Between Mosque and House: An Archaeology of Swahili Open Space

Between Mosque and House: An Archaeology of Swahili Open Space

Jeffrey Fleisher

Rice University

Introduction

Almost thirty years ago, Thomas Wilson (1982: 207) argued that open spaces in Swahili towns “played an important part in the social, political and economic life of the community.” Mark Horton (1994) also explored this theme for Shanga, on the northern Kenyan coast, suggesting that Swahili towns were oriented around a central enclosure, with first millennium origins, that was “fundamental to the spatial organization” of the later town.

1

Central Open Area, Songo Mnara, Tanzania

However, despite these assertions, most suggestions about the nature of open space in Swahili towns have been based on guesses and suppositions, and only rarely has open space been a specific object of study. This derives, I argue, from an emphasis on houses and mosques, the investigation of which has taught us much about social and religious life in Swahili towns. However important these architectural spaces have been in understanding the institutions of Swahili urban life, they remain somewhat adrift within urban space, and require the grounding that a serious investigation of urban space can provide. Along these lines, this paper discusses recent research in the open spaces of Songo Mnara, and aims to describe how such research might proceed, and the specific issues that it raises.

2

confined and delimited space

Alongside recognizing the importance of urban space, Wilson (1982:207) also made the useful distinction between ‘confined’ and ‘delimited space’; the former is coterminous with interior, architectural spaces, while the latter is “formed by the arrangement or occurrence of structures or natural features upon the settlement landscape.” This definition captures well the important linkage between interior and exterior space, especially in urban contexts where much of the ‘open’ or ‘empty’ space is itself defined—or delimited—by architecture. This is a useful distinction, and in this paper I’d like to begin exploring the importance of delimited space in thinking about the functional, political, and commemorative aspects of Swahili towns. After introducing an archaeology of public or open space, I will summarize arguments offered about Swahili urban spaces. Then, I will present recent research on the open spaces of Songo Mnara, an important medieval settlement on the southern Tanzanian coast. These spaces were investigated through a combination of geophysical surveys, geochemical sampling, and excavations. Preliminary findings from this work suggest that Songo Mnara’s open spaces were important zones of commemoration, industrial production, and possibly agricultural fields. These data hint at the rich material practices associated with Swahili open spaces, and suggest that they offer important contexts in which to explore issues of social memory, political authority, and town‐based production and consumption.

Thinking about Open Space

Increasingly, a concern with understanding the putatively ‘open’ and ‘empty’ spaces of urban and other settlements has helped frame new understandings about political and economic uses of space (Fox 1996; Hutson et al. 2007; Low 2000; Moore 1996; Robin 2002; Robin and Rothschild 2002; M. Smith 2008). Much of this current research rejects a notion of space as a “backdrop or container for action” (Robin 2002:248), and now recognizes that the production of space and place is the result of social practices (Ashmore 2002; Blake 2006; A. Smith 2003). Studies of open or empty space have followed two related directions: one focuses on public spaces and the way they serve as contexts for political action; the other examines the way peoples’ everyday actions (practices) were integral to the construction of spatial meaning. For example, open public spaces, such as Mesoamerican plazas (Low 2000; Moore 1996) and ball courts (Fox 1996), have been shown to be important ritual arenas, in which political order was negotiated (and not simply materialized). Thus, the public spatial practices of complex societies are not just symbols of authority, but rather constitutive of it (Ashmore 1989; A. Smith 2003:77). More recently, M. Smith (2008:220) has described how the “use and disuse of empty space provided ample opportunities for the generation of conflict and consensus about the material realm of urban habitation.” Other studies have explored open spaces “outside of houses” (Robin 2002) as a way to break down divisions archaeologists often assume between interior and exterior spaces. This research has paid careful attention to the way outdoor spaces were locations of a host of daily activities, from gardening to ritual acts. This work challenges the notion that domestic life can be connected primarily to the house (Robin 2002:261).

Suggested activities in open spaces

• open‐air meeting places (Garlake 2002)

• market areas (Horton 1996)

• protected space for future town growth (Kusimba 1993:122, 1996: 711);

• gardens and/or orchards (Garlake 2002; Kusimba 1993)

• impermanent architecture

• industrial production (Kusimba 1993:122; Garlake 2002; Gensheimer 1997: 328‐339).

Swahili Open Spaces

Open spaces within Swahili towns commonly include areas within and against town walls; areas surrounding central congregational mosques, especially near entrances and outside the north‐facing mihrab or prayer niche; and central open spaces delimited by domestic architecture. Previous interpretations of these open spaces have addressed them in two ways: as functional spaces and settings for particular types of activities, and as central parts of a spatial plan that represents the social organization of Swahili society.

