|
home | franco frescura | architecture | urban issues | lectures | graphic work | postal history | historical archive
THE PRE-INDUSTRIAL ARCHITECTURE OF THE TRANSVAALFranco Frescura ABSTRACTThe region now known as the Transvaal is probably the last portion of southern Africa to have been colonialised by European immigrants. Consequently, during the nineteenth century, it was spared the main thrust of missionary and trader incursions upon its indigenous culture, and it was only after the 1940s that any significant changes began to be noticed in its rural habitat. It is true that, although a succession of land measures and labour laws did much to provide its rural workers with access to urban technologies and materials, relative to other parts of southern Africa the process of breaking down the historical values of its indigenous communities was slow in starting. So slow, in fact, that until the 1970s the homestead architecture of parts of Sekhukhuneland, Venda, Swaziland and Botswana was still recognisably linked to the traditions and building technologies of previous eras. Much as in other parts of the country, what changes as did take place up to 1976 did not follow a pattern of radical and revolutionary transformation, as had been sought by missionaries and Government Agents. Instead they were integrated into the larger framework of rural values and social behaviour inherent in the lifestyle of local communities. In this essay I focus upon the indigenous dwelling form and its historical patterns of distribution in the Transvaal up to 1925, and examine some of the social, economic and environmental parameters which have guided its development. INTRODUCTIONThe creation of a historical model describing the development of indigenous architecture in the southern African interior can present a number of problems. Because of the complex and sometimes sensitive nature of the subject matter involved, it is not easy to impose a framework which will be both accurate and free from value judgments. Previous researchers in this field, for example, have attempted to explain the local built environment in terms of culture, group affiliation and language. However, criteria based upon group structures and social values are in themselves notoriously variable and open to interpretation. They are also liable to evolution and change over relatively brief periods of time and, as such, fail to offer a constant continuum within which a historical and comparative study may take place. It would be difficult, for example, to describe the inhabitants of the western Transvaal four centuries ago as being "Tswana". Such generalised nomenclature is largely irrelevant to southern Africa's rural residents who prefer to describe their group affiliations in terms of clan and family name. Anthropologists only adopted this term during the early 1900s to categorise a group of people sharing in the same language patterns. I have attempted, wherever possible, to avoid such terminology, and when it has been used, this has been done purely to describe the ancestors of a particular group, or the people who occupied a particular geographical region whose descendants today are known by that generalised name. By comparison, factors of a geographical and environmental nature are relatively more stable and, barring any radical changes in the prevailing climate of a region, tend to provide a more reliable if, at times overlapping, set of parameters within which the study of human activities can take place. This is made all the more desirable by the dynamic interaction existing between vernacular architecture and its physical environment Another problem which was encountered involved an exact geographical definition of "Transvaal". While it is true that the modern boundaries of this region are relatively well-defined, these are artificial colonial constructs which were not necessarily observed as barriers by previous residents of the region. For the purposes of this essay therefore I have included into my considerations Swaziland and those portions of the northern Cape, Botswana and Natal and the Orange Free State which adjoin the study area. REGIONS OF TECHNOLOGYIt would be wrong if a description of the local built environment were to regard rural builders as mere passive agents responding to their physical context in a pragmatic and ad hoc manner. Judging by the quality of their habitat, quite the opposite is, in fact, true and it is evident that, in their provision of shelter, they are active manipulators of their surroundings. It is also true, however, that the availability and quality of natural building materials can act as a limiting factor in the creation of an architecture. It would be difficult, for example, for a timber-based built tradition to develop in a treeless savannah, such as that originally found on the southern African highveld. The physical environment could therefore provide general guide-lines as to the nature of building methods employed in any one particular region. Also, because climate, soil and vegetation conditions are seldom radically altered from one generation to the next, the physical environment can provide a useful link between present building traditions and those which preceded them. By using only the broadest of environmental criteria, it becomes possible to hypothesise that, in the past, three major regions of technology have arisen in southern Africa (illustration 1). These, in their turn, have given rise to three major building cultures whose development was guided as much by their physical surrounds as by the economic activities of their builders. Although each region was not stylistically homogeneous and gave rise to a number of different dwelling forms, these were usually linked by a common technology. These were:
