98 Pula: Botswana Journal of African Studies vol.16, no.2 (2002) 'Mapping the land' in Gudigwa: a history of Bugakhwe territoriality Michael Taylor, United Nations Development Programme. E-mail: taylors@botsnel.bw. Drawing largely from the experiences of a land mapping exercise in the ancestral lands of the residents ofGudigwa in northern Ngamiland, this paper explores the changing ways in which land ownership have been structured. Itfollows the decline of tenure over land and natural resources by Basarwa, a process that began after the first Bantu-speaking immigrants became established in the area. This analysis challenges the fallacy that Basarwa only had a loose sense of territoriality, and examines in detail the nature of local conceptions of land ownership, and the salience of these conceptions to present struggles over land. I On most maps of Botswana, the large green triangle of the Okavango Delta sits conspicuously in the centre ofNgamiland. On larger scale maps, the complex topographical features are faithfully displayed in great detail; the myriad perennial and annual waterways, as well as floodplains, tracks, islands, safari camps, along with the variously-ascribed names of each of these places. The large expanse of sandveld north of the Okavango Delta -which I term the northern sandveld-is in contrast conspicuous only by its emptiness. Apart from a few sites on old trade routes, and old tourist camps, the northern sandveld is represented as featureless and without name. To those with the power to make maps, the northern sandveld may not carry the hydrological or tourist interest of the adjacent Delta. Nonetheless, its intricate surface of sands and soils waterholes trees and animal routes, is intimatelyknown and named by those who have liv~d in it. ' The northern sandveld ofNgamiland was the location of the author's doctoral fieldwork for 18 months in 1997-8 (Taylor 2000). The northernmost village in this area is Gudigwa, whose residents are almost entirely Bugakhwe San. This paper is based on an exercise the author undertook in June 1998 with the residents of Gudigwa called 'mapping the land'. 'Mapping the land' involved attempting to map the names of their ancestral lands, an area of about 7,000 square kilometres. This was the land in which the ancestors of the residents of Gudigwa had lived, before they started congregating over the past two generations. As we drove and walked through the land, what we discussed and recorded was much more than an inert series offeatures that can be reduced to colours and lines on a map. We also started ~ncovering a 'landscape of semiotics', to use Moore's (1993:396) phrase--a land that carnes.a web of shared and contested meanings between the many different people who have expenence and interest in it. This pa~er uses the experiences of the 'mapping the land' project as a basis for discussing the changmg patterns of land tenure by Bugakhwe (Khwe-speaking Basarwa) in ~he ~orth~m sandveld through different historical dispensations. Augmented by extenSIve mtervtews alongside historical and archival documentation I trace how land tenure has changed through. different historical dispensations, particuiarly with the immigration of Bantu-speakers m the nineteenth century. Alongside this historical commentary ~e contemporary statements, showing the way that this history is understood and related m contemporarycontexts. 99 Introduction to Gudigwa The village of Gudigwa was created in 1988, through the Remote Area Development Programme's policy of providing service centres in remote areas, to promote the aggregation of scattered settlements. It was made up mostly of people who moved there from IIGamlwi, and Letshaobe, both nearby settlements. Each of these settlements was in turn made up of people who had moved together at different times in the preceding decades, from their different family areas that had covered much of the northern sandveld. Although it has a population of over 600, Gudigwa does not give the impression of being a village of this size (Plate 1). From anyone place in the village, no more than a handful of huts were visible. Gudigwa could be more accurately described as a cluster of small villages spread over several kilometres, with each section (of the ten families that came together to form Gudigwa) separated by a band of trees. The layout of Gudigwa is therefore a spatial expression of both the centripetal forces arising from the advantages of living together, and the centrifugal forces inherited from a history of smaller units of social organisation. Gudigwa is now the only settlement remaining on the land that its residents claim historically as their own. As such, it represents the culmination of a process of gradual agglomeration that has been taking place over the past two generations. Kinship bonds extend between these family groups, uniting all the residents of Gudigwa, so that each is able to claim at least some form of relation to the others. All the residents of Gudigwa consider themselves Bugakhwe, making Gudigwa probably the largest wholly Basarwa village in Botswana. 