Читать книгу Cartography and the Political Imagination - Julie MacArthur - Страница 12

1

The Geographies of Western Kenya

Slippery ground does not recognize kings.

—Abaluyia proverb¹

FROM THE shores of Lake Victoria to the foothills of Mount Elgon, the landscape of what would become western Kenya undulates with immense ecological and topographical variety, from gentle hills and flat-bottomed valleys to deeply gorged rivers fringed with lush tropical rainforests. For centuries, African settlers moving through this area greeted each other with the term “Mulembe.” In its many variations, “Mulembe” asked visitors where they had been, where they were going, and entreated them to come and go in peace.² A Luganda-English dictionary defined the term as “may he bring you in peace to your community.”³ Community members struck the murembe tree, with its brilliant red flowers, to undo taboos or to consecrate peace between warring communities.⁴ Despite its wide use across eastern Africa, Luyia elders in western Kenya claimed the term as central to their formulations of community, migration, and home: anywhere they traveled, “to be Muluyia was to say ‘Mulembe.’”⁵

The environmental diversity northeast of Lake Victoria invited a host of African settlers to cultivate the land and graze their cattle. Diverse migratory routes and relatively recent patterns of settlement created a linguistically and culturally mixed region: in the words of the doyen of Kenyan history, Bethwell Ogot, “lying in an ancient migration corridor, the traditional history of the district is one of the most confused and complex in the whole of East Africa.”⁶ This traditional history was a continual source of contention and argumentation, complicated by regional networks of trade and cultural exchange across complementary ecological zones. Migrating communities practiced a particular form of itinerant territoriality, a portable ideology of territorial control and belonging that linked networks of clans to particular tracts of land. African communities settled into niche environments and developed a diverse range of small-scale and defensive clan structures characterized by multiple and overlapping systems of authority. The precolonial space of the lake region represented a dynamic area of expanding frontiers, heterarchical social formations, and ethnic interdependence.⁷

Colonial conquest enclosed these frontiers and mapped new lines of exchange, community, and power. First circumnavigated by Henry Morton Stanley in 1875, the lake region was known to early traders, explorers, and colonial administrators as the Nyanza basin and its people as Kavirondo.⁸ Colonial conquest brought all the imperial instruments of state fixity—map, census, and tax—to bear on these diverse and decentralized African communities. Colonial geographers enlisted local inhabitants as amateur surveyors, guides, and porters to track rivers and navigate forests, to set out survey beacons, and to “beat the bounds” of their newly demarcated boundaries. The new maps introduced by imperial cartographers at the turn of the century imposed a top-down geographic vision that overwrote local conceptions of space and fixed named African communities to particular mapped territories. However, Africans around and across these newly mapped boundaries turned their spatial strategies to resistance, sabotaging the work of colonial surveyors, adopting new cartographic practices and symbols, and purposefully mistranslating their own geographic conceptions to countermap colonial geographies. The construction of colonial boundaries fostered new geographic imaginations of labor, obligation, and resistance and precipitated the remapping of authority, moral community, and competing territorialities in the making of western Kenya.

ENVIRONMENT AND SOCIETY: MAPPING PRECOLONIAL COMMUNITIES

As John Iliffe argued in his seminal text Africans, the history of the African continent and its peoples must be understood against the backdrop of a diverse and difficult environment.⁹ In his study of the Taita in Kenya, Bill Bravman put the point succinctly: “Geography is not exactly destiny . . . but it has powerfully influenced the history of the people who live there.”¹⁰ In the Great Lakes region, the ability to control and profit from this rough and potentially rich environment provided the primary source of tension for precolonial societies.¹¹ This geographic space, however, was more than just a backdrop against which diverse African political systems, social relations, and material cultures evolved: this space, both physical and cognitive, was dynamic and ever changing, inspiring and constraining the geographic imaginations of these communities over time.

Northeast of one of eastern Africa’s Great Lakes, named Lake Victoria in the nineteenth century by British explorer John Hanning Speke, lies a land of immense ecological and topographical variety. Created by a faulting in the geological crust, two main reliefs dominate the topography: the fertile lake plateau composed of a broad belt of granite soils studded by massive granite boulders and the drier highlands that flank the region to the north and east, where cool, moist air nurtures healthy volcanic soils. Altitudes range from 3,770 feet (1,150 meters) along the shores of the lake to over 7,050 feet (2,150 meters) on the slopes of Mount Elgon.

FIGURE 1.1. Topographical map of Kenya and magnification of western Kenya. World Resources Institute, Natural Benefits in Kenya.

This geomorphology has been described as “pockets”: although predominantly grassland savanna, bordered by equatorial forest to the north and tropical rainforest to the east, a series of interlocking ridges, hills, and valleys alternate to create an undulant plain. Two distinct seasons keep the region fairly regularly well watered: the dry season, from mid-December to mid-February, and the rainy season, from March to December, with annual rainfall between 61.7 and 76.3 inches.¹² A tilt in the African Plateau allows water to flow through numerous rivers and streams from northeast to southwest and drain into the lake. This varied relief created a great diversity of agro-ecological zones with niche economic and settlement potentials.¹³

This distinct and varied terrain formed a migratory corridor in eastern Africa that invited a long history of human settlement.¹⁴ In the “classical age” of African settlement in the Great Lakes region, from 1000 BCE to 400 CE, new developments in ironwork and crop production encouraged more intensive agricultural practices and the development of more settled and complex social and political systems. Early evidence of scattered settlements of southern Cushites, Bantu, and southern Nilotic groups existed in the southern regions around the lake, joined by an influx of Kalenjin- and Maa-speaking settlers further north at the turn of the millennium. Scholars have dated the current settlements to the fourteenth and fifteenth centuries.¹⁵ Though still a point of heated debate, most historians concur that Bantu settlements existed before the arrival of the Nilotic Luo from the Sudanese north in the late fifteenth century.¹⁶ Luo migrations dislocated these Bantu settlements, creating two distinct northern and southern blocks of Bantu communities around Lake Victoria: the Bantu communities of the north and the Gusii, or Kisii, of the south.

FIGURE 1.2. Southern view from Mount Elgon. Photo by author, March 2007.

FIGURE 1.3. Photos of North Kavirondo. Wagner, Bantu of North Kavirondo, vol. 1, plate 1.

Those Bantu populations pushed north by the arrival of Luo settlers settled into the ecologically diverse and productive region northeast of the lake extending up to Mount Elgon. The sixteenth century marked an explosion of new settlements and population growth that put pressures on cultivation, cattle, and land. The arrival of the Nilotic Teso in the mid-seventeenth century marked another important shift, as Teso-Bukusu/Bagisu wars over the next few centuries created a wedge between these two communities around Mount Elgon.¹⁷ Bantu and Luo settlers continued to arrive well into the eighteenth century, penetrating the outline of this region from multiple routes and settling into productive agricultural and fishing niches around the lake.

