Coastal East Africa and the Western Indian Ocean: Long-Distance Trade, Empire, Migration, and Regional Unity, 1750–1970

IN RECENT YEARS there has been a growing awareness that the disciplinary boundaries of traditional area studies often do as much to obscure the contemporary world as they do to illuminate it. In a world where Indians, Pakistanis, and Koreans might work in the Persian Gulf or Silicon Valley, where North Americans run cafes in Katmandu, where Nigerian immigrants live in the United States, the United Kingdom, and Persian Gulf, where Dearborn, Michigan is home to thousands of Yemenis, and all of these people live in a world awash with transnational flows of culture and trade, making sense out of one part of the world without reference to others seems increasingly quixotic. As a result, people who teach and do research on the contemporary world have been at the forefront of rethinking the geographical or spatial units of analysis they use. Historians have been slower to move away from this old analytical framework because, at first glance anyway, the boundaries used in area studies seemed to make sense in the past, even if they have shortcomings in the present.

Historians have, nevertheless, used area studies boundaries in several ways. First, much of the historical work done on the non-western world is done by people who are trained in area studies, teach in area studies, and publish primarily for an area studies audience. Second, historians have used the geographical units employed by the area studies as units of analysis for the teaching and writing of world history. Furthermore, while world historians as a group have long been fascinated by the sorts of cross-cultural interactions that anthropologists have only recently rediscovered, in practice they have spent as much time using civilizations as units of analysis as they have looking at interactions. Often the civilizations in question are derived at least in part from area studies. Thus our textbooks and syllabi have headings like “Early States of Africa,” “East Asia to 200 CE,” and “Ancient Middle East.” World historians have used these area studies units in two ways. They sometimes appear as the units of comparative analysis. For example, a course lecture might explore why African farming practices are different from or similar to South Asian farming practices. Or they be the fundamental units used to discuss interaction between cultures. Thus we might talk about what happens when Hellenic culture meets Middle Eastern culture following the Macedonian conquest of Persia. In both instances an essentialized view of the civilizations or cultures in question is needed to make the analysis work. The field of African Studies, for example, depends on the existence of a cultural Africa somehow linked to the African continent. World historical analysis, in turn, relies on the existence of bounded and discrete entities like “Hellenic civilization” or “Persian civilization.”

All units of analysis in history are imperfect. They are models, and models are by their very nature simplifications of reality. So while I am about to point out what I think are the weaknesses of the current units of analysis, I recognize that my counter-proposal has shortcomings too. Furthermore, I am not proposing that we entirely discard area studies units of analysis. Rather, by lighting our topics from different angles, I think it is possible to see into the hidden corners of the past that might otherwise escape our notice had we used only one source of illumination.

In what follows I will look closely at the recent history of the East African coast and the Swahili people who live there. In doing so I hope to highlight the inadequacy of the area studies/continentalist approach to studying and teaching the Swahili and to propose a different way of considering their past. My contention is that one can better understand the history of the coast by dropping the traditional Africa/Arabia and Africa/Asia dichotomies and instead view the coast and islands of western Indian Ocean as a region. Further I will argue that this historical region existed at some times in the past, but not at all times. This is another factor we must remember about the regions we choose as units of analysis or instruction. There is a critical link between periodization and geographical units of analysis. The lands touched by the Atlantic Ocean, for example, may serve as a good unit of analysis in 1700, but not in 700. Below I will argue that for the period between 1750 and 1960 the western Indian Ocean makes a valid and useful geographical unit of analysis. 1750 serves as starting point in this essay for several reasons. The decades following 1750 saw the rise of new patterns of imperialism as the English East India Company, and, to a lesser but still significant extent the Omanis, became more assertive in the region. Those same decades also saw new economic developments as the industrial revolution in Europe began to call for new types of goods to be brought from the Indian Ocean to Europe. These economic and political developments in turn triggered large-scale movements of people within the region. After 1960, the commercial and imperial structures that lent the region its coherence collapsed.

Who are the Swahili?

In the last twenty years African historians, linguists, and archaeologists have demonstrated that East Africa’s Swahili civilization is African in its origins. Nurse and Spear, Chapurukha Kusimba, Mark Horton, and others have all made a convincing case that the Swahili civilization, which created the port cities of the East African coast, developed out of African traditions and the efforts of African people.[1] In doing so they have overturned an earlier body of work on the Swahili coast that saw its cities and civilization as a sort of alien growth. Colonial era archaeologists concluded that the ruins they found on the coast must have been Arab colonies, because Africans—they believed—did not build stone cities.[2] So the revisionist work of the last several decades has had the salutary effect of giving East Africans credit for their own civilization. At the same time, these revisionists have all recognized that external elements are present in the culture. Islam is central to Swahili identity. The Swahili language, though structurally Bantu, includes much Arabic, Persian, and South Asian vocabulary. In fact some dialects of modern Swahili derive as much as fifty percent of their vocabulary from these sources. Classical Swahili poetry is difficult to understand without some knowledge of Arabic. Prestige objects, notably Chinese ceramics, were obtained though trade. Noting all this, Kusimba has likened Swahili culture to a cloth “of local manufacture woven from threads spun from local fiber as well as imported yarn.”[3] So even the revisionists accept that the Swahili were influenced by non-African sources even if they staunchly maintain that these influences do not detract from the essential Africanness of the Swahili.

I do not disagree fundamentally with their conclusions, but I do see some room for argument. First, the assertion that the Swahili are fundamentally African is an argument based on origins that is then typically extrapolated into much later periods. Thus instead of viewing culture as a process of constant creation, it sees culture as essential and static. The argument goes that the roots of Swahili culture are African, therefore the Swahili are and remain African. It leaves little room for factoring in the dramatic shifts in oceanic trade and the population movements that accompanied it which took place in the Western Indian Ocean between 1200 and 1960. Second, their position posits what I think is a false opposition between Africa and Southwest Asia and India. It assumes that people and culture ought to fit into categories based primarily on which continent they inhabit. Cultural geographers have been attacking this idea for a while now. Most recently Lewis and Wigen have critiqued this notion and have in particular questioned the cultural unity of sub-Saharan Africa.[4] Thus the revisionists have been at pains to show that the Swahili are of Africa but not of Asia and so their position is in some sense dependant on the argument that there is a cultural “Africa” that correlates with the continental entity called “Africa.”

I will argue below that at some times (though not all) the Swahili world is best viewed as part of the western Indian Ocean region. This is not to deny its indigenous roots (African or whatever you wish to call it). Nor is it an effort to present the Swahili coast as a recipient of alien culture or as a passive partner in the commercial world of the Indian Ocean. Rather it is an attempt to see the Swahili, at a particular time in their history, from the sea rather than the land, and at the same time to let that maritime gaze sweep to the other shores of the western Indian Ocean and to see how and when they relate to the Swahili world. I will argue that between 1750 and 1960 the western Indian Ocean rim had a common history and that the people who inhabited that region were involved in a common process of cultural creation. The Swahili were part of that regional process and at every level from the political, to the economic, the cultural, and the religious, it affected them and they it.

Regions and Periods: Bentley and Manning

Regional analytical units are much like historical periods. They are hard to pin down, easy to find fault with, and always a bit nebulous. At the same time, like historical periods, they are critical to the process of doing history. Also like historical periods, regional units of analysis come and go in time. What makes a perfectly viable analytical unit in the age of sail may be woefully inadequate in the age of the cell phone. Most of the rest of this paper will look at this interplay between period and region in the western Indian Ocean. My premise is that the same criteria that make a valid historical time period also help to define a geographical unit of analysis. What are those criteria? I would submit that cross-cultural interaction offers the best way to judge when the western Indian Ocean comprised a valid historical unit. By examining the changing quantity and quality of these interactions we can begin to define periods whose relevance extends beyond just the component sub-regions to the larger region itself.