'Mapping the land' Mapping their ancestral land was an idea that arose from discussions on how to motivate for the allocation of an area of land under the new Community Based Natural Resource Management (CBNRM) programme. The residents of Gudigwa were unhappy at being combined in a Community Trust with four other villages; Seronga, Gunitshoga, Ereetsha and Beetsha, each of which is dominated numerically and politically by Hambukushu and/or BayeL The Trust, named Okavango Community Trust (OCT), was allocated two Controlled Hunting Areas (CHAs), NG22 and NG23, which were a distance from Gudigwa itself (Map 2). The residents of Gudigwa felt that the other villages did not respect them, being Basarwa, as equal partners in OCT, thus denying them a fair share of the benefits. They felt this especially sharply, as they had given up their Special Game Licences, which allowed each family to hunt a quota of animals, in exchange for a Community Quota, given to OCT as a whole. All in all, people in Gudigwa felt that they had sacrificed more, and benefited less, than the other villages in OCT (Taylor 2002). They therefore wanted to secede, in order to form their own Community Trust, and be allocated their own CHA. Debating which areas they could be allocated under the CBNRM Programme became a discussion of the different boundaries that had been drawn across their land, and their potential entitlements within them. They spoke of their own historic family boundaries, of the ten families that had come together over the last two generations to form Gudigwa. In places these boundaries were crosscut by the colonial ones now marked between Botswana and the Caprivi Strip by an electrified triple fence. There were the boundaries on the map drawn up in 1968 that divided Ngamiland into concession areas for citizen hunting and safari hunting, determining where residents could hunt the animals listed on their Special Game Licences. These fell away and were replaced with new divisions in 1991 when Ngamiland's concession areas were redrawn, dividing the whole district into 49 CHAs, some of which could be allocated to villages for management under the CBNRM Programme. Finally, there was the very real boundary created by the northern buffalo fence, 100 passing six kilometres east of Gudigwa (Plate 3). Built in 1991, and extended in 1997, it divides land in which cattle are allowed (north and west), from land in which they are not (south and east). The fence also affects livelihood potentials (from both subsistence hunting and tourism) by blocking migration routes of ungulates, thus reducing wildlife numbers on both sides ofthe fence, but particularly on the cattle side (Albertson 1997). These were all boundaries that Gudigwa had to consider in their struggles for land. Doing so involved, in part, going back to the only boundaries that they had themselves instituted; those of their own family lands. This was initially an oral process, one that arose in discussions of entitlement to land. Such discussions often led to people referring to the names of their ancestral lands, listing the names of the different areas as they mentally walked through them. Expounding such oral maps was a local strategy akin to more formal techniques of drawing maps known as 'counter-mapping' (Peluso 1995). Counter-mapping has come to be a popular tool by which political movements and community groups attempt to counter dominant representations of property regimes and landuse practices (Poole 1995).Most residents of Gudigwa had never seen any maps of their area to be aware of the big expanse of blankness around their village. Neither were they aware before our discussions of the potential of creating their own map to promote their own land rights or provide the basis for a possible management plan for their area. While paper maps may not have otherwise meant much in a generally non-literate society, the residents of Gudigwa were very aware of the power of written information and images about their land. Thus was the idea of producing a 'counter map' of Gudigwa's ancestral land born. Masarwa Community, the committee they had set up themselves to motivate for their own CHA, arranged for members of each family to go out with me to their respective ancestral lands and plot important sites of settlement, boundary, subsistence, water and travel. Three intense weeks of driving and walking produced a map covering an area of about 7,000 square kilometres, in which we recorded 454 place names. Driving through land with no roads or tracks was a difficult process. Most of the time we fol.lowedthe well-used elephant paths that formed a network linking each waterhole to adjacent ones. Elephants did the work of maintaining a gap through the bush, which made these paths the easiest routes for animals and people on their journeys. Known as 00 dao [00 - gap, dao - path], these paths, with waterholes at their intersections, were the spines that made up the mental maps of the landscape, upon which hung other areas of importance (P~ates ~ and 5). Except for a few isolated safari hunting camps, this land is now umnhablted. Yet, people who have now moved to Gudigwa lived in parts of it recently enou~ so that there are still sites where wells have not caved in, roof support poles are standmg, and where the debris of life, such as shoes, knives and ploughs, can be found (Plate 2). By sharing knowledge of the land-not only its names, but also the location of fruitful areas and sites where water was close to the surface--I was inserted into Gudigwa's ne~ot~ationsof power over land; given knowledge that could be used against, as well as for, their mte~est~.The issue .of land was without doubt a very sensitive one locally, illustrat~g not only Its Importance m the present, but also their painful history of alienation from It. Basarwa were very aware that in showing the land to others in the past who had then taken control . of S::sadm"InIstrators,. hunters, conservationists-they '. it-ch' lei:, . I'Icated had been Imp In the alienatIOnof their own land.2 . Maps represent views of the land, sites and channels of interest to those with the power to m~ence mapmaking. They therefore speak of authority and power not just to name, but to m e those names heard. This is a form of power generally held by the state which can use mapstom ffi' I ' ore e ectIve y control a peasant or tenant population (Harley 1988:284, Scott 101 1998). Creating a map for Gudigwa's own use thus became an exercise in appropriating the 'power to nominate' (Parkin I982:xlvi), and asserting the rights that should come from their own power to name the land. Travelling through their family lands also presented numerous opportunities to discuss their relationships to land, both past and present, which