AN INTRODUCTION TO TSWANA ARCHITECTURE

Franco Frescura

PREAMBLE

Since 1753, when Marc-Antoine Laugier proclaimed that All the splendors of architecture ever conceived have been modelled on the little rustic hut (1977: 12), our understanding of vernacular architecture has been subjected to a variety of theories and ideologies which, like Laugier, have sought to explain its "fixed and unchangeable laws". Most early works were guided by a strong cultural and environmental predeterminism which, like Viollet-le-Duc (1876), were often coloured by the romantic idea that such environments were built by Rousseau's "noble savages" (1959). More recently, writers such as Rapoport (1969, 1977, 1982), Hall (1973) and Marc (1977) have discounted such theories and have postulated that human environment is the product of a wider cognitive language deeply rooted in cultural cosmology.

Regardless of their philosophical content, all such works have two major factors in common. Firstly, they are the construct of Western intellectuals who, taking a number of isolated case studies, usually drawn at second or third hand, have sought to formulate "universal truths" on the subject. It is then left for their acolytes and faithful followers to go forth and validate these "truths" by "discovering" supporting data in the field. Secondly, the people most concerned with these theories, the builders of vernacular architecture themselves, have little or no say in their formulation. They are still regarded as the "noble savages" who provide the grist to the intellectual mill and it is questionable whether they could ever recognise their own work, once processed and sanitised, for the consumption of a predominantly European and North American public.

Thus any essay which sets out to discuss the vernacular architecture of a people in terms of a pre-stated ideological stance ceases to be a discussion about that architecture and, by definition, becomes a discussion about the writer's ideology. Worse still, any response by outsiders to this debate can only be made in terms of other ideologies, thus becoming a debate between ideologies, not about architecture.

This paper sets out to discuss the vernacular architecture of the Tswana. Much of its data is based upon the historical accounts of nineteenth century travelers and early twentieth century anthropologists and is therefore the product of European and colonial observers whose objectivity was often coloured by a variety of cultural, religious and political prejudices. Therefore the first and major task of this essay was to bring together these historical sources, most of which are already well known individually, and present them as architectural accounts unencumbered by colonial ideology and jargon. This was done by re-interpreting them in terms of current field work conducted amongst the descendants of those groups visited by white travelers nearly two centuries ago. Where contemporary information is given, this is done in order to give the architecture concerned a sense of continuity and transition and not to reinforce any current social or political stereotypes which may have developed in the intervening years. Although the writings of many early and contemporary theorists have been taken into account, their work serves as a background to this paper and its research does not set out to justify any particular philosophy. Where such terminology as "left and right" has been used in the context of settlement patterns, this has not been done as a reference to Adam Kuper's structuralist writings (1980) but because this concept is used currently by many rural inhabitants of southern Africa to explain their own habitat and inheritance laws (Seymour, 1982).

HISTORICAL BACKGROUND

It is probable that the term "Tswana" dates its origins from the early years of the nineteenth century when it was used by white visitors to southern Africa, such as Dundas (Barrow, 1801 and 1804: 114-116) and Lichtenstein (1812 and 1815: 2: 373-379), as a collective noun to describe those indigenous groups inhabiting the region's hinterland. Although more recent times have witnessed the development of a "Tswana" state, the name should be seen originally to have represented a cultural identity rather than a political unity. Indeed the history of the Tswana is characterised by a series of internal schisms giving rise to a number of independent chiefdoms. Among the most important of these were the Kwena, the Hurutshe, the Kgatla, the Tlhaping, the Ngwato and the Rolong. The development of separate Pedi and southern Sotho identities subsequent to the Difaqane of 1822-1836, a period of widespread social upheaval and economic hardship throughout central southern Africa, should also be viewed in terms of this fissionary trend (Maylam, 1986: 54-63).

Although these groups are known to have originally inhabited parts of the present-day northern Cape, the central-southern Transvaal and the northern OFS, as a result of the Difaqane many moved westwards and their descendants may be currently found in the western Transvaal, the northern Cape and Botswana. A small group of Rolong, which is also identified with the larger Tswana polity, moved south in 1833 and settled in the Thaba 'Nchu area of the OFS (Omer-Cooper, 1978). (figure 1)

PHYSICAL ENVIRONMENT

Generally speaking those areas currently inhabited by the Tswana lie in the highveld and semi-arid regions of southern Africa. They are typified by warm rainy summers and cool dry winters, with large diurnal temperature ranges and periodic droughts increasing in severity as one moves westward (Cole, 1961: 64-92). These conditions are inimical to tree growth and the region may be termed a true grassland, an open savannah with an admixture of thorny acacia trees and bushes (Acocks, 1975). When Campbell visited the northern Cape in 1820 he stated that he "... entered upon another extensive plain, also without tree or bush, and which seemed to be only bounded by the horizon." (1822: 1: 152)

A measure of timber, mostly acacia, could, at one time, be found growing further to the north and east towards the Harts valley, but the development of a diamond mining industry in and about Kimberley during the 1870's had the practical effect of deforestating this area. It has been estimated that during this era the Diamond Fields were consuming an average of one wagon-load of firewood per white family every month (Shillington, 1985: 66-68). The exhaustion of such natural resources was to have important repercussions in later years upon the folk architecture of the region for, like rural builders elsewhere in southern Africa, the Tswana rely upon their natural environment as a quarry for their building materials. Writing in 1805, Lichtenstein reported that:

"A Beetjuan tribe always fix their abode in the midst of a wood of mimosas, the stems of these trees, where they grow to a considerable height, being one of the most essential articles in building their houses. Twelve or fourteen of these stems are requisite to build a good house; eight or ten of these must be nine foot high, the rest three or four feet higher." (Lichtenstein, 1812 and 1815: 2: 373-379)

Yet it would be wrong to describe the architecture of the Tswana as a series of responses to their physical environment. It is true that an availability of natural building materials and prevailing climatic conditions are important limiting factors in the creation of a folk architecture anywhere, and the dwellings of the Tswana are prime examples of how a community has adapted the grass, the clay, the wood and the stones about it to create dwellings which are at once cheap, durable and attractive. However an analysis of the forms, textures and decorations of the individual dwelling unit reveals that the built environment of the Tswana is a reflection of the cultural values, sexual mores, social hierarchies, religious beliefs and cosmological norms of the people as a whole. It is a habitat which is subject to a series of patterns, rules and constraints which, once its language has been learned and understood, gains in substance and significance and demands from the visitor recognition as an architecture in its own right.

HISTORY OF THE TSWANA DWELLING BEFORE 1820

It is not easy to piece together a detailed picture of the architecture of southern Africa before 1820. Because of the nature of the structures concerned, very little direct knowledge of indigenous architectural forms and building technologies has come down to the present day and much of what we do know is the direct result of archaeological reconstruction and the interpolation of data. It is fortunate that the architect, like the historian, is able to draw upon the richness of local oral and folk traditions to a large extent and although much of this information is open to interpretation, part of it may be shown to be historically valid.

