Introduction

Theorising modernity and understanding the ways it has structured historical writing have benefited from the consideration of indigenous meanings (Cobb and Loren this volume, Sahlins 1994). Similarly, broader approaches to world systems theory, developed with its focus on European global expansion beginning in the fifteenth century (Wallerstein 1974), offer intriguing alternative models of other ‘world systems’ when older and more geographically far-ranging systems are considered (Abu-Lughod 1989; Choudhury 2003; Frank 1998). Cores and peripheries, terms that we have come to take for granted as framing the landscape of the unambiguous exploitation of the colonised by the coloniser, are richly complicated by historical and cultural contexts where such spatial and hierarchical constructs are not easily defined, and in which the agency of those in the supposed periphery may be far greater than the model presumes (cf. Prestholdt 2004). In this paper, I focus on the Swahili of Eastern Africa (Figure 1). Eastern African coastal peoples were long interacting within a large African and Indian Ocean world system, from the late first millennium B.C. onward. Their increasing ‘entanglements’ with trade and eventually Islam offer unique perspectives on a rethinking of power and self-definition on a traditionally defined periphery. Although I was led to this discussion through the question of indigenous modernities in the context of post-medieval colonisation, the Eastern African coast poses a series of different questions, one where contact between coastal dwellers and other Indian Ocean peoples did not result in colonisation per se for well over the first thousand years of such contact. Rather, I hope to develop the idea that coastal peoples’ increasingly deep engagement with the Indian Ocean world system involved internal transformations which for them, defined an identity highly distinct from the numerous societies that lay behind the Swahili coastal region. Indeed, the Swahili maintained their deep distinctions from other Indian Ocean societies as well.

Figure 1
figure 1

Map of the Eastern African Coast (from Connah 2001:183)

Full size image

The Swahili have lived along 2,500 km of the Eastern African coast since the mid-first millennium A.D. The term Swahili is ambiguous in its usage for different times in history. It refers to the culture and society that emerged on the coast, characterized by distinct material culture and lifeways (including a mixed farming/fishing economy and early and ongoing commitment to ocean-based trade), but it is a label whose content has changed over the many centuries it has been in use. In the seventh century, for example, practicing Islam was not part of what defined coastal culture, but eventually it very much was. Locally the term Waswahili refers to the people practicing this coastal culture; Kiswahili refers to the Bantu language they speak. Historically their greatest geographic range was, north to south, from Somalia in the Horn of Africa to southern Mozambique, and including the Lamu archipelago, the major offshore islands of Pemba, Zanzibar, and Mafia, the northern tip of Madagascar and the Comoros Islands (Horton and Middleton 2000). They began as a village-based society, with a few larger settlements appearing mid-millennium (e.g., Juma 2004), and a string of independently governed urban centers appearing throughout the coastal region as early as the late first millennium A.D. Village life on the coast became overshadowed, in terms of archaeological and historical research, but remained an integral part of the polities that developed there. Even in their earliest centuries of village living, Swahili people engaged in international trade, and management of long-distance trade interests has remained central to Swahili wealth throughout their history. Small groups of coastal dwellers began converting to Islam in the eighth century, and by the fourteenth a Muslim society flanked that entire coastline. Today Swahili culture occupies a similar extent of the coast.

It is my argument that at the heart of the coastal identity known as Swahili that emerged in the mid to late first millennium was an embracing of worldliness, of a cosmopolitanism asserted eventually in every tier of society. As a working definition of cosmopolitanism, I draw on Hannerz (1996:103), for whom it is a disposition, an attitude:

A more genuine cosmopolitanism is first of all of an orientation, a willingness to engage with the other. It entails an intellectual and aesthetic openness toward divergent cultural experiences, a search for contrasts rather than uniformity...cosmopolitanism can be a matter of competence, and competence of both a generalised and a more specialised kind. There is the aspect of a state of readiness, a persona; ability to make one’s way into other cultures, through listening, looking, intuiting, and reflecting. And there is cultural competence in the stricter sense of the term, a built-up skill in maneuvering more or less expertly with a particular system of meanings.