provides a basis for the remaining sections of this paper: the historical nature of Basarwa land tenure; the impact on land tenure systems of non-Basarwa immigrants; and contemporary attitudes to land tenure. Historical land tenure The nature of Basarwa land tenure has been a contentious issue, especially as it has come to bear on issues of contemporary land rights. In facing the very real issue of the erosion of land rights for Basarwa, Wily (1994:8) argues that 'the most urgent need at this point is simply for the state to recognise, once and for all, that contrary to local opinion, Basarwa did own land through their customary tenure system'. However, concepts such as 'own' and 'territory' are socially related concepts that have little meaning unless embedded in the social systems of which they are a part (Carstens 1983:60). The process of writing land tenure systems as an institution can give the impression that they existed as a coherent set of rules and practices, which they did not (cf Peters 1994, Neumann 1997). There were, nonetheless, commonalities-subject to some debate--in the way land tenure in the past was spoken about, and it is both these commonalities and the contemporary debates over them that are examined here. While there are differences in the tenure systems of the many people that have become subsumed under the umbrella 'Basarwa' (Barnard 1992a:223ff), the most important common element is that entitlement to land has been mediated through social relationships (e.g. Cashdan 1977:22-4 for Gllana; Heinz 1972 for !X60, Barnard 1979, 1980 for Naro; Silberbauer 1981:99 for G/ui; Lee 1979:333-43 for JuI'hoansi). People therefore became associated with geographic space through their position in a social network. Ownership was therefore neither individual nor absolute, but negotiated by individuals through relationships at a community level; a principle that applied to tenure systems across southern Africa, Basarwa and Bantu-speakers alike. These similarities made tenure systems between Basarwa and their neighbours mutually intelligible, and, to some extent, complementary, so that their landholdings could overlap where necessary. Figuring entitlement through a social matrix that extended over many hundred of kilometres enabled migration over wide areas. The origins of some of the elderly residents of Gudigwa, for example, range from southern Angola and western Zambia to the north, and beyond the Okavango River to the west. Many such people gained, or concretised, entitlement to lands in the northern sandveld through marriage, which-as with other Khwe speakers (Barnard 1992a: 127, Silberbauer 1981:I48}-tended to be with cross-cousins. The land that people in Gudigwa considered their own was divided into ten family 'territories', as illustrated on the map produced from the land mapping (Map 3) These were not, however, absolute territories belonging to absolute families. Our land mapping exercise captured land ownership as it was figured in a specific historical moment; memories at the end of the twentieth century of lived patterns from several decades previously. Relations were, and are, negotiable, and consequently entitlements to land, and the way land was divided, also changed. The average size of the seven territories whose complete borders we were able to map (counting Amos and Sangando's land as two) was 915 square kilometres (Table 3.1). Although this figure is significantly larger than the 300-600 square kilometres that Lee (1979:334) estimated for JuI'hoan n!oresi in Ngamiland's western sandveld, the per capita 102 area of territory is similar; roughly 30.5 square kilometres per person for Bugakhwe in Gudigwa, as compared to Lee's 24 square kilometres.3 The larger overall territory size for Bugakhwe may be partly due to the gradual agglomeration of land that has taken place. For example, the largest of the territories mapped was that of Amos, which in the past was considered separate from Sangando's land. Due to the close relationship of the two families, they have used each other's land, and over the past generation, have congregated for extended periods in common villages such as Ghoi and Letshaobe. A similar process appears to have occurred in the Kgalagadi District, where Wily (1974:21) reported an average territory size of 1,500 square kilometres, but each with about 150 claimants. Land was also divisible, as relationships changed. IIAe/exo's land, lying to the west of Kharakhwe's, provides a case in point. In the early 1900s, his father, Borakanelo, migrated down from Bwabwata on the border between West Caprivi and Angola. Borakanelo took Kharakhwe's classificatory sister as a second wife. Kharakhwe in turn took Borakanelo's daughter, and IIAe/exo's half sister, Kwima, as a wife. As a result of this double alliance, Kharakhwe gave Borakanelo the western portion of his land, which is today still regarded as belonging to Borakanelo's descendants, rather than Kharakhwe's. The Bugakhwedam and Ts'exadam word for land is ngu, which Cashdan (1977:22-4) translates for IIGanakhwe as 'place' or 'territory'. This is similar in meaning to the JU/'hoansi word n!ore, which Wilmsen (l989a:51-54) argues carries connotations of belonging, and translates as 'place in land/country' (l989b:162). However, the Bugakhwedam ngu is closer in meaning to the Setswana lefatshe, meaning simply 'land', rather than 'territory'. There is no specific, and commonly used, word that implies land belonging to particular people.4 When referring to a family territory, Bugakhwe use the possessive construct 'my land' (ti da ngu a), or more commonly, 'the land of my father' (mha m da ngu a), emphasising the relational element of establishing rights to specific territorie~. Rather than being framed in terms of 'my land', claims to specific tracts were communIcated to me as, 'This is the land my mother's brothers and grandfathers showed me when I was young, saying, "See, my nephew, this