This is hardly a revolutionary proposal. Students of the region have long been interested in contacts between the various sub-regions and in the cultural continuity of the Indian Ocean rim—something that must be in part the result of regional interaction. Witness the title of the conference for which this paper was originally written. It called on its participants to consider “connections, commonalities and cosmopolitanism” in the region. Nor is it just students of the Indian Ocean rim that have been using interaction as the cornerstone of their study of the past. World historians typically use “cross-cultural interaction” as the focus of their work. Indeed Jerry Bentley recently proposed a system of periodization for world history that uses changes in the quantity and quality of cross-cultural interaction as a basis for constructing historical periods.[5] His proposal sparked a response from Patrick Manning that accepted the premise that world history can best be periodized by looking at cross-cultural interaction, but warned that the terms “cross-cultural” and “interaction” are highly problematic.[6] I agree with Bentley that the best chance of creating a truly regional periodization lies in interaction. (Bentley was trying to create a global periodization, but I think his model works for regions too.) And I also agree that the terminology we use needs closer examination, and that the word cross-cultural may need to be dispensed with.

Bentley’s periodization of world history need not detain us here, but his discussion of the nature of interaction is interesting. Bentley defines three major forms of cross-cultural interaction as central to the process. These are long-distance trade, imperial conquest, and mass migration.[7] These seem well suited as starting points for a periodization of our region. Trade has been an issue in the Indian Ocean since the earliest times, and in the recent centuries that concern us most, trade has run riot. Imperial conquests, notably the extension of British power into the western Indian Ocean in the second half of the 19th century, would seem to fit, as would the Omani, French, and Portuguese imperial ventures. Likewise we have migrations: Hadhramis, Indians, and Omanis have moved on a more or less permanent and voluntary basis. (One wonders though whether any migration is totally voluntary. There is almost always some sort of economic or political compulsion behind these journeys.) East Africans have migrated too, though often (but by no means always) through enslavement. So Bentley’s model has much to offer as a starting point for thinking about periodization.

And what of Manning’s objections? These too have much to offer. First, Manning says that interaction is a fairly obtuse term, and we need to identify what we mean when we talk about “interaction.” We need to be specific about the actors and the location of their interactions. In particular he is concerned about the relationship between “interaction” and ideas like “diffusion” and “dominance,” both of which tend to privilege one party in an interaction. The other objection to interaction as a focus of historical inquiry is that Bentley’s chosen forms of interaction tend to focus on the elite. Trade especially is the province of the merchant and of the elite, although that does not mean that the affairs of merchants do not affect others. Braudel tells us that history is not driven by the “reasonable rules of universal suffrage. There are plenty of reasons for arguing that the minority had a greater influence over the course of history than the majority.”[8] But Manning suggests that by looking beyond just the three forms of interaction Bentley chooses to use, we might see a range of interactions and historical actors that would be obscured if we focus primarily on trade, conquest, and migration. Interestingly, one of the examples Manning offers of an important of interaction that would be missing in Bentley’s schema is the exchange of food crops across the Indian Ocean. The movements of African varieties of millet to India and the South East Asian crops like sativa rice and bananas to East Africa were probably not the result of any of the big three types of interaction, according to Manning.

Manning’s first objection is of greater concern for the more recent centuries that interest us here. Migration, trade, and conquest in the last few centuries have been quite democratic in their effects in the western Indian Ocean, more so perhaps than they were a millennium ago. Of greater worry to us is a way of characterizing the nature of these interactions. As noted above, earlier work on the western Indian Ocean region has typically ignored East Africa or seen it as a passive recipient of Arab culture. More recent work often minimizes the Arab influences—a position that may have some merit for earlier centuries, but hardly works for the last two hundred years. But the fact remains that interactions across cultures shaped the rim of the western Indian Ocean in profound ways and that some of those interactions did not take place between equals.

Perhaps of greatest interest to us is the problem of “culture” in cross-cultural interaction. Manning points out that as far back as 1952 there were more than 160 different definitions of culture in circulation, and now we are probably even more circumspect or even confused as to what exactly we mean when we talk about cultures.[9] This is doubly true when we talk about cultural change. How much can a culture change and still be itself? Clearly a world as cosmopolitan as the Indian Ocean cries out for analysis in terms of the interactions of cultures, but defining those cultures has proven difficult.

Chaudhuri is the only person who has really tried to think in detail about what it means to talk about cultures in the Indian Ocean region. He defines the four civilizations of the region as Islamic, Sanskritic, Southeast Asian and Chinese. Recognizing the huge overlap between these civilizations, he employes mathematical set theory to show that one could be part of two or more of these civilizations at the same time. So Zheng He, the eunuch admiral who led the early Ming expeditions into the Indian Ocean, could simultaneously be part of a set that included Muslims and a set that included the Chinese, and perhaps even a smaller set of eunuchs.[10] I probably do not do justice to Chaudhuri’s rather complex attempt to make sense of the spaghetti tangle of ethnicity and culture in the Indian Ocean area, but my sense is that he has done little more than to show how complex and slippery a concept culture/civilization can be. He leaves unresolved (probably because it is unresolvable) the problem of what the core essence of Islamic or Indian culture is except in so far as he falls back on the usual high culture/great books definition of a civilization.

Many of Manning’s objections to Bentley’s use of the term cross-cultural interaction and many of the complaints of Africanists who worry about the implications of seeing recent Swahili culture as a product of regional influences, can be met by changing the way we view culture. Instead of seeing culture in the essentialist and often continentalist way we are accustomed to, we would be better off seeing culture as the product of human interaction. I suppose an utterly isolated person would still have a culture of sorts, but as a practical matter, culture matters primarily in human interaction. It is created whenever people communicate with each other. Thus regional interaction creates a regional process for making culture. The more interaction there is in a region, whether through trade, migration, or conquest, the more a regional culture will emerge. Thus in a cosmopolitan culture like that of the western Indian Ocean rim it is virtually impossible to figure out the origins of common items like the futa/sarong/kikoi (wrap skirts worn by men in East Africa, South Arabia and Western India coastal areas). Are they Arab? African? Indian? Or do they perhaps belong to a system of cultural interaction and creation that transcends Arabness, Africanness, and Indianness?

A recent White Paper produced by the Globalization Project at the University of Chicago called on teachers and researchers to stop seeing cultures as clusters of “traits” fixed to geographical space. Rather they urged that we “explore an architecture for area studies which is based on process geographies, and sees significant areas of human organization as precipitates of various kinds of actions, interaction, and motion—trade, travel, pilgrimage, warfare, proselytization, colonization, exile, and the like. These geographies are necessarily large-scale, changing, and tend to highlight variable congeries of language, history, and material life; regions, in our approach, are taken to be initial contexts for themes which generate variable geographies, rather than as fixed geographies marked by set themes.”[11]

Regions and the Historiography of Coastal East Africa

The role of cross-cultural interaction has been a major theme in the literature on coastal East Africa; explicit discussion of regionalism has not. The Swahili origins debate, which has already been outlined above and will be considered in more detail below, is a debate about the role of “outsiders” in the rise of Swahili civilization. To this extent, it is relevant to a discussion of the wider west Indian Ocean region. Early writers on the topic—most notably James Kirkman—took the position that Arab invaders and migrants played a central role in the creation of Swahili civilization.[12] In this interpretation, it was Arabs who brought urbanism and Islam and thus “civilization” to the coast. In the last twenty-five years there has been a major reaction against this position, and while the details are still debated, the general consensus is that the Swahili cities were the creation of local (i.e. “African”) people and not Arabs. No one denies that Arabs, Indians, and probably Persians made contributions to the Swahili world, but they are no longer considered the driving force behind the creation of an urban civilization in East Africa.[13]