This linear narrative of migration and settlement in the region, however, belies the urgent and contentious argumentation at stake in the telling of these histories. Narratives of how and when particular groups arrived in the region have been recast to fit partisan projects of ethnic imagining throughout the colonial and postcolonial era. Bethwell Ogot’s pioneering text History of the Southern Luo, for example, was the product of ongoing partisan history projects among Luo intellectuals.¹⁸ Ogot’s dual volume Luo Historical Texts created an “archival patrimony” that obscured the intellectual work of other Luo historians.¹⁹ While early partisan historians such as Shadrack Malo and Samuel Ayany “consciously and unconsciously suppressed the signs and substance of discoordination and contention” to achieve a harmonized Luo past, Ogot went a step further, offering “the careful and patient reader a view of a people, a nation, Canaan, constructed out of critical tensions and conflicts over land, political domination, and domestic insecurity.”²⁰ Luo historians drew their constituents in line through direct descent from their fifteenth-century founding ancestor, Ramogi. Across the various projects of ethnic imagining in Kenya, patriotic historians wrote of unified and purposeful migrations to subsume divergent claims, to draw their partisans in patrilineal descent to a mythic founding father, and to promote the past as a model for contemporary threats to their moral communities.²¹

Further north, Gideon Were produced a similar archival reference entitled Western Kenya Historical Texts, cataloguing the oral clan histories of Bantu and Kalenjin settlers.²² In his history of these migrations, Were drew a sharp distinction between the largely settled and agriculturalist Bantu communities and their pastoralist neighbors, who, in the nineteenth century, were “still busy expanding northwards into their present territories.”²³ In this way, Were glossed over the multiple and complex migrational routes of Bantu agriculturalists to lay claim to a longer history of settlement. However, unlike Ogot, Were was harder pressed to create an overarching historical narrative out of the stubbornly diverse clan histories of these Bantu settlers, with no “Ramogi” common ancestor or common linguistic grounding. While Ogot could provide a unifying thrust to Luo migrations, Were was continually obliged to acknowledge the diverse origins of his subjects. At stake in these partisan histories were urgent concerns over land rights, community membership, and the future moral discipline of the history teller’s audience. Indeed, many of these early patriotic histories have been, unintended or not, transformed into evidence in contemporary land disputes and familial conflicts over custom.²⁴

In the texts of Were and others, the diversity of the communities in this lake region refused to be aligned. The multitude of migratory routes visualized in their maps reflected the diverse linguistic and historical backgrounds of these new settlers (figs. 1.4, 1.5). Clan histories of these migrations varied over time and space and reflected the variety of social organizations and political objectives of their tellers. Under pressure from overcrowding, disease, warfare, and would-be state builders, migrations often followed localized and uncoordinated clan lines that gave birth to new community formations. Many Bantu clans told of their historic origin in Egypt (“Misri” among Bantu speakers), then traveling down through Bunyoro and Buganda, providing a mythic point of origin and hinting at a stated or royal past.²⁵ Others emphasized their diverse migratory origins traveling through carefully recalled landmarks and difficult terrains to defend their cultural distinctiveness and political sovereignty. These “narrative maps” sketched a history of complex spatial relations, movements, and important sites such as forts and shrines along a migratory journey that strategically culminated in the construction of a regional homeland in their current settlements.²⁶

The histories of these migrations and settlements, whether Bantu, Kalenjin, Luo, or Maasai, must be understood not as the forward march of coherent ethnic groups but rather as complex histories of social interaction and multiple movements over centuries. These small-scale migrations led to fluid regional patterns of interaction and integration across ethnolinguistic divides: as Were argued, “The Abaluyia owe their origin to the interaction of many diverse cultural and linguistic groups stretching back to over one thousand years.”²⁷ Individual clans in the southern areas of Buhayo, Kisa, Marama, and Tiriki self-consciously trace their lineages to mixed Bantu, Kalenjin, Luo, and Maasai origins.²⁸ The clans of the Tachoni claimed their roots among both the early Uasin Gishu settlers and the later arrivals of Bantu clans passing around Mount Elgon.²⁹ In the space northeast of Lake Victoria, clans of various linguistic backgrounds formed strategic alliances through intermarriage and absorption, making every clan multiethnic from their very arrival. As Luyia historian John Osogo put it, “The history of our people has been at the local level, the story of the interaction of the clans.”³⁰ The residue of this regional culture of exchange persisted in the systems of trade that developed, in shared linguistic features, and in the assimilation of cultural practices.

FIGURE 1.4. Migrational map by Günter Wagner. Bantu of North Kavirondo, 1:23.

FIGURE 1.5. Migrational map by Gideon Were. Western Kenya Historical Texts, back cover.

The uneven and varied environment encouraged niche settlements and the development of specialized agronomic practices according to environmental capabilities and regional patterns of exchange. In the south, rich soils and compact valleys proved ideal for the intensive cultivation of major crops such as sorghum, sweet potatoes, beans, and bananas. In the more expansive plains of the north, a mix of cultivation and cattle ranching provided a middle ground of trade and political interaction between agriculturalists and pastoralists. A small body of research on precolonial trade networks in western Kenya suggests the influence of interethnic interaction on the making of precolonial communities.³¹ Markets developed early in the eighteenth century to facilitate trade across economic specializations. In the 1930s German anthropologist Günter Wagner recorded a lengthy history of precolonial markets from a Logoli elder: “The people of many different tribes assembled there . . . and in those years everybody who wished to obtain anything he liked could go to that market.”³² Northern Bantu settlers frequented these markets in search of fish and livestock, while the Luo came for the grains brought by Bantu agriculturalists, mostly millet and sorghum. From their new agricultural neighbors, Kalenjin-speaking pastoralists adopted a number of terms involved in cultivation and food production, including words for “beans,” “flour,” and “to weed.”³³ Northern Bantu populations similarly borrowed cattle-keeping terms and practices from their Kalenjin and Nilotic neighbors. As Jean Hay found, even between the divergent Bantu and Dholuo language groups, the terms for “homestead, wooden hoe, sorghum, maize, beans, and a number of other crops are essentially the same . . . suggesting extensive cultural contacts and influence in economic matters.”³⁴ Economic interdependence and trade fostered cultures of exchange and integration across ethnolinguistic divides.