is our land, this is where you must look for food"'. In.referri~g to the 'owner' of the land, the Bugakhwedam possessive construct di ma (m) or dl sa (t) IS used. Ownership was ascribed to a living member of the family group (only ?ne of whom in this case - Taetso - was a woman), as Silberbauer (1981) reported of Gllana m central Kalahari. These owners themselves however often referred to their land as belo~ging to relatives of their parents', grand~arents' o~ great-grandparents' generation. Movmg through their land provided a continued link with these ancestors through the op~ortunity of asking for their assistance and protection. For example, as we camped in t~elr land, Amos gave a mug of water to his older classificatory cousin Sangando, and told him to phekola (bless) the land. Sangando sprinkled the water around him, saying: I/Gilwa Ilgawe yalle lie ngiyae tshoo a 1100da oro ke 1100tuwoko tshoo /weiWe. IAo tee ngu hi 1100ko xo Iu ko. Tee hi ngu o. llEe lieu ko l/we uu lie yoo XQ ka ku, 1T17Tr1' 1T17Tr1'. Sekundeko, Puduhudu. O~r ancestors, come. We are going to bless the water in your waterhole from which we drink, let the ram f~1 and fill this waterhole. We are now in your land, the one in which you ate. We are in your land. Go WIth us and look after us tomorrow ... Sekundeko [Sangando's classificatory uncle], puduhudu [Sangando's father]. In talking about these ancestors, they were generally not referred to as Ilaxa (chiet), but di ~ ngu.a (owner of the land), or di XII //'ae a (owner of the village). These people had.a relationship to their land referred to by the verb II'ae in Bugakhwedam, and oil In 103 Ts'exadam. The closest translation is the Setswana tenn go rua, literally meaning to possess, but with the implication of benefiting from the productive nature of what is owned. If one moved away and stopped using the productive assets of a particular tract of land, title to it was lost, and it was free to be taken over by another family group. Boundaries between family lands were, and are, not without contestation. Perhaps disagreement is intrinsic to territorial claims, and has certainly been common in other Basarwa claims to land (e.g. Heinz 1994:94 for !X60 Lee 1979:334 for Ju/'hoansi). Even as we mapped their areas, arguments arose in adjoining areas as to which land was whose. The boundaries between territories were often imprecise (cf. Lee 1979:334-5), and sometimes overlapped. At other times, clear topographical features functioned as borders, such as elephant paths (00 dao) or fossil riverbeds. Such boundaries, however, were not presented as lines of separation, but lines of meeting. The Setswana word used to refer to them was mokopano, meaning 'place of meeting', rather than the more common tenn kgaolo ('district'), from the verb go kgaola ('to divide/separate'). Important resource areas along these boundaries, such as waterholes and fruit collecting areas were shared (a common attribute of Basarwa territories; see Barnard 1992a:235), and neighbours would sometimes move together for a period at such places. These physical 'boundaries' were thus-as Barth (1969) argued of social boundaries-places of social connection; spaces for the construction of relationships rather than places of separation. Water was paramount in figuring patterns ofland ownership. Virtually without exception, all the names in the land are of waterholes. In addition, each family territory bordered a supply of permanent water, either the lagoons and floodplains of the Okavango and Kwando/Linyanti river systems, or the pennanent groundwater of Dishokora, also known by its Bugakhwedam name, BienlwlI. Situated in one of the east-west valleys just south of the Caprivi border, Dishokora was the meeting point of about five different family lands; the centre from which the 'petals' of territories radiated, giving each one access to its water. It was a place of congregation when all the other waterholes in their land had dried up. Early commentators in the Okavango were struck by the high degree of exclusivity that marked Basarwa territories. 'No Mokuba [Moyei] or River Mosarwa is allowed by his neighbours to poach outside his own district', wrote Stigand (1923), and Doman (1925:85) observed that, 'each family group had its own hunting ground and bitterly resented the intrusion of others, either native or European'. As these comments imply, what was guarded was not so much the land itself, but the useful resources on it, which included waterholes (or lagoons in the swampland), gathering areas, honey and wildlife. As with G//ana in central Kalahari (Cashdan 1984:447-9), and Tyua in northeastern Kalahari (Hitchcock 1995: 177), some of these resources could also be individually owned. These could include dispersed resources such as melon groves, or specific point resources, such as an anthill that Amos pointed to south ofN/omn/om that had belonged to his father. Using the resources of another's territory was possible if permission was sought and granted, which it often was, especially for water. Water was so essential that it had to be held loosely and shared with anyone who, if necessary, would go through the formality of asking. Less essential resources, such as groves of fruit or nuts, or honey were often more exclusively held. Informants today maintain that individuals who asked such permission, whether Basarwa or Bantu-speaking, were accompanied by a member of the family who owned the land to ensure there was no abuse, such as hunting juvenile animals, picking unripe fruit, or over- harvesting. Such practices are termed f'ae u (/'ae-destructive, u-hunt[ing)), and individuals who made this a habit were denied continued access to these resources. Both Basarwa and Bayei informants stated that, in the distant past, any Bayei and other 104 (unrelated) Basarwa found hunting or gathering without permission could be beaten or killed (confirmed by Schapera 1930:155-9). While specific