The literature on the Swahili world after 1500 has two notable characteristics. First, it is very thin between 1500 and 1750. Second, the works on the post-Portuguese Swahili world are concerned in part with figuring out who the “real” Swahili are. There are few major works on the Portuguese period in East Africa. The classic work, by Justus Strandes, is over 100 years old. The most recent work on this period was done by M. N. Pearson, whose specialty is Portuguese India. While he brings a broad knowledge of the Portuguese sources to the topic, he has been criticized for being unfamiliar with the local materials and with the literature on East Africa. Jeremy Prestholdt is working on a dissertation that will treat this period, but it is not yet complete.[14]

The literature on the more recent history of the Swahili coast is far more extensive than it is on the Portuguese period. There are a number of political histories of this period, most notably those of Norman Bennett and Michael Lofchie. Bennett is primarily concerned with relations between the Arab rulers of the Omani empire in East Africa and the European colonialists who eventually supplanted them. Lofchie describes the conditions that ultimately led to a bloody and racialized revolution in Zanzibar in 1964. Lofchie is concerned with defining Swahili identity in the context of the Zanzibar revolution. His work shows how the actions of the colonial state and the internal dynamics of coastal society created conditions wherein some elements of coastal society came to define themselves as local and African as opposed to Arab or Indian and alien.[15]

Other works on this period have taken a more economic slant. Edward Alpers and Abdul Sheriff have written major works on the role played by the emerging global economy in the 18th and 19th centuries in undermining both the economic and the political independence of the East African coast. Alpers’ work focuses mostly on Mozambique while Sheriff is concerned with Zanzibar. Both reach similar conclusions about the detrimental effects of the trade between the industrialized West and East Africa.[16]

Another body of work considers the internal dynamics of coastal society. It is concerned with the growth of the plantation system on the coast and with the status of slaves in coastal society. Frederick Cooper’s work has looked at the plantation system and the changing status of laborers on the coast as slavery was abolished and free labor took hold during the colonial era. Cooper’s work, which includes interesting and wide-ranging references to labor ideology in the Americas, in Europe, and other parts of Africa, shows astonishingly little interest in the broader Indian Ocean region.[17]

More recently Jonathon Glassman and Laura Fair have looked at the class structure of coastal society in the 19th and 20th centuries and especially at the role of slaves and former slaves in redefining their own status. Glassman’s earlier work was focused on the role of slaves in a rebellion against German rule on the coast. His more recent work has focused on the use of claims of external origin in debates about nationalism in late colonial Zanzibar. Laura Fair’s work argues that slaves and former slaves, in trying to create a social niche for themselves in post-abolition Zanzibar, created many of the cultural elements that we have come to call Swahili. Both Fair and Glassman consider to some degree the question of regionalism because class on the coast is defined in part through claims of outside origins. (The elite have long claimed Arab or Persian origins.) But neither considers that the Swahili world may at this time have been part of the regional process of culture creation.[18]

Recently there has been more interest in the idea of a regional unity in the western Indian Ocean. Alpers has published several articles looking at the role of East African migrants in the region, Janet Ewald has looked at the movement of sailors through the region and Alpers recently edited a collection of articles for Comparative Studies of South Asia, Africa and the Middle East that considers African participation in a western Indian Ocean region in the colonial period.[19]

Periodizing Western Indian Ocean History

While the idea of a western Indian Ocean region is a relatively new one for Africanists, it is less novel to South Asianists. It is Chaudhuri who has done the most to think about the broader periodization of an Indian Ocean regional history. In Trade and Civilization in the Indian Ocean and Asia before Europe, he argues that there is a “historical unity” to the period between the rise of Islam and 1750. That period can be further subdivided into a time before 1000 when merchants made long multi-year voyages, and a time after 1000 when a more developed and fully articulated trade let merchants make short voyages between emporia. But 1750 marks a “symbolic meridian in time” between the ancien regime of Asia’s trading world and the new commercial world that would be predicated on the steam engine. Steam power would drive the English mills that encouraged cotton production in Gujarat. It drove the north Atlantic economies that sought ivory, coffee, and dyestuffs in East Africa and Arabia. It brought new merchants to new places, created new towns, only to have them melt away in later years.[20]

Chaudhuri is right that 1750, which is the starting date of the first period I will discuss, marks the end of something. But it does not clearly mark the beginning of anything new either, at least in the western Indian Ocean. Rather, 1750 to 1850 is a period of transition. During this time we see the slow death of the old Indian Ocean and the beginnings of the new steam-powered Indian Ocean trade that would emerge after 1850. It was also during this century that the first tentative stirrings of large-scale migrations from India and Arabia to Africa, and from Arabia to India began, and the period also saw the precursors of the major imperial conquests of the next century.

From 1850 to 1960 marks another period, not this time one of transition but rather the full flowering of the latent trends visible in the preceding century. This was the age when steamers plied the sea-lanes accompanied by sailing ships and when the great colonial ports of Aden, Karachi, Mombassa, and Bombay came to prominence. It was a great age of commerce. Unprecedented high volumes of shipping traffic moved in steamers between the major ports, but also in sailing craft between the minor ports and the major ports. Indeed, steamships helped create a boom in the older “traditional trade” (i.e. trade carried in “dhows,” the lateenrigged sailing craft that were the standard locally produced vessels in the western Indian Ocean), a branch of maritime trade that would prove important for interactions between peoples of the region. Steamers and dhows facilitated great migrations of people from one area to another, while the steamer played a role in shaping a major new imperial conquest.

The period 1960 to the present marks the end of the great age of steam. The British departed, the dhow nearly disappeared from the long-distance trade, and the exploitation of oil in the Persian Gulf redirected old migration patterns. Shipping containers made sea transport cheap but also favored a few capital-intensive ports. Aden fell off the map and Dubai/Jebel Ali, closer to the oil money, took off. Air travel made movement much easier and people from outside the region began to crop up in unexpected places: Nigerians in Jeddah and Mecca, Koreans and Thais in the Gulf.

Each of these periods seems to have a historical integrity to it that transcends a single sub-region of the western Indian Ocean. Interestingly, there seem to be changes in all three of Bentley’s classes of interaction that correspond to my proposed periods. In the next few pages I will look at the quantity and qualities of interaction in each of these three periods. Ships and their movements serve as the main barometer of change. Until the most recent period, ships moved almost all of the trade goods, migrants, and imperialists that concern us. I am especially interested in the movements of dhows or native vessels. While in some respects marginal to the larger economic trends of the last two centuries, they play a much more prominent role when looked at as vectors for regional interaction. Dhows carried huge crews compared to their size because of the difficulties of handling the unwieldy spars of their lateen rigged sails, and in some cases the size of the crews exceeded even the demands of sail handling. Lots of extra sailors came along as much to participate in petty trade on their own accounts as to serve the sail. The result was that during the dhow season, ports would be filled with sailors, many of them from far off places who were there for months. East African ports filled with sailors waiting for the monsoon winds to carry them home. Indian and Arabian ports must have likewise been filled with sailors at certain times of the year. By contrast, however, steam ships could carry a great deal of cargo, but they needed very few people to operate them. They also were less bound by the monsoons, and so did not congregate in certain ports at certain times of the year. Some ports—Aden and Bombay for instance—must have had a significant number of steamer crews roaming them, but steamer crews typically passed through quickly and slept aboard their vessels. Dhow sailors lingered and typically slept ashore. As a result they did a lot more interacting than their counterparts on steamers.

