Studies of space within the towns have tended to draw on the rich ethnographic record available for the towns of the twentieth-century coast, and particularly Lamu, on the northern coast of Kenya (el-Zein 1974; Ghaidan 1975; Middleton 1992; Prins 1971). This somewhat anachronistic approach to the exploration of ancient town plans is nonetheless valuable in delineating certain principles that may have been present over the long term. The influence of Islam on Swahili urban spaces is, for example, an interesting avenue because the period of greatest town growth in the fourteenth to ifteenth
Centuries was also a time of large-scale mosque construction. It is therefore necessary to look to the Islamic world for the kind of "high-level" structuring principles, or "normative urban theory" (M. E. Smith 2011:180) that would account for the overall form of a town according to cosmology or world view (cf. Kelly and Brown, Chapter 9 in this volume).
SWAHILI URBAN SPACES Of THE EASTERN AFRICAN COAST
At this level, Swahili towns might be seen as part of the world of Islamic cities that have been explored in detail elsewhere. The extent to which Islam creates a distinctive pattern has been widely debated: the idealized (and orientalized) "Islamic city" based around the mosque, market, and public baths has been sought across the Islamic world (Lapidus 1969; al-Sayyad 1991). In Swahili towns - which do not contain these latter two features - the influence of Islam has instead been seen through the centrality of the mosque, the evocation of concepts of ritual purity, and on the seclusion and control of women.
The most extensive treatment of the structures of Swahili urbanism draws on Islam as an overarching cosmology that provides context and content for the speciic patterns seen on the eastern African coast. El-Zein (1974) considers contemporary Lamu, and his account is based largely on the testimony of informants of the 1970s. This has important ramiications for the ways that Swahili tradition is presented in the immediate postcolonial context, with much weight given to historicity; a sense of Lamu society as timeless and enduring, as well as an emphasis on notions of pedigree, entitlement to land, and to local identity. Nevertheless, it represents an important study for the exploration of the structures of Swahili life here, based around concerns of Islam and historical memory. El-Zein's structuralist approach seeks the deep grammars of Swahili social interaction, explained through a series of conceptual oppositions traced through the oral histories related within the town. The Lamu "myths of creation" are given particular precedence (el-Zein 1974:167-220), shown to be based upon the monotheistic tradition with certain local adaptations and interpretations; through these myths, el-Zein draws a binary distinction between light and dark, white and black, angels and jinns (unseen spirits that were the irst inhabitants of the earth), and life and death. For him, these structure every aspect of social life, including spatial prescriptions, and they are "not only a logical model but also part and parcel of the social reality shared by the people" (el-Zein 1974:172). El-Zein thus suggests that the particular
Ways that Islamic beliefs were structured in the town of Lamu were of fundamental importance in the shaping of the urban milieu. This sense echoes throughout architectural histories of Swahili towns, with Gensheimer's study of Swahili urban spaces concluding with the suggestion that it was the:
Acorporeal world of spirits and dead ancestors, which were used to distinguish the palaces, mosques, houses and tombs within the Swahili city. This figureless and formless construct of the city which gave meaning to the built environment and guided its construction, this city of the mind, was the essence of the SwaHili city (Gensheimer 1997:359).
The notion of purity identified by el-Zein is also picked up in an architectural study of Lamu conducted at roughly the same time (Ghaidan 1975). Despite eschewing the notion of an overarching plan for the town, Ghaidan (1975:61) claims that Swahili concepts of space are based around systems of behavior that require specific "shells." He identifies these shells in three guiding principles for the architecture of the town, which could be extended to the town plan: purIty, involvement, and pedigree. Purity refers largely to ritual purity and the avoidance of pollution. Like el-Zein, Ghaidan sees Swahili urban spaces as being structured around the concerns of maintaining this purity through the strict segregation of clean and unclean activities and persons; elsewhere, he links this explicitly to the stonehouses and associated notions of privacy (Ghaidan 1971). Again, this is linked back to Islam and to the specific rituals that accompany observance within Lamu. Involvement and pedigree are more worldly concerns, associated with the establishment of social relations through spatial proxemics (involvement) and to the maintenance of exclusivity for elite groups (pedigree). For Ghaidan, these relations of kinship and distinction are key overarching principles of the same order as the concern with privacy, demonstrating the overlapping nature of "high-" or "mid-level" meanings and the fact that in practice, these are not separated out. Pedigree leads, he suggests, to the establishment of certain delineated areas associated with particular social classes, and is the reason for the maintenance of historical tradition and the importance attached to ancestral places. Thus, the bipartite division of Lamu (Figure 4.2) into the wards of Mkomani and Langoni (see also Prins 1971) reflects an ongoing concern with social distinction between patrician and commoner, a belief about
Figure 4.2 Schematic plan of twentieth-century Lamu, showing bipartite division and locations of mosques (adapted from el-Zein 1974:16).