These three regions must not be considered to have been mutually exclusive, and a certain amount of overlap is known to have existed. We must also take into account the existence of immigrant groups from other regions who, for reasons of cultural inertia or political identity, preferred to retain the built forms and technologies of their former habitats. Although the Transvaal will be perceived to fall predominantly into the third, or central hinterland, its boundaries also overlapped with other regions. It therefore included small groups of migrant pastoralists who had either moved up to the south bank of the Vaal, or the Ki-Gariep as it was then known, or who had settled in parts of present-day eastern Botswana. The eastern littoral region intruded along the eastern flank of the Transvaal Lowveld and of Swaziland, where a hemispherical grass dome building tradition survives to the present day. Small groups of immigrant dome builders could also be found in the Middelburg area and, for a short time between 1825 and 1836, in the vicinity of Zeerust. Although research has shown that groups of hunter-gatherers have populated southern Africa for at least the past 26,000 years (Lewis-Williams and Dowson, 1989), their impact upon the building culture of the region appears to have been minimal. For this reason it is not intended to discuss their habitat in any great detail except where it falls within the ambit of a larger Khoisan mat building tradition. SOURCES OF DATAIt is not easy to piece together a detailed picture of the architecture of the Transvaal before the Difaqane. Because of the nature of some of the structures concerned, very little direct knowledge of indigenous architectural forms and building technologies has survived to the present day. Much of what we do know is the direct result of archaeological research, which is often open to a number of interpretations. The major sources of early data may be summed up as follows:
Ideally, the piecing together of a historical picture such as this should draw equally upon archaeological, historical and current research. Unfortunately this has not always been possible and a measure of interpolation has been unavoidable (illustration 2). After the 1820s, the nature of the data available to architectural historians changes as the input from archaeological sources is reduced in volume. This means that from this time onward, the process of historical reconstruction must rely increasingly upon the data provided by the eyewitness accounts of visitors to the region. This includes travelers, such as Baines, Backhouse and Holub, and missionaries, such as Campbell, Mackenzie and Casalis. After the 1880s reports originating from Government and other official sources also becomes available (illustrations 3-6). However, all of these suffer from a patent cultural bias, necessitating a process of re-interpretation in their texts and silences. THE MIGRANT PASTORALIST BEEHIVE REGIONThis was an extensive area which, at one time, encompassed the major part of the Cape and southern Free State. However, the only parts of concern to this essay are the banks of the Ki-Gariep, the northern Cape and eastern Botswana. Although the hemispherical mat structure was, at one time, the region's predominant domestic form, it was by no means its only one, and a number of other dwelling types are also known to have been built here concurrently. They all held in common the fact that they were designed to meet the needs of migrant society and were thus either impermanent or of a highly transportable nature. Apart from the more obvious clustering of settlements about the region's major water sources, there does not appear to have been a definite pattern to their distribution. Thus, in spite of the fact that hemispherical mat dwellings were associated with both of the region's two predominant social groupings, the Khoikhoi and the San, historical records fail to make a clear distinction between the habitats of the two. Early travelers, for example, tended to associate the hemispherical form (illustration 7) with Khoikhoi migrant pastoralists, while the curved mat shelter (illustration 8) was attributed to San hunter-gatherers (Barrow 1801 and 1804). Difficulty arose however when San pastoralists or Khoikhoi hunter-gatherers were encountered. Sparrman (1875) attributed both types of dwelling to the Khoikhoi whilst Burchell (1822) ascribed a Khoi-like structure to the San (illustration 9). It is probable that differences between the two groups were founded upon perceived levels of economic activity (Elphick 1985), and that both shared in a common mat dome building tradition which found different applications under different environmental conditions. The mat dome consisted of saplings planted into the ground in a circle or oval 3m to 4m in diameter covered over with reed matting tied down with leather thongs or grass ropes (illustration 7). Animal skins could also be used instead of mats, but these were normally reserved for the lower skirt of the hut where this more flexible material was better able to seal off a hard and uneven edge with the ground (Kolbe 1727). In some structures the position of the doorway could be moved by the simple expedient of repositioning the matting. The height of the average hemispherical structure does not seem to have exceeded 1.2m to 1.5m, but, in some cases, the hut floor was slightly excavated thus increasing the interior head-height (Campbell, 1815). The dwelling of a headman or chief was normally built along similar lines but was more generous in its proportions. By the time of Difaqane, from 1822 to 1836, the concept of a homogeneous mat dome region had became something of a misnomer. Not only had the process of land alienation proceeded so far to drive the last of