These complementary environments also fostered the development of specialized skills. The availability of great deposits of iron ore in the Samia hills prompted its settlers to specialize as blacksmiths, peddling their skills across the region. The Banyore, located in the well-watered hills of Bunyore, were recognized as rainmakers and consulted as experts throughout the region. The sharing of professionals and specialists across clans traced a regional network that further encouraged cultural exchange. In border regions, many informants pointed to the adoption of place names and linguistic features from their Kalenjin neighbors. The Tiriki adopted Kalenjin age-set names, arguably one of the defining features of male-community membership.³⁵ The Tiriki, frequently at war with their Logoli neighbors in the nineteenth century, often allied with Nandi clans.³⁶ The adoption of Kalenjin age-set names may have allowed young men to identify allies in warfare despite linguistic and cultural divisions. The relative practice of circumcision, both male and female, among these groups also attested to complex historical exchanges. The practice of circumcision varied greatly despite contemporary popular beliefs that all Bantu communities in this region practiced male circumcision, in contrast to their uncircumcised Luo neighbors, but not female circumcision, in contrast to their Kalenjin and Maasai neighbors. Early European explorers noted the falsity of such common beliefs, recording a variety of practices ranging from elaborate male circumcision ceremonies to the absence of any form of circumcision.³⁷ These exchanges of experts and customary practices blurred the environmental and social lines seemingly dividing these groups. Ideas of community and production were not confined in this era by notions of enclosed ethnic communities but rather grew out of geographic interdependence, pragmatic comparative work, and strategic cultural borrowing.

OF CLANS AND TRIBES: TERRITORY, AUTHORITY, AND BELONGING

It was this “slippery ground,” this environment of rich diversity, interdependence, and fluidity that, in the words of one proverb, did not “recognize kings.”³⁸ Alongside lessons in agronomy and social interdependence, this environment taught its first settlers to develop independent small-scale polities characterized by heterarchy, unique languages of political and social organization, and distinct cultural practices in land and authority.

Precolonial studies of eastern African societies often emphasized the flexibility of local identities as necessary for “establishing relative positions on the cognitive map of expanding frontier societies.”³⁹ Early oral historians pinpointed the clan as the most stable and fruitful source of historical information and genealogical periodization: unlike the “invented” tribes of 1970s historians, clans were “out there,” in the words of Jan Vansina.⁴⁰ Recent scholarship has criticized these evolutionary and lineage-based models and has begun to interrogate the moral and intellectual organization of precolonial African communities. In their classic study on the Luo identity in Siaya, David Cohen and E. S. Atieno Odhiambo argued that “friendship fortifies kinship,” a challenge to mainstream anthropology that treats kinship as “an enclosed autonomous locus of structure.”⁴¹ In her study of the Bagisu, close relatives of the Bukusu on the Ugandan side of Mount Elgon, Suzette Heald demonstrated that segmentary lineage systems and descent-based kinship ideologies were not inherently at odds with the practices of territoriality or creative reformulations of kinship.⁴² Less the basic and innate structure of community envisioned by colonial administrators and European anthropologists, clans were the historical products of imaginative social work and provided a locus of belonging amid multiple shared sites of identification and interaction.

Among those who settled northeast of Lake Victoria, the clan was often the largest and most constant source of identification and support. The clan acted as a unit of agency, a tool in the management of social relations that offered cognitive structure to mappings of community and territory.⁴³ Clans were often a heterogeneous mixture of families and small groupings linked together by common ancestry, common migration, or common settlement. However, the terms of membership and size of the clan were as flexible as the new environments in which they settled required. Exogamy, the practice of marrying outside the clan, and patrilineal descent underlined the gendered nature of community membership. Wagner found that “each tribal community (particularly its male half) derives its ‘group consciousness’ first and foremost from the belief that all or the large majority of its constituent clans have descended in an agnatic line from a mythical tribal ancestor.”⁴⁴ However, as earlier alluded, the primacy of descent and ancestral myths proved more illusory than this unqualified assertion.

As with partisan histories of migrations, the nomenclature of “clan” and “tribe” represented terms of imagined political communities continually contested and reworked within larger cultural projects. Some larger “tribal” associations of clans, such as the Wanga, the Logoli, and the Bukusu, did emerge in the eighteenth and nineteenth centuries; but these communities were the products of the same imaginative work as clan communities on a larger scale. For the purposes of this study, these “tribal” names suggest not articulated structures of political authority but rather larger communities of clans that were recast in the colonial era to defend precolonial autonomy.

Beyond changing ethnonyms and constraining lineage-based models of kinship, the basis of clan formations in western Kenya relied on the ability of the group to effectively settle and civilize new ecological niches. Neil Kodesh has argued that the variety of terminology for clan structures in the Great Lakes region, including ubwoko in Kinyarwanda, kika in Luganda, and ruganda in Lunyoro, suggests that “the ideology and practices of clanship developed along different lines in various settings within the region.”⁴⁵ Among the Bantu communities of this compact area, terms for clan organization varied from oluhia and olugongo to ibula and ehiri.⁴⁶ These terms reflected not only varying forms of political organization but also the varied environments that constrained as well as sheltered.

Environmental conditions taught these inhabitants lessons in how to structure political and social relations. Ecological niches prompted the development of specialized economic production and skills that fostered distinct communal identities. Some, like the larger groups of the Tiriki and the Bukusu, became frontier communities as they settled on steep hills and beside rich forests that offered opportunities for expansion.⁴⁷ The expanding Bukusu promoted strong military leadership (in the figure of the Omugasa), built fortified, walled villages into the landscape, and transformed territorial msambwa ancestral spirits into “ancestral ghosts,” akin to those in Buganda, so that they could inhabit the caves and springs of their new territories.⁴⁸ Throughout eastern Africa, msambwa sites on the edges of settlements provided important sacred spaces linking the “health of the land” to the health of residents and their descendants.⁴⁹ Living on the imagined edge of this landscape, these frontiersmen would pose distinct problems for both colonial officials and ethnic architects seeking to territorialize ethnic identity.