The functional interpretations of open spaces in Swahili towns include: open‐air meeting places (Garlake 2002) or market areas; protected space for future town growth (Kusimba 1993:122, 1996: 711); gardens and/or orchards (Garlake 2002; Kusimba 1993); areas of impermanent architecture; or areas of industrial production (Kusimba 1993:122; Garlake 2002; Gensheimer 1997: 328‐339). Ethnohistoric and ethnographic data support some of these hypotheses: early 16th century Portuguese accounts indicate that the sultan of Kilwa was crowned in an open area adjacent to the palace, while specially designated open spaces in Comorian cities, called fumboni, were contexts for social gatherings, weddings, feasts, and game playing (Gensheimer 1997:334). Archaeologists have also explored certain aspects of these spaces; on Pemba Island, Tanzania, and at Gede, Kenya, excavations have demonstrated some ‘open areas’ were actually dense with earth‐and‐thatch houses; at Shanga, evidence of possible trade kiosks has been located in the central open space of the town. In addition, the feasts and communal events mentioned in the ethnohistoric data have been hinted at in the archaeology of Chwaka (LaViolette and Fleisher 2009) and Vumba Kuu (Wynne‐Jones 2010).

Idealised plan of early Shanga(from Horton 1994)

Horton has argued, many ancient Swahili towns are structured around a central area or enclosure, containing on a congregational mosque and suggests that this spatial arrangement had social meaning. In the initial stages of its development, Shanga contained a central enclosure, with seven gateways, surrounded by the domestic architecture of the town. Another wall enclosed this settlement, with only four gateways on the cardinal axes. Horton argues that the central area would have been the site of the market, as well as of communal rituals or activities, and where the congregational mosque was constructed in wattle‐and‐daub, and replaced in stone in the 10th century. Horton’s model of Swahili space is thus one in which the spatial layout of the town reflects, or materializes, established social relations; as he suggests (1994:147), the planning of the settlement was “a map with which to express elements of their social and kinshipstructures.”

6

The Open Spaces of Songo Mnara

Due to Songo Mnara’s remarkable architectural preservation it is most commonly invoked as the premier example of town planning (Garlake 1966; 2002; Horton 1984: 179‐180; Gensheimer 1997:336‐338).

Figure 3: Songo Mnara site plan with structures and areas referred to in text

Songo Mnara, Tanzania

Most references to Songo Mnara discuss the central open space, which contains a small mosque, a walled cemetery, and dozens of graves marked by simple headstones.

Monumental staircase, Songo Mnara

The most elaborate houses at the site, in the southern and eastern areas, overlook this space, many with monumental staircases leading directly into it, set up in a way that provides an unbroken vista north across the open area.

View from staircase into central open area, Songo Mnara



However, the site also contains two other delimited spaces, less formally defined, in the western part of the site, flanked by a town wall and the shoreline, as well as smaller space in the northern extent, set among smaller buildings and bound by the northern town wall.

Research on the open spaces at Songo Mnara in 2009 included geophysical surveys, discussed today by Kate Welham, as well as geoarchaeological testing across them by Federica Sulas. Seven trenches were dug in open and outdoor areas: three based on anomalies from the geophysical surveys, two against the exterior walls of domestic structures, and one each surrounding a tomb and well. We also mapped all visible headstones and tombs, providing the first spatial understanding of this important cemetery, which dominates the central space. The results suggest that open space at Songo Mnara was used in multiple ways, including serving as the location of impermanent architecture and industrial production, possible garden or arbor plots, and formal commemorative spaces.

Impermanent Houses and Daily Activities

Two trenches excavated in the zone between the western and central area contained possible domestic debris.

Trench 11

In trench 11, in particular, the density of artifacts seemed to indicate the presence of impermanent houses. This was quite surprising, given the proximity of this area to the formal entrance of the Palace complex, to the south. Additionally, artifacts from this trench provided a rich assemblage of material, including a high style oil lamp, eight copper coins and hundreds of glass and shell beads. Another trench to the north of 11 may also relate to domestic debris but the deposits were more ambiguous.

Trench 13base of staircase of House 23

A number of trenches suggest the presence of activity areas in outdoor spaces. In two trenches against the exterior walls of structures, clear cultural surfaces were identified. While each contained a spread of materials that may be understood as part of a general sheet midden, the debris found in them was quite variable. Trench 13, excavated at the base of a monumental staircase, contained a clearly defined midden layer with a rich assemblage of materials. Trench 2, however, related to a more humble structure, contained deposits less dense with materials. While both contained the same range of materials, the density of artifacts from Trench 13 was much greater than that in 2. These excavations point to the possibility that investigations of outdoor spaces will help reveal distinctions between houses, as well as the range of activities carried out within and without the houses themselves.