localities belonged to specific families, kinship ties created larger 'clusters' of territories into which access was guaranteed for members, and a united front could be presented against incursions. Doman (1925:85) was told that '[Basarwa] clans combined in former times to resist other tribes who threatened their control of land'. Such rigid control, however, began to decline with the immigration of Bantu-speakers. Immigration and the decline of tenure The entry of different waves of Bantu-speaking immigrants into Ngamiland from about 1700 marked the beginning of the end of the strong controls over land by Basarwa described by the likes of Doman. This was a common fate of Basarwa land tenure systems throughout southern Africa. For example, Heinz (1994:94) documented this of !XCIland tenure, once their land became an important cattle trek route. In the eyes of many Basarwa, this was the beginning of a single continuous process of gradual land alienation that has continued until today, thus affecting perceptions of current programmes, like CBNRM, that touch upon issues of resource tenure. Take, for example, the impassioned account of changes in land tenure given to me by Petros, an old blind man from Khwai (which I have arranged in chronological order): Longago, if people wanted to hunt in my [ancestors'] area, they carne and asked first. We told them not to finishthe animals, but to take a few, then go back home. When Bayei came into our land, we showed them our animals, but they started finishing them. So we refused to let them hunt anymore in our area, to finishoff the few animals that were left. Ifwe found a Moyei hunting, we would beat him with sticks so that he wouldn't come back, then let him go. Whenblack people came into our land, they asked for land to keep their cattle in. We showed them and they kept it...Batawanacame and took our land and animals without asking. They let us keep our food [Le.access to land and wildlife], they just said they would eat it too, not like the government [of today]that has taken them away altogether. Today all our land and all our food has been kidnapped by the government, without thought for how we will live and stand up. Today the government has struck us down. Many of the initial Bantu immigrants into the northern sand veld, particularly Bayei, ?ITived via the waterways of the Okavango-Linyanti-Kwando system (Tlou 1985: 15). The Immigrants lived along the waterways in semi-permanent villages with summer sites and winter sites, to avoid the annual floods. //Anikhwe who also lived along the riverine fringes,S felt the impact of immigration most. In cont;ast, Bugakhwe at Gudigwa maintain th~t their preferred residence has always been the sandveld. Both Gibbons (1904:207) and Semer (1909) confirmed this pattern, with their observations that the sand veld from O~avango to Kwando was the domain of Basarwa. On the map he produced of his travels, Semer (1909) labelled the northern sand veld Hukwe- Ve/d. 'Xukhwe' is the Bugakhwe name for Basa,:"a. Like Ju/'hoansi in western Ngamiland (Lee 1965: 198) and //Ganakhwe near the Botetl (Cashdan 1986), Bugakhwe spent the wet season dispersed in the sandveld, and moved to permanent water Sources (springs or rivers) in the dry season. For some Bugakhwe gro~p.s, Dis~okora was the point at which they congregated (which by 1900 had ~ambukus~u hvmg at It), for others it was along the riverine fringes. Villages along the fIvers prOVIded an opportunity to trade skins, honey, meat and wild fruit in return for toba~co, ~annabis, agricultural produce, pots and iron. In many respects, therefore, the r~latlonshIP. ~~tween Bugakhwe in the sand veld, and their Bantu neighbours along the fIvers, was InItIally mutually beneficial. 105 By the time of the fIrst census of Ngamiland in 1921, the riverine fringes around the northern sandveld were scattered with villages. The census estimated 1,500-2,000 Hambukushu lived down the eastern bank of the Okavango from Mohembo to Gabamukuni, most likely an underestimate considering the difficulty of access (Stigand 1923:412). By this time, however, northern Ngamiland was a mix of not only Hambukushu, but also Bayei, Bakgalagadi, Basubia, Banajwa, and, of course, Bugakhwe and //Anikhwe. The largest settlement was Gabamukuni itself. Situated in the middle of the northern arm of the Delta, it was a cosmopolitan cluster of villages where Bayei, Hambukushu, Bakgalagadi and Bugakhwe lived in close proximity (Stigand 1923). Hambukushu and Bayei villages also sprang up along the southern banks of the Kwando and Linyanti, where a 1934 census enumerated 474 Bayei and Hambukushu (including nine Batawana) in 14 villages (BNA 1934). The influx of Bantu-speaking immigrants did not at fIrst substantially change the spatial layout of existing territories. Basarwa continued to adhere to these, and a patchwork of tenure arrangements grew, forming a complex set of overlapping rights. Areas were delineated that were open access and uncontested, others that were managed to restrict access to some degree, and yet others that were effectively private (cf. Scoones 1995). As Bantu-speakers became more established, their rights along the rivers superseded those of Basarwa. Nonetheless, principally by virtue of their intimate knowledge of the sandveld, the strength of Basarwa tenure in the sandveld was not so easily eroded. Even the Batawana chiefs, who claimed the entire sandveld as their own hunting grounds, would hunt in conjunction with the Basarwa owners of the land in which they were hunting. In part this was a logistical necessity; people unfamiliar with its repetitive terrain could easily die of thirst. But it was also recognition of their unique form of power as fIrst people in the land, a supernatural power ascribed to occupying a liminal state between society and nature (cf. Gordon 1992:212-5). Batawana hunters (according to Basarwa informants) would ask Basarwa owners of the