Others, such as the southern clans of the Bahayo, Banyore, Batsotso, and Idakho settled into pockets isolated by hills and ridges and well suited for intensive agriculture. Among these more dispersed settlements, patriarchal clan heads and councils of elders provided only symbolic leadership linked to their ability to amass material and human wealth and maintain peaceful relations.⁵⁰ Among the large but scattered clans of the Logoli, the weng’oma, or “one of the drum,” would beat a drum across the hills to gather clan heads together in times of war.⁵¹ Early anthropologists such as Wagner viewed these systems of political authority as “inarticulate,” not “linked up with clearly defined rights and privileges, such as usually associated with institutionalized chieftainship.”⁵² Even in warfare, as Sir Harry Johnston noted, precolonial raiding patterns were individualistic and “almost entirely defensive.”⁵³ However, despite these varied structures, political authority was neither “inarticulate” nor merely “defensive”: political power among the communities of western Kenya was heterarchical, organized along horizontal lines that allowed for flexible patterns of interdependence, defense, environmental management, and multiple sites of authority.⁵⁴

One glaring exception existed alongside this depiction of decentralized political life. The Wanga crafted the most hierarchical political culture in the region, having a royal family and a king, the nabongo, who ruled over Wanga clans. Much has been written on the Wanga royal family and the power struggles between different clans over leadership.⁵⁵ The Wanga kingdom functioned somewhere between the large and powerful lake kingdoms further west and the more horizontally organized communities of the east.⁵⁶ Simon Kenyanchui better described the Wanga political structure as “a confederation of co-equal clandoms.”⁵⁷ Later political accounts would map the extent of the Wanga kingdom from Lake Victoria to Lake Naivasha, but little evidence supports this expansive claim and the numerous clans described above jealously guarded their autonomy against claims of rule or tribute by Wanga kings.⁵⁸ In the nineteenth century, Wanga monarchs used Maasai mercenaries to extend their range of tribute and territorial rule, prompting these independent communities to seek refuge in their protective environments and buttress their own political structures against these monarchical state builders.

The first Europeans to arrive in the region were greeted by the recently ascended Nabongo Mumia, who quickly offered the new arrivals hospitality and aid. For many colonial administrators and later historians, the Wanga kingdom provided a recognizable political structure ideally suited to their ends and interests. Colonial administrators saw in the Wanga kingdom a hierarchical system of authority that could be usefully extended over the decentralized communities of the area. Later local politicians saw in the Wanga’s monarchical history a useful narrative of precolonial political sovereignty and organization. By the end of his study, Were became preoccupied with the hypothetical future of the Wanga kingdom had the British not interrupted its consolidation and expansion.⁵⁹ The centralization of the Wanga kingdom, however, has not only been overstated by colonial officials and partisan historians but has also overshadowed the complex political interplay and almost defiant tradition of decentralization among the majority of western Kenya’s inhabitants.

Relations among these diverse but interdependent communities were not “of coercion and control but of separate but linked, overlapping yet competing spheres of authority.”⁶⁰ Heterarchy allowed for multiple religious, political, and economic sites of power and identification to exist in parallel. Each clan maintained autonomous control of their own political affairs and yet depended on wider networks for economic, spiritual, and social exchange. Within and outside the reach of the Wanga kingdom, heterarchy also provided a measure of accountability, as it did within the kingdom of Buganda, further west.⁶¹ Among the Wanga, clans organized themselves in circular spatial patterns around the royal family, at once buttressing and constraining the power of the king. Succession was not determined necessarily by descent but rather by a council of clan elders who considered multiple factors.⁶² Strategic intermarriages and political negotiations with the various clans of the Wanga ruling elite allowed non-Wanga clans to maintain a great deal of autonomy and protection.⁶³

Outside the reach of Wanga tribute, clans used horizontal and complex systems of social organization, strategic alliances, and spatial encampment. The term oluhia, from which later cultural entrepreneurs would find a name for their imagined community in the 1930s, reflected this heterarchy and the primacy of place in the social formations of the region. Sometimes translated as clan or clansmen, oluhia in many of the languages in the region referred to the “fire-place on a meadow,” where the heads of associations of clans would meet.⁶⁴ The oluhia served as a sort of assembly site for initiation rituals, for political negotiations, and for the burial of clan heads: it was a “microcosm . . . the place of practical everyday life.”⁶⁵ This term embodied the horizontal coming together of representatives from different clans in one symbolic and physical space, reflecting the close relationship between place, belonging, and communal identity among these diverse communities.

Despite these multiple sites of identification, land and the practices of territoriality played crucial roles in the spatial organization of belonging. For Osogo, limited tribal structures represented loose linguistic and cultural affinities, while clans performed their most important function as “owners and bequeathers of property,” or, more accurately, land use.⁶⁶ Land tenure practices varied greatly across the region, producing “different systems from a common background.”⁶⁷ A common term for the clan and clan territory, olugongo, literally translated in many linguistic traditions to “a ridge.”⁶⁸ Within each olugongo, the clan leadership determined how to allot land to each family and how to absorb and manage the land claims of “strangers.” Although clan heads were responsible for negotiating internal boundary demarcations, the limits of their olugongo were “known by their natural boundaries.”⁶⁹ Uncultivated virgin bush land, oluangeraka, or “what is beyond” among the Logoli, allowed clan lands to expand and contract in response to seasonal environmental changes, demographic pressures, and interclan disputes.⁷⁰ This practice facilitated crop rotation and the strategic fallowing of lands to avoid the overuse of any one area of cultivation. The edzinzalo, a common term for uninhabited mile-wide buffer zones, similarly provided an important precolonial territorial strategy for managing conflict over land and resources. During war times, these buffer zones provided fields of contest and a space for the meeting of warring factions; however, “in times of peace, grazing [was] communal.”⁷¹ Wagner noted that these buffer zones acted as political frontiers, as clans did not enforce the “subjugation of neighbouring” groups but rather enacted a strategy of “political integration through territorial continuity of clan.”⁷² The terms and practices of land tenure in the nineteenth century reflected the processes of segmentation—of fission and fusion—that characterized social formations on still-expanding frontiers.⁷³

Niche settlement patterns translated directly into distinct practices of territoriality. Among the Banyore, the hills they inhabited became central to their “situational identity.”⁷⁴ The Bunyore hills, from whose geographic characteristics clans found their names, represented not only the place where their forefathers first settled but also a protective barrier that defined friends from foes. In the other direction, place-names taken from clan names or important individuals also told stories of migrations and interethnic exchange.⁷⁵ Terms for lineage often overlapped with a sense of geographic enclosure. The terms enyumba, eshiribwa, and indzu all could be translated as both lineage and the physical enclosure of the clan or gateway of a homestead.⁷⁶ In Marama, minor clans who had only recently joined the larger network were referred to as emikuru, veranda poles that propped up the household structure of the larger clans.⁷⁷ As Christopher Gray has argued, the mistranslation of these terms into simply “land,” “clan,” or positions of authority by later colonial administrators would elide “the whole complex series of obligations and duties owed to lineage heads by their dependents, and as such . . . the relations of production for these societies.”⁷⁸ This linguistic variety revealed the intimate connection between community formations and the territory they inhabited.