A wide range of dwelling forms is known to have been built in the southern African hinterland before the years of Difaqane. However, as one would expect from so large an area, their spread was by no means homogeneous, and if present-day evidence is anything to go by, the region probably supported a number of different building technologies.

The earliest evidence found, to date, of a Tswana architectural tradition was uncovered by Mason at his Broederstroom site. There he excavated hut floors some 3,000m in diameter which have been dated at about 450 CE (1974; 1981: 401-416). As at most other proto-Tswana archaeological sites, the dwellings are thought to have been in the form of a cone on cylinder with verandah running about their perimeter. More conclusive, if later, data of cone on cylinder and full verandah structures have been found by Maggs in the northern Orange Free State, dated at 1550 CE (1976: 238-239); by Mason at Olifantspoort, dated at 1600 CE (1974; 1981: 401-416); and by Taylor at two separate sites, at Buffelshoek near Parys and at Suikerbosrand near Heidelberg, both being dated at about 1700 CE (1979). (figure 2)

The picture being presented to date by archaeologists of the early architecture of this region is therefore somewhat incomplete. Fortunately from 1801 onwards this meagre data is supplemented extensively by the accounts of early travellers to southern Africa. Men like Truter, Lichtenstein, Burchell, Campbell, Backhouse and, later, Holub all traveled among the Tswana extensively and their reports tend to confirm the findings of current archaeological researches. They also described an architectural tradition which, to European eyes, was both exciting and unique. When Burchell arrived at Dithakong, the Tlhaping capital in the northern Cape located near present-day Kuruman, in 1812, he was met with:

"... the most gratifying sight which my journey had yet afforded, presented itself; and part of the Town of Litakun now appeared before me.  As we advanced nearer, and gained higher ground, the multitude of houses which continued rising into view as far as I could see, excited astonishment; while their novel form and character seized my whole attention." (1953: 254-256)

Although these early visitors generally 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 in substance from that described by twentieth century archaeologists. The Dundas report of 1801 stated that:

"... (the chiefs) house, like all the rest in the town, was built in a circular form, being about sixteen feet in diameter. The bottom part, to the height of four feet from the ground, was stone laid in clay, and wooden spars erected at certain distances. On the east side of the circle, about the fourth part of the house was open, the other three-fourths entirely closed. A round pointed roof covered the whole in the form of a tent, well thatched with long reeds, or with the straws of the holcus. From the centre to the back part of the house, a circular apartment is made off, with a narrow entrance into it, where the head of the family takes his nightly rest; the other members of the family sleep in the fore- part, or between the large and small circles of the house. All of the houses were enclosed by pallisades; and the space between these and the dwelling serves for a granary and store for their grain and pulse. These granaries were constructed in the form of oil jars, of baked clay, ... and they were supported on tripods, composed of the same material, which raised them about nine inches off the ground. They were covered with a round straw roof erected on poles" (Barrow, 1801 and 1804: 114-116)

The drawings of Daniell, probably originating from the same expedition, aptly illustrate these points (1820). Subsequent travelers to the region were all in agreement with the Dundas report, describing the Tswana to be residing 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 (figure 3). Most homesteads appeared to follow a "bilobial" pattern, being two courtyards, or "lobes", created by means of walls or screens, and set about a central dwelling unit (figure 4). The fore- court was usually associated with the family's semi-public or leisure functions whilst the rear court served as a work and cooking area. This is a settlement pattern which has become particularly identified with the architecture of the Sotho/Tswana on the highveld (Maggs, 1976) and which, in a slightly modified form, has been transmitted through to the present day.

Most accounts also made many valuable comments about local building technology and wall decorations. Lichtenstein, writing in 1805 (1812 and 1815: 2: 373), 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.

The first accounts of Tswana wall decoration also date from this time. Burchell visited the Tlhaping homestead of "Serrakutu" at Dithakong in 1812 and recorded that this man's younger wife:

"... exhibited her paintings in a manner which evinced that she was well satisfied with her own performance. They were, the figures of several animals, rudely drawn, with a paint of white earth, against the front-wall of the house." (1953: 321-322)

A year later Campbell visited the same homestead and "... found them very rough representations of the camel leopard (giraffe), rhinoceros, elephant, lion, tiger and stein-buck, which Salakootoo's wife had drawn on the clay wall with white and black paint." (1815: 194)

Then, as now, this work appears to have been the province of Twana women. Regrettably both Burchell and Campbell failed to reproduce in their work more accurate graphic illustrations of such wall decorations, an omission which was rectified by Campbell seven years later when he visited the Hurutshe town of Kaditshwene, located immediately to the north of present-day Zeerust. There he visited the local chief's homestead and found that:

"Sinosee's house was neatly finished; it was circular like all the others, having not only the wall plastered both within and without, but likewise the inside of the roof. The wall was painted yellow and ornamented with figures of shields, elephants, cameleopards, etc. It was also adorned with a neat cornice or border painted of a red colour." (1822: 1: 227-228)

He also pointed out that the houses of Kadidtshwene were "... plastered on the outside and painted yellow. One we observed painted red and yellow with some taste." (1822: 1: 222-224)

These travelers also gave details of other types of structures found in this region which, because of the nature of their technology, are not likely to have left many traces behind for modern-day archaeologists to excavate. The residents of Dithakong, for example, were recorded by Burchell in 1812 to have built temporary huts at a few days' travel from their town to serve as a winter cattle grazing station (1953: 360-372). Campbell, in 1820, noted the use of huts raised on stilts, as well as the more conventional verandah cone on cylinder dwellings, among the Hurutshe of Kaditshwene. That same year he also visited Maribogo, in the northern Cape, where the Thamaga, a Kgalagadi and hence predominantly Tswana group, were building hemispherical structures, covered with matting in the Khoikhoi style but using some form of extended entrance (1822: 1: 279). (figure 5)

One structure which may provide a valuable insight into the processes of Tswana architecture during the early nineteenth century, but which has gone unremarked until comparatively recent times, was depicted by Daniell in or near Dithakong in about 1801. It is entitled "A Boosh-Wannah Hut" and shows what can be positively identified as a hemispherical structure surmounted by a conical roof carried by a series of verandah posts (1820) (figures 6 and 7). Similar structures were subsequently described by Burchell at Dithakong in 1812 where he found that some Tlhaping dwellings:

"... have a small inner apartment which occupies the centre of the building ... intended as the bed-room for the parents while the outer apartment is for the children. This inner or central apartment is frequently built in the shape of a cone, or a half-ellipsis, the point of which reaches to the height of the roof which it serves to support and strengthen. In other instances ... its form is cylindrical;  and this appeared to be an improved construction." (1953: 360-372)