Trench 6space around communal well

Similarly, Trench 6, a large unit that surrounded an open air well, provided a rich assemblage of materials, suggesting that this area was home to a host of activities and possibly a neighborhood meeting area.



Ritual and Commemorative Space

One of the main questions that we hope to address through our research at Songo Mnara is the nature of the central open area. The plan of this space, and the features associated with it, seem to support Horton’s notion of a ritually significant central zone within the settlement. What is different about the central area at Songo Mnara, however, is that it contains a large and extensive cemetery. Burials are commonly associated with mosques on the east African coast, often located in the space north of the mihrab; in fact, the congregational mosque at Songo Mnara contains just such a cemetery. However, both the size and central location of the cemetery at Songo Mnara is something different altogether.

central cemeterySongo Mnara

By mapping each headstone and tomb, we have been able to recognize at least five different areas of this cemetery. At present it is not clear if these are actual groupings of tombs, or simply the result of limited visibility of headstones. Horton, however, has argued that similar types of tomb groupings at the main cemetery at Shanga represent different clan groups, and this may be a possible explanation.



It is becoming increasingly clear that the central open space at Songo Mnara was constructed as a public zone of commemoration. Most Swahili cemeteries are located on the edges of town, outside the stone walls, as is the case at Shanga and Kilwa. And Songo Mnara itself contains three other recognizable cemeteries outside the central open space: the small one associated with the main mosque, and two outside the town walls, including the densely packed Necropolis cemetery and an area of tombs and headstones just outside the northeast wall. Far from being tucked away, the cemetery in the central open space at Songo Mnara was situated to be seen and observed. The most monumental houses at the site define the south edge of this open space, and are built on a coral outcrop such that the houses are raised above the rest of the settlement.

Monumental staircase, Songo Mnara

At least four monumental staircases lead from the houses into this open area, each providing a full vista of the area itself.


Based on geoarchaeological research we can also begin to reconstruct the vegetation: from the preliminary phytolith study, we know that it was likely covered in grasses, and was distinguishable from the northern open area which contained more evidence of woody plants. This may suggest that this area was kept clearer than others.

Tomb excavations

Excavations in the area around a stepped coral rag tomb (Trench 12) were aimed at examining whether archaeologically detectable practices were associated with the tombs. This tomb had indications of offering niches on the eastern and western sides, and the remains of what appeared to be ceramic vessels placed as offerings at the surface.

The excavated deposits represent the built up fill around the tomb, which itself rested on subsoil; two other headstones, one each on the northern and southern sides were located, in situ, during the course of excavation.

quartz pebbles

copper Kilwa coins

The finds from this trench offer an intriguing set of materials that are likely associated with ritual offerings. Quantities of local and imported pottery, as well as seven copper coins and five quartz pebbles might be evidence of offerings left next to or on the tomb.

There are ethnographic and historical analogs for these offerings: coins are currently left on tombstones in the Necropolis cemetery, where rounded quartz pebbles are also found scattered amongst the headstones. These stones were described by Burton (1872:359) when he visited the site in the late 19th century, who described an “ancient custom…[in which] small stones are washed, perfumed and sundried…finally, they are strewed with prayers upon the tomb.”

Local and imported ceramics might also be evidence of food left at the tomb, or feasts carried out in their vicinity, both ethnographically‐known practices, as will be discussed by Tunde in a moment. Geochemical samples taken from the excavated deposits at the tomb suggest that there was an abundance of burned plant material indicative, perhaps of food offerings or incense. Additionally, phytolith evidence from these excavations may hint at particular burial practices: in the fill surrounding the headstones north and south of the tomb, there were very high levels of palm phytoliths, which were completely absent on the eastern side of the tomb, and lower levels. This may indicate the use of palm fronds to demarcate these burials, or as parts of other ritual activities associated with the tomb.

Finally, two excavations within the central open area hint at the types of activities that were carried out there. Trench 5, located on the basis of the geophysical survey, contained a dense accumulation of materials associated with iron smithing, including scale and droplets, as well as tuyere fragments. This surprising find seems to define an area on the western edge of the cemetery, surrounded by tombstones. The linkage between smithing activities and the central graveyard is intriguing based on the common relationship between blacksmiths and sacred power in east African archaeologies (e.g., Schmidt 1997). More work will need to be carried out in this area to understand the stratigraphic relationship between mortuary and industrial contexts.