Despite the multiplicity of precolonial social relations, the practice of territoriality was central to the functioning of these communities. This was not the “aterritorial kinship ideology” Gray found practiced in Gabon.⁷⁹ Aterritorial practices did provide social entities with the flexibility to account for complex trade and exchange patterns as well as pressures on the land, whether from the environment or warfare. However, African communities in this region practiced a form territoriality that developed out of their complex migrations and niche environmental settlements. As introduced at the outset of this chapter, the greeting term mulembe carried within it this itinerant territoriality that linked who one was with where one came from and where one was going. Addressing the territoriality of these clans, Wagner lamented that “the extent of the geographical ‘horizon’ of the various sub-tribes in pre-European days I found very difficult to discover. The traditions of some tribes refer to places which are hundreds of miles away from their present homes.”⁸⁰ Wagner’s frustration rested in the mobile form of territoriality practiced by these settlers, carried within kinship ideologies and modes of political authority. Clan representatives often recounted their histories as a people on the move: they “moved in clans” and carried into their new settlements a sense of territorial community membership.⁸¹ Despite long migrations and new environmental settlements, clans often repeated former spatial configurations in their new settlements, reestablishing the spatial ordering of families and figures of authority.

By the nineteenth century many of the communities northeast of Lake Victoria lived as agriculturalists in defined territorial settlements, despite constant pressures from expanding agriculturalist neighbors and continuing cultural and economic trade with surrounding pastoralists. However, the nineteenth century brought with it a time of political and environmental upheaval in eastern Africa, what Gideon Were termed the “age of confrontation.”⁸² Population growth and continued migrations in the first half of the century caused almost constant warring. Although on the periphery of long-distance trading emanating from the coast, by the late nineteenth century Swahili traders were regular visitors to the area and stories of the infamous slave trader Sudi of Pangani circulated widely, though actual slave trading in the region seems to have been limited.⁸³ From 1890 the Nyanza basin suffered a series of droughts and diseases that decimated much of eastern Africa.⁸⁴ The devastating rinderpest outbreak that caused widespread stock loss, as well as a smallpox epidemic in the 1890s, compounded by the arrival of colonial conquest, undermined traditional practices of land and bush management that had kept disease at bay.⁸⁵ These factors forced communities to push into the bush in search of new lands, thus unleashing the threat of trypanosomiasis, or sleeping sickness, spread by tsetse flies across the region.⁸⁶ By the end of the century the political and territorial sovereignty of the communities northeast of Lake Victoria was under threat—from expanding neighbors, from disease and drought, and from the arrival of imperial surveyors.

BEATING THE BOUNDS: THE MAKING AND UNMAKING OF IMPERIAL GEOGRAPHIES

From the earliest explorations of Henry Morton Stanley, Joseph Thomson, and Frederick Jackson, explorers, administrators, and missionaries wrote widely of the rich diversity and warm hospitality they encountered northeast of Lake Victoria.⁸⁷ These first European visitors universally commented on the plentiful food production, the variety of languages, the diversity of cultures, and the “confusion” of political structures among the African inhabitants they called Kavirondo. Early explorers described the Kavirondo as “industrious” and the “most moral of all tribes.”⁸⁸

Morality and gender relations seemed to preoccupy these early European visitors to the lake region. They marveled at the Kavirondo men and women working in the fields together in complete nudity. Comparing the naked Kavirondo women to the covered “ladies of Lamu,” Sir Charles Eliot mused, “In Africa, female respectability is in inverse ratio to the quantity of clothes worn, and the beauties of Kavirondo, who imitate the costume of Eve, are said to be as virtuous as she was when there was no man but Adam in the world.”⁸⁹ As attested to in postcards and photographic collections, the “naked Kavirondo” became an alluring tourist attraction in these early years of imperial travel in eastern Africa (fig. 1.6).⁹⁰ Images of naked women standing in fields, pulling a hippo across the shore or sunbathing on rocks became collectibles for travelers from coastal traders to Theodore Roosevelt.⁹¹ Early explorations and imperial travels pictured “Kavirondo” as an untouched landscape, remote and unknown, filled with culturally exotic and morally “naked” people.

The term Kavirondo first appeared to administrators as a distant mapped space on E. G. Ravenstein’s maps near the end of the 1870s.⁹² In 1884, Joseph Thomson noted with surprise that “Kavirondo does not at all occupy the place which has been assigned to it on the map.”⁹³ Bewildered by the sheer “number of very distinct tribes,” the term Kavirondo was applied to both the Nilotic and Bantu communities of the lake region, though it progressively came to more specifically denote “all those natives speaking Bantu dialects west of Busoga and north of Kavirondo Bay.”⁹⁴

Bringing this confusion under colonial control required the remapping of local geographies. In 1890 the Imperial British East Africa Company (IBEAC, or IBEA) signed treaties with “Sultan Mumiya” of “Upper Kavirondo,” with the nabongo, or king, of the Wanga, and with the Bukusu clan head “Majanja, Sultan of the Kitosh.”⁹⁵ In signing these treaties, these two leaders ostensibly “ceded to the said Company all his sovereign rights and rights of government over all his Territories, Countries, Peoples and Subjects.” Ravenstein’s map for the IBEAC visualized the area of Kavirondo as an important thoroughfare in the economic trading path between the coast and the large kingdoms of the Uganda Protectorate, centering on the town of Mumias. Nabongo Mumia provided the British with a base for trade running between the coast and Uganda and ample supplies of food and manpower for their journeys (fig. 1.7). The space northeast of Lake Victoria, however, remained barely surveyed. Early maps traced complex webs of escarpments and rivers amid whole sections of undifferentiated landmass. Treaties and other informal arrangements led to the incorporation of the region as the Eastern Province of the Buganda Protectorate in 1894.⁹⁶ Misreading the political landscape of this region, colonial officials pictured its communities as “unsettled,” with no recognizable systems of political structure aside from the Wanga kingdom. In 1895 it became C. W. Hobley’s task to “gradually establish an administration over the various sections of the turbulent collection of tribes, collectively known to the coast people as the Kavirondo.”⁹⁷

FIGURE 1.6. Mombasa postcard, postmarked 1899.

FIGURE 1.7. IBEAC map of East Africa, 1891–92, by E. G. Ravenstein. Lugard, “Travels from the East Coast.”