The structure of society rendered in spatial form. Likewise, the predominance of trade-related structures (market stalls and craft workshops) in Langoni is linked to their lesser status, while the patrician ward of Mkomani is instead purely residential, and home to eleven of Lamu's nineteen mosques (Ghaidan 1975:62-64). Pedigree is therefore seen to be mapped fairly straightforwardly onto the town plan of contemporary Lamu, reflecting the social order in spatial form. Involvement, by contrast, creates the character of the Swahili town: Ghaidan argues for an extremely sociable form of urbanism (contrasting strongly with the Weberian model of alienation as intrinsic to urban life). The high involvement ratio has, for Ghaidan (1975:71), "set the scale of streets and open spaces at an intimate level. . . Involvement is also responsible for the absence of any expression of grandeur from Swahili architectural patterns."
Although Ghaidan and el-Zein locate the concept of purity within the realms of overarching cosmology or the "city of the mind," their development of the concept sees it played out through the more quotidian concerns of what have been termed "mid-level" meanings (Rapoport 1988; Smith 2007). These are more bound up with the social negotiation, communication, and power struggles of the inhabitants as they seek to inscribe their vision onto the urban spaces. The notions of involvement and pedigree, although elevated to the level of cosmology or worldview by Ghaidan, also clearly fall into this category. For these writers, the centrality of the mosque reflects its central position in the Islamic worldview, but is also interpreted through the social negotiations and communications of the
Residents of Swahili towns. Thus, for el-Zein (1974:14-15), the mosque represents the unity of the town beyond the multiplicity of ethnic and social backgrounds found in the various mitaa or quarters, but also, in its Mkomani position, as part of the power negotiations between the khatih or religious leader and the waungwana or mercantile elite, who are most associated with the Pwani mosque near the foreshore. The separation of worldly and religious power is seen reflected in this separation of the mosques themselves and their locations within the town. This conflation of spatial and conceptual centrality is also assumed for the earlier towns (Garlake 1966:3), as will be discussed.
Apart from the prominence of mosques, however, suggestions of how Islam might have influenced town planning focus around the ways that daily life was structured and movement constrained by the concepts of purity and the kinship relations of contemporary Lamu inhabitants. This is similar to the ways that the Islamic city has been visualized elsewhere, as analysis has moved away from the more rigid deinitions of speciic institutions, and toward a sense of the ways that Islam structures daily life and - through that - creates a distinctive urban form. Islamic cities are deined by Abu-Lughod (1987:172) as "processes, not products," born of an interplay between the prescriptions of religion and the ways that this is exacted in diverse environments and social settings. Wheatley (2001) has aptly demonstrated how the same fundamentals of Islamic practice led to diverse and functionally distinct urban forms in different areas - the centrality oF the mosque often the only point of commonality.
The literature on high-level meaning within Swahili towns is therefore actually quite varied. Despite a claim to speak of structuring principles, and the ways that these controlled daily life in and around the town, most of the cases we examined assume that spatial differentiation will reflect social differentiation without much analysis of how this process occurs. The classic model of this is the didemic Lamu model, with the fundamental division between Mkomani and Langoni seen as reflecting a social division that is further broken down into mitaa or clan distinctions. Although this demonstrates the interplay between top - and mid-level meaning at all stages, it is actually at odds with the kind of process-based analysis el-Zein anD Ghaidan suggest, by which activity is structured through overarching principles, and then creates a unique Islamic urban form: instead, the urban plan is seen simply to mirror the
Structure of society. The difference is between urban plans based on high-level meaning - pedigree, purity, Islam - and one where urban plans reflect sociaL divisions.
World History