My goal here is to trace archaeological evidence for the emergence of such an engagement with others, to explore aspects of the creation, maintenance, and transformation of Swahili culture within an ongoing encounter between Eastern African coastal peoples and those from other shores of the Indian Ocean world.

I examine this history through the evidentiary lens of objects and practices that became characteristic of widespread Swahili self-expression. I argue that a sense of cosmopolitanism became part of Swahili cultural expression as the Swahili sought to solidify their many modes of engagement with the Indian Ocean world, and to distance themselves from hinterland societies, whose own engagement with the Indian Ocean was brokered, for the most part, through the Swahili themselves (see e.g., Glassman 1995; Kusimba 1999b). This is not in any way to say that hinterland societies were not modern, or as modern as the Swahili, but rather that the Swahili diverged explicitly from hinterland peoples in a number of ways that tended to underscore cosmopolitan and Islamic values. The result was a distinct coastal lifeway, firmly embedded in a local and regional political economy, but which kept an eye to the Indian Ocean world. In the early centuries of this phenomenon, much was shared with neighboring mainland peoples, but this became increasingly less the case, as urbanism and Islam in particular became fundamental to coastal life, and as neither did in adjacent mainland regions until the nineteenth century.

In the context of nineteenth-century European colonialism in the region and the ensuing scholarship of the twentieth century, the Swahili were identified not as an African society per se, but as African-Asian, the direct result of what was presumed to be early and ongoing Arab colonisation and large-scale settlement, and of intermarriage between Arab men and African women (e.g., Chittick 1974; Kirkman 1964). However, recent work from multiple disciplines has questioned the early Arab contacts as colonial domination (Allen 1993; Horton 1996; Nurse and Spear 1985), although the early interaction of Eastern Africans and Asians which resulted in the spread of Islam in the region was itself a complicated and elusive process (Horton 1996; Horton and Middleton 2000; Pouwels 1987). The creation of a heterogeneous modern Swahili identity in the context of modern European colonisation and post-coloniality is itself the subject of much discourse (e.g., Arens 1975; Eastman 1971; Glassman 1995; Mazrui and Sheriff 1994; Prins 1967; Willis 1993). But as I hope to show from an archaeological perspective, Swahili society’s image of itself as cosmopolitan, past and present, served to frame its long-term contacts with outsiders both positive and negative. In this history, the Swahili share much with other societies colonised over the last 500 years. What I emphasise here, featuring archaeological data, is the centuries of interaction between Eastern African coastal dwellers and other peoples, and the self-construction of Swahili life which emerged from that interaction, prior to the later waves of Arab and European colonial encounters from the late fifteenth century onward.

The Swahili Context

In the late first millennium B.C. and early first millennium A.D., savannah-based Iron Age farming peoples, and savannah and highland pastoralists, began moving into the long, coastal plain of Eastern African. Their economy was based largely on fishing and farming, but part of the attraction to settling on the coast was the opportunity to engage in long-distance trade with Classical Mediterranean and later, Indian Ocean merchants, who were making landfall on the coast for this purpose. As far as archaeologists can tell, these ancestral groups to the Swahili came mostly from the hinterlands of the coast (Allen 1993; Horton and Middleton 2000). They were not the first inhabitants of this region, which had long been lightly settled and traversed by hunter-gatherer populations, some of whom remained in the hinterland areas as the coastal plain itself became more densely populated (e.g., Kusimba and Kusimba 2005; Kusimba 2005). Despite low population densities and mostly impermanent settlements, a few trading centers emerged on the coast in these centuries, the most famous being Rhapta, still not located archaeologically but thought to be on the central Tanzanian coast, and mentioned in Classical documents beginning in the first century A.D. (e.g., Casson 1989). As early as the last centuries B.C., Roman merchant vessels had obtained items from people they encountered on the Eastern African coast. By the centuries when a distinct coastal culture began to emerge, ca. A.D. 700, permanent villages had become numerous along the coast, and with this had come a commitment to engaging a wider world of new contacts on the Indian Ocean rim. Some would call this the beginning of Swahili culture, and some would apply the term when the larger towns emerge around A.D. 1000 (see Fleisher and LaViolette 2004). In either case, the period ca. A.D. 700 marks the spread of numerous village settlements along the coast. The efficacy of sailing using annual southeast/northwest monsoonal winds made regular trade, travel, and communication possible between these villagers and others around the Indian Ocean. The main route, a circuit of about fifteen months’ duration, was roughly counterclockwise from the western shore of India, across southern Persia, the Gulf region, and southern Arabia, to the eastern African coast, and from what is now Mozambique back to India; a second southern route extended from Java east to Eastern Africa and back to Sumatra (Casson 1989; Horton and Middleton 2000:215, n. 27). This traffic allowed mutual exposure, interaction, and trade, with numerous questions remaining about the scale and nature of all three.