Dhows also went to places that steamers could not go. They served the minor ports—Mandvi, Lamu, Shihr, Mukalla—that steamers visited infrequently if at all. In doing so they brought alien objects and people into places that would otherwise have been little exposed to regional influences. In East Africa, for instance, the mangrove trade brought Kuwaiti and Omani dhows into sleepy fishing villages on Pemba Island and into remote corners of the Rufigi delta. These boats and their crews would sometimes spend as much as a month in these places. Dhows also brought passengers, often from places without steamer services to places without steamer services. Until the opening of the port at Dubai/Jeblel Ali there was little steamer traffic between the Persian Gulf and East Africa. Most of the many migrants from the Gulf got to East Africa in dhows. In short, dhows and their movements tell us much about where and how the interactions that interest us take place, and so they will be at the forefront of my analysis of the proposed periods.

1000–1750: The Emergence, Efflorescence, and “Decline” of the Swahili

In discussing the centuries that precede the focus of this essay, it is necessary to note something more about the contentious historiographical debate about the origins of the Swahili. A recent review article on the question noted that in the last twenty years over 240 books and articles have come out dealing with this issue.[21] The core of the debate centers on whether the Swahili civilization, with its coastal city-states, is the product of a local historical dynamic or whether outsiders (i.e. people whose origins were in a wider region, not the African continent) played a crucial role in the creation of these cities and the culture that flourished in them. Earlier historians and archaeologists, relying on Swahili oral and written traditions that speak of immigrants from Shiraz in Persia or from various parts of Arabia founding the Swahili cities and working in a world that saw Africans as essentially uncreative and historically passive, concluded that the Swahili cities were Arab or Persian colonies that sprang up ex nihilo.[22] Since then, archaeologists and linguists have significantly revised this assessment. They have shown that the Swahili cities were a natural extension of earlier developments on the coast. Instead of being cities that sprang up out of nowhere, archaeologists have shown that the Swahili cities that emerged in the period between 1000 and 1250 were the result of several centuries of growth that began with fishing villages that slowly grew into towns that eventually became trading cities. It was these already established towns and cities that welcomed visiting merchants, rather than visiting merchants who brought urbanization to East Africa. In the case of Shanga, one of the best excavated of the Swahili towns, it is clear that the first Muslims there were a small minority who arrived in an already existing town. Eventually their religion became the majority religion and one of the cornerstones of Swahili identity, but they cannot be credited with creating Swahili cities or culture.[23]

Archaeologists have also shown that the people who created the Swahili town were of the same cultural and physical stock as the people of the coastal hinterlands. Pottery found in the villages that were the precursors of the Swahili cities, called Tana ware, is also found at sites not on the coast. The implication is that the people who created the Swahili towns and cities were just the coastal branch of a larger group of people who occupied a region that stretched from the coast several hundred miles inland. Examinations of the skeletons found in burials in coastal cities have also supported the idea that the earliest city dwellers in East Africa were local people rather than immigrants.[24]

Linguists too have gotten into the game. Although modern Swahili is shot through with words that derive from Arabic and other Asian languages, linguists have pointed out that Swahili (or “Kiswahili” as the language is properly called) belongs to the North Coast branch of the larger Bantu family of languages.[25] They point out that some of the earliest written accounts of the coast make reference to words that are of Bantu origin. Thus, when the great traveler Ibn Battuta visited the Swahili city of Kilwa in the 14th century, he reported that the king was called falme, a word very close to the modern Swahili word for king, which is mfalme, and clearly a word with Bantu roots rather one that derives from Asia.

Thus, the roots of Swahili civilization have been convincingly shown to be local. If we wished to be continentalist about it, we would say that the origins of the Swahili were African rather than Asian. But what was the relationship between this coastal civilization and the western Indian Ocean world? And what was it relationship to its own hinterland? First, we should note that virtually all of the Swahili cities of the Swahili classical age were coastal. The one major exception to this is the city of Gedi, which is about twenty miles from the sea, though it may have once been closer. That these towns were ports of one type or another says something about the importance of trade to Swahili life. To be a merchant was the standard occupation of Swahili patricians. The stone houses constructed by Swahili patricians were intended not just to be comfortable and beautiful places to live, but were also symbols of their credit-worthiness and places where they could host visiting foreign merchants. As for Islam, which by 1250 was the majority religion of the Swahili, it is clear that the earliest mosques found on the coast were tiny and obviously intended for the use of a minority of the population, perhaps even just by visiting Muslims, although by the 13th century the cities had elaborate congregational mosques in addition to numerous smaller mosques. They are clearly Muslim cities. While Islam is no more alien in Africa than Christianity is in Europe, its presence on the Swahili coast is a clear sign of the movement of people and ideas within the region. The Swahili are Shafii Muslims, the same school that dominates most of the western Indian Ocean.

The importance of trade and Islam to the Swahili cities at the height of their political and economic independence in the period between 1000 and 1500 suggests that there was a degree of regional integration during this period. But if we examine this era from the perspective of long-distance trade, empire building, and migration, it is only in the realm of long-distance trade that the Swahili were part of a larger western Indian Ocean. There is no evidence of large-scale migration to or from the coast nor is their any evidence of empire building.

The relationship between these coastal city states and their hinterlands is poorly understood. That the coast was Muslim and the people of the hinterlands were not is quite clear. It is also fairly certain that the Swahili did not politically dominate the people of the interior, but because trade was central to the Swahili towns, presumably they had trade connections with the people of the interior. The extent and range of those trade connections are debated. Michael Pearson has proposed that the mutual influence of the Swahili and the people of the hinterlands on each other was greater in both its cultural importance and extent than earlier works have allowed.[26] But most of the trade goods that the Swahili sold—ivory, rhino horn, timber, slaves, turtle shell, ambergris, and so forth—could have been obtained quite near the coast. The one exception to this was gold. The richest Swahili towns were those that controlled the gold trade, and that gold came not from the coast but from the Monomotapa, far up the Zambezi River. Thus, at least in south-central Africa the influence of long-distance trade stretched 400 miles into the interior. But again there seems to have been no empire building by the Swahili and no significant migration between the coast and the interior or visa versa. Even the forced migration of the slave trade seems to have been a relatively minor affair. Thus, although trade may have stretched as far into the interior as modern day Zimbabwe, Islam and other Swahili cultural habits did not follow.

A new influence entered East Africa in 1497 with the arrival of the Portuguese who brought a radically new approach to trade. Their innovation was to militarize the Indian Ocean trade. They sought to control the sea lanes of the Indian Ocean and to control most of the region’s major ports. They seized Hormuz, Goa, and Diu and made an unsuccessful attempt to control Aden and thus the Red Sea. In East Africa the Portuguese were able to take control of some important cities. Mombasa, Kilwa, and Sofala all came under Portuguese control. At Mombasa, the Portuguese built Fort Jesus, their first major fortress in the Indian Ocean. In Kilwa their depredations resulted in the abandonment of the city. The sixteenth century was thus a difficult time for the Swahili. Kusimba maintains that the Portuguese incursion set in motion the decline of the Swahili, and while his assessment may be a bit too gloomy, he is right that from the 16th century forward the Swahili coast had to deal with aggressive outsiders bent on a type of hegemony that was not seen before 1500.[27]

The degree of Portuguese control in the western Indian Ocean has been a subject of debate, but the general consensus is that they failed to fundamentally transform the nature and patterns of western Indian Ocean trade. Instead they superimposed themselves on a pre-existing trade system without changing it very much. The degree to which the Portuguese can be said to have contributed to creating or reinforcing a regional system depends on one’s perspective. If one were looking outward from Goa, then one might be inclined to say that the Portuguese presided over a reasonably integrated Western Indian Ocean. However, their attempts to control Indian trade were not terribly successful, and by 1600 they had learned that it was more profitable to facilitate trade than to interfere with it.[28]

From an African perspective the Portuguese period looks different. South of Kilwa, in Mocambique and Sofala, the Portuguese created a degree of regional integration. Trade between Mocambique and India flourished, Indian merchants moved in small numbers to Mocambique, and the same imperial power that ruled in Goa ruled in Sofala and Mocambique. However, north of Sofala, in the region from Kilwa to Lamu that was arguably the core of the Swahili world, the picture was grim. With major cities languishing, some forced to pay tribute to the Portuguese, and trade in decline, the Swahili coast was poorly placed to participate in whatever might have existed of a regional system.