the migrant pastoralists to the more arid and uninhabitable fringes of southern Africa, but many of the Khoi who remained within the boundaries of the Cape no longer built hemispherical dwellings as of old. However, the mat-covered dome had, by this time, become firmly established in the mind of White observers as a Khoikhoi architectural stereotype (Backhouse 1844). This is faintly ironic, for although this type of structure was still commonly built in some areas, these were mostly on the perimeter of the Cape Colony, such as Lishuani, Mirametsu (Backhouse 1844) and Modder River (Smith 1975) in the northern OFS, and Griquatown, Daniels Kuil and Afrikaners Kraal (Backhouse 1844) in the northern Cape. The only area where the mat dome appears to have survived to any great extent was in Namaqualand and southern Namibia. Although little is known of the San and their movements during this time, half-dome shelters associated with this group were recorded near the confluence of the Gariep and Ki-Gariep in the northern Cape; and roughly thatched temporary shelters were noted near Makwatling in the northern OFS (Backhouse 1844). From 1840 onwards, the migrant pastoralist region, to all intents and purposes, disappeared as an architectural entity. Scattered examples of mat huts or maantje huise, as they were now called (Dower 1902), were still being built, but not as frequently as before. Holub visited the northern Cape and the western Transvaal between 1873 and 1877 and reported Korah settlements in these areas (illustration 10), mostly at Kimberley, Klipdrift and Christiana, as well as a mixed Rolong/Korah village further north at Mamusa, where dwellings in both the Khoi and Tswana manner were being built (Holub 1881). During the 1850s isolated examples of hardbieshuise (illustration 11) and kapsteilhuise (illustration 12), previously recorded further south in the Cape and OFS, were also erected by some Khoi families in the northern Cape. However, little mention of these dwellings being built in the Transvaal has been found to date, and it is probable that, from this time onwards, they fell into general disuse throughout southern Africa as a whole. After the 1840s little is also known of the architecture of hunter-gatherer groups. Holub found some roughly built temporary structures belonging to Masarwa, near Shoshong, in 1873 (Holub 1881), but this is only an isolated reference. Generally it must be assumed that the San, under increasing pressure from colonial settlement, continued their withdrawal into the more arid and less densely inhabited areas of southern Africa. THE LOWVELD REGIONHistorically the domestic architecture of this region has been overwhelmingly orientated towards the thatched dome structure (illustration 13), a tradition which has been maintained up to comparatively recent times. Unlike the mat-covered dwellings of migrant pastoralists, which were easily dismountable and transportable, grass-covered domes were permanent structures of substantial construction, designed to last for a number of years. It is true that, theoretically, these were also theoretically portable, but they were seldom moved over any considerable distances. They were built by agrarian pastoralists with strong ties to their planting lands and even though the removal of a household from one location to the next was by no means unknown, their settlements were of a generally more permanent nature. The earliest evidence of a grass-covered dome building tradition found in this region to date, was recorded in the Tukela basin (Maggs 1982) and has been carbon dated at about 600CE. By the eighteenth century it appears to have become the preferred habitat of Nguni-speaking groups resident in the Lowveld and the eastern littoral of southern Africa, and its building range included a number of regional variants. During the period of Difaqane the geographical extent of this region in the Transvaal became particularly difficult to determine. Not only were there a number of indigent Nguni, and hence presumably dome building groups, moving about on the highveld during this time, but there is every indication that a number of Sotho/Tswana tribes, displaced by these events, had also temporarily turned to dome construction. In spite of the disruption caused by the Difaqane to the social and economic fabric of southern Africa, after the 1840s the architectural character of northern Natal and Swaziland remained relatively unaffected. This was abetted by the establishment in both areas of a political system which used the military as a strong unifying social force, and vested executive power in the hands of a centralized kingship. Its effect upon local architecture was two-fold. In the first place it gave rise to larger regional identities which sought to undermine clan loyalties and replace these with wider Zulu and Swazi nationalisms, as the case may be. Part of the creation of such identities was the willful adoption and maintenance of a style of housing which, in time, became recognised as belonging to that particular group. The success of such a policy can best be attested to by the fact that today the dome dwelling is still widely associated in the public mind, both Black and White, with Zulu or Swazi polity. Further evidence is also provided by Hall and Maggs (1979), whose archaeological research has shown that pre-Shakan cattle byres had their entrances facing uphill towards the Great Hut whilst in post-Shakan times these were turned downhill towards the settlement entrance. The former practice can be associated with pastoral custom and the nightly inspection of the family's wealth; the latter, on the other hand, had to do with military and ceremonial access. Secondly, a centralised leadership led to the implementation of a uniform foreign