The Swahili confined the vast majority of settlements to the coast itself, but participated in an enormous African hinterland/ Indian Ocean interaction sphere. This sphere was fundamental to the growing wealth and scale of life on the coast. The pattern of village settlement that dominated the first few centuries eventually changed, as numerous urban centers developed from the ninth and tenth centuries onward. Kiswahili emerged from Northeast Coast Bantu roots in this period, Bantu in syntax and the majority of its lexicon but exhibiting numerous loan words as well (Nurse and Spear 1985). The exact fit between the early native speakers of this language and the people who became known as the Swahili is imperfectly understood, but they grew to be broadly coterminous by about A.D. 1000. It is notable that Swahili is derived from the Arabic sahil, for coast or port-of-trade. It was ultimately popularised by foreigners such as Arab travelers, who distinguished this coastal culture and society from those of other Eastern Africans (Tolmacheva 1976). It is a widely used ethnonym today, but one which homogenises numerous local geographically-based terms (e.g., Waamu, Waunguja, Wapemba: people of Lamu, of Unguja, of Pemba). It was not always embraced, being contested politically over the last centuries along with other aspects of identity, but has returned to more common usage from a period of rejection by coastal people. I use the term here to refer to the particular coastal culture which we identify from the late first millennium onward, with full acknowledgement that it underwent additions and transformations over the ensuing millennium, and has gone in and out of use throughout this time.

Iron Age coastal village life, prior to the rise of towns, showed some distinctive characteristics though shared much with the hinterland. Such villages from the sixth to seventh centuries exhibit the use of Tana Tradition or Triangular Incised Ware ceramics (Horton 1996; Chami 1998), technologies such as iron smelting and boat construction; the building of both round and rectangular earth and thatch houses, household economy based on millet agriculture (Walshaw 2005), animal husbandry, and fishing (Horton 1996). Tana tradition ceramics are most closely associated with the long, linear coastal region, but have been located at a number of interior sites as well. The basic economic pattern is not unique to the coast, though boating and fishing were obviously more common there. One of the most marked early characteristics of this period of coastal village life is the presence of imported goods including glazed ceramics (some arriving with organic contents), glassware containers (some transporting perfume and oils), silver and copper and copper-alloy jewelry, and stone and glass beads from the Persian Gulf, Indian subcontinent, and beyond. These goods continued to be sought after by coastal peoples throughout the period in question.

Also significant is that by the eighth century A.D. there is evidence of conversion to Islam. This culminated in widespread adherence to Islam among Swahili peoples by the twelfth and thirteenth centuries and continuing thereafter. The earliest evidence to date for its practice is seen in the excavated remains of an eighth-century timber mosque at Shanga in the Lamu archipelago of Kenya (Horton 1996). A series of mosques of increasing size and formality were subsequently superimposed, a sequence attesting to contacts, by at least this time, between Swahili and Muslim traders and/or missionaries. As mentioned above, earlier understandings of the formation of Swahili society presumed that large-scale Arab and Persian immigration colonised the coast for the purpose of developing trading opportunities. Conversion of coastal peoples was seen as part of this presumed wave of immigrants. Archaeological as well as linguistic evidence now points convincingly to a model of the African origins of Swahili culture. However, this African-origins model freely acknowledges that Indian Ocean immigrants arrived throughout Swahili history, and that an openness to select foreign influences was part of early Swahili life, and ultimately fundamental to Swahili identity.