Sanderson, writing subsequently in 1858-9, described a dwelling of similar construction among the Pugeni, in the Magaliesberg. He said that:

"Every hut or house is divided from the others by a fence or wall, and is surrounded by a broad eaves or verandah. The hut itself is circular, built of stone plastered over with clay, and in the better class polished inside and out with beeswax. The roof is thatched with straw or reeds, conical, and 20 or 30 feet in height. The courtyards surrounding the huts are plastered, and kept scrupulously clean. The front half of the veranda is usually enclosed with a dwarf wall, and the semicircular doorway, 18 or 20 inches in height, closed by a board sliding behind a couple of pilasters forming a frame. Opposite the door inside, and extending a third of the way round, is a platform or dais raised about 6 inches, in which are sometimes planted the stems of one or two small trees, the branches of which serve to hang articles upon. The inside of the hut is in the middle, about 6 feet high, oval in section, and without light or ventilation, except from the doorway." (1860: 233-255)

It becomes clear therefore that the dwelling drawn by Daniell represented a structure such as those described by both Burchell and Sanderson, but in a state of incompletion. In this context also it becomes clear that the strange painted structure drawn by Campbell at Kaditshwene in 1820 was no more than a decorated inner chamber belonging to the Hurutshe chief (figure 8).

A further group of indigenous structures whose status is the subject of some debate are the corbelled stone huts found in the northern Orange Free State and some parts of the southern Transvaal. These shelters are a feature of early Iron Age settlements whose construction was originally attributed, up to comparatively recent times, to an indigenous group called the "Lekoya" or "Ghoya" (Walton, 1965). Early White travellers to the region first recorded their existence, but found them mostly uninhabited as a result of the Difaqane. Lord and Baines described their construction as follows:

"A circle of blocks is laid on the ground, then another on them, with the edges projecting a little inward, so that the circumference of each course is less than that of the one immediately beneath it; a large slab covers the top and finishes the building." (1976: 246-250) (figure 9)

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 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 (1976). Corbelled stone structures were generally built to such a small scale that it is difficult to conceive of them as adults' dwellings. This, and the fact that such shelters were often incorporated into the walling of cattle byres leads us to the conclusion that their function was to act as herdboys' huts. This supposition is supported by Walton, whose photographs of herdboys' huts in Lesotho show structures very similar in nature and intent to the corbelled stone shelters of the highveld (1956: 133).

TSWANA ARCHITECTURE DURING THE DIFAQANE

The Sotho/Tswana word "difaqane", or to give it is Nguni equivalent "mfecane", means, literally, "the scattering of the people" (Wilson and Thompson, 1975: 1: 391). It has come to signify a period in southern African history between 1822 and 1836 when the major part of the sub-continent entered a period of upheaval and turmoil which transformed the whole nature of rural pre-industrial society and paved the way for changes in the region's economy and system of government.

General opinion among historians has attributed the beginnings of the Difaqane to the emergence of a Northern Nguni state under the leadership of one ambitious and charismatic leader (Omer-Cooper, 1978). Although outwardly true, it is also probable that the rise of a Zulu kingdom was also precipitated by a need on the part of southern African rural society to meet, in the foreseeable future, the increasing demands for land being made by White settlers to the south. This would explain why the Zulu Royal Court persisted, right up to 1879, in treating the British as allies and the Dutch as potential enemies. It would also explain how groups from Sotho, Tswana and Nguni language backgrounds were able to coalesce into the centralised state of the Basotho in order to resist, with a high degree of success, the incursions of Zulu, British and Dutch upon their settlements. Certainly the resultant massive population shifts, the concentration of people into some areas and the depopulation of others was to set the pattern for the future growth of the modern states of southern Africa as well as provide the inception of many of its later problems.

The period of Difaqane, far from being an isolated event in the history of the indigenous people of southern Africa, was in many ways also interlinked with parallel events in the sub-continent's White immigrant community. Although scattered missionary infiltration had begun some time before, it was given considerable impetus by the publication of Campbell's accounts of his travels through the region during 1812-13 and again in 1820 (1822). As a result, missionary activity in southern Africa increased rapidly between 1822 and 1836, most particularly in the northern Cape, the Eastern Cape and Lesotho, all three areas being the backdrop for some of the most momentous events of this era. This period was also to witness an increasing British military involvement; the introduction of immigrant British farmers into the Eastern Cape; the extension of a trading infrastructure into the southern African hinterland; and finally the northward migration of itinerant Dutch farmers and pastoralists from the Cape. All four groups, missionaries, military, traders and farmers, were to have a powerful influence upon the subsequent history, and hence the architecture, of rural southern Africa.

The descriptions of Tswana architecture which have come down to us from this time differ little from what we already know of the region. Many of the patterns outlined previously are now confirmed and brought into greater focus and the first signs of an infiltration by foreign dwelling forms and building technologies are now perceived to occur. However owing to the events of the Difaqane, there is a noticeable shift in population patterns away from the central highveld and the cores of what were subsequently to become known as the Tswana, the south Sotho and the Pedi groupings become apparent during this time.

It is probable that during this period the region's historic style of dwelling construction, the cone on cylinder, was maintained unaltered, reports of such structures being made by Burrow at Dithakong in 1835 (1971: 35), Backhouse at Motito in 1839 (1844) (figure 10) and Smith at Motito and Mothibi in 1835 (1975: 195), all of the above being located in the northern Cape.

Tswana cone on cylinder dwellings were also noted by Smith at Bethulie in 1834 (1975: 38), a report which was confirmed by Backhouse five years later (1844: 354-355); and Casalis recorded structures of an apparently similar nature in 1836 at Makossane (1861: 143) (figure 11). The first observations concerning the Pedi of the central-northern Transvaal were also made at this time by Smith in 1835 (1975: 245) and Arbousset and Daumas in about 1836 (1846: 273-274). Both recorded the use of circular plan verandah dwellings, thus confirming the archaeological and historical evidence of previous eras (figure 12).