land they were hunting in to bless or charm (phekola) them to ensure their success, in the same way that Sangando did for himself and his companions as we mapped his land. Despite being able to maintain control over sections of the sandveld (especially those with no permanent water sources) there were times of crisis, when Basarwa were forced to enter servitude in order to survive. When Reid (1901) travelled along the Kwando in 1899, he described the Basarwa he encountered living with Hambukushu as 'the lowest of the low'. Perhaps in the wake of the rinderpest epizootic of 1897, the vulnerability of some Basarwaas they lost the relative independence afforded by the sandveld, was taken advantage of by their non-Basarwa neighbours. Through such experiences, and with the rise in political power of Batawana and select subject tribes, the patchwork of overlapping rights gradually tipped against the favour of Basarwa, to the extent that practices such as asking permission to hunt in another's territory, fell away. For Basarwa in the northern sandveld, therefore, the enduring consequence of Bantu immigration was dispossession: from their land, their resources, and often their labour. Contemporary attitudes to historical territories Basarwa live today in the northern sandveld in a very different context than that which gave rise to the pattern of family territories which we constructed as we mapped their land. Demographically, the most important change is that they now live in one single village, rather than each in their own family's territory. As such, there has been a corresponding shift from emphasising the ownership of individual families, to a more inclusive sense of ownership of the land as a whole, by the village as a whole. These two levels of locally 106 figuring ownership; on a family as well as a village-wide basis, overlap. Most people are familiar with, and use, both. However, the middle aged and elderly people who had life experience in these lands, were often those most keen to emphasise family territories, and they spoke with fondness of the land that they called their own. The gradual agglomeration of Bugakhwe at Letshaobe and //Gam/wi, and then at Gudigwa from 1988, was due to a variety of factors. The lower than average rainfall in the past three decades made reliable supplies of water harder and harder to find. This. was coupled with promises by the Remote Area Development Programme that, if they congregated in one place, they would receive essential services such as water, a health post and a primary school.6 Although no one was forced to move to Gudigwa, some families alleged that harassment by the Anti Poaching Unit of the Botswana Defence Force (BDF) left them with little choice. For example, Two-Boy's land lay between Gudigwa and Khwai, now in a eRA leased by the government directly to a tourism operator. He explained to me one day, as we sat in the hunting lodge where he worked as a tracker, how he had ended up moving from Four Rivers to Gudigwa: Ian Khama [thenheadof the BDF]cameto my house and asked for me, sayingthat the soldierswould end up shootingme, as they said I was harassingthe animals.But I thought, •If I am harassingthem, whyaretheystillaround?'Nonetheless,I brokemy villageand movedto Gudigwa Todaywe giveourselvesnothing,living insteadat the hands of white people. I didn't wantto fight withthesoldiers~ I am illiterateand don't knowhow to protectmyself:or my youngerbrothers andchildrenwhowerewithme. If they had not harassedme, I would stiIl be there. I want assurance fromthegovernmentthatI wiIlnot be harassed,and I wiIl return there. Todaywe can only liveby honey,a bit of work, and kiIlsfrom predators.But as a Mosarwa,I am theownerof the wildlife.The governmenthas raped us. Hambukushuand Bayei were givensorghum, butKhara!'urna[progenitorof ailBasarwa]was givenanimalsto look after. The land is my inheritance. If I had the choice,I would send my children to hunt for me. But today the government has kidnappedthe wildlife, and if we try to hunt we go to jail. Despite continuing to hold an attachment to the land they still considered their own, most resi~en~ realise that, despite their sentiments, they are unlikely ever to live in their family temto~les.again. The new set of official principles, priorities and laws over land, settlement and wIl~hfe make agglomeration in villages a virtually irreversible process. Although none of Gudlgwa's ancestral lands fall within land zoned as national parks or game reserves, much of it is zoned as Wildlife Management Areas, which restrict the growth of 'new' settlements. As Brown, a young member of Khwai's Interim Management Committee, told me, 'Ifyo~ want to move, you can move to an existing settlement, but not out into the bush. The bush ISfor wildlife'. ~~o-~oy'S response to the restrictions imposed by the government were aimed. at legitImating continued claim over 'his' land, as well as making a more generalise~ c~alm over the land as a whole and wildlife in it ('As a Mosarwa, I am the owner of the wIldhfe.,'. Kharal'~a w~ given animals to look after. The land is my inheritance'). It is thiS generahsed claIm of land ownership that is preferred by the younger people, who, alth~ugh they are often familiar with the boundaries of their family territories have a life expenence more rooted in the contemporary dynamics of living in a single viilage. One such person was Starvation, a younger member of both the Okavango Community Trust committee, and the newly formed 'Masarwa Community' committee. He was of the opinion that, 'Those who say different areas belong to different people are ba bog%g% [of the old ways], the ~h?le land belongs to me [as a Mosarwa]'. Such notions by Basarwa of land and their pIa« 1D It also reflect the position that they have come to occupy in an overarching political 107 economy that encompasses both them and their land, transcending the social and spatial boundaries of individual