In the middle of the 17th century the Portuguese were expelled from Muskat and a new and aggressive Omani dynasty emerged. This new dynasty set out to drive the Portuguese from the East African coast. They did so by allying themselves with the rulers of the surviving Swahili towns and by 1700 had driven the Portuguese from the coast north of Sofala. The Omanis attempted to build an empire on the East African coast, but their own society was so fractious that in practice their initial efforts at empire building were not much more successful than the Portuguese effort. Omani governors in East Africa, like the Portuguese captains that had preceded them, were more interested in pursuing their own interests than in keeping a tight imperial control in East Africa. However, under the Omanis the cities of the coast began a sort of renaissance. The Omanis may not have been ideal colonial masters and the inhabitants of the coast soon came to resent them, but they were fellow Muslims and could more easily be incorporated into local social structures. The city of Kilwa was rebuilt on a new site, and peoples called the Yao and the Nyamwezi opened new trade routes from farther in the interior. These trade routes brought ivory and slaves from the interior to the coast, but it should be pointed out that they were created in response to regional economic forces largely on initiative of the Yao and the Nyamwezi, not the Swahili.[29]

Thus the period from 1000 to 1750 was a long one and, in itself, probably does not make a coherent analytical unit. I have discussed it primarily to contrast the much greater regional unity of the next periods, where empire building, migration, and long-distance trade occurred on a much more dramatic scale. All three of these elements clearly occurred during this time, but, with the possible exception of long-distance trade, they did not occur on a regional scale. The Portuguese built an empire, but it was loose one that cannot be said to have controlled the whole region or even the coasts of the whole regions. Likewise, the Omanis built an empire, but its reach was limited to East Africa. By comparison with the British Empire of the 19th and 20th centuries they were both minor affairs. Migrations certainly took place during this era also. But migration too lacked the scale that it would attain in the 19th and 20th centuries. Some Omanis moved to East Africa, some Indians moved to East Africa, some East Africa slaves were moved to India and the Middle East, and the ever peripatetic Hadhramis found their way to India to serve as mercenaries, but the scale and coherence of these events pale in comparison to the huge movements of people that were to be facilitated by the Pax Britannia.

1750–1850: Period of Transition

The decades following 1750 mark beginning of the Industrial Revolution and the year 1757 saw the bellwether battle of British imperialism in Asia at Plessey. In the next 100 years, Britain took a steadily growing imperial role in India, and her imperial interests spread outward from there to the Persian Gulf, to Aden, and to East Africa. European (and after 1800 American) sailing ships continued to ply the waters of the western Indian Ocean, but they did not enjoy any huge advantage over local dhow shipping. The square-rigged ships that Europeans favored were more efficient sailing craft than the local lateen-rigged craft, but like the local ships they were dependent on the monsoon winds to get around. While it would be easier to tack a square-rigger against the monsoon than a lateen-rigged boat, it was still not easy and most of these vessels timed their voyages to coincide with the favorable winds. Moreover, Europeans did not enjoy a monopoly on square-riggers. The Omanis owned many of them, often buying up prize ships during the Napoleonic Wars. Indian shipyards also produced square-riggers. The result was a world that moved at a monsoon pace.

But move it did. Using rather limited evidence—I have not had a chance to look at port statistics for India—it appears that there was a modestly active commercial world during this period.[30] The French bought slaves in East Africa for their Indian Ocean sugar plantations. The Omanis carried on a trade between the Persian Gulf, India and East Africa. But the volume of this trade is hard to judge. American ships went to Mocha to buy coffee, and began to stop en route at Zanzibar in the 1830s. However, until someone does a survey of the trade statistics from this period it will be hard to talk about the trade of the era except impressionistically.

In East Africa the implications of this trade stretched from the coast well into the interior. New trade routes into the African interior emerged to meet the demands of an emerging regional economy. Trade routes also shifted and routes that had once reached the coast further south toward Portuguese territory on the Zambezi now began to lead to Kilwa. Under the Omanis Kilwa became a major outlet for slaves and ivory, both of which trades grew in importance during this period. The demand for ivory was driven, as were many other aspects of trade in this period, by the nascent industrial revolution. In the early 19th century it became possible to cut ivory mechanically and “soft” East African ivory was uniquely suited to mechanical cutting. As ivory came to be used for more and more industrially produced goods—piano keys and billiard balls stand out as two examples—the demand for ivory soared. Yet even in the late 18th century, before this new technology emerged, there had been a marked increase in the price of ivory, so other forces must also have contributed to this increased demand.[31] At the same time demand for slaves for use on the date plantations of Oman and in the various (mostly French) sugar islands of the Indian Ocean led to a major increase in the volume of the Indian Ocean slave trade. As Kilwa became a major outlet for these trades both the Yao and Nyamwezi caravan routes in the interior began to focus on Kilwa.

At the same time a new Omani dynasty—the Busaidi—sought to make Zanzibar the focus of East African trade. They did this by undermining the autonomy of Omani governors in the major coastal towns and by creating a system of variable tax rates that encouraged trade to pass through Zanzibar. Thus Zanzibar emerged as the capital of a revived Omani empire in East Africa. The man behind all this was Seyyid Said, the ruler of Muscat. He oversaw the consolidation of Omani power in East Africa in the early 19th century and the creation of Zanzibar as the locus of that power.

The early 19th century also saw a new development in the relationship between the coast and the interior. While it is clear that up to 1800 the people of the interior controlled virtually all of the trade to the coast, after 1800 people from the coast began to organize trade caravans themselves and to venture into the interior. By 1850 Swahili speakers had reached the great lakes of the interior and there was a sizable Swahili speaking population in the town of Tabora in what is now central Tanzania. These “Swahili” were a diverse and cosmopolitan group, including people of Indian, Arab, and local coastal extraction. Their presence, hundreds of miles from the coast, is a sign of the extent to which the changing patterns and increased intensity of trade had changed the nature of the interactions between East Africans and the larger western Indian Ocean region.

It is also worth noting the role of the British in all this. Seyyid Said’s empire building took place in the shadow of an expanding British Empire in the western Indian Ocean. The British supported the Busaidi in Oman as a bulwark against Wahabi expansionism, but in East Africa British agents constantly interfered with the Omanis. These British agents were most often interested in suppressing the slave trade. They were able to assert control over Indians in Omani territory, claiming them as British subjects. As British subjects the Indians were not allowed to own or trade in slaves. That the British could get away with this is evidence of the degree to which Omani sovereignty was compromised. Later, after the death of Seyyid Majid, Said’s successor, the British were able to orchestrate a division of Omani territory into two separately ruled sections—one based in Zanzibar, the other in Muscat. Thus while one might see in Seyyid Said’s East African empire an extension or revival of older Omani imperialism in East Africa, there were two new elements. First there was the new presence of coastal people in the interior, many of them connected in some way to the Omani state. Then there was the new presence of the British, not yet in control of East Africa but nonetheless working on it.