The majority of coastal dwellers were undoubtedly villagers (LaViolette and Fleisher 2005), but the Swahili urban centers were well-known in the Indian Ocean world and beyond. The centers are known today as stonetowns, due to the presence of elaborate stone-built houses, mosques, and tombs, standing among earth-and-thatch buildings. In the stonetowns, merchant activity with the African interior and with Indian Ocean traders took place; Koranic scholarship and arts and letters flourished, craft production was carried out, domestic spaces were tailored to changing notions of gender and privacy, a variety of foreign goods was made available, and a majority of people continued to engage in food production. And despite the broadly shared culture which extended along this vast stretch of coastline, no centralised government held it together politically (Horton and Middleton 2000; Kusimba 1999a; Wright 1993). Certain towns sought to extend their political power over others at different times, but the coast remained organised around the individual larger stonetowns whose variable political structures we have some understanding of through both archaeology and ethnohistory (Allen 1993; Strandes 1899; Wright 1993). The height of Swahili wealth and influence is usually placed in the fourteenth and fifteenth centuries A.D. After that period the sovereignty of the region was challenged by European and Arab colonial invaders with specific political and economic motivations, but important Swahili centers, including Malindi, Mombasa, and Pate, continued to thrive and exist to the present. The coast’s lack of political unity remained until the Omani Sultanate brought it under colonial rule in the 1830s, which lasted until 1964.

The Archaeology of Swahili Cosmopolitanism

I suggest that there are three aspects of Swahili culture most critical to the Swahili-defined separation between coast and hinterland, and relevant to the embodiment of the Swahili notion of cosmopolitanism. They are: (1) an active appropriation and Swahilisation of certain foreign material culture, style, and practice; (2) conversion to Islam and Islamic practice; and (3) a society organised around urbanism, an urbanism which retains ties to its rural hinterlands. To explore this, in the remainder of this paper I present evidence for each of these. Swahili archaeology which emphasised the indigenous adoption and/or transformation of the exotic can perhaps be best traced to James de V. Allen’s and Mark Horton’s 1980s studies of early Swahili history and archaeology (e.g., Allen 1981, 1982, 1993; Horton 1984, 1986), where they were the first to argue for the continuity of the Swahili with other African groups, and to reject the mid-twentieth-century notions of Arab trade contacts as large-scale colonisation. They have been followed by a handful of other archaeologists who explored further the notion of an indigenous Swahili identity (e.g., Abungu and Mutoro 1993; Chami 1994; Fleisher 2003; Helm 2000; Kusimba 1999a, b).

With Jeffrey Fleisher and Bertram Mapunda, I have been directing a long-term project on Pemba Island, Tanzania. It focuses on the large seventh-to-tenth-century village of Tumbe with some 30 hectares of scattered earth-and-thatch houses, and the later stonetown of Chwaka, dating to the eleventh to sixteenth centuries, which was built up adjacent to Tumbe after a half-century gap in settlement (LaViolette et al. 2001, 2003, 2004). We have carried out extensive survey and excavations focused on a single region showing a polity at different stages of its development. The goal of the research was to reveal how the regional domestic economy interfaced with the large-scale trade networks of the African interior and Indian Ocean worlds, through analysis of individual households from a range of settings in urban and rural settlements, and over time. With the contribution of Mark Horton, we have also studied the mosque architecture at Chwaka, and are beginning to contextualize this in the changing political economy. This work provides insights into the current question of how Swahili peoples expressed their social and cultural identity in opposition to how they came to perceive interior peoples, and in the context of their foreign interlocutors.

Domesticating Exotic Material Culture and Foodways

Archaeological evidence reveals that long-distance trade in luxury goods was a feature of coastal life before it was densely settled, and before Islam became a factor in trade relations (e.g., Casson 1989; Chami and Msemwa 1997; Juma 2004). So too was local manufacturing, including the working of metals (e.g., gold, copper and its alloys), stone (e.g., rock crystal, carnelian), and fibers (e.g., cotton) part of coastal life, such that for a number of classes of elite objects, local craftsmanship which developed from a knowledge of the import must be considered (see Table 1).