family territories (cfSuzman 1997:90). By the late 1990s, the boundaries between family territories of Gudigwa received little more than lip service, seen, for example, in Amos' compliments to Taetso for allowing, without complaint, the whole village of Gudigwa to live on her land. In practice, residents of Gudigwa hunted, grazed their cattle, and used the veld with no restraints to stick to specific band territories. The boundaries of family lands had been overtaken by the more recent boundaries imposed by the government. When I asked Ts'ima, the headman of Gudigwa who had lived and hunted throughout the northern sandveld, about hunting in another's territory, he responded, 'We didn't ask each other's permission to hunt. We just used each other's areas freely because we are relatives. It is the government that makes us ask to go places.' Conclusion: the centrality of land The 'mapping the land' project opened up unique opportunities to gain an insight into the changing tenure patterns of Basarwa in the northern sandveld over the land on which they have lived, and the way that these changes are understood and represented by its inhabitants today. In tracing these processes, one factor stands out: control over land is an absolutely central issue in negotiating development options today. In the words of Roy Sesana, then Chairman of First People of the Kalahari, to a delegation from the northern sandveld who visited his organisation, 'Our human rights are our land. They cannot do anything for us if they take us off our land'. Basarwa in the northern sandveld have, over the last two centuries, experienced a gradual loss of control over the land in which they live. Nonetheless, they enjoy a privilege that few other Basarwa have: being able to call at least some land their own. Such land remains common property, but they face comparatively low levels of competition for its use from non-Basarwa agro-pastoralists. Having land that is at least customarily recognised as under their jurisdiction (because there are no other non-Basarwa locally to claim it) gives them a sense of standing in the wider social economy that landless Basarwa cannot achieve. This principle is confirmed by Wily (1976:16), who observed that relations between Basarwa and non-Basarwa at Bere (in Kgalagadi District) improved after they were officially allocated a small tract of land, as it gave them a standing by being able to declare, 'we have a place'. Woodburn (1997) also observed higher levels of discrimination against landless Hadza in Tanzania than against those who were still able to assert a measure of control over land. Thus, the land of the northern sandveld is a tangible representation of the history and identity of many residents of Gudigwa. It is also a reminder of their alienation from not just their physical space, but from many of the markers by which they have come to defme themselves as Basarwa. Notes I. I am grateful to the University of Botswana and University of Tromso Collaborative Basarwa Research Programme, Wenner-Gren Foundation, Association of Commonwealth Universities and Conservation International for their support of the research from which this paper is written. 2. Most of the early white travellers through Ngamiland gave credit to the Basarwa guides on whom they were so reliant. The Resident Commissioner, for example, described Basarwa on a visit to Mababe in 1906 as 'the most useful to meet of all people' (BNA 1906a). 3. Based on an average occupancy size of thirty (my estimate) for Bugakhwe in a territory of 915 square kilometres, as compared to 18.9 for JuI'hoansi (Lee 1979:60) in a territory of 450 square kilometres. 108 4. The word n;#QQ is also used to refer to a place or area, but is not used in a possessive sense. Xom (literally meaning 'soil') can be possessive, but is not commonly used to refer to family lands (I am grateful to Matthias Brenzinger for a discussion on this topic). 5. Both Gibbons (1904) and Seiner (1909), who travelled along the northern rivers in 1899 and 1905- 06 respectively, noted that many of the small villages along the Okavango River were //Anikhwe. 6. Ten years later, however, the only one of these they had received was a regular supply of water. References Albertson, Arthur 1997. Northern Botswana Veterinary Fences: Critical Ecological Impacts. Unpublished report, Okavango People's Wildlife Trust, Maun, Botswana Bamard, Alan (1979) Kalahari Bushmen Settlement Patterns. In P. Burnham and R. Ellen (eds) Social and Ecological systems. ASA Monographs 18:131-44. Bamard, Alan (1980) Basarwa Settlement Patterns in the Ghanzi Ranching Area. Botswana Notes and Records 12:137-48. Bamard, Alan (1992) Hunters and Herders of Southern Africa: a Comprehensive Ethnography of Khoisan Peoples. Cambridge: Cambridge University Press. Barth, Frederik. 1969. (ed) Ethnic Groups and Boundaries: The Social Organisation of Cultural Difference. Oslo: Scandanavian University Press. BNA (Botswana National Archives) 1906a S3217. RC's visit to Ngamiland, RC to HC, Johannesburg, 28th June. BNA (1934) S96/4. Census of Natives by St Fox, 9th August. BNA (1940) S396/4. Caprivi Strip-Bechuanaland Protectorate Boundary. Question of Policing of.. Carstens, P. 1983. The Inheritance of Private Property among the Nama of Southern Africa Reconsidered. Africa 53:58-70. Cashdan, Elizabeth. 1977. Subsistence, Mobility and Territorial Organisation among the G//anaque of the Northeastern Central Kalahari Game Reserve, Botswana. Unpublished report submitted to the Ministry of Local Government and Lands, Gaborone. Cashdan, Elizabeth(1984).