An interesting building tradition which appears to have existed in the region during this time, alongside the more conventional cone on cylinder, was that of the Hurutshe of the western Transvaal who raised their dwellings above ground level by means of a series of stilts. This practice was first reported by Campbell when he visited Kaditshwene in 1820 and was confirmed subsequently by both Smith (1975: 195) (figure 13) and Burrow in 1835 when they visited the Hurutshe in the "Kashane" mountains, the present-day Magaliesberge. Burrow stated that:

"Some of the inhabitants of this part live in houses perched on the tops of poles, to protect them, I suppose, from the Lions ... These pole-houses are all enclosed in a kraal like the others, but they never come down if they can avoid it at night, as the Lion will sometimes lie in wait the whole night ..." (1971: 53-55)

A "structure" having similar intent but of a somewhat different form was recorded by the traders Schoon and M'Luckie in 1829 at Kaditshwene (figure 14). They went one better than their predecessors when they claimed to have found:

"... a large tree containing seventeen conical huts. These are used as dormitories, being beyond the reach of the lions, which, ... have become very numerous in the neighbourhood and destructive to human life. The branches of these trees are supported by forked sticks or poles, and there are three tiers or platforms on which the huts are constructed. The lowest is nine feet from the ground, and holds ten huts, the second about eight feet high, has three huts, and the upper storey, if it may be so called, contains four. The ascent to these is made by notches cut in the supporting poles, and the huts are built with twigs thatched with straw, and will contain two persons conveniently." (Steedman, 1835)

The validity and common sense of such structures is difficult to establish. Stilt architecture elsewhere in the world has usually been associated, in the past, with low-lying marshy lands, a condition which can hardly be said to have prevailed in the semi-arid regions inhabited by the Hurutshe. 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. However these predators do not seem to have presented undue terrors for the Hurutshe's neighbours nor, for that matter, for the rest of southern Africa. Hence the existence of such a building tradition must be regarded as a curious enigma in the history of indigenous architecture.

An interesting development dating from this era is the early introduction of immigrant building forms and technologies into this region. The first such example was recorded in June 1820 when Campbell visited the homestead of Seretse at Phatwane, in the northern Cape. There he recounted that:

"We ... visited ... his house, which is built in the fashion of the country; but he has one behind constructed after the European manner, having taken the plan of it from those of the Missionaries that he had seen at Lattakoo (Dithakong). I made him a present of two saws, two chissels and four gimlets. Never did a hungry man receive food with more avidity than Seretz received these tools." (1822: 2: 81-82)

The exact form taken by this dwelling built "after the European manner" has not come down to us but it may be assumed that it was a rectangular plan structure, single or two cell, probably gabled at either end. It is also not known whether it was the result of indigenous effort or, as in other contemporary cases elsewhere in southern Africa, the work of some itinerant European builder. However it should be noted that although square plan dwellings begin to appear in Tswana architecture henceforth with increasing frequency, their use was generally limited to their leadership, principally their chiefs and the headmen. The adoption of immigrant building forms and technologies did not become more widespread until over one century later.

THE TSWANA DWELLING DURING THE COLONIAL ERA

When viewed from an architectural standpoint, the years between 1840 and 1925 were to prove an important period in the history not only of this region but of southern Africa as a whole. 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 Mzilikazi's Matabele to Zimbabwe early in 1837. This may finally have brought to an end fifteen years of turmoil on the highveld, but it also 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, an international flood of miners, land speculators, fortune seekers and assorted camp followers 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 social and physical environment.

  1. It created areas of urbanisation which continue to provide, to the present day, opportunities for cross-cultural contact and mixing. The beginnings of a breakdown in historical patterns of rural culture throughout southern Africa can be traced back to this period.
  2. It created centres of industrial development which were to bring into the region new technologies and materials. With time, these were to filter through into the rural areas.
  3. These centres of industrial development also effectively plundered the surrounding countryside of much of its timber resources thus forcing Tswana architecture away from its historical technologies and building forms.
  4. It created a demand for labour which, at its onset, was only partly met by local communities. The political subjugation of Black rural groups, the subsequent imposition of hut taxes and the creation of markets for consumer goods in the rural areas were to provide the stimulus which was ultimately to turn rural southern Africa into one vast pool of migrant labour (figure 15).
  5. It created markets for the sale of surplus rural food production. Initially this was predominantly in the hands of Black farmers but, with the imbalanced development of a support infrastructure in White rural areas after 1925, this position was to change considerably. Important factors after this date which should also be considered are the curtailment of Black land ownership and the increasing impoverishment of formerly fertile Black rural lands through overgrazing, subsistence farming and overpopulation.

It is obvious that the factors listed above present only a limited view 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.

Accounts of Tswana architecture made by White immigrants and travelers to the region during this era confirm the general picture gained up to and including the Difaqane (figure 16). The dominant dwelling form, the cone on cylinder, often surrounded by a circular verandah about its perimeter, survived through this period and well into the twentieth century with relatively few changes. Such dwellings were recorded by Baines near the Ki-Gariep or Vaal river in 1850 (1961 and 1964: 180-181), by Holub in the northern Cape, western Transvaal and eastern Botswana in 1873-77 (1881: 1: 372)(figure 17), by Wangemann at Thaba 'Nchu in 1867 (1871-75: 1: 89), and by Merensky at Sekhukhuni's capital in 1882 (1875: 84-86). Ordinary cone on cylinder structures were also recorded by Holub at Kimberley and among the Rolong in the western Transvaal, both in 1873 (1881: 1: 409).

Further evidence of a Tswana tradition in wall decoration was recorded by Baines, when he visited a homestead on the Vaal River in December 1850. He told how the courtyard walls were:

"... of mud about six feet high, also circular, nicely levelled and polished, more particularly on the inside, where it was fantastically marked off into square, triangles and diamonds, marked either by straight or waved lines in light drab clay upon a darker ground." (1961 and 1964: 180-181)

The introduction to this region of the flat-roofed, square plan dwellings, more commonly known today as the "highveld" or "parapet" house, can probably also be dated to this period. This is a dwelling form which, over the past two centuries, has been identified with a variety of urban as well as rural environments. Its origins may be traced back to Cape Town as early as 1717 (Dennis Radford, pers comm) but by the end of the eighteenth century its use had become associated with the local Malay community. Originally this consisted of a central doorway giving onto a living/cooking area with a sleeping room located to one side of it. Windows were set equally about the door opening giving the front facade a symmetry which suggested a balanced disposition of plan which was seldom retained. The door and window reveals were often emphasized with broad plaster surrounds as were the façade corners. The parapet walls rose on three sides of the low-pitch roof allowing rainwater drainage to occur to the rear and away from the facade. The front parapet was often corniced although in some examples the plaster mouldings were heavily ornamented (Lewcock, 1963).

The flat-roofed dwelling was soon found to be highly suited to the semi-arid conditions of the southern African interior and during the nineteenth century its use spread to the towns and farms of the Karoo. Subsequently, as Dutch farmers emigrated into the OFS and Transvaal regions, they took the knowledge of this architectural form with them, leading Holub to remark in 1873 that:

"In its general aspect Fauresmith is very like the other Towns in the Free State ... and the clean, white-washed residences, flat roofed as elsewhere ... looked altogether pleasant enough." (1881: 1: 43)

Despite the availability of corrugated iron sheeting in the southern African interior from the 1870s onwards, the construction of flat-roofed dwellings was slow to spread into indigenous Tswana architecture. There is sufficient evidence to show that, outside of certain urban areas, this domestic form did not meet with immediate acceptance and it is probable that its use in rural areas did not occur much before the 1940s.