For Hobley, this reconnaissance work was one of both topographical surveyance and ethnographical investigation (fig. 1.8). The ethnonyms identified by Hobley curved and contorted amid a mess of complex topographical features, aiming to organize space in ethnic terms. The crowded “Isukha” branched out from and over rivers and forests; the “Marama” curved and fitted into a pocket of empty space; the “Ketosh,” the name known to Hobley for the Bukusu clans, stretched and expanded across vast territories in the north. In this early map, the territory of these ethnonyms remained amorphous and undefined either by geographic markers or internal boundaries. In his 1896 report on Kavirondo District, Hobley noted, “The fact of the county being split up into so many sub-tribes without any really powerful Chiefs has induced a more complicated situation than would otherwise be the case.”⁹⁸ As Sean Hawkins has argued, the failure to read topographical features that functioned as “mnemonic devices” in local cognitive mappings led the British to believe these communities had no “inscribed past” and no geographic system of social organization.⁹⁹

FIGURE 1.8. C. W. Hobley, map of Kavirondo, 1898. Hobley, “Kavirondo.”

More than mere lines on a page, colonial mapping practices worked to transform previously relational geographies of exchange and community into top-down, scientific, and measurable demarcations. Local geographical concepts proved difficult to translate, as they were relational and relative rather than constant abstract points of reference.¹⁰⁰ The fixed compass points of north, south, east, and west had no corresponding terms in any of the local languages. The term masaba for most communities translated as north but referred specifically to Mount Masaba, known by colonial officials as Mount Elgon, the northern frontier of this region.¹⁰¹ In the Trans Nzoia, an area parallel to the mountain, masaba translated as west. Group names like Isukha and Idakho translated as forward and backward or lower and thus similarly reflected situated, geographic relationships and histories of migration and settlement.¹⁰² The Luo referred to the Banyore as “those people from the other side of the hill.”¹⁰³ Spatial relations thus shifted depending on the positioning of the subject. Colonial conquest not only imposed new and ever more fixed territorial boundaries but also brought with it a new vision of geographic relations.

Conquest and pacification expeditions from 1894 to 1908, numbering fifty separate military operations as catalogued in John Lonsdale’s authoritative account, helped fill in Hobley’s map and establish internal boundaries around his curving ethnonyms.¹⁰⁴ The hierarchical Wanga kingdom provided the British with a ready army of Wanga and Maasai troops mobilized not only against their Bantu neighbors but also in battles against the Sudanese mutiny of 1897 and Luo and Nandi uprisings.¹⁰⁵ Other communities in the region reacted to colonial incursions with varying degrees of interest, trade, alliance, and resistance. Many viewed the British as pawns of Wanga territorial ambitions.¹⁰⁶ The construction of colonial sovereignty required a great deal of violence and the use of what Christopher Vaughan has called a “hybrid regulatory order,” an ambiguous and not always controlled devolution of power and the means of violence to local chiefs, particularly in border regions.¹⁰⁷ The most protracted resistance came from the Bukusu, whose battles at Lumboka and Chetambe’s Fort received vivid portrayals in both Hobley’s writings and local oral narratives.¹⁰⁸ Lonsdale pointed to the relative geographic isolation of the Bukusu population, in the northernmost extent of the region around Mount Elgon, as key to their protracted resistance to outsiders.¹⁰⁹ As in precolonial conflicts, the slopes and caves around Mount Elgon provided safe haven for those escaping the grasp of would-be state builders.¹¹⁰ After winning the final major battle against the Bukusu, in 1895, “Hobley had no more to fear from the Luyia. He had been greatly aided in their pacification by their historic disunity.”¹¹¹ This “historic disunity,” however, can in actuality be understood as a strategy, an alternative way of thinking and practicing geographic and social relations, and a means of resistance that while failing to prevent colonial conquest persisted in dogging colonial rule.

Christian missionaries arrived by similar paths as explorers and colonial surveyors from earlier positions in Uganda and on the coast at the end of the nineteenth century. Tales of Bishop James Hannington’s visit, in 1885, and Mumia’s warning to him not to travel to Buganda, where he died soon after, have taken on legendary status. The North Kavirondo District was unique in the region for the sheer number of missionary groups that gained footholds, numbering as many as ten by the 1920s. The American Quakers with Friends African Mission (FAM) were the first to establish themselves, in 1902. In quick succession the Catholic Mill Hill Missionaries, the Church of God (originally known as the South African Compounds and Interior Mission), and the Anglican Church Missionary Society (CMS) all founded missions in these early years. Colonial officials hoped mission influence would help bring the “most independent and unruly” Bantu communities under colonial control.¹¹² Missionaries made deals with local elders and competed for land claims to secure territorial “spheres of influence” as first laid out in the 1885 Treaty of Berlin.¹¹³ These negotiations for evangelizing rights in particular areas added yet another layer of mapping onto the multiple colonial processes of territorial demarcation. However, unlike the territories of colonial administration, religious spaces were not always contiguous. Mission grounds and Christian villages crossed, jumped, and intersected different community spaces, creating hybrid sites of cultural contact and imposition, refuge, and exchange. This more dispersed spatial organization would have distinct consequences for later political and social movements.

Colonial conquest brought with it a flurry of mapping projects, from the railway that arrived at Kisumu in 1901 to the carving out of roads, trading posts, and administrative sites. The administrative mapping of the Uganda Protectorate created internal boundaries and delimited top-down spaces of administration and authority that could now be imposed and altered by British bureaucrats, even at great distances (fig. 1.9). In 1902 a decision from the Foreign Office in London to dramatically redraw the eastern boundary of the Uganda Protectorate prompted a more concerted mapping of colonial boundaries in eastern Africa. With the boundaries of the Uganda Protectorate already mapped, British officials in London could, by the stroke of a pen, transfer a large portion of the Central Province and all of the Eastern Province of the Uganda Protectorate to the East Africa Protectorate, renamed the Kenya Colony in 1920.

FIGURE 1.9. Map of the Uganda Protectorate, 1902. Johnston, “Uganda Protectorate, Ruwenzori.”

The motivations behind this transfer remain difficult to assess.¹¹⁴ The arrival of the railway at Kisumu, in 1901, provided at least one factor in the reorientation of the province, aiming to maintain the whole railway from the coast under one authority (fig. 1.10). As the arrival of the railway and road surveys demonstrated, Kavirondo was increasingly being positioned on the map not as a midpoint on the way to Uganda but as a terminus of trade and administration emanating from the coast. Further, the concurrent arrival of white settlers in the highlands prompted calls for land and labor within the boundaries of the East Africa Protectorate. Nyanza Province, formed out of this enlarged western frontier, effectively placed one of the most agriculturally productive and densely populated areas in the region at the disposal of this rapidly developing settler colony.

FIGURE 1.10. Map of road making and surveying in British East Africa. G. Smith, “Road-Making.”