policy towards the Cape Colony and regulated the influx of White immigrants into these areas. The Swazi, for example, were antagonistic to White settlement right up to the 1880s, and an early attempt to establish a mission station at Mahamba in 1844 failed dismally (Merensky 1875). The Zulu, on the other hand, adopted a broad range of policies in their dealings with Whites. They were openly antagonistic to the itinerant Dutch, whilst seeking an alliance with the British; traders were accorded many privileges, were allowed to travel freely and to establish a settlement at Port Natal, and were dealt with as independent chiefs and allies (Wilson and Thompson 1975); missionaries however were, in the main, only allowed to settle on the coastal belt of Natal (Frescura 1986) and were perceived as a potential threat to the authority of the King. This means that whilst the one group totally excluded Whites, the other controlled their influx and were selective in their choice of settlers. As a result, those Whites who were allowed to settle in this region largely conformed to local norms. One such person was the trader John Dunn who, prior to his betrayal of his patron and benefactor, King Cetshwayo, in 1879, had taken local wives, adopted local customs and architecture and had been rewarded with the rank of Chief and Adviser to the King (Guy 1982). In the long run however the tactics of both Zulu and Swazi proved to be inadequate to the task and, at best, only managed to delay the impact of foreign influences upon their material culture. As a result, the region's grass dome tradition remained almost unaltered right up to the 1940s. Although its first cone-on-cylinder structures were signalled by Wangemann (1871-1875) in 1875 at the Emangweni mission station, subsequent pictorial evidence by Duggan-Cronin in the 1920s as well as Walton in the 1940s indicates that the influx of this form in both Zululand and Swaziland was scattered and of a very limited nature. It could therefore be concluded that the construction of this dwelling type in these areas is a recent development. The building of square plan dwellings in the Lowveld and Swaziland is largely identified with either White settlement or with trader and missionary activity, and only became manifest in significant numbers after the 1940s. THE CENTRAL HINTERLANDHistorically this area ran from the northern and central highveld in the south through to the Limpopo in the north, and from the eastern Lowveld through to the less arid stretches of Botswana and the northern Cape in the west. Its vegetation was somewhat varied, alternating from a treeless grassland in the south, to a savannah in the midlands, through to an arid mopane veld in the north. Extensive sources of indigenous timber suitable for building were found in Venda, in the Lowveld, and on the south-western reaches of the Transvaal and the northern Cape. Because it was cut off from the coastal regions by the arid Karoo to the south and the Kahlamba mountains to the east, the central hinterland was not penetrated by White travelers to any great extent until the last years of the eighteenth century. Thus few written accounts of this area exist before 1800, and a description of its early architecture only becomes possible through extensive reference to its archaeological record. As might be expected from so large an area, its spread of dwelling forms was by no means homogeneous, and if present-day evidence is anything to go by, it probably supported a number of different building technologies. Unfortunately the architectural picture which is presented by archaeologists is incomplete and, in some aspects, inconclusive. Most evidence seems to indicate that the predominant dwelling form was the cone-on-cylinder, often with a peripteral, or circular, verandah running about its external perimeter (illustration 14). However, there is sufficient data to show that the grass dome was not unknown in these areas. Loubser (1981), for example, has excavated beehive huts at Eiland which have been dated at 800-1100CE. Unfortunately these samples are few and historically scattered, and do not allow a meaningful pattern to emerge. After 1800 the use of dome-type structures for domestic as well as certain temporary and specialised functions is well documented, but the full extent of their spread is still to be established by archaeological research. The Cone-on-Cylinder TraditionThe earliest evidence of an architecture found to date in this region was uncovered by Mason (1974) at Broederstroom. There he excavated hut floors some three meters in diameter dated at 450CE. As at most other archaeological sites in this region, the dwellings are thought to have had peripteral verandahs although the extent of their construction has not been fully established. More conclusive, if later, data of cone-on-cylinder and full verandah structures have been found by Hanish (1980) in the Shashi Valley of the northern Transvaal, dated at 840CE; by Fouche (1937) and Gardner (1963) at the K2 and Mapungubwe sites, dated at 1000CE and 1100-1240CE respectively; and by Evers (1982) at Lydenburg, dated at 1700CE. Partial or front verandah cone-on-cylinder dwellings have been documented by Mason (1974 and 1981) at Olifantspoort, dated at 1600CE; and by Taylor (1979) on two separate sites, at Buffelshoek near Parys, and at Suikerbosrand near Heidelberg, both being dated at 1700CE. The relatively minor differences in verandah building methods between the northern and southern parts of this region may be attributed to variable environmental factors, but they could also be owed to conscious stylistic changes made by the builders themselves. Although visitors to the region during the early years of the nineteenth century limited their travels to the more populated