TRANSITION AND CHANGE IN TSWANA ARCHITECTURE

Previous discussion has shown how, from 1866 onwards, the development of major urban and industrial centres in the southern African hinterland had major repercussions upon the cultural and environmental fabric of rural Tswana society. Their impact was aggravated by a general deforestation of the region and a prolonged period of drought which was followed successively by plagues of locusts, the rinderpest epidemic of 1897 and the South African conflict of 1899-1902. As a result, architectural traditions of the region were forced to undergo a number of adaptations, usually involving the application of new building forms, materials and technologies. Amongst others, these included:

  1. The introduction of rectangular and square plan dwellings. Despite the implication that these represent an obvious example of cross-cultural borrowing between immigrant Whites and indigenous Blacks, the development of such architecture must be seen in the context of a growing cash economy among the rural Tswana. This made available to them a wide range of mass-produced artifacts, including furniture, which, being based upon a right-angle geometry, made their inclusion in circular living spaces difficult and inefficient. It should also be borne in mind that most Tswana homesteads currently still include circular as well as square plan dwellings. The latter therefore should not be perceived as supplanting older and more traditional Tswana domestic forms but rather as supplementing and enriching their range of architectural expression.
  2. The introduction of a sun-dried or "green" brick technology, more commonly known throughout southern Africa as "Kimmerli" or "Kimberley" bricks. As indicated by their name, these probably originated in the 1870s in the diamond fields where a lack of infrastructure made the large-scale transport of such building materials from the coastal areas uneconomical. However it must also be pointed out that the Tswana already had a brick technology of their own in existence before Whites ever reached the highveld. The Tswana brick, called "polwane" differed from the European "green" brick, locally referred to as "ditena" (from the Afrikaans "steen"), in that it was smaller, flatter and made by hand.
  3. The replacement of the Tswana's historical timber-orientated wall building technology with one using sun-dried bricks. This was forced upon local builders through the depletion of the region's already meager timber resources.
  4. The breakdown of historical Tswana roof thatching technologies. This was one of the unforeseen results of a series of land acts enacted in South Africa during the 1920s and 1930s which limited the ownership and settlement of the Tswana upon prescribed reserves. With subsequent population expansion many of these areas became overcrowded and overgrazed, thus increasing the competition already existing between rural communities and their cattle for an already limited resource.
  5. The depletion of natural building resources, timber and grass, which at one time had been freely available from the environment about them, forced the Tswana to seek out alternative sources of materials. This need was partly met by the availability, from the 1870s onwards, of corrugated iron sheeting. Although this material was initially expensive and limited in use to the urban areas, with time it also began to percolate through to the rural countryside. The introduction of corrugated iron as a roofing medium also had the effect of forcing a change of floor plan upon Tswana builders. An introduction of square plan dwellings should therefore also be seen as an environmental response quite divorced from the idea that it was the result of direct borrowing from an immigrant European culture.

Although missionaries had been urging changes upon the architecture of the Tswana from as early as the 1820s, the side-effects of industrial development did not begin to manifest themselves upon the countryside, until well into the twentieth century. The records of Duggan-Cronin, who traveled widely among the Tswana in the 1920s, fail to show much western impact upon local building customs up to that point in time (1929). It is probable that what changes have since been wrought in their built environment, only began to take overt form from the 1940s and 1950s onwards. Even so, a survey of most Tswana towns will show that historical building forms and techniques are still predominant over foreign influences.

CONTEMPORARY TSWANA DOMESTIC ARCHITECTURE

Despite the fact that the Tswana have, for the past nine generations, been presented with powerful incentives to change the nature of their dwelling forms and building technologies, they have remained steadfast in their attachment to the aesthetics and techniques of previous pre-colonial eras. Where modern materials have been adopted or new forms devised, these have been incorporated into the built environment without abandoning the essential character of Tswana rural architecture. Today their homes bear a distinct regional identity which proclaims their origins just as visibly as if they were to attach banners to them.

It is true that, since they were visited by Daniell in 1801, some Tswana dwelling forms have been abandoned or have become archaic. This is most particularly true of the verandah cone on cylinder which may still be widely found in Botswana, but of which only scattered examples are still being built in parts of South Africa. Its construction however varies little from earlier times. In the western Transvaal and northern Cape a series of timber posts or columns some 2,100 m high are set in a circle and support a timber ring beam which, in its turn, carries the conical roof structure, battered and thatched with the grass bundles laid seed-end upwards. hatching takes place from the eaves upwards. The wall is constructed in monolithic earth or sun-dried brick plastered over with "daka", a mixture of clay and cow dung, and usually stands 600mm to 1,000m within the external perimeter of posts, thus providing the dwelling with a verandah (figure 18). In Botswana construction is similar but eaves are often lower and verandahs deeper thus allowing a number of domestic function to take place beneath them. Thatching in Botswana also appears to follow an older tradition: the thatch is not sewn down (referred to as the "Boer" style in South Africa) but is held down by means of an overlaid network of grass ropes. In all cases the inner drum or sanctuary appears to have been dispensed with altogether, although many examples were noted where the circular living chamber was partitioned into two or even three compartments.

Several variations on the theme of the verandah cone on cylinder were recorded in the northern Cape. In the region immediately west of Mafikeng several examples were noted where the circular plan was retained except for the frontage of the dwelling which had been "flattened" at the doorway. Another, at Genesa, had been similarly treated but the room had been extended into the rear verandah space to create two storage rooms. In most examples recorded the central post had been retained (unusual in folk architecture elsewhere in southern Africa) and, in a few isolated instances the dwelling had two doorways, one to the front and one directly opposite, to the rear.

In most cases it was found that the concept of a verandah had fallen by the wayside and that this dwelling form had been supplanted by the simple cone on cylinder. However the verandah tradition was found to have survived in many locations in the form of square plan dwellings either surrounded by a small overhang or, more commonly in the Transvaal, with a wide verandah running the full length of the dwelling facade (figure 19).

Although the use of square plan dwellings is not uncommon in contemporary Tswana architecture, their use is also by no means widespread. Isolated rectangular structures, either gable-ended or more often covered with a hipped roof, have been recorded in various locations both in South Africa and Botswana. However the use of parapet or flat-roofed structures appears to be a feature of more urbanised areas such as Winterveld, north of Pretoria, and parts of the northern Cape. Relatively few rural examples have been recorded to date.