Negotiations over the drawing of this new boundary revealed the competing knowledge systems involved in the colonial mapping projects. While some argued for the use of the “scientific” determinants of topographical features as “natural” boundaries, others argued ethnographic considerations should be paramount.¹¹⁵ The director of surveys, Raymond Alan, later laid out the “scientific” argument against Sir Frederick Jackson’s preference for foothills that often corresponded to the environmental ridges of clan boundaries: “As a surveyor, I prefer watersheds as the latter are definite and ascertainable and the former [foothills] cannot be determined by anyone and are, therefore, purely artificial.”¹¹⁶ Conflicts over the meaning and placement of boundaries often pitted imperial geographers against the “men on the ground” responsible for local governance.

At the turn of the century, Sir Harry Johnston, special commissioner of the Uganda Protectorate, and many other men on the ground favored the gradual amalgamation of the two protectorates and thus petitioned for a boundary that would entail the least disturbance possible to ethnolinguistic groupings. A flurry of correspondence in 1901 argued for the “well-known boundaries” that had been secured through barazas, meetings with African elders and clan leaders.¹¹⁷ For reasons obscured in the historical record, in 1902 the Foreign Office backed the “natural frontier” proposed by Sir Clement Hill in London, despite protests from the men on the ground that this new frontier “did not readily coincide with tribal boundaries.”¹¹⁸ Although the scientific arguments for “natural” boundaries prevailed, even these features remained disputed. In his later economic study of the region, Hugh Fearn argued that there was greater territorial logic in using the Nandi escarpment as the new boundary, placing the Bantu tribes in Uganda and the Luo in Kenya.¹¹⁹ Imperial debates over the logic of boundaries highlighted the conflicts and contradictions of colonial rule in eastern Africa.

It again fell to C. W. Hobley to demarcate the new boundary of the North Kavirondo District. In theory, this new district would contain all the Bantu tribes northeast of Lake Victoria, bordered by Uganda to the west, the Luo to the south, and the Kalenjin and Maasai in the Rift Valley to the east and north. In reality, the “Hobley line” ran through Lake Victoria, along the Sio River in the south, and jaggedly over Mount Elgon in the north, effectively dislocating the Samia community around the Sio River and severing the closely related Bagisu and Bukusu around Mount Elgon.¹²⁰ Indeed the Hobley line cut through and across many of the curving ethnonyms he himself originally mapped in 1898 (fig. 1.11). This reorientation transformed the Nyanza region into a borderland, a space thoroughly caught between two colonies that were rapidly differentiating in terms of local governance, European settlement, and African rights. Interterritorial disputes over the exact limits of this border persisted throughout the colonial period and well into the postcolonial era.¹²¹ This process of remapping would have profound effects on the alignment and geographic imaginations of political communities in both territories.

Despite the 1902 boundary agreement, as late as 1927 the boundaries of North Kavirondo remained in flux.¹²² The creation of these boundaries first occurred on paper, redrawn over the detailed maps of the Uganda Protectorate, with written descriptions of boundaries circulated in British proclamations. Surveyors and administrators then set out on boundary tours, enlisting African laborers to carry heavy stones for cairns, erect large stone pillars, set beacons, and dig trenches across the new interterritorial boundary with Uganda and against the expanding white highlands.¹²³ Colonial administrators would then take elders, headmen, and local villagers out to “beat the bounds” of the new boundaries, using drums and ornamental ceremonial flair.¹²⁴ In this way, the imperial instruments of territoriality imposed a top-down cartographic spatial ordering while enlisting local communities to invest in the construction of these boundaries.

FIGURE 1.11. Map of the East Africa Protectorate, 1902. Beachey, History of East Africa, xii.

Officials used survey maps, the construction of roads, and the collection of taxes to construct and consolidate meaningful boundaries. As the colonial administration envisioned Nyanza Province as a potentially rich source of labor and cash crop production, the enforcement of boundaries became a tool in the territorial control of work and local production. With the arrival of the railway, the colonial administration introduced cotton and maize as new export commodities. While cotton production foundered due to local resentment and low prices, maize proved a profitable and exportable cash crop, quickly overtaking traditional crops such as millet, sorghum, and cassava as the staple food of African diets, particularly for laborers.¹²⁵ North Kavirondo was fast becoming the “granary of East Africa” and the largest pool of potential laborers in the young colony.¹²⁶

Colonial officials also used borders and taxes to direct the flow of labor. Roads in North Kavirondo emanated out east and north from the administrative base in Mumias to direct the flow of labor and commerce away from Uganda and toward work on the railways, new European settlers’ farms in the highlands, and government public works projects within the limits of the new colony. Colonial commissioners worked to “impress all people with the necessity for their young, unemployed men going out to work.”¹²⁷ Much to the frustration of colonial officials, African laborers from North Kavirondo negotiated seasonal contracts to fit agricultural cycles and displayed a preference “to work month by month and their dislike of definitely binding themselves by a written contract.”¹²⁸

Many defied colonial demands on their labor and resources by migrating across the still poorly defined Ugandan border. Farmers on the Ugandan side of the border similarly evaded forced cotton cultivation and military conscription by crossing into western Kenya.¹²⁹ These movements were not without risk and insecurity. Kenyan and Ugandan officials responded with joint taxation collections and punitive actions against clans on either side of the border.¹³⁰ In 1917, as Kenyan officials raised hut taxes to fund World War I efforts and pressured chiefs to supply constant labor for the Carrier Corps, fifty Wamia families were forced to return to North Kavirondo from Uganda, where they had fled.¹³¹ The administration also put strict restrictions on the movement of cattle, seeing any form of pastoralism as “opposed to social or political advancement” and using boundaries to control the spread of sleeping sickness and rinderpest.¹³² Administrators introduced “census books” for the registration of residents within a given boundary to help control movement and enforce the payment of taxes.¹³³ While chiefs went around their territories counting huts, they also subverted these processes by offering refuge to competing communities. In the case of Chief Sudi, his entire census had to be thrown out as he was found to have collected taxes and census data on villages on Mumia’s side of the border.¹³⁴ Border patrols and the introduction of identity permits in the form of the kipande reinforced territorial boundaries as a central feature of colonial governance.