Table 1. Representative sample of coastal imports from A.D. 600–1500 indicating place(s) of origin and approximate time of introduction, based on archaeological and historical sources (Chittick 1974 1984; Horton 1996; Juma 2004)
Full size table

In either case, the taste for objects of foreign origin or styled after the former is documented from the time of early coastal populations and settlement, and then increasingly throughout Swahili history. Although archaeological exploration of Swahili village settlements is still in its infancy (see Fleisher 2003), those villages underlying later stonetowns and thus which have been partially excavated show the presence of imports including glazed ceramics, beads, and copper-alloy objects in their earliest levels (Table 2). Most of these originated in the Arab world and Persia in the earliest centuries of regular trade (Figure 2).

Table 2. Imported ceramics dating from the sixth-tenth centuries A.D., combined from two excavated earthen house contexts, Tumbe, Pemba Island
Full size table
Figure 2
figure 2

A crushed Sasanian-Islamic pot on the floor of an earth-and-thatch house, Tumbe, Pemba, eighth-ninth century A.D. Photo by J. Fleisher

Full size image

Around A.D. 1000 there is a boom in the appearance of stonetowns, and it is at this time that some argue for the crystallisation and maturation of Swahili urbanity per se, followed by centuries of urban development (e.g., Fleisher 2003; Horton 1996). The array of trade goods grew to be enormous, still dominated by Arab and Persian items, but including numerous goods from the Indian subcontinent, and East and Southeast Asia (see Table 1 and Figures 3 and 4). Some other major trade items are known through documents but not from the archaeological record, including cloth from Egypt and India, and written documentation related to Islamic scholarship. Traded out of Africa by the Swahili were materials from the interior, brought to the coast through networks involving numerous other societies, small-scale and large, and included ivory, animal skins, gold, and semiprecious stone. Coastal products being processed and exported by the Swahili included vast quantities of mangrove poles for building, shell beads, and iron (Horton and Middleton 2000). Mangrove poles, harvested directly from the coastal perimeter and sought after for construction projects in the timber-poor Middle East, have been a mainstay of coastal trade for two millennia. Shell beads are suspected as the primary manufacturing activity being carried out on the so-called bead grinders found at the majority of early settlements in large numbers (e.g., Horton 1996). Shell beads were traded to both the interior and to foreign locations, while glass and stone beads were held largely on the coast until quite late. Iron was in widespread use in the Indian Ocean by the late first millennium, but there still seems to have been a market beyond the coast for iron manufactured there (Kusimba 1999b).

Figure 3
figure 3

Selection of thirteenth-fourteenth-century Persian Gulf and Chinese ceramics, and fragments of copper, Chwaka, Pemba Island, Tanzania. Photo by A. LaViolette

Full size image
Figure 4
figure 4

Examples of Chinese blue on white pottery, fourteenth-fifteenth centuries, Chwaka, Pemba Island, Tanzania. Photo by A. LaViolette

Full size image

Although it is difficult to know the extent to which the residents of smaller towns such as Chwaka dealt directly with foreign traders, they too show consumption of imported goods in virtually every house context, whether stone or earth-and-thatch, and every midden (see Table 3). At the very least, Swahili traders in the larger towns gaining access to imports were trading them to people in the smaller towns, or imports were moving out from the towns through some other mechanism such as gift exchange. Glassware and glazed and unglazed ceramics in particular seem to have made their way to villages too, from the few village examples we have (Fleisher 2003; Fleisher and LaViolette 1999). The long-distance trade in luxury items is not in itself evidence for coastal cosmopolitanism per se. But coastal traders made it possible for the majority of goods that came to the coast from the Indian Ocean to stay on the coast. Through their own craft production and the commodification of local resources, Swahili traders exchanged coastal goods with peoples of the interior for things they would themselves export to foreign traders, keeping the majority of the Indian Ocean imports in their own settlements, and allowing only a small trickle of imported goods to the interior. They adorned themselves with imported cloth, jewelry, make-up and perfume. They collected glazed ceramics for domestic use and display, and incorporated them into both their domestic and public architecture, in some cases securing them in to mosque, house, and tomb walls with mortar and plaster.