The steady growth of the British imperium in East Africa matched its growth elsewhere in the western Indian Ocean. The East India Company took a clear interest in the Persian Gulf in the early 19th century, negotiating treaties in Trucial Oman as well as the aforementioned alliance with Seyyid Said in Muskat. The Company acquired Perim Island and Aden and established agents in Muscat and then in Zanzibar. Bombay and Aden began their climbs to major port status. Britain, though, was not the only imperial power at work in the region. Oman revived its claims in East Africa, and in 1838 Seyyid Said, the Sultan of Oman, moved his residence to Zanzibar, bringing with him Atkins Hamerton, the East India Company’s agent in Muscat, and Britian’s first offical representative in East Africa.

Each of these imperial efforts resulted in changes among the Swahili. In the case of the British, their interactions took place mostly at the elite level, where they coerced various interested parties in India and the Gulf into restricting the slave trade to India. They also began the process of suppressing piracy in the Gulf (they would soon bring their own brand of piracy to the slave trade suppression effort). The Omani’s imperial efforts were perhaps more significant, in that they initiated what would grow into a major movement of people from the Persian Gulf and South Arabia to East Africa. Seyyid Said’s policies encouraged Omanis to establish plantations in East Africa and to engage in trade there. However, many of the “Omani” immigrants were actually Indians who had been residents of Oman prior to coming to East Africa.[32]

The nature of these migrants is interesting. They came from a huge variety of backgrounds. Richard Burton’s description of the various ethnic groups he found in Zanzibar in the late 1840s is fascinating. There were several types of Arabs, Baluchis, and a number of different Indian religious and ethinic groups. What is most interesting is his assertion that some of the Indians spoke Swahili as their first language.[33] This suggests that these Indians were at least second-generation immigrants and that they were not just “visitors.” Furthermore, these people were on some level becoming Swahili. While they would not call themselves such, their use of the Swahili language made them in some respects “Swahili.” For other immigrants, the process went even further. They adopted not just the Swahili language, but Islam, and urban living too. Thus immigrants from places as disparate as Central Africa and the Wadi Hadhramaut in Yemen found themselves participating in a process of culture creation on the coast that drew on both internal and external elements. A regional culture was in the making, a process triggered by commercial opportunity and Omani conquest of the coast.

1850–1960: High Imperialism and the Age of Steam

All of the trends we saw in the period of transition were dramatically magnified during the long century from 1850 to 1960. Changing technology explains this transition better than anything else. Steamers made their appearance in the Indian Ocean quite early if only to provide mail service and to do military duty on rivers. But not long after 1850 steamers were moving regular cargoes rather than just the high value and strategic cargoes like the mails. This became increasingly true after the opening of the Suez Canal in 1869. And steamers changed everything. They could travel in defiance of the monsoons. They also kept their cargoes dry and safe compared with sailing ships. Above all, they drove a region-wide commercial expansion. For East Africans this meant further integration into the regional economy and culture. The trade routes into the African interior that had been opened in the earlier part of the century continued to flourish. Zanzibar-based traders moved farther and farther into the interior. The most famous of these was Tippu Tip, who by the 1870s had carved out a trading empire beyond Lake Tanganyika in what is now the Democratic Republic of Congo (the former Zaire). Tippu Tip made his fortune acquiring, often at gun point, slaves and ivory that he exported to the coast. (His real name was Muhammad al-Murjebi. His nickname “Tippu Tip” is an onomatopoeic rendering of the sound of gunfire.) Many others did likewise, though few did so on the scale that Tippu Tip managed. The slaves they brought to the coast were exported to Arabia or employed locally on the coast. The suppression of this trade was to be a major point of entré into coastal affairs for the British.

There were many buyers for the ivory that made it to the coast. Americans from Salem and New York had offices in Zanzibar where they bought ivory, as well as hides, dried chilies, and gum copal. Like the ivory, the hides and gum copal were raw materials sought by industry in North America. Indian merchants also bought ivory and transported it to India, often in steamships. Some of this was cut and carved in India and some was re-exported to Britain. In addition to exporting slaves and ivory, the coast also exported agricultural products. On the islands of Zanzibar and Pemba this meant cloves; on the mainland, it meant grain and coconuts. Grain and, in less favorable locales, coconuts flourished in the lands surrounding the coastal towns. Tanga, Pangani, and Mombasa all produced millet and sorghum for export to the grain starved regions of south Arabia and the Persian Gulf. The laborers that did the work on these plantations were slaves, most of whom came from the interior.

In Zanzibar, the clove plantations that Seyyid Said had encouraged his followers to plant became more central to the economy and as the Omani state acquiesced to British demands that slaves not be sent overseas, these plantations absorbed more and more of the slaves that might otherwise have been exported. The clove industry went through several cycles of boom and bust in the long century we are looking at, but the general trend was toward its expansion. Like the grain plantations on the coast, the clove plantations had a voracious appetite for labor. Much of this labor came from the mainland interior. Even after the legal slave trade ended in the 1870s, an illegal slave trade continued to feed labor to the plantations. Once the British stamped out this illegal slave trade, however, they found themselves in an awkward position. British rule in Zanzibar was even more dependent on revenue from the clove industry than the Omanis had been. Thus in their zeal to end the slave trade, they undercut the clove industry’s labor supply. In the end the British began to recruit workers for the Zanzibar clove industry in many of the same areas where Zanzibaris had “recruited” slaves a few years before. Plantation agriculture on the coast—whether it produced grains on the mainland or cloves on the islands—drew labor from the mainland interior throughout the period in question. The regional economy created enough wealth to bring migrants, voluntary and involuntary, to the coast.

The steamship revolution brought in its wake other new technologies. Telegraph lines began to traverse the region making it possible to move commercial intelligence even faster than steamships could travel. Ship captains could now know worldwide prices for the cargoes they bought, ship owners could easily follow the movements of their ships, and commerce exploded. Ports like Aden that specifically catered to the steamer trade boomed, especially after the opening of the Suez Canal. Aden went from a struggling town mostly dependent on having wrested the coffee trade from Mocha to one of the busiest bunkering ports in the world. Bombay surpassed Calcutta as India’s leading port because it better served the needs of steamers. Steamers and the telegraph, coupled to the cheaper goods now produced by the industrialized economies of the north Atlantic, created conditions that rapidly increased the rate of interaction in the region.[34]

Under these conditions Britain’s empire grew by leaps and bounds. British control over and economic direction of India increased in the aftermath of the Sepoy Rebellion, Oman came under much tighter control, and by 1890 most of the East African Coast belonged to one European power or another. Whatever the failings of imperialism, the British Empire facilitated trade and the movement of people around the region. Indians moved to East Africa and the sugar islands of the southern Indian Ocean in droves. They brought with them business habits, culinary habits, styles of dress, and vocabulary that profoundly shaped urban life and, to a surprising extent, rural life in East Africa.[35] Indians also settled in the Aden colony, where they ran businesses that dealt with Somalia and Ethiopia as well as working as an administrative class for the British. For example, the Cowsaji Dinshaw firm, based in Aden with branches in Zanzibar and Bombay, even ran a steamer service between Aden and East Africa. The family that owned the firm also helped pay for the construction of a Zoroastrian fire temple in Aden, which is in itself a good example of the types of encounters with the alien that occurred during this period. There had not been a Zoroastrian presence in South Arabia since the Sassanian conquest, but the British Empire swept up in its train just enough Parsis to again bring a Zoroastrian presence to South Arabia and also to Zanzibar too.