The temporary hut has always had a place in Tswana architecture, being used in the past for such varied functions as the shelter of herd-boys, travelers and harvesters.  Latter-day temporary structures have been recorded more recently in the northern Cape, to the north and west of Mafikeng. These were essentially lean-to buildings consisting of two major timber posts placed some 4,000m apart and bridged by a timber beam laid horizontally at a height of about 2,000m. Sheets of corrugated iron were then leant against this rudimentary timber frame from either side and the structure was braced by two triangular end-walls, one of which included the doorway opening. This resulted in a triangular faced dwelling whose form was not unlike that of the hardbieshuis, a structure originally used by Dutch settlers to the region up to the 1840s.

SOME GENERAL PRINCIPLES OF RURAL SETTLEMENT

The idea of "rural settlement" as a series of huts placed haphazardly over the southern African countryside is plainly mistaken. Not only is the indigenous homestead subject to a number of principles which govern the distribution of its constituent parts but such principles may be interpreted to represent a cognitive and unspoken language which involves not only the social, political, economic, sexual, inheritance and religious values of the people concerned but, in many ways also acts as a mirror to their cosmological beliefs as a whole (Frescura, 1985: 349-363). As such therefore, it may be concluded that the usage of space within the domestic settlement represents a more reliable guideline to cultural pattern and group identity than other elements, such as dwelling form and decorative motif, which historical research has proved to be open to pragmatic manipulation (Frescura, 1987: 263-271).

When applied to a wide range of examples, research has also shown that the southern African region is currently the home to two major architectural groupings: the Venda, a small minority inhabiting the northern reaches of the Transvaal, and the larger Nguni/Sotho/Tswana grouping, or virtually the remaining indigenous population of southern Africa (Frescura, 1985: 364-367).

The principles or, if we wish, the vocabulary of such a language of settlement are based upon the following factors:

  • General architectural form and governing axes.
  • Aspect and direction of approach and orientation.
  • Hierarchical determination and the relationships existing between husband and wife (or wives), parents and children, family and visitors, domestic unit and community.
  • Inheritance patterns.
  • The architectural expression of territorial statements, common spaces, work areas and privacy.
  • Areas of interaction between the individual family unit and the community as a whole.
  • Position of byre and other animal enclosures.
  • Position of cooking facilities and attendant work functions.
  • Location of burial areas and places of religious significance.
  • Other considerations such as seating hierarchies, and spatial divisions determined by sex and age.
HISTORY OF TSWANA SETTLEMENT

Our current understanding of the principles which underlie the structuring of Tswana settlement was only gained in relatively recent times. Early travelers to the region were initially unable to perceive any but the most general of forms (figure 20) and significantly the majority of their reports concentrated upon the individual dwelling unit. Some, like Burchell, who described Dithakong in the northern Cape in 1812 as follows:

"The buildings were nowhere ranged in the form of streets, nor placed according to any regular plan; but were scattered about, in some places far apart, and in others standing so closely together as not to admit a passage for my waggons between them." (1953: 254-256)

could find little sense in their layouts and although he later stated of the same town that it "... may be considered as a collection of little villages, each under the superintendence of its own chieftain ..." (1953: 360-372), he was unable to make the link in his own understanding between the "collection of little villages" and the extended family structure of the Tswana. On the other hand, most early visitors to the Tswana were to express surprise and astonishment at the size and scope of their towns. Dundas reported of Dithakong in 1801 that:

"Passing through several large tracts of ground, that were laid out and cultivated like so many gardens, we arrived about noon at the city of Leetakoo, not a little astonished to find, in this part of the world, a large and populous city." (Barrow, 1801 and 1804: 114-116)

The size of such towns varied quite considerably. Lichtenstein estimated the population of Old Dithakong in 1805 at about five thousand inhabitants, a figure confirmed by Burchell seven years later (Lichtenstein, 1812 and 1815: 2: 379; Burchell, 1953: 360-372), whilst Campbell, who visited the Hurutshe town of Kaditshwene in 1820, said:

"It was impossible to number the houses in Kurreechane, but probably the population may amount to sixteen thousand, it being at least four times the size of New Lattakoo." (1822: 1: 220-277)

Despite their relative size, Tswana towns were by no means permanent settlements. Dithakong, the capital of the Tlhaping, is known to have been moved three times between 1802 and 1814, probably as the result of internal schisms within the community and the need to find fresh sources of grazing and building materials. However these dislocations are not thought to have involved more than ten to twenty kilometres on each occasion (Burchell, 1953: 360-372; Campbell, 1822: 1: 67).

Travelers through this region were also led to comment on the quality of urban life enjoyed by the Tswana. Burchell visited "Kramoni's" homestead at Dithakong in 1812 and found that "It was one of the largest houses; nor could it be excelled by any, in neatness and in the cleanliness and good order of every part." (1953: 315-316), whilst Campbell remarked a year later of the same town that:

"One thing which makes an African town appear to a European as dull and deserted is, the almost total absence of smoke; but while we were walking, about sunset on the brow of a hill opposite to the city, the fires were lighted up, it being a feast day, and the cloud of smoke which in consequence hovered over the city, reminded me of London." (1822: 1: 220-277)

The same author also commented in 1820 that:

"The stillness that prevails universally over (Kaditshwene) in the night, particularly strikes a stranger; indeed it is so great, that, when there happens to be no wind, if a person coughs loud, all the dogs around bark." (1822: 1: 229-277)

However the first comprehensive analysis of the Tswana built environment was only made by Schapera in the 1930s. In 1935 he defined the Tswana "ward" as a settlement containing households belonging to segments of one or more agnatic lineages whose constituent families are either closely related to the headman or otherwise related through the male line to one common male ancestor (1935: 203-224). Subsequently, in 1953, he stated that:

"The Tswana live in large compact settlements ... In appearance the Tswana village is typically a cluster of small circular hamlets separated from one another by narrow lanes or broad roads. Each hamlet is inhabited by a single ward or sub-ward and their number varies with the size and organisation of the population. Villages are normally situated on the banks of rivers or at other places where water is readily accessible, and preference seems to have been given to sites where hills afford protection from enemy raiders.

"(The Tswana Homestead) is the smallest well-defined social unit. It consists basically of a man with his wife or wives and their unmarried children, but often includes one or more married sons, brothers or even daughters, with their respective families. Polygamous households were fairly common ..." (1953: 35-36)

It becomes obvious from the above that although our understanding of the structuring of Tswana settlements has increased considerably since the first white travelers entered Dithakong in 1801, the nature of these environments has changed little in the intervening 150 years (figure 21). Although present-day settlements are no longer as compact as they were 100 or even 30 years ago, their builders are still being guided by the same basic principles. This does not mean to imply that these, or the society which gives rise to them, are static: merely that they are evidence of a larger cultural framework within which changes are possible without altering the essential nature of that society.