Agricultural Space

Geoarchaeological research in the open areas also helps us think about the possible ways that they were used for garden or orchard plots. Within the northern and western open areas, there are defined areas of nutrient rich soil, soil that appears to be naturally‐occurring on the coral terrace above the site where fields exist today, but not on the sandy flats where the site sits. It is possible that these pockets of soil served as the basis for garden plots. Phytolith evidence offers some guidance on this, suggesting that the northern open area contained more woody plants than the southern one, possibly indicating orchards in that zone. This is consistent with historical observations, as Garlake noted “the stone houses were probably surrounded by orchards and plantations; several visitors, from Ibn Battuta to Vasco da Gama noticed these with admiration” (2002:181). Geochemical sampling also indicates high levels of phosphate in areas against buildings and these might suggest animal penning or other domestic activities in these zones.

Conclusion

Conclusion

Recent research at Songo Mnara thus offers a glimpse of the rich insights that can be learned about Swahili urban life by examining the organization and activities of open space. First, following from my long‐term research in northern Tanzania, we can see that some of the putatively open spaces at Songo Mnara were ones in which non‐elite members of the town lived, and we thus need to be sensitive to these very ephemeral deposits, if we aim to understand the full range of Swahili society. The process of investigating open space also indicates evidence of prosaic activities occurring outside the house, and urges us to unlock the domestic sphere from the confines of the house itself. A more synthetic household archaeology needs to include areas around and near houses if we are to understand the spatial extent of domestic life. Although the coral walls of Swahili houses seem to impose a strict boundary between interior and exterior spaces, they were likely much more permeable in practice. The possibility that circumscribed open areas at the site, particularly to the north and west, were locations for gardens or orchards, also suggests a domestic sphere that has been invisible to us until now. By examining these spaces, we can begin to reconstruct a more complete sense of urban space itself, a sense not easily apprehended through visits to the ruins today.

Perhaps the most interesting parts of the work at Songo Mnara, however, are the initial understandings of the way that the central open area may have served as a commemorative space. The physical placement of the cemetery at the core of the settlement, with vistas from elite houses, suggests that this area was not simply the functional resting place for the dead. Tombs and burials were to be seen and experienced daily, as people walked through the cemetery, made offerings at the edges of tombs, and reaffirmed connections with ancestors. The physical placement of tombs within the core zone of settlement was itself an act of inscribing memory, making public a particular past. This area was very likely a place where social memory was continually engaged, through the maintenance and preservation of the space itself, and the ritual acts performed at tomb’s edge. Unlike the more organic development of the central area at Shanga, the central open space at Songo Mnara was created, maintained, kept clear, and offered—throughout the life of the settlement—as a defined zone of social memory. These acts of social memory would have been an active part of the ongoing constitution of power and authority in Swahili towns. Thus, rather than understanding the organization and use of space within Swahili towns as reflecting a timeless social structure, the acts that defined and constituted the central area at Songo Mnara were likely part of the ongoing negotiation of status, hierarchy, and elite identities at any given period. Decoding the evolution and use of this open space and the development of the cemetery and monuments, will thus offer a rich way forward to think about the performance of power in a Swahili town, and the way social memory became bound up with public, commemorative practices.

BibliographyAshmore, W. 1989. Construction and Cosmology: Politics and Ideology in Lowland Maya Settlement Patterns, in W.F. Hanks and D.S. Wright (eds.) Word and Image in Maya Culture: Explorations in Language, Writing and Representation, pp. 272 – 286. Salt Lake City: University of Utah Press.

Ashmore, W. 2002. “Decisions and Dispositions”: Socializing Spatial Archaeology, American Anthropologist 104(4): 1172 – 1183.Blake, E. 2006. Space, Spatiality, and Archaeology, in L. Meskell and R. Preucel (eds.) A Companion to Social Archaeology, pp. 230‐254. Oxford: Blackwell Publishing.Burton, R. 1872. Zanzibar: City, Island, and Coast. London: Tinsley Brothers.

Fleisher, J. and S. Wynne‐Jones 2010. Kilwa‐type coins from Songo Mnara, Tanzania: New Finds and Chronological Implications. Numismatic Chronicle.

Fleisher, J. and S. Wynne‐Jones 2010. Archaeological Investigations at Songo Mnara, Tanzania: Urban Space, Social Memory and Materiality on the 15th‐ and 16th‐century Southern Swahili Coast. Preliminary Report Submitted to the Department of Antiquities, Republic of Tanzania. Available at http://www.songomnara.rice.edu/Fox, J. 1996. Playing with Power: Ballcourts and Political Ritual in Southern Mesoamerica. Current Anthropology 37: 483‐509.