Arabs also migrated in increasing numbers. The best documented of these movements is the Hadhrami diaspora.[36] Hadhramis carried on their tradition of leaving the Hadhramaut—a place that could not support an increasing population—and migrating to lands of the Indian Ocean rim. Hadhramis were a factor in East Africa where they found work doing everything from selling prepared coffee to being religious leaders. Hadhramis also went to India as merchants and as soldiers. In their military capacity they served the Nizams of Hyderabad. Omanis and other Gulf Arabs came to East Africa, often as sailors staying for the austral winter and then returning home in August and September. But some also came as permanent migrants. When Villiers rode a Kuwaiti boom to Zanzibar the passengers included a whole sub-clan of Bedouin who disembarked at Mombassa, apparently fleeing famine at home.[37] Perhaps the most salient feature of migration in the western Indian Ocean area during this period is the degree to which East Africa was the focus of that migration. East Africa was the land of opportunity; the Gulf and South Arabia were impoverished places that people sought to escape.

Only some of these migrants traveled in steamships. Passengers have always chosen steam over sail when they have had a choice, but choice is often constrained by economics and availability. For most of the period in question getting from the Gulf to East Africa was best accomplished in a dhow. Passengers coming from India had more options, but costs were still higher in steamers. Passengers who desired to travel routes for which there were few steamers available coupled to a steady regional demand for bulky low-value cargoes—dried shark, coconuts, mangrove wood, dates, onions, ghee, and matting—made this a busy time for dhow sailors. At least as far as my data for East African ports (and a tentative look at the data for Aden) indicates, dhow traffic boomed during the long century from 1850 to 1960.[38] If we accept the dhow as the barometer of regionalism then this was the heyday of regional interaction.

Looking just at trade between East Africa and the Gulf, we see mangrove wood from East Africa (the two by four of the Indian Ocean) going to the Gulf where it was a standard part of the building material. Dried shark from the Gulf was a staple of the poor on the East African coast. In both cases not only did common people of the area use these trade goods in their everyday lives; there were also a substantial number of people engaged in producing these goods for export and in moving them from place to place. It was a trade that helped to create and define a region and a historical period.

If the Swahili coast was once “African” and local, the period from 1850–1960 was the era in which it ceased to be so. The huge influx of immigrants, many of whom became “Swahili,” coupled with the dramatic integration of regional trade with everyday concerns—food, clothing, shelter—indicates that in this era the Swahili became more cosmopolitan and regionally integrated than they ever were in the past or have been since. Therefore, in this period the coast is best seen from the sea rather than viewed from the land.

1960 and Beyond: Oil and the Airplane

Three main factors have redefined the most recent period of western Indian Ocean history: decolonization and its aftermath, the Gulf Oil boom, and the transportation revolution created by cheap airfares and containerized shipping. I know that decolonization came to India a bit earlier than 1960 and a bit later to much of East Africa and the Gulf. I chose 1960 because it seems to be a genuine watershed in terms of migrations and trade, and is right in the middle of the regional decolonization process. One effect of decolonization was to make some of the migrants from the previous era feel distinctly unwelcome in their adopted homes. In Zanzibar this took the form of the elite increasingly seeing itself as Arab, while some non-elites sought to reemphasize their African identity. The earliest manifestation of this comes in post-colonial India when the Hadhrami mercenaries who worked for the Nizam of Hyderabad were sent packing in 1948. In Zanzibar, after the revolution of 1964, Arabs and Indians left in droves, some for the mainland but many for Oman. On much of the East African coast being an Arab or an Indian became a political liability.

At virtually the same time, oil revenues began to pick up in the Persian Gulf. Labor, skilled and otherwise, came into hot demand. By the early 1960s Gulf dhow captains were having difficulty recruiting crews.[39] Sailors found they could make more money in the oil fields, doing safer work that kept them closer to home. Rising prosperity in the Gulf made catching and drying shark look less appealing, and at the same time killed the market for mangrove wood as most new construction was done with concrete and re-bar. Long distance trade conducted in sailing craft between East Africa and the Gulf dropped off after 1960. The dhow crews that had once clogged the streets in East African ports stopped coming. Indian vessels continued to come for a few more years, but after a while they too dropped from the scene. The action was now in the Gulf. Where once migrants from India and Arabia came to East Africa, now everybody wanted to get to the Gulf. A great reversal had occurred. Once the poorest part of the region, the Gulf rapidly became the place to make one’s fortune. Indians flooded into the Gulf. So did Koreans, Thais, Americans, Britons, and Arabs from non-oil producing regions.[40]

Airplanes and shipping containers have also reshaped the nature of the interactions that take place in the region. Up to 1960 the average person could not afford air travel. However, since the 1970s the types of encounters that once took place in port cities now happen in the restless transience of airports, and probably are more superficial even if they are more frequent. Containerization has also changed ports and hence the types of regional interactions that take place in them. Containers move goods with incredible efficiency. As a result the number of people needed to move them from place to place has dropped. So while trade has increased around the region, the sailors who once moved with the goods no longer do so. Container ships have tiny crews compared to the huge volumes they carry and ports now seem dehumanized because of it. Containers also call for lots of capital investment. As a result small labor-intensive ports have conceded much of their business to larger capital-intensive ports. A few big, highly mechanized ports get most of the traffic. Smaller ships and even sailing craft still serve the smaller ports, but the days when a ship from some far off place might call in at a small regional port are gone. Only local boats ply the waters between the big international ports and the small regional ports.

In the last forty years the locus of regional interaction has changed. The sea no longer is a major source of interaction even though sea borne trade has grown dramatically. If you want to meet someone from “away” as they say in Vermont, skip the port, go to the airport. Or go to the movies. But watching someone on the screen, whether silver or TV, is no substitute for direct human encounters. Nevertheless, air travel has widened the range of possible interactions. A recent article by Susan O’Brien indicates that the Hausa practice of bori—a spirit possession cult—has spread to Mecca. Nigerian hajis, some of whom were also bori practitioners, have been staying on after the haj in Mecca where they work illegally providing bori services—mostly fertility-related—to Saudi clients.[41] Air travel has made the Hijaz as accessible from West Africa as it is from East Africa.

We are probably too close to these events to be able to see them in historical perspective. But it looks to me as though the concept of a regional unity of the western Indian Ocean may not work as well for more recent years as it did before 1960. At least if we accept Bentley’s trinity of migration, conquest, and trade, the events of this most recent period may display too wide a scope to constitute western Indian Ocean history. Meanwhile the Swahili coast itself has stagnated economically. While there is some trade between the Gulf and the old Swahili ports, tourism—much of it based on nostalgia for the pre-1960 trading world—has replaced regional trade as the region’s lifeblood. The coast now seems caught between its links to mainland Africa and the jet ports that link it to the seemingly inexhaustible supply of young westerners with backpacks. But the western Indian Ocean no longer seems to be a “place.”


Regions, for the purposes of historical analysis, come and go in time. The Swahili coast has at times been a part of a western Indian Ocean region. When the Swahili civilization was in its earliest stages, it was certainly affected by regional forces, but those forces were not sufficiently central to the life of the coast for us to say that a valid regional unit of analysis existed. Even with the arrival of the Portuguese—who had region-wide ambitions—the coast, without being isolated, continued to exist in a world of its own.

