Jonathon Glassman

Slower Than a Massacre: The Multiple Sources of Racial Thought in Colonial Africa

Educated Africans are continually agitating to be given more responsibility, but I submit to you ... that you will be unable to take that part unless and until you have inculcated in your own people a pride of race. Without this, education is useless.

A. W. Norris, in Mazungumzo ya Walimu, 1930.¹

Soap and education are not as sudden as a massacre, but they are more deadly in the long run.

Mark Twain, quoted in Mazungumzo ya Walimu, 1957.²

The above insights appeared in Mazungumzo ya Walimu (Teachers' Conversations), a magazine published by schoolteachers in colonial Zanzibar, an island sultanate just off the East African coast.³ While unusual for having been written by Westerners rather than Zanzibaris, the quotes represent Mazungumzo's overall faith in the power of education to advance the goals of moral improvement. But Twain's aphorism also fits in a way that was unintended, and is chilling. It appeared in 1957, the opening year of the Zama za Siasa or "Time of Politics," when the sultanate became gripped by mounting racial tensions as nationalist parties vied in elections meant to prepare it for independence. Even though the schoolteachers believed their thirty-year project of uplift and nation-building had been as salutary as the teaching of modern hygiene, their rhetoric in fact contributed to the atmosphere of racial hatred that culminated during the early 1960s in pogroms against the so-called Arab elite, including those accompanying the racial revolution that overthrew the sultanate in January 1964, a few weeks after independence.⁴ Outsiders viewed the massacres as emblematic of the "primordial attachments" that blocked the way of those who wished to build nations in the former colonies.⁵ But, in many ways, the obstacle had been created by the nation-builders themselves.

The schoolteachers and their rhetoric typify the elite intelligentsia who dominated mainstream nationalism throughout Africa in the middle decades of the twentieth century. In suggesting that Zanzibar's postwar racial hatred was rooted in the political discourse they authored, I am going against several currents in the literature on exterminationist ethnic violence in the colonial and postcolonial world. Much of that literature assumes that ethnic conflict arose more or less automatically from social structures that had been bolstered or even created outright by colonial rule. Its emphasis, then, is not on indigenous political thinkers but on European policymakers who defined and divided their subjects by race and ethnicity.⁶ Historians who do take cognizance of indigenous racial thinkers usually portray them as marginal figures, the tools of colonial mentors. The result is that the intellectuals who incited dehumanizing violence are treated as aberrations, "sub-nationalist" demagogues isolated from the anti-colonial mainstream of nationalist thought.⁷

Authors who give any consideration to the role of local thinkers in the rise of Zanzibari racial politics (they are remarkably few, for reasons to be explored below) generally bypass the intelligentsia altogether and focus instead on the poorly educated ideologues of the African Association and its successor, the Afro-Shirazi Party (ASP), whose virulent racial populism informed most of the pogroms. These subaltern intellectuals came late to the political scene, as did the migrant workers and urban poor who constituted the core of their constituents, and their rhetoric is usually assumed to have been derived from external sources, especially the abolitionist history taught in colonial schools. The intelligentsia, in contrast, were the pioneers of nationalism, first in the Arab Association and later in the islands' first viable political party, the Zanzibar National Party (ZNP). For decades, they had advocated an inclusive form of national identity that bridged racial and class divides by enjoining common loyalty to the sultan and to the longstanding traditions of Islamic civilization that he supposedly represented. This unifying message steadily gained support throughout the 1940s and 1950s (according to the standard account), only to run aground on the racial fears whipped up by the ASP during the Time of Politics of 1957–1963.⁸

Although it is undeniable that ASP demagoguery lay behind much of the violence of the Time of Politics, the racial thinking from which it emerged, and which made so many Zanzibaris susceptible to its seductions, was fairly pervasive and as such is unlikely to be traced to a single cause. On the contrary, as Ann Laura Stoler has observed, a "scholarly quest for origins"—a quest for the moment of "original sin" when "the die of race was cast"—can only obscure a full understanding of the etiology of racial thought.⁹ Such quests assume what Loïc Wacquant describes as "the logic of the trial," with investigators seeking to name "victims and culprits" rather than understand complex historical processes.¹⁰ A pointed illustration of the dangers involved can be found in the literature on Rwanda, a case that parallels Zanzibar's in many ways.¹¹ In a recent synthesis, Mahmood Mamdani focuses on the processes by which notions of difference became racialized during the colonial era. This concept of racialization can be indispensable, for it prompts us to ask how diverse forms of ethnic and national thought can become invested with racial meanings.¹² Yet, like many authors, Mamdani traces the racialization process back to a single source, the actions of the colonial state.¹³ The result is a view of colonial intellectual history in which Europeans are the only actors, inventing and imposing identities as prompted by administrative needs.¹⁴

The literature on colonial Africa is rife with such interpretations. To understand why this is so, one must confront a cluster of misapprehensions about the nature of race and associated forms of ethnic and national thought, some specific to the study of Africa, others more general. Only then can one craft a strategy that does not underestimate the role played by African thinkers in the construction of race.

The first of these misapprehensions is a lingering tendency toward what Robert Miles calls the "conceptual inflation" of race into an element of social structure; this is especially pronounced in studies of the colonial world. Most sociologists now reject that tendency, preferring instead to understand race as a mode of thought—in constant interplay with social structures and political processes, to be sure, but best approached as a topic of intellectual history. This understanding stresses that the history of race has involved the "production and reproduction of meanings," specifically, meanings concerning particular ways of categorizing humanity.¹⁵

However, most studies of the history of racial thought limit themselves by regarding their subject only as a specific corpus of ideas: a "doctrine" that categorizes and ranks humanity in terms of biology. Racial thought is thus usually approached as a school of Western science ("raciology," as it once was called), which realized its classic distillations in nineteenth-century Europe.¹⁶ To be sure, some authors have recognized the limits of such a view, since, even at the height of raciology's academic respectability, few pogromschiki or lynch-mob members would have been conversant with the writings of Joseph-Arthur de Gobineau or Houston Stewart Chamberlain.¹⁷ More pointedly, over the past twenty years, a substantial literature has taken cognizance of post-scientific forms that have flourished in the wake of raciology's postwar academic demise. These "new racisms" demonstrate that racial thought need not be manifested in a scientific idiom or entail a comprehensive ranking of racial categories. There now exists a "racism without races," writes Etienne Balibar, "a racism whose dominant theme is not biological heredity but the insurmountability of cultural differences."¹⁸

Yet even the literature on the "new racism" portrays it as a holdover from classic raciology or, more precisely, as a deteriorated version of that ideology, which once existed in a pristine, originary state.¹⁹ Such literature is mistaken both in its depiction of the