Table 3. Sample of imported ceramics (showing sherd counts), dating from the eleventh to sixteenth centuries A.D., from four superimposed earthen houses excavated at Chwaka, Pemba
Full size table

The multi-storied stone houses which became characteristic of many of the larger Swahili stonetowns after about A.D. 900 merged principles of Swahili floor plans emphasising privacy, with a building technology imported from the Red Sea shore (see Garlake 1966). This style comprised building with rough blocks of limestone, called coral rag, quarried from surface and near-surface deposits along the coastal plain. The corners, doorways, and other finishing details were accomplished by using carved fine-grained coral elements mortared in place with the coral rag. Lime plaster, typically stark white but sometimes dyed to shades of pink and red, coated the interior rooms and certain external surfaces. Stone became the hallmark of the most elite dwellings and public and ritual structures. This architecture—elaborated, multi-storied—may have been critical to anchoring relationships of certain Swahili lineages with foreign traders as mentioned above, as well as contributing to the prosperity of towns and regions at large. Such buildings were unique in eastern Africa and in the Muslim world, and were an active way elite Swahili expressed a sense of taste and accomplishment that marked them as cosmopolitan. They also attracted the attention of foreign visitors. Even Swahili villages which contained no stone houses usually boasted a small stone mosque, or a few above-ground stone tombs (Wilson 1982). Stone buildings created an atmosphere in the coastal towns different from that of hinterland non-Swahili settlements where less permanent materials were used for architecture. Foreign visitors to the coast such as Ibn Battuta, who traveled and stayed there in the fourteenth century, noted these stark differences, and commented on the finesse and urbanity of Swahili towns such as Kilwa, Mombasa, and Mogadishu (Gibb 1962).

Although imported wares were small in number in comparison to the overall artefact assemblages, elite objects were high-quality and widespread. Brokering long-distance trade became a major commercial interest of the Swahili, such that stone houses in stonetowns came to act as sites of hospitality and trade negotiation, where foreign traders were hosted (for ethnohistoric examples see e.g., Donley 1982; Donley-Reid 1990). Sailing boats and ships capable of going medium and long distances were built and operated by the Swahili themselves. Their elite were major actors in the Indian Ocean trade and sought to surround themselves with whatever imported goods as they could (Tampoe 1989); the elites set standards which those of more modest means emulated.

Although subsistence evidence is still quite rare on the coast, our project in northern Pemba Island yielded extensive archaeobotanical evidence from seventh-to-tenth-century Tumbe and eleventh-to-sixteenth century Chwaka. Food preferences can be shown to have changed as this particular region became increasingly engaged in the Indian Ocean world. Tumbe’s subsistence economy relied heavily on pearl millet, an indigenous cultigen in Africa and in widespread use in sub-Saharan Africa by this time. Tumbe’s residents were also using coconut (Walshaw 2005), an Indian Ocean import which quickly became a staple of coastal diets. Based on the archaeobotanical work of Sarah Walshaw (2005) at these settlements, we know the later stonetown dwellers of Chwaka began cultivating imported Asian rice in the eleventh century. Although comparative botanical evidence from other Swahili settlements is rare, ethnohistorically and ethnographically we see rice become the staple grain of choice as stonetowns increase in number and influence. This suggests, along with the ongoing consumption of imported serving vessels, a transforming pattern of foodways involving feasting which further served to distinguish Swahili coastal patterns from those in hinterland societies (Fleisher 2007; see also Glassman 1995).

Conversion to Islam and Islamic Practice

The Swahili began converting to Islam no later than the eighth century A.D. as noted above, within a century of the birth of the faith itself. The earliest evidence for Islamic practice is found at Shanga, an eighth-to-fifteenth-century settlement in the Lamu Archipelago of Kenya which began as a village and eventually blossomed into a significant stonetown. There, a series of seven superimposed timber and, later, stone mosques was excavated (Horton 1987, 1996), the oldest of which dates to ca. A.D. 780 (Horton and Middleton 2000). The human mechanism of early conversion is unknown, but an assumption is that Muslim missionaries and/or Muslim merchants came from Arabia to the coast at places such as Shanga in small numbers and initiated conversion directly or unintentionally. The archaeological evidence for the rate of spread in the earliest centuries is not abundant, as it is buried in the pre-stonetown deposits that have not been prioritized by archaeologists. However, our knowledge of the stonetowns shows that the Swahili became increasingly Islamicised, with what appears to be nearly complete participation in Islam by the mid-1300s at the latest (Insoll 2003). Swahili practice of Islam varied tremendously along the coast and over time, with sectarian preferences detectable in mosque architecture, and in the considerable historical evidence on the topic available (Horton 1996; Insoll 2003; Pouwels 1987). Non-Islamic spiritual practices continued to be carried out on the coast and are widely in evidence today (Insoll 2003; LaViolette 2004).