Garlake, P. 1966. The Early Islamic Architecture of the East African Coast. Nairobi and London: British Institute in Eastern Africa. Garlake, P. 2002. Early Art and Architecture of Africa. Oxford: Oxford University Press.

Gensheimer, T. 1997. At the Boundaries of the Dar al‐Islam: Cities of the East African Coast in the Late Middle Ages. Unpublished PhD thesis, University of California, Berkeley.

Horton, M.C. 1994. Swahili Architecture, Space and Social Structure, in M. Parker‐Pearson and C. Richards (eds.) Architecture and Order: Approaches to Social Space, pp. 132‐152. London: Routledge.

Horton, M.C. 1996. Shanga: the Archaeology of a Muslim Trading Community on the Coast of East Africa. Nairobi and London: British Institute in Eastern Africa.

Hutson, S.R., T.W. Stanton, A. Magnoni, R. Terry, and J. Craner 2007. Beyond the buildings: Formation processes of ancient Maya houselots and methods for the study of non‐architectural space, Journal of Anthropological Archaeology 26: 442–473.

Kusimba, C.M. 1993. The Archaeology and Ethnography of Iron Metallurgy on the Kenya Coast. Unpublished Ph.D. Thesis, Bryn Mawr College.

Kusimba, C.M. 1996. ‘Spatial organization at Swahili archaeological sites in Kenya, in G. Pwiti and R. Soper (eds.) Aspects of African Archaeology, pp. 703‐714. Harare: U of Zimbabwe.

Kusimba, C.M. 1999. The Rise and Fall of Swahili States. Walnut Creek: Altamira Press.

LaViolette, A. and J. Fleisher 2009. The Urban History of a Rural Place: Swahili Archaeology on Pemba Island, Tanzania, 700–1500 AD, International Journal of African Historical Studies 42(3): 433 – 455.

Low, S.M. 2000. On the Plaza: The Politics of Public Space and Culture. Austin, TX: University of Texas Press.Mathew, G. 1959. Songo Mnara, Tanganyika Notes and Records 53: 154‐60.

Moore, J.D. 1996. The Archaeology of Plazas and the Proxemics of Ritual: Three Andean Traditions, American Anthropologist, 98(4):789‐802.

Pradines, S. 2001. Une mosquée du XIIe siècle à Gede (Kenya, Mission de Juillet‐Août 2001), Nyame Akuma 56: 23‐28.Pradines, S. 2004. Fortifications et Urbanisation en Afrique Orientale. Oxford: British Archaeological Reports, S1216.

Pradines, S. 2005. Archéologie et préservation du patrimoine: le projet franco‐tanzanien de Kilwa, 2002‐2005, Nyame Akuma 63: 20‐6.

Pradines, S. and P. Blanchard 2005. Kilwa al‐Mulûk. Premier bilan des travaux de conservation‐ restauration et des fouilles archéologiques dans la baie de Kilwa, Tanzanie, Annales Islamologiques 39: 25‐80.

Robin, C. 2002. Outside of houses: The practices of everyday life at Chan Nòohol, Belize, Journal of Social Archaeology 2(2): 245–268.

Robin, C. and N. Rothschild 2002. Archaeological ethnographies: Social dynamics of outdoor space, Journal of Social Archaeology 2(2): 159–172.

Schmidt, P. 1997. Iron Technology in Africa: Symbolism, Science and Archaeology. Bloomington, IN: Indiana University Press. Smith, A.T. 2003. The political landscape: constellations of authority in early complex polities. Berkeley, CA: University of California Press.Smith, M.L. 2008. Urban empty spaces. Contentious places for consensus‐building, Archaeological Dialogues 15(2): 216–231.

Sulas, F. 2010. Intra‐site geoarchaeology at Songo Mnara, Tanzania: Preliminary Report. Unpublished report, January 2010. University of Cambridge.

Wilson, T.H. 1979. Swahili Funerary Architecture of the North Kenya Coast, in J. de V. Allen and T.H. Wilson (eds.) Swahili Houses and Tombs of the Coast of Kenya, pp.33‐46. London: Art and Archaeology Research Papers.

Wilson, T. 1982. Spatial Analysis and Settlement Patterns on the East African Coast, Paideuma 28: 201‐19.

Wynne‐Jones, S. 2010. Remembering and Reworking the Swahili Diwanate: the Role of Objects and Places at Vumba Kuu, International Journal of African Historical Studies.

Documents

Between Mosque and House: An Archaeology of Swahili Open Space