THE CONCEPT OF LEFT AND RIGHT

This is based upon the hierarchical relationship perceived to exist between the "first" and subsequent wives of a polygamous marriage. The interpretation of which hand is assumed to be ascendant varies from group to group and is the subject of numerous historical anecdotes or myths. Generally speaking it may be said that those groups who hold "right" to be superior to "left" explain this by means of a metaphor which reflects an old rural belief that a warrior wields his spear with his right hand and his shield with his left. The right is therefore assumed to have an ascendance over the left being "active" and "aggressive" whilst the left is "passive" and "defensive". The Tswana are known to have built their settlement in a roughly circular shape with a large space, being the men's area and cattle byre, located at its centre. The domestic unit of the first wife of the senior man or pater familias was sited at the head of the homestead on the central axis of the settlement directly opposite the main entry to the central space, with those of subsequent junior wives being located alternately to her left and her right. The homesteads of his brothers or other members of his retinue would then also be located alternately to the left and right of his abode according to their descending order of status in relation to him. They, in their turn, would also follow a left-right hierarchy in the distribution of domestic units for their wives within their own individual residences. Today, although it has been found that the circular fan pattern has fallen largely into disuse and individual homesteads tend to follow a linear form along the lines of land contour, the same considerations of left and right have prevailed and are still being maintained by succeeding generations.

THE CONCEPT OF FRONT AND BACK

A second hierarchical differentiation which may be said to exist in Tswana settlement arises between those domestic units and homesteads closest to the central common, the residence of the parents, and those located on the external perimeter, the residences of married children. Although strictly speaking such a differentiation can best be described as one existing between "centre" and "perimeter", this is only valid for as long as the settlement form remains circular. Once the built environment breaks up into individual homesteads which tend to follow a linear pattern along contour lines, a trend in Tswana settlement which has accelerated in more recent times, then it will be seen that the traditional centre-perimeter apposition will become translated into one based upon the concept of front and back. In the context of the individual traditional Tswana household, the opposition has always been one between "front" and "back" where the dwelling of the parents is located in the fore court or lobe of the homestead whilst the areas of privacy, of cooking and of children's residence have been located to the rear.

THE BILOBIAL DOMESTIC UNIT

Historically, the planning of the Tswana Iron Age domestic unit presents some interesting features. Its form, which archaeologists currently describe as being "bilobial", is known to have been built in the southern African hinterland since the sixteenth century and possibly earlier (Maggs, 1976). It consisted of two circular courtyards, the front being a socialising space and the rear being given over predominantly to work functions. Approach was made through a low semi-circle of stones set on edge into the ground, probably to demarcate the line of a reed screen such as may still be found in some Sotho or Ndebele homesteads. The dwelling was located centrally at the intersection of the two circles and access to the rear work space was made either through the back of the dwelling itself or via a gap left in the courtyard wall to one side of the hut. In many recorded instances the rear court was defined by means of a dry-stone wall of about 1,650m height. The masonry work was usually of a high standard and included the detailing of such problem areas as the conjunction of walls and the provision of outlets for surface water run-off. Wall-ends were neatly pointed and returned (Taylor, 1979) (figure 22). The concept of a bilobial domestic unit may be seen to be the product of the Tswana political system, whereby a polygamous society was concentrated into densely populated villages, possibly through reasons of infrastructure, defense or government. The breakdown of this form of domestic unit, which has taken place in more recent times, can be attributed to an increase in monogamous families and a decentralisation process in larger rural settlements. This has given the Tswana domestic unit more room for lateral expansion than was hitherto available to it. As a result the bilobial pattern has undergone a number of evolutionary changes with the work area to the rear being extended to enclose a number of service courtyards to the back and side of the dwelling unit, thus losing its distinctive "bilobial" form (Frescura, 1981: 154-157). (figure 23)

CONCLUSION

Although this paper has been formulated from a predominantly historical and anthropological viewpoint, this was deemed necessary to enable the lay reader to establish both the depth of Tswana building traditions and the manner in which Tswana architecture acts as a mirror to their social values and cultural mores. It is true that their built environment is not as well documented as that of the Zulu or as photogenic as that of the South Ndebele, but this does not detract from the fact that their settlements and dwelling forms probably represent the longest surviving architectural tradition in southern Africa. The complexity of Tswana settlement, the textures of their dwelling forms and the variety of their painted motifs "excite the astonishment and seize the whole attention of the visitor" as much today as they did nearly two centuries ago when they were visited by Burchell. It is unfortunate that a lack of space has not made it possible to explore other important areas of Tswana architecture such as their use of granaries, their adoption of White vernacular forms like the kapsteilhuis, and a more comprehensive discussion on their cognitive use of settlement space. The question of wall decoration is also worthy of longer and more detailed analysis.

POSTSCRIPT

This paper owes its origins to a colleague who, in 1987, was working in Bophuthatswana, and decided to bring out a journal on local building developments. I am not certain that a second edition of the journal was ever published (The Folk Architecture of the Tswana: A Historical and Current Overview. Motlhatlhana, No 1, March 1987, Pretoria). It was extensively revised and extended, and appeared in the SA Journal of Art and Cultural History (Vol 3, No 2, 148-164: 1989) under the title of An Introduction to Tswana Architecture.

