people's varied lived experiences to a universal and false conception of what slavery invariably is and means.

Another axis of differentiation not so far explored in this article is that of gender. Written and oral historical accounts of Kikole's inhabitants emphasize the presence and activities of male followers. However, Kikole's persistence decades after bin Masoud's departure demonstrates women's participation in the reproduction and cultural reproduction of the settlement. Archaeology makes these women more visible—through, for example, the beads and pottery recovered at the site, both artifact classes being strongly gendered in local contexts. However, our investigations have done little so far to unpack the varying status differentiation of different women living at Kikole. Certainly, the patriarchal structure of Omani and Swahili societies (and Eastern Africa more generally) delimited the statuses that women could potentially reach at Kikole. However, it is equally unlikely that their statuses were internally undifferentiated.

A Long Emancipation, a Different Slavery

Walcott's (Reference Walcott2021) The Long Emancipation convincingly demonstrates the persistence of structures and inequities that prevented African diasporan people from achieving true freedom and autonomy in the Caribbean and the United States. What Kikole offers us is a push to consider anew the definition of slavery itself. Africanists have long argued that slavery was a long-standing, dynamic, and varied practice on the continent that can inform and challenge Americanist universalizing approaches to the institution (e.g., Miers and Kopytoff Reference Miers and Kopytoff1977). In nineteenth-century Eastern Africa, Sarah Croucher (Reference Croucher2015) built on this work in her analysis of a nineteenth-century clove plantation on Pemba, Zanzibar's sister island. Croucher complicates Americanist conceptions not just of slavery but also of plantations and capitalism. What we would like to do here is to return to the words of elderly Kikole descendants, those who variously described bin Masoud's followers as mafundi (technicians or specialists) or askari (soldiers or guards). Earlier in the article, we showed how the word mafundi was applied in the nineteenth century to a class of semi-independent enslaved artisans. The enslaved status of mafundi on the coast and their relative social and economic independence is pertinent to our discussion.

Enslaved mafundi generally made their own money from their work while sending remittances to their owners on a regular basis (Glassman Reference Glassman1991:291–292). As semi-independent skilled urban workers, they were exclusively male. Some mafundi even acquired female and junior male dependents—including slaves of their own—and attained ranked titles in an effort to establish legitimacy as coastal gentlemen (Glassman Reference Glassman1995:89, 92). From an Americanist perspective, one could reasonably ask, how could a slave own a slave? And, if they could, should mafundi be considered enslaved at all? Do we need to find a different word for this type of dependency, as has been suggested by some Americanists to us? Along the same line, can bin Masoud's mafundi at Kikole be considered enslaved if they also participated in slaving and the slave trade? What these questions underscore is the limits in applying a single Americanist definition of slavery globally. No, we do not need a different word for “slavery” in Eastern Africa. We need instead awareness of the flexibility of the term as well as the different patterns of dependence that may exist under slavery at any particular time and place. Building on Walcott's long emancipation, the protracted history of abolition and emancipation in Eastern Africa prompts us to propose a concurrent pattern of broad emancipation, a concept that emphasizes the coeval struggle of many individuals achieving many differing degrees of freedom in the same historical moment.

In 2002, the historian James Giblin interviewed Saidi Ali, an elderly resident of Mpitimbi, Songea region, near Kikole. Ali traced his ancestors to Pangwa people, who had been enslaved and resettled by their enslavers in the area. In speaking of Rashid bin Masoud and Kikole, he related, “And their business? Slavery! Slavery! (How in text to convey the disgust [with] which the narrator spits out the words, ‘Biashara yaooo!? Utumwa! Utumwa!’?) . . . Of all the evils there is none that exceeds that of this Arab” (Giblin Reference Giblin, Zimba, Alpers and Isaacman2005:272–273; his aside and translation). The difficult history of Kikole—of slaving and of helping to supress an anti-colonial movement in the Maji Maji War—continues to haunt relationships between Kikole descendants and other local peoples settled nearby. However, this dissonant heritage still has much to teach us about slavery, dependency, and identity. We have used archaeological evidence to analyze Masoud's followers and grapple with contradictory representations of them in oral and written histories. Instead of simply answering the question of whether Kikole residents were slaves, we have sought to use Kikole's multivalent and contested heritage to unravel the definition of slavery itself. As Walcott (Reference Walcott2021) shows, achieving freedom was and is something achieved only by degree. In an emancipation that was not just long but also broad, what was a “slave” in nineteenth- and early twentieth-century Eastern Africa was clearly a question disputed and contested on the ground by subordinated people and those who sought to control them.