By the 1750s economic and technological change began to create a stronger regional dynamic and the East African coast became increasingly a part of a regional system. This regional system, which was at its peak in the late 19th and early 20th centuries, was characterized by British imperial power along the virtually the entire western Indian Ocean rim, by extensive long-distance trade carried by steamers and sailing craft, and by large scale migrations from Arabia, western India, and central Africa to coastal East Africa. The result was a change in coastal East African life and identity. The Swahili language was adopted by the new arrivals, who in turn added new vocabulary to it. New clothing styles emerged that reflected both regional and local concerns. While scholars may use the term Swahili to identify the people of the coast during this period, it is clear that the people in question were not the same as the people whom we call Swahili in 1000. Many of them may have been the descendents of those original Swahili, but their world had become dramatically bigger, more regional, and more cosmopolitan, than that of their ancestors. Increasingly, more of these “new” Swahili were probably not the descendants of the ancient Swahili. Rather, their ancestors could be found in places as disparate as Malawi and the Hadhramaut. But the notion that one’s ancestors, or for that matter one’s skin color or the continent of one’s birth, determine a person’s social and cultural identity is a dangerous and obscurantist relic of folkish history. The Swahili world of this period belongs more to the western Indian Ocean region than it does to the huge, varied and essentialized continent of Africa—if such a place can be said to exist in any sort of cultural or historical sense.

By the 1960s, however, the unity of the western Indian Ocean was collapsing. Newly independent states sought to limit the migration that characterized the earlier period. Dhow movements were intentionally limited and controlled. Zanzibar, for example, outlawed port calls by vessels that had called previously in Arab ports. The new states of Kenya and Tanzania were ruled by governments that saw nothing positive in the western Indian Ocean region, and instead sought to identify themselves with a continental Africa. The coast, with its history of regionalism and cosmopolitanism, was either ignored or treated as a threat to the “African” identity of the new states. Since the early 1960s coastal East Africa has become less cosmopolitan and less linked to a western Indian Ocean regional culture and history than it once was.

The lesson in all this is that units of analysis are contingent upon a variety of historical forces. There is no historical “Africa” or “Indian Ocean” that can be said to exist at all times. These units of analysis come and go in time as the historical forces that bring people together or separate them come and go. Likewise ethnic and cultural identities are less the product of origins—historical or continental—than they are the product of shifting and contingent historical forces. If coastal East Africa was “African” in 1000 and in 1990, that does not mean it was in 1900.


1. Mark Horton, Shanga: a Muslim Trading Community on the East African Coast (Nairobi: 1996) inter alia; Derek Nurse and Tom Spear, The Swahili (Philidelphia: 1985), Chapurukha Kusimba, Rise and Fall of the Swahili States (Walnut Creek, CA and London: 1999).

2. Neville Chittick, “Kilwa and the Arab Settlements on the East African Coast,” Journal of African History 4(1963), 179–90, and James Kirkman, The Arab City of Gedi (London: 1954).

3. Kusimba, Rise and Fall, 26.

4. M. W. Lewis and K.E. Wigen, The Myth of Continents (Berkeley: 1997).

5. Jerry Bentley, “Cross-Cultural Interaction and Periodization in World History,” American Historical Review (1996), 749–770.

6. Patrick Manning, “The Problem of Interactions in World History,” American Historical Review (1996), 771–782.

7. Bentley, “Cross-Cultural Interaction,” 752.

8. Quoted in K.N. Chaudhuri, Asia Before Europe (Cambridge: 1990), 6.

9. Manning, “Problem of Interactions,” 778.

10. Chaudhuri, Asia Before Europe, 36.

11. Area Studies, Regional Worlds, white paper produced by the Globalization Project at the University of Chicago, 1997.

12. Kirkman, Arab City, 1954.

13. See for example: Thomas Spear, “Early Swahili History Reconsidered,” International Journal of African Historical Studies, 33:2(2000); Mark Horton, Shanga (London: 1996); Kusimba, Rise and Fall; and James Allen, Swahili Origins (London: 1993).

14. Justus Strandes, Portuguese Period in East Africa (Nairobi: 1961); M. N. Pearson, Port Cities and Intruders (Baltimore: 1998); and Jeremy Prestholdt, “Portuguese Conceptual Catagories and the ‘Other’ Encounter on the Swahili Coast,” Journal of Asian and African Studies 36:2(2001).

15. Norman Bennett, History of the Arab State of Zanzibar (London: 1978) and Arab versus European (New York: 1986); Micheal Lofchie, Zanzibar, Background to Revolution (Princeton, NJ: 1965).

16. Edward Alpers, Ivory and Slaves in East Central Africa (Berkeley and Los Angeles: 1975) and Abdul Sheriff, Slaves, Spices and Ivory in Zanzibar (London: 1987).

17. Frederick Cooper, Plantation Slavery on the East African Coast (New Haven: 1977), From Slaves to Squatters (New Haven: 1980) and On the African Waterfront (New Haven: 1987).

18. Jonathon Glassman, Feasts and Riot (London: 1995), and “Sorting Out the Tribes,” Journal of African History 41:3(2000), 395, and Laura Fair, Pastimes and Politics (Athens, OH: 2001).

19. Edward Alpers, “Indian Ocean Africa: The Island Factor,” Emergences, 10:2(2000), 373, and “Recollecting Africa: Diasporic Memory in the Indian Ocean World,” African Studies Review, 43:1(2000), 83, Janet Ewald, “Crossers of the Sea,” American Historical Review, 105:1(2000), 69, and see the special section entitled “Trade, Politics, and Identity in the Colonial Indian Ocean,” edited by Edward Alpers in Comparative Studies of South Asia, Africa and the Middle East 19:2 (1999).

20. Chaudhuri, Asia Before Europe, 40–41, and Trade and Civilization in the Indian Ocean (Cambridge: 1985), 221.

21. Spear, “Early Swahili History,” 257.

22. Kirkman, Arab City, and G.S.P. Freeman-Grenville, Medieval History of the Coast of Tanganyika (London: 1962).

23. Horton, Shanga, and Middleton and Horton, The Swahili.

24. Kusimba, Rise and Fall, 109–114.

25. Derek Nurse and Thomas Spear, The Swahili (Phildelphia: 1985).

26. Pearson, Port Cities, Chapter 3.

27. Kusimba, Rise and Fall, 155.

28. See M. N. Pearson, Portuguese in India (Cambridge: 1987), and Niels Steensgaard, Asian Trade Revolution of the Seventeenth Century (Chicago: 1973).

29. Alpers, Ivory and Slaves, 172–203.

30. See for example, Sheriff, Slaves, Spices, and Ivory.

31. Alpers, Ivory and Slaves, 176.

32. M. Reda Bakr, Trade and Empire in Muscat and Zanzibar (London: 1992).

33. Richard Burton, Zanzibar: City, Island, and Coast (New York: 1967) [1860].

34. Daniel Headrick, Tools of Empire (New York: 1981).

35. One effect of this period was to change not just the scale of interaction but also the scope. The area drawn into the Indian Ocean region’s influence grew as Indians and Arabs moved inland in East Africa following first trade routes and then the colonial flag. At the same time Yemeni highlanders began to be recruited into the Port of Aden’s work force, bringing that region a more direct taste of salt water than it is used to.

36. U. Freitag and W.W.G. Clarence-Smith, Hadhrami Traders, Scholars and Statesmen in the Indian Ocean (Leiden: 1997).

37. Allan Villiers, Sons of Sinbad (New York: 1940), 72–75.

38. Erik Gilbert, The Zanzibar Dhow Trade: An Informal Economy on the East African Coast, 1860–1964, (unpublished Ph.D. thesis, Boston University: 1997), 88–101.

39. Gilbert, Zanzibar Dhow Trade, 246.

40. For a brilliant evocation of this world see Amitav Ghosh, In an Antique Land (New York: 1992).

41. Susan O’Brien, “Pilgrimage, Power, and Identity: The Role of the Hajj in the lives of Nigerian Hausa Bori Adepts,” Africa Today 46 (2000), 12–40.


BY: Erik Gilbert