BIBLIOGRAPHY

ACOCKS, JPH. 1975. VeId Types of South Africa. Editor: DJB Killick. Botanical Research Institute, Department of Agricultural Technical Services, South Africa.
ARBOUSSET, T, and DAUMAS, F. 1846. Narrative of an Exploratory Tour ... etc. Cape Town: AS Robertson.
BACKHOUSE, James. 1844. A Narrative of a Visit ... etc. London: Hamilton, Adams.
BAINES, Thomas. 1961 and 1964. Journal of Residence in Africa, 1842-1853. Cape Town: Van Riebeeck Society.
BARROW, John. 1801 and 1804. An Account of Travels ... etc. London: Cadell and Davies.
BURCHELL, William J. 1953. Travels in the Interior of Southern Africa. London: Batchworth Press.
BURROW, John. 1971. Travels in the Wilds of Africa. Cape Town: Balkema.
CAMPBELL, John. 1815. Travels in South Africa. London: Black and Perry.
1822. Travels in South Africa ... Being a Narrative of a Second Journey. London: Francis Westley.
CASALIS, Eugene. 1861. The Basutos. London: James Nisbet.
COLE, Monica M. 1961. South Africa. London: Methuen and Co.
DANIELL, Samuel. 1820. African Scenery and Animals, 1804-1805. London: W Daniell.
DUGGAN-CRONIN, AM. 1929. The Bantu Tribes of South Africa; Vol 2, Section 1, The Suto-Chuana Tribes: The Bechuana. Cambridge: Deighton, Bell. Kimberley: Alexander McGregor Memorial Museum.
FRESCURA, Franco. 1981. Rural Shelter. Johannesburg: Ravan Press.
1985. Major Developments in the Rural Indigenous Architecture of Southern Africa of the Post-Difaqane Period. Unpublished PhD Thesis, University of the Witwatersrand.
1987. Folk Architecture in Transition: A Southern African Case Study. SA Journal of Cultural and Art History, Vol 1, No 3, 263-271.
HALL, Edward T. 1973. The Silent Language. New York: Anchor Books.
HOLUB, Emil. 1881. Seven Years in South Africa. London: Sampson Low, Marston, Searle and Rivington. Vol 1, p 372.
KUPER, Adam. 1980. Symbolic Dimensions of the Southern Bantu Homestead. Africa. Vol 50, No 1: 8-23.
LAUGIER, Marc-Antoine. 1977. An Essay on Architecture. Translated by Wolfgang and Annie Herrmann. Los Angeles: Hennessey and Ingalls.
LEWCOCK, Ronald. 1963. Early Nineteenth Century Architecture in South Africa. Cape Town: Balkema.
LICHTENSTEIN, Henry. 1812 and 1815 Travels in Southern Africa ... etc. London: Henry Colburn.
LORD, WB, and BAINES, Thomas. 1976. Shifts and Expedients of Camp Life, Travel and Exploration. Second Edition. London: H Cox.
MAGGS, Tim MO'C. 1976. Iron Age Communities of the Southern Highveld. Pietermaritzburg: Council of the Natal Museum.
1972. Bilobial Dwellings - a persistent feature of Southern Tswana Settlements. South African Archaeological Society. Goodwin Series, No 1.
MARC, Olivier. 1977. Psychology of the House. London: Thames and Hudson.
MASON, Revil. 1974. Background to the Transvaal Iron Age: New Discoveries at Olifantspoort and Broederstroom. Journal of the South African Institute of Mining and Metallurgy. Vol 74, No 6, January 1974.
1981. Early Iron Age Settlement at Broederstroom 24/73, Transvaal, South Africa. South African Journal of Science, Vol 77, No 9, 1981: 401-416.
MAYLAM, Paul. 1986. A History of the African People of South Africa. Cape Town: David Philip.
MERENSKY, A. 1875. Beitrage zur Kenntnis Sud-Afrikas. Berlin.
OMER-COOPER, JD. 1978. The Zulu Aftermath - A Nineteenth Century Revolution in Bantu Africa. London: Longman.
RAPOPORT, Amos. 1969. House Form and Culture. New Jersey: Prentice-Hall.
1977. Human Aspects of Urban Form. Oxford: Pergamon Press.
1982. The Meaning of the Built Environment. Beverly Hills: Sage Publications.
ROUSSEAU, Jean-Jacques. 1959. Oeuvres Completes. Edited by B Gaguebin and M Raymond. Paris.
SANDERSON, J. 1860. Memoranda of a Trading Trip into the Orange River (Sovereignty) Free State and the Country of the Transvaal Boers, 1851-52. Journal of the Royal Geographical Society, Vol 30, 1860: 233-255.
SEYMOUR, Wilfred Massingham. 1982. Customary Law in Southern Africa. Cape Town: Juta.
SCHAPERA, I. 1935. The Social Structure of the Tswana Ward. Bantu Studies, Vol 9 No 3 September 1935. pp 203-224.
1953. The Tswana. London. pp 35-36.
SHILLINGTON, Kevin. 1985. The Colonisation of the Southern Tswana, 1870-1900. Johannesburg: Ravan Press.
SMITH, Andrew. 1975. Andrew Smith's Journal of his Expedition into the Interior of South Africa: 1834-36. Cape Town: Balkema.
STEEDMAN, Andrew. 1835. Wanderings and Adventures in the Interior of Southern Africa. London: Longman.
TAYLOR, MOV. 1979. Late Iron Age Settlements on the Northern Edge of the Vredefort Dome. Unpublished MA Dissertation, University of the Witwatersrand, Johannesburg.
VIOLLET-LE-DUC, EE. 1876. The Habitations of Man in All Ages. Translated by Benjamin Bucknall. London: Sampson Low, Marston, Searle and Rivington.
WALTON, James. 1956. African Village. Pretoria: Van Schaik.
1965. Early Ghoya Settlement in the Orange Free State. Bloemfontein: Memoir No 2, Researches of the National Museum.
WANGEMANN, Dr.1871-75. Lebensbilder aus Sudafrika" and "Geschichte der Berliner Missions Gesellschast und Ihrer Arbeiten in Sudafrika. Vols 1 to 3. Berlin: Im Gelbstverlag des Ev Missionshauses in Berlin.
WILSON, Monica, and THOMPSON, Leonard. 1975. The Oxford History of Southern Africa. London: Oxford University Press.

LIST OF ILLUSTRATIONS
  1. Language map of southern Africa: after Van Warmelo, 1952.
  2. Distribution of rural dwelling forms before 1800.
  3. Dwelling of Kramori, recorded by Burchell at Dithakong in 1812.
  4. Dwelling of Moleni, recorded by Burchell at Dithakong in 1812.
  5. Distribution of rural dwelling forms, 1801-1820.
  6. Tlaping homestead: Daniell, 1805. The inward slope of the inner drum wall is distinctive.
  7. Pencil studies for Tlaping homestead: Daniell, undated c1801. The beehive form of the dwelling below the verandah roof is clearly evident.
  8. Interior of Hurutse chief's residence at Kaditshwene: Campbell, 1820.
  9. Herdboy's corbelled stone hut: Baines, 1876.
  10. View of Motito Mission Station: Backhouse 1839.
  11. Section through a Rolong dwelling: Casalis c1833.
  12. Distribution of rural dwelling forms, 1821-1840.
  13. "Houses where lions abound." Stilt dwellings recorded by Smith, 1835.
  14. "The Inhabited Tree": Steadman, 1835.
  15. A group of migrant workers returning from the Diamond Fields encounters a group of new recruits on their way there. The material "benefits" to be derived from this work is emphasised by the nature of the "European" clothes worn by the men on their way home.
  16. Distribution of rural dwelling forms, 1841-1890.
  17. Mangwato homesteads at Serowe: Holub, 1874.
  18. Dwelling at Ntshteleng, western Transvaal, September 1976.
  19. Dwelling at Groot Marico, western Transvaal, September 1976.
  20. Campbell's view of Dithakong in 1813.
  21. Aerial photograph of Tswana cluster homesteads, c1976. The resemblance between this and Campbell's view of Dithakong is striking.
  22. Excavated plan and reconstruction of Tswana bilobial homestead, c1750. After Maggs, 1974.
  23. Tswana bilobial, c1960.
Copyright @ francofrescura.co.za