We also see, not surprisingly, Arabisation in certain kinds of material culture and practice that accompany increasing interaction with Muslims and conversion to Islam (such as the abovementioned elevation of rice as a staple). The number and size of mosques increased dramatically up and down the Swahili coast. Burials were in simple Islamic style, with bodies laid to rest on their right sides, facing north toward Mecca. Above-ground stone tombs often bore Arabic inscriptions. A tomb style known as the pillar tomb became a Swahili architectural idiom (Figure 5). From the twelfth century onward, the Swahili held the position of being the long southwestern frontier of the Islamic world. Living in Swahili stonetowns provided access to literacy in Arabic, and Swahili language was written in Arabic script, all aspects which defined the cosmopolitan modernity of this time and place.

Figure 5
figure 5

Eleventh-century pillar tombs and mosque (in background), Ras Mkumbuu, Pemba Island, Tanzania. (Photo by A. LaViolette)

Full size image

The practice of Islam has continued to be strongly represented on the coast, although many practitioners in modern urban centers would not consider themselves Swahili per se, and Muslims can be found in numerous other inland locations. For nearly a thousand years, however, Islam and coastal settlement were exclusively intertwined. It has been argued that Swahili conversion to Islam was largely pragmatic, a way to join forces with Indian Ocean merchants who themselves were increasingly Muslim after the seventh century. A shared understanding of the rules of economic engagement was surely an asset to everyone involved. But there has been an unflagging commitment to Islam in every aspect of Swahili life such that it is certainly one of its defining features (see Insoll 2003).

Urbanism

In the study of indigenous African urbanisms, the Swahili coast is one of fewer than a dozen known regions on the continent which exhibited pre-twentieth-century urban formation (Connah 2001; Coquery-Vidrovitch 2005). With stonetowns as the defining settlement type of the Swahili coast, there has been a tacit assumption of how Swahili urbanism worked, coast-wide: that cities were the centers of tiered settlement hierarchies, providing specialised goods and services to hinterland villages, which provided them with agricultural and animal products (e.g., Kusimba 1999a). As the work currently being conducted on Pemba has shown, the growth of the major settlements took place at the expense of smaller ones; that is, as some settlements grew to much larger proportions, the hinterland villages were largely abandoned (Fleisher 2003). Furthermore, production in the settlements was at the household level, from pottery and iron production, to food production (LaViolette et al. 2003, 2004; Walshaw 2005). Stonetowns looked and acted somewhat differently at different places on the coast, mirroring various internal political styles and proportions of political, merchant, and religious leaders (Allen 1993). But despite these differences, their combined presence is a dominant feature of Swahili society. Stonetowns that predate the nineteenth century are within five kilometres of the coastline, and most are directly on the shoreline, with buildings sited to take advantage of sea views wherever possible.