//Gana Territorial Organisation. Human Ecology 12:443-63. Cashdan, Elizabeth(1986) Competition between Foragers and Food Producers on the Botletli River. Botswana. Africa 56:299-318. Doman, S.S. (1925) Pygmies and Bushmen of the Kalahari. London: Seeley Service. Gibbons, Alfred St Hill. (1904) Africa from South to North through Marotseland (2 vols). London: John Lane, the Bodley Head. Gordon, Robert (1992) The Bushman Myth: the Making of a Namibian Underc/ass. Boulder: Westview Press. Harley, J. (1988) Maps, Knowledge and Power. In D Cosgrove and S. Daniels (eds) The Iconography of Landscape. Cambridge: Cambridge University Press, pp277-312. Heinz, H. (972) Territoriality among the Bushmen in General and the Bushmen in Particular. Anthropos 67:405-16. Heinz, H. 1994 [1966]. Social Organisation of the !Ko- Bushmen (ed. Klaus Keuthmann). Koln: Rudiger Koppe Verlag. Hitchcock, Robert K. (1978) Kalahari Cattle Posts: a Regional Study of Huntergatherers, Pastoralists, and Agriculturalists in the Western Sandveld region, Central District, Botswana (2 vols). Gaborone: Ministry of Local Government and Lands. Hitchcock, Robert K. (1995) Centralisation, Resource Depletion and Coercive Conservation among the ~yua of Northeastern Kalahari. Human Ecology 23(2): 169-98 .. Lee, RIchard (1965) Subsistence Ecology of the !Kung Bushmen. Unpublished PhD thesis, UniversIty of California. Moore, Donald (1993) Contesting Terrain in Zimbabwe's Eastern Highlands: Political, Ecology, Ethnography ~d Peasant Resource Struggles. Economic Geography 69:380-401. Ne.um~, Roderick P. (1997) Primitive Ideas: Protected Area Buffer Zones and the Politics of Land In .Africa .Development and Change (28):559-82 .. Parkin, DaVId.(1982) Introduction. In David Parkin (ed) Semantic Anthropology. London: AcademiC Press, A.S.A. Monographs 22. 109 Peluso, N.L. (1995) Whose Woods are These? Counter-mapping Forest Territories in Kalimantan, Indonesia. Antipode 27 (4):383-406. Peters, Pauline (1994) Dividing the Commons: Politics, Policy and Culture in Botswana. London: University Press of Virginia. Poole, Peter (1995) Indigenous Peoples. Mapping and Biodiversity Conservation: an Analysis of Current Activities and Opportunities for Applying Geomatics Technologies. Washington DC: Biodiversity Support Programme. Reid, P.c. (1901) Journeys in the Linyanti Region. Geographica/Journal 17:573-88. Schapera, Isaac (1930) The Khoisan Peoples of Southern Africa: Bushmen and Hottentots. London: Routledge and Kegan Paul. Scoones, Ian (ed) (1995) Living with Uuncertainty: New Directions in Pastoral Development in Africa. London: Intermediate Technology Publications. Scott, James (1998) Seeing like a State: How Certain Schemes to Improve the Human Condition have Failed. London: Yale University Press. Seiner, F. (1909) Karte des Gebiets zwischen Okawango und Sambesi (Caprivi-Zipfel) nach den Jahren 1905-1906. Mitteilungen aus den deutschen Schutzgebieten Band XXII Karte I. Silberbauer, George (1981) Hunter and Habitat in the Central Kalahari Desert. Cambridge: Cambridge University Press. Stigand, A.G. (1923) Ngamiland. Geographical Journal 62:401-19. Suzman, James (1997) Things from the Bush: Power and Colonialism in the Making of JuJ'hoan Identity in the Omaheke Region of Namibia. Unpublished PhD thesis, University of Edinburgh. Taylor, Michael (2000) Life, Land and Power: Contesting Development in Northern Botswana. Un- published PhD thesis, University of Edinburgh. Taylor, Michael (2002) The Shaping of San Livelihood Strategies: Government Policy and Popular Values. Development and Change 33(3):467-488. Tlou, Thomas (1985) A History of Ngamiland 1750-1906: The Formation of an African State. Gaborone: Macmillan. Wilmsen, Edwin (I 989a) Those who have each Other: San Relations to Land. In Edwin Wilmsen (ed) We are Here: Politics of Aboriginal Land Tenure. London: University of California Press, pp 43- 67. Wilmsen, Edwin (1989b). A Land Filled with Flies: a Political Economy of the Kalahari. Chicago: University of Chicago Press. Wily, Elizabeth (1974) Bushmen in the Kgalagadi District: Report on a Trip made ~ the Bushmen Development Officer, Ministry of Local Government and Lands, 2200 August - 26 August, 1974, and 5'h September - 19thSeptember, 1974. Unpublished report. Wily, Elizabeth (1976) Bere and Kalgae. Unpublished manuscript, Botswana National Archives MICRO 809. Wily, Elizabeth (1994) Hunter-gatherers in Botswana and the Land Issue. Indigenous Affairs 2:6-19. Woodburn, James (1997) Indigenous Discrimination: The Ideological Basis for Local Discrimination against Hunter-gatherer Minorities in sub-Saharan Africa. Ethnic and Racial Studies 20(2):345-61. MAPS AND PHOTOGRAPHS FOR THIS ARTICLE ARE TO BE FOUND FROM PAGE 178 Pula: Botswana Journal of African Studies vo1.16. no.2 (2002) Maps and Photographs for '''Mapping the Isand' in Gudigwa: a history of Bugakhwe territoriality" by Michael Taylor (pp. 98-109) N - -- -- "1 ,,":; Shokomoll;wa I L- I I 0 25 50 75 100 Kilometres Northern sandveld Buffalo fence Swamplfloodplain Settlement Game reserve! National par1l boundary - -- n,tlllI'"'''''''''''' r~ Sand ridge Surtaced road - ... - .. .----- International boundary -------- ...... -. Track ........................... Dry riverlled MtIp 1: The Okavango Delta and northern sandveld 179 ---.-------- _--r __ ---- I tl ,,~ ---- " \ 6 \,.. ----- ---:. " tl " .-- --- 60 ,. S "' _--- Dishokora (Bienlwa) - IIXl 0 c: Q;' - 0 NG13 NG14 N NG11 NG20. NG18 10 30 Map 2: Gudigwa with other Okavango Community Trust villages and surrounding Controlled Hunting Areas (CHAs allocated to OCT shaded). NG15 En E:X: EO Em Em E:Z: b> NG18 5-1 EO E:z: E:X> 5r- E-o Mapako a 9a Moremi E:X> E::O r' 10 20 30 40 KiIooletres Map 3: Map of the northern sandveld produced from the land mapping exercise Name Territory size (km2) Taetso 587 Khwarako 719 //Ae/exo 761 Rapula 974 Goo 1387 Amos and Sangando 1978 (combined) Average 915 Table 1;Family land sizefor seven of the ten mainfamilies that now live in Gudigwa. Plate 1:Entering Gudigwa. Plate 2: Old village site in the northern sandveld (Tsekugwa). 182 Plate 3: Buffalo fence near Gudigwa. 183 Plate 4: Elephant path through the bush (00 dao) 184 Plate 5: Amos drawing maps in the sand; a network of waterholes (oro) connected by elephantpaths (00 dao).