After quashing the final throes of outright resistance, in 1908, British officials set about demarcating internal boundaries and an effective local administration. That same year, Geoffrey Archer, acting district commissioner of North Kavirondo, began his demarcation tour.¹³⁵ Archer found the undulating landscape of North Kavirondo, with its “many fixed points and much open rolling grasslands,” a “good training ground” for practicing his surveying skills.¹³⁶ Archer enlisted local inhabitants as porters, guides, and aides and taught them lessons in cartography as he triangulated locations and called out instructions on the placement of markers. Archer gained a strong reputation for managing clan disputes across the interterritorial boundary and securing local support for “unsatisfactory” boundaries, earning him the difficult job of delimiting the Northern Frontier District and later the governorship of British Somaliland at the young age of thirty-two.¹³⁷ Although officials professed a desire to make administrative boundaries coincide with the “tribal” areas mapped by Hobley, Archer struggled to collect accurate clan numbers and to consolidate boundaries along the lines of “native laws and customs.”¹³⁸ Archer complained that the “Kavirondo are the most pronounced land grabbers.”¹³⁹ Like Francis Fuller among the Asante in the Gold Coast, Archer believed that boundaries fixed to a topographical map would make sense of the confusion of local customs;¹⁴⁰ and yet, as Sara Berry argued, “However precisely they were drawn on paper, boundaries could be remarkably elusive in practice.”¹⁴¹

The Wanga royal family and their emissaries invested early in the ideology of boundaries and acted as surveyors in the creation of the basic units of authority and territorial control within the district. Nabongo Mumia and his half brother Murunga were the only Africans officially consulted on Archer’s demarcation tour.¹⁴² Archer sent out Wanga chiefs as territorial agents to construct “locations,” the smallest administrative unit in the district. In this initial demarcation, only eight locations were drawn around the multiple communities of the region. Throughout the demarcation tour, the Wanga proved not only their usefulness but also their ability to benefit from the colonial processes of mapping. As later reported by Provincial Commissioner C. M. Dobbs, Archer gave “Mumia the biggest sub-district as . . . he alone had ‘capable men who are fit to be appointed as headmen over the various sub-divisions of this area.’”¹⁴³

Between 1904 and 1909 the British elevated their Wanga allies to the position of chiefs over the diverse range of clans never previously subject to Wanga power. On 15 November 1909, the colonial government confirmed their alliance with the Wanga by appointing Mumia “paramount chief” over the entire district. This official title would come to haunt the British administration as later African politicians attempted to claim the legitimacy and authority of a paramount chief. Colonial officials reinforced the new “native authority” of Wanga chiefs with the power to arrest, issue orders on the movement of people, compel labor, preside over local disputes, and collect taxes.¹⁴⁴ In the Wanga the British found the local source of “indirect rule” they needed to enforce colonial boundaries.

However, written into the very processes of boundary demarcation and colonial authority were the tools of its subversion for local actors. Countermapping strategies reflected both local geographic practices and the adoption of colonial technologies of mapping. The most commonly used tactic, as had been the case for centuries, was evasion. Local inhabitants, particularly along the interterritorial border, strategically moved throughout geographic networks that extended around and beyond these new borders to evade tax collection, to defy new authorities, and to confuse colonial officials. These cross-border movements took on cyclical patterns and created new networks later used in illicit trade and movements of rebellion. Others, particularly in the north and east, found refuge in the hills and mountains that provided geographic protection from the reaches of the state. While these movements built on and created networks of clan associations, they also produced new sites of conflict, as different communities responded to colonial impositions and incentives in diverse ways.¹⁴⁵ Border clashes in the early years of colonial rule continually disrupted the work of colonial surveyors and often arose strategically to pressure colonial officials, who often only addressed the complete demarcation of locational boundaries as a result of these local “boundary riots.”¹⁴⁶ The construction of colonial boundaries was a messy affair, continually interrupted and confused by the activism of local communities.

Local activists also used previous geographic knowledge and their early lessons in cartography to subvert the work of colonial surveyors, through subterfuge, trickery, and outright manipulation. Africans destroyed cairns, moved beacons, and uprooted pillars to move the colonial boundaries set by surveyors. Often the same men who carried and placed the boundary beacons for the British during the day were suspected of, or boasted of, removing them by night.¹⁴⁷ In one example, Wanga chief Murunga was said to have warned the Bukusu that the British wanted to cut them off from land in the northern area of Kamakoiya. Murunga advised the local population to take the beacon and move it, using the British officers’ lack of mastery over pronunciation of river names to trick them into accepting the new boundary.¹⁴⁸ The arrival of new European settlers in the Trans Nzoia lands north of the Kamakoiya River after 1912, however, would again push this boundary back and constrict the northern expanses of the district.¹⁴⁹ The environment, too, revolted against the symbols of boundary demarcation as grasses and shrubs overgrew beacons and hid boundaries from view.¹⁵⁰ These countermapping strategies became so pervasive that in 1911, British officials proclaimed harsh penalties for the “destroying or moving or diminishing the utility of any Land Marks fixed by public authority.”¹⁵¹ Colonial officials suggested the use of increasingly permanent markers such as “iron posts sunk in cement” and cairns alongside survey pegs within African farmlands that more starkly mapped the landscape.¹⁵² However, the symbols of surveyors continued to prove susceptible to the countermappings of African activists.

Colonial surveyors thus encountered what Raymond Craib has termed the “fugitive landscapes”: territories characterized by multiple and overlapping geographic systems that were not landscapes at all but “places created and recreated through the prisms of memory, practical wisdom, use and collective decision making rather than the lens of instrumentation.”¹⁵³ This confrontation between competing geographic imaginations of place and space would not result in a total victory for imperial cartography. Hungry for novel technologies and adept at adapting them to their own purposes, Africans found in mapping just such an instrumental tool to make claims to new landscapes of power and resources. By World War I, the number of locations had expanded to eighteen, mostly as a result of protests and the disaggregation of recognized ethnic communities, though more than half the lands and diverse communities of North Kavirondo remained under Wanga rule. These boundaries did not succeed in imposing a “master narrative” of authority and exclusionary rights.¹⁵⁴ The lines drawn by Hobley and Archer, while fugitive and continually contested, did, however, provide a new cartographic grammar for future debates over governance and the territorial horizons of ethnic communities.

. . .

IN THE making and unmaking of competing territorialities in western Kenya, early settlers built autonomous clan structures within niche environments and mapped larger networks of interaction, exchanging goods, services, people, and ideas across this uneven terrain. While the fixed and rigid maps of colonial geographers imposed new fields of conflict and debate onto the diverse communities of North Kavirondo, they also introduced new ways of “writing the world,” new tools for competing over local resources, and new strategies for resisting imperial impositions. Local activists combined these early lessons in cartography with local geographic knowledge and spatial strategies to countermap the hegemonic project of imperial geography. In an ironic twist, it was perhaps Hobley’s first map—with its curving, overlapping, and expanding ethnonyms—that most directly reflected the heterarchy and shifting geographic relations of communities in western Kenya. As colonial officials worked to consolidate their rule, the need for more clearly demarcated lands and more hierarchical forms of political authority prompted local political activists to rework and reimagine their traditions of geographic and political community. The grounds of authority and community in western Kenya, it seemed, would remain “slippery,” and defiant to the recognition of “kings.”