areas of the northern Cape, eastern Botswana and the western Transvaal, the picture they presented did not differ too radically from that described by archaeologists. Dundas in 1801 (Barrow 1801 and 1804), Lichtenstein (1812 and 1815) in 1805, Burchell (1822) in 1812, and Campbell (1815 and 1822) in 1813 and subsequently in 1820, all uniformly described residents of the region to be living in cone-on-cylinder dwellings, surrounded by circular verandahs which at times were quite substantial, and may or may not have been partly enclosed to give a number of storage or sleeping chambers about a central drum (illustration 15). Most homesteads appeared to follow a bilobial layout which is thought by archaeologists to have been developed from about 1550CE (Maggs, 1976). This is a settlement pattern which has now become identified with the architecture of the Sotho/Tswana on the highveld (Maggs, 1972) and which, in a slightly modified form, has been transmitted through to the present day (illustration 16). In many cases valuable comments were also made about the local building technology and wall decorations. Lichtenstein (1812 and 1815), writing in 1805, told us that the drum wall was stopped short of the timber roof structure, a detail which we now know to be a measure against termite spread into the thatch. Early Tswana architecture was also well-decorated. When Burchell (1822) visited Dithakong in 1812 he was hosted at the homestead of "Serrakutu" whose younger wife had painted on her dwelling:
Eight years later Campbell (1822) visited the Hurutse of Kaditshwene where he recorded that the dwelling of "Sinosee":
During the Difaqane the extent of a cone-on-cylinder building tradition in this region remained essentially unchanged with the possible exception of its south-eastern reaches which extended into the southern OFS and western Lesotho, probably as a result of migrations of groups fleeing the northern Cape. Smith (1975) wrote of Tswana cone-on-cylinder dwellings at Bethulie in 1834, a report which was confirmed by Backhouse (1844) five years later; and Casalis (1861) recorded structures of an apparently similar nature in 1836 at Makossane. After the Difaqane the region's historic habitat remained largely unaltered. Reports of cone-on-cylinder dwellings, often with peripheral verandahs, were given by Burrow (1971) at Dithakong, Backhouse (1844) at Motito in 1839, and Smith (1975) at Motito and Mothibi in 1835, all of the above being located in the northern Cape. The first observations concerning the Pedi of the central-northern Transvaal were also made at this time by Smith (1975) in 1835 and Arbousset and Daumas (1846) in 1836. Both recorded the use of circular plan verandah dwellings, thus confirming the archaeological evidence of previous eras. In the latter's case, they are not known to have visited this area and it is probable therefore that they were only relaying the accounts of other travelers. One structure which may provide a valuable insight into the processes of Tswana architecture during the early nineteenth century, but which has gone without note until comparatively recent times, was depicted by Daniell (1820) in or near Dithakong in c1801. It is entitled "A Boosh-Wannah Hut" (also known as "A Tlapin Homestead") and shows what can be positively identified as a hemispherical structure surmounted by a conical roof carried by a series of verandah posts (illustration 18). The verandah superstructure is generally unremarkable, being no different from those built in this region for the past couple of centuries. The dome below however is of some considerable interest. We know from Lichtenstein (1812 and 1815) in 1805 as well as from current field work that the inner drum of the dwelling in this area need not necessarily bear any of the roof load. Therefore technically, it could take up any form, including that of a dome. We also know from recent research that the two architectural elements, the roof and the inner drum, can be, and often are, built independently of each other. The existence of this dwelling form was confirmed by Sanderson (1860) who, writing in 1858 or 1859, described a structure of similar construction among the Pugeni, in the Magaliesberge. He said that:
Burchell (1822), writing at Dithakong in the northern Cape some 47 years earlier, reported that:
If this was indeed the case, then it is obvious that this kind of structure does not represent a transitional stage in the construction of the cone-on-cylinder, but is a distinct building element in its own right. The Grass Dome TraditionAlthough much of the data for this region seems to indicate that its architecture was dominated by the cone-on-cylinder dwelling, usually surrounded by a peripheral verandah, a number of interesting exceptions to this norm have been recorded. The first came to light as the result of Loubser's archaeological excavations near Pietersburg where he uncovered the remains of beehive-type structures dated to 1650CE (1981). These were found in settlements whose builders migrated into this area more than three centuries ago. It is thought that, with time, they merged into a larger group which has subsequently become known as the "North Ndebele". His researches in the same area however, also show that by about 1850 this same group had conformed to the prevailing building technology of the region and had switched over to cone-on-cylinder dwelling forms. Loubser's findings are interesting for they contradict two long-held anthropological assumptions made by Van Warmelo: that both northern and southern Ndebele groups were part of the same migration from the eastern Nguni region; and that all Transvaal Ndebele originated from Nguni stock. Van Warmelo did not attempt to explain the mechanics of such a migration, nor