The majority of Swahili towns were allegedly founded between the tenth and twelfth centuries. The first stone buildings date to the tenth century in Shanga and to the end of the eleventh century in Kilwa and Gedi (Horton 1996; Pradines 2010: 27). The growth of the population in the fourteenth century created demographic pressure, which led to the abandonment of some sites and the creation of new towns. From the fifteenth century onward, Swahili towns were surrounded by stone walls, as notably at the Kenyan sites of Pate, Ungwana and Gedi (Pradines 2004: 328-334).
These urban fortifications gave the developing political power some security and reinforced the distinction between town and country. The emergence of domestic architecture in stone differentiated from traditional African houses symbolized hierarchy among the Swahili community. Only rich merchants and individuals from powerful lineages could live in stone houses. This change in habitation is confirmed by observations at Gedi where architecture profoundly altered at the beginning of the fifteenth century.
The restructuring of the cities on the east coast is related to the apex of Swahili trade. This widespread phenomenon has been observed from the Lamu archipelago, to northern Kenya and all the way to Kilwa in southern Tanzania. In the sixteenth century, many Swahili ports were abandoned by commercial traffic because of the Portuguese intrusion in the Indian Ocean. The towns were then protected by forts, such as the Portuguese fort of Mombasa or the Omani fort of Kilwa. From the eighteenth century, the Omani presence led to a renaissance in Swahili architecture and urbanism. The towns gained height with multi-storey buildings.
The genesis of the Swahili house is still subject to discussion, between supporters of a local evolution, including Mark Horton and John Middleton (Horton & Middleton 2000: 119) and those favouring exogenous influences, including Linda Donley-Reid, Abdul Sheriff and the current author (Donley-Reid 1990: 114-126; Sheriff 2002: 76; Pradines 2004: 111-112). This new technology was first reserved for mosques and palaces, before becoming more widespread, being used for a group of large private houses in the fourteenth century. Stone buildings became emblematic of the power of the wa-ungwana patricians/nobility and represented the very notion of the Swahili city.
Mark Horton has argued for continuity in the evolution of Swahili houses, from traditional earth-built dwellings to large houses in stone, in terms of consistent traits in the plans of the buildings and in the use of space. Unfortunately, examples prior to the fourteenth century are virtually non-existent. A further problem is the existence of very similar architecture in Yemen, the occupation of which by Indian merchants is confirmed by historical sources. Horton also asserts that stone architecture is a local invention reinforced by Fatimid technologies from the Red Sea. He cites multiple examples of architecture in coral limestone in the Dahlak Islands, Er Rih, Aydhab and Suakin. However, this technology was not only found on Egyptian and Sudanese shores, with examples existing on the coast of the Hadramaut and in the Persian Gulf, in Bahrain and in Qatar where coral blocks are called hagar al-bahr, or “stones of the sea”.
It is more likely that domestic Swahili architecture from the fifteenth century was strongly influenced by Indo-Persian cultures between the thirteenth and fourteenth centuries. Two elements reflect the role of India in the genesis of Swahili habitations: the division of space into narrow, parallel rooms and the use of niches for wall ornamentation. Linda Donley-Reid has identified three types of Indian house, which are very similar to Swahili habitations. The southern region of Gujarat, from Mumbai to Broach, exhibits houses with two or three parallel rooms and a lateral circulation zone. Women live at the back of the building, which has an exit separate from the main entry situated in the façade. Houses to the north of Gujarat, from Broach to Cambay, have three parallel rooms with a central passage. There is no door at the back and the walls are decorated with niches in a style very close to that found in Lamu.
In Gedi, the plans of the houses in the north-eastern area are very significant: they present narrow, elongated rooms, all parallel except for one forming a side corridor. Some rooms have symmetrically arranged niches in the walls. As just noted, the room layout and the wall decorations are characteristic of Gujurati homes. Domestic units in Shanga had niches during the fourteenth century (Horton & Middleton 2000: 118). These cavities were organized in a symmetrical fashion, with a balancing of space still found in Swahili houses today. From this we infer that the organization of central secluded rooms (ndani) and of niches in modern, traditional Swahili homes dates back to the fourteenth century. The origin of this form of niche comes from Gujurat and not from the African coast, as ornamentation with niches is found in Zabid and Mocha in Yemen, towns strongly influenced by Indian architecture from Surat. Niches found inside homes in Gujurat and Kutch obey a certain harmony, a regularity of space. This balance is based on symmetry of architectural elements. The plaster niches in Indian Bohras’ houses are like miniature mihrabs; the alcoves, called gokala, are often associated with magical squares containing the name of Allah.
In conclusion, it must be said that archaeological research on Swahili settlements is still in its infancy. To progress further, what is needed is collaboration between specialists working in Africa, Arabia, Persia, India and even China.
References on urban studies and the Swahili house
2003 – Islamization and Urbanization on the Coast of East Africa: recent excavations at Gedi, Kenya, Azania 38, BIEA, Nairobi, p. 180-182.
2002 – La bipartition des cités swahili : l’exemple de Gedi (Kenya), in Southern Africa and the Swahili World, Studies in the African Past n°2, 2002, University of Dar es Salaam Press, p. 66-75.
2002 – Architecture domestique swahili, les fouilles de Gedi : la mission de 2002, Nyame Akuma, n°58, Bulletin of the Society of Africanist Archaeologist, University of Alberta, p. 13-18.
1999 – L’influence indienne dans l’architecture swahili, Swahili Forum VI, AAP n°60, Institut für Afrikanistik, Université de Köln, p. 103-120.