Whatever their individual size or degree of wealth, stonetowns were extravagant relative to coastal villages, with their large central districts of stone houses such as at Lamu and Mombasa, or their handful of stone buildings and enormous, densely settled neighborhoods of earth-and-thatch, such as at Chwaka. Some were surrounded by town walls. They were stages for the acting out of urban tastes and practices, many of which were informed by Islamic ideology and cosmopolitanism. The stonetowns were home to the larger congregational mosques and often additional, smaller neighborhood ones, flanking stone tombs, and sometimes other stone public buildings. Swahili cities had cultural capital, through the relative concentration of wealth and cutting-edge style, access to imported goods, and concentrations of specialised knowledge. This came in many forms, from literate religious elites, to traders, to craftspeople, and was prioritized differently at various settlements. At Chwaka, for example, urbanity was heavily defined by the presence of four elaborate stone mosques spanning the site’s history, with two to three coexisting at any time. One of these was among the most expensively built of any on the coast, comparable to those at wealthy and powerful Kilwa (Figure 6, Chittick 1974). Yet Chwaka contained few stone buildings beyond its mosques and tombs. Wealth in the town was invested in its religious architecture, rather than individual stone houses. Although stonetowns can be said to be consistent with a general Islamic urban model (Lapidus 1969), their organisational and production patterns continued to resemble other African models (Fleisher 2003; Kusimba 1999a; LaViolette and Fleisher 2005; Sinclair and Håkansson 2000; Walshaw 2005; Wright 1993). The Swahili achievement of melding those models together is what created the dynamic coastal frontier.

Figure 6
figure 6

Fourteenth-century mosque at Chwaka, Pemba Island, Tanzania, excavated by Mark Horton, 2004. Photo by M. Horton

Full size image

Conclusion

I and others have argued elsewhere for many continuities between coast and hinterland, and for the indigenous forces which shaped the developments we see on the coast over the last two thousand years. In that argument, Swahili culture was not born in the late first millennium, the offspring of intermarriage between Arab settlers and coastal African women, but the product of longer-term indigenous processes. But just as undeniable is the extent to which the early coastal dwellers and particularly the Swahili sought out and incorporated foreign inspiration—specifically in the Indian Ocean and related to Islam—in their self-realization. Jeffrey Fleisher and I have argued elsewhere (Fleisher and LaViolette 2007) that Swahili elites constructed themselves consciously with elegant architecture, piety, privacy, cosmopolitan dress and décor, creating idioms which defined coastal life. The richest array of data is available for making this argument for elites, but it is increasingly clear that the trend began early, before the stonetowns, and that it became widespread in stonetown life, in elite and non-elite households, and extended into rural village settlements as well. Echoing Hannerz’s definition once again (1996:103), Swahili exhibited openness, a willingness to engage with people outside their own societies, “cultural competence...a built-up skill in maneuvering more or less expertly with a particular system of meanings.” They did this in a way that attracted the ongoing attention of outsiders, their small-scale and eventually large-scale immigration, so much so that it helped set in motion the recent centuries of debate about Swahili origins in Arabia and Persia, and Swahili discontinuities with neighboring peoples, with whom they in fact shared deep ancestry.

It is a concern that in revisionist Africanist historiography, as we have focused on elucidating the indigenous roots of many developments in precolonial Africa, that we risk denying particularly outward-looking cultures the cosmopolitan modernity which they created in the face of ongoing encounters with other societies. We continue to be challenged by the task of understanding and explaining the motivations and worldviews of African peoples who built such societies without removing them from Africa, and without considering them less African than societies which chose seemingly less outward-looking ways of expressing their own modernities (see Piot 1999; Stahl 2001). It is perhaps time, or long past time, that we discuss the African past without having to typologise and defend what is indigenous and what is imported, whether we are speaking of an artefact, an idea, or a religious conversion.

The turn of the sixteenth century marked the beginning of several centuries when the Swahili faced regional attempts at political and economic colonisation by Arab and European invaders, a history of colonial successes and failures beyond the present concern. Starting in 1832, the coast faced its most widespread effort at colonisation, when the greater part of the coast was colonised by the Persian Gulf-based Sultanate of Oman, whose Sultan Sayyid Said moved his capital city from Muscat to Zanzibar. The coast remained an Omani colony ruled by his family line until 1964. In the nineteenth century, large numbers of Omanis and other Arab peoples settled on the coast, and after a thousand years of engaging with foreign peoples, domesticating into Swahili life elements of Indian Ocean and Muslim cultures, the coast and especially the urban centers became Arabicised in ways far beyond anything that had happened previously. Whatever heavier mix of Omani culture is now present in coastal life, however, the Swahili continue to define themselves as different from their hinterland neighbors with their particular mix of material culture, specific styles of both stone and earthen architecture, deeply rooted and regionally specific practice of Islam, and an outward-looking urban culture over a thousand years old.