the subsequent schism he perceived to have occurred between the two groups. He did however acknowledge that linguistic differences existed and that "nothing but pure Sotho is spoken nowadays" among the North Ndebele whilst the southern Ndebele "have retained the customs and language of their (Nguni) ancestors to a remarkable extent." (Van Warmelo 1930) Had Loubser been able to show that the North Ndebele were indeed the descendants of an original Nguni migration, then his revelation that this group had built and lived in dome-shaped dwellings would not have been difficult to accept. However his research has shown precisely the opposite: that the North Ndebele are the product of a number of separate population movements, none of which are recognisably Nguni. This means that many of the morpho-centered concepts and cultural stereotypes which have grown up about southern African rural and indigenous architecture (Walton 1956) needs to be re-examined and revised. One of these held that the Nguni build in grass, the Sotho/Tswana in clay. However, current research has shown that not only did the Sotho/Tswana also build in grass, but did so some two centuries before travelers like Backhouse (1844) and missionaries like Arbousset and Daumas (1846) recorded similar building technologies among Sotho groups further south. The existence of an indigenous grass dome building tradition on the highveld must therefore be recognised as part of its wider local architectural tradition. A second exception to the region's norm of cone-on-cylinder construction is thought to have occurred amongst the South Ndebele, an Nguni group who, according to their own oral history, moved into the Transvaal in about 1630-1670 (Van Warmelo 1930). Unfortunately comparatively little is known of their architecture before the turn of this century. However during the course of field interviews one of Van Warmelo's informants described the construction of what was obviously a hemispherical structure and claimed that this type of dwelling had been common among his people before the 1883 conflict between the Ndzundza and the ZAR. If this was indeed the case then it would not be unreasonable to assume that this group could have built this type of structure long before 1883, and perhaps as far back as the late seventeenth century. It is ironic that the region's best documented grass dome building tradition was not indigenous to it, and did not survive longer than a dozen years. It was brought into the southern African hinterland by the Matabele, or the amaKumalo as they were more accurately known at the time, a Nguni-speaking group who, under the leadership of Mzilikazi, migrated from Zululand in 1823. Their first stop was the upper Oliphants River valley in the Lowveld, where they remained until their move to the Crocodile River, immediately north of Pretoria, in 1825. By 1832, they had moved again into the Marico valley of the western Transvaal, near present-day Zeerust. When they were visited by Smith (1975) in 1835, thirteen years after their migration from Zululand, they had already undergone three separate relocations, with a fourth one still to come two years hence. Despite these upheavals, the dwelling forms and building methods they adopted throughout this period, and well into the 1890s, differed little from those used contemporarily in Zululand (Cooper-Chadwick, 1975) (illustration 20). Their towns are also known to have achieved a large measure of permanence (illustration 21). Arbousset and Daumas (1846) told how, in 1827 or 1828, Pedi captives erected for Mzilikazi a palisade about his wives' quarters:
These accounts are significant, not only because, by that time, the Matabele had ceased to be of a predominantly Nguni composition, having assimilated large numbers of highveld Sotho and Tswana into their polity (Wilson and Thompson, 1975), but also because their wanderings had taken them through a number of different vegetation and climatic regions. Yet, despite the cultural and material considerations which must have been brought to bear upon their building traditions, they steadfastly maintained their attachment to a northern Nguni dwelling form without appearing to have made any significant technological or stylistic adaptations to it. Other Building TraditionsDuring the early years of the nineteenth century travelers to the southern African interior also gave details of other types of structures found in this region which, because of the building technology used, are not likely to have left many traces. The residents of Dithakong, for example, were recorded by Burchell (1822) in 1812 to have built temporary huts at a few days' travel from their town to serve as a winter cattle grazing station. A second example of temporary abodes was recorded by Campbell (1822) in 1820 when he visited "Meribohwhey" (Maribogo), in the northern Cape, where the "Tammaha" (Thamaga), a Kgalagadi and hence predominantly Tswana group, were building dome-type structures, covered with matting in the Khoikhoi style but using a type of extended entrance. Another group of indigenous dwellings whose status is the subject of some debate is the corbelled stone structures found in the northern Orange Free State and some parts of the southern Transvaal (illustration 22). These shelters are a feature of early Iron Age settlements whose construction, up to recent times, was attributed to an indigenous group called the "Lekoya" or "Ghoya" (Walton 1965). Their existence was described by early travelers to the region, who found them mostly uninhabited as a result of the Difaqane. Smith (1975) commented in 1835 that:
while Arbousset and Daumas (1846) were also struck by the desolation that they encountered in the northern Orange Free State in 1836, stating that:
Forty years later Lord and Baines (1876) gave a more detailed description of their construction:
Recent archaeological research has attributed their construction to either the Taung or the Kubung, both being Sotho groups who inhabited this region up to the time of Difaqane. Maggs (1976) has linked their building to that of both his Type V and, on a few occasions, Type Z settlements, thus dating them to the early sixteenth century. Corbelled stone structures were generally built to such small proportions that it is difficult to conceive of them as dwellings intended to house adults. This, and the fact that such shelters were usually associated with the walling of cattle byres leads us to the conclusion that their function was to act as herd-boys' huts. This supposition is supported by Walton's research, whose photographs of similar shelters in Lesotho show structures very similar in form and technology to the corbelled stone domes of the highveld (Walton 1956). Another building tradition which appears to have existed in this region up to and during the Difaqane was that of the Hurutse of the western Transvaal who raised their dwellings above ground level by means of a series of stilts. Campbell (1822) in 1820, noted the use of huts raised on stilts as well as the more conventional verandah cone-on-cylinder dwellings among the Hurutse of Kaditshwene, in present-day western Transvaal near Zeerust. This practice was subsequently confirmed in 1835 by both Burrow (1971) (illustration 23) and Smith (1975) (illustration 24) who located the Hurutse in the Kashane or Magaliesberge. A structure having similar intent but of a somewhat different form was recorded by the traders Schoon and M'Luckie in 1829 near Kaditshwene, in the district of present-day Zeerust (illustration 25). They claimed to have found:
The common sense of such structures is difficult to establish. Stilt architecture elsewhere in the world is usually associated with settlement in low-lying marshy lands, a condition which can hardly be said to have prevailed in the semi-arid regions inhabited by the Hurutse. All of the travelers quoted above were in agreement that such architecture was the direct result of lions hunting in these parts, terrorising the local population. This is a strange assertion, considering the fact that lions were also common in other parts of southern Africa where buildings on stilts are unknown. We must conclude thus, that either the Hurutse were an uncommonly cautious people or, more likely, that these foreign visitors were wrong in their assumptions. Subsequent research seems to indicate that these persons were probably describing granaries or poultry hocks, none of which were intended for human habitation. Finally, some mention should also be made of the building of granaries in this region (illustrations 26 and 27). Despite the many references made in local archaeological literature to the presence of flat stones which could have supported a grain storage facility, little is known of their construction or form outside the accounts of early travelers. Early Tswana granaries were recorded in c1801 by Daniell (1820) who gave an excellent pictorial rendition of both their building technology and final form, and by Campbell (1815) who, in 1812, told of the Ngwaketse building large "store-houses for containing provisions". This, of course, refers only to those structures specially designed and built to serve the function of granaries and not to the existence, in some areas, of an old hut which had been relegated to this function and which would be outwardly indistinguishable from its surrounding dwellings. The importance of granaries to a study of this nature should not be underestimated. Current field work has shown that, despite their humble and apparently expendable nature, such structures tend to reflect, on a smaller scale, the form and building technology of their larger and more permanent residential counterparts. In many cases it was found that their construction, particularly in those areas where grass dome dwellings were formerly built, reflected the building modes and forms of (locally) archaic structures. If this could be accepted as a guide-line, it could then also be argued that if granaries are indeed a scaled down version of a dwelling unit, past or present, then, judging from Daniell's evidence, the Tswana have not built permanent hemispherical residential units in their immediate and traceable past. THE WHITE INVASIONFrom an architectural stand-point, the years after 1840 were an important period in the history of this region. It was made notable not so much by any direct developments which may have occurred in this field, as by events of a political and economic nature which, with the passing of time, were to exert an increasingly powerful influence upon local building traditions. The political stage was set by the defeat and subsequent emigration of the amaKhumalo, now more commonly referred to as the Matabele, to western Zimbabwe early in 1837. This created a power vacuum which White immigrants were not slow in filling. At first this was limited to migrant Dutch farmers but, with the discoveries of diamonds in the northern Cape in 1866, and of gold in the Transvaal in the 1870s, a flood of miners, land speculators, fortune seekers and an assortment of camp-followers from numerous countries were also attracted to the region. The resultant internal tensions between Black and White, as well as between White and White, were to manifest themselves in periods of sporadic violence which had the larger effect of slowly dispossessing more and more indigenous groups of the land they had occupied for centuries. The establishment of primary industries in the southern African interior was to have some important effects upon its architecture.
It is obvious that this presents a limited view of the process of early urbanisation on the southern African highveld. Other important elements of a socio-political nature do exist but largely tend to fall outside the scope of an architectural study such as this one. The influx of White immigrants and other travelers into the region during this time produced the first detailed accounts of the architecture of the central and northern Transvaal. Generally these confirmed the picture up to the 1700s which has been portrayed by archaeologists, with the cone-on-cylinder, often surrounded by a circular verandah about its perimeter, as the dominant dwelling form. Such dwellings were recorded by Baines (1867) near the Ki-Gariep in 185O, by Holub (1881) in the northern Cape, western Transvaal and eastern Botswana in 1873-77 (illustration 28), by Wangemann (1871-1875) among the Venda and Lovedu in 1875, and by Merensky (1889) at Sekhukhuni's capital in 1882. Ordinary cone-on-cylinder structures were also recorded by Holub (1881) among the Rolong in the western Transvaal in 1873, and by Wangemann (1871-1875) at the Galekalekale, Makopane, Malokung, Blouberg, Makchabeng and Wallmannsthal mission stations in about 1877 (illustration 29). Little mention was made by travellers of any substantial grass dome building tradition in this region during this time. The Matabele had long gone and, with them apparently, also went their "Zulu" style of dwelling construction. The South Ndebele could have built hemispherical dwellings up to the time of the defeat of the Ndzundza by the ZAR in 1883, but no eyewitness records of such structures have been discovered to date. After 1883, they apparently conformed with the region's norm of cone-on-cylinder construction, some of which still survive to the present day (Frescura 1981). The only pictorial record of hemispherical building found for this region was made by Wangemann (1871-1875) at Botshabelo in 1872 (illustrations 30 and 31). These structures would appear to have been built, possibly as temporary dwellings, by Pedi converts to Christianity seeking missionary protection during Sekhukhuni's leadership struggle with his younger brother Mampuru (Delius 1983). The introduction to this region of the flat-roofed, square plan dwelling, more commonly known today as an iplata or iflat, can probably also be dated to the latter part of the nineteenth century (illustration 32), although its use in indigenous settlement did not become more widespread until the 1950s (illustration 33). CONCLUSIONSThe visual and aesthetic nature of indigenous architecture is largely dependant upon the form, technology and materials of its primary constituent, the individual dwelling unit. Its construction gives scope to local builders for the development of new structural systems and technologies; its materials create the richness of surface textures commonly associated with vernacular architecture; its walls provide rural artists with a ready canvas for their decorative motifs; and its superimposition of forms and roof lines create a bulking which is both interesting and sensitive to context. It is not for nothing that when William Burchell entered Dithakong in July 1812 he was led to exclaim:
It is true that this essay has had to cover a comparatively large amount of ground, both physical and chronological. For this reason it has not attempted to present a definitive picture, but has rather focused upon those aspects of the indigenous built environment, such as dwelling form, building technology and decoration, which were both obvious to past travelers and accessible to modern field researchers. However the indigenous architecture of the Transvaal and, indeed, the southern African region, is not defined by the forms and textures of the individual dwelling so much as the principles which guide the planning of the larger homestead. Rural settlement and the distribution of its living space is governed by a number of complex considerations, including family hierarchy, social values, the location of sacred areas, gender and age group access to resources, food preparation, the storage of surpluses, and the control of wealth belonging to the larger group. For a number of reasons, not the least being a question of space, I have chosen not to enter into such a specialized debate, and the Reader wishing to explore these aspects further is referred to the relevant texts. POSTSCRIPTThis paper was commissioned as the introductory chapter for a book entitled “Architecture of the Transvaal”. (Frescura, Franco. 1998. Pre-Industrial Architecture: Historical Patterns and Distribution. ARCHITECTURE OF THE TRANSVAAL, Editors Roger C Fisher and Schalk le Roux. Pretoria: University of South Africa, 1998. 1-23) BIBLIOGRAPHYANDERSON, Andrew A. 1888. Twenty-Five Years in a Wagon. London: Chapman and Hall.
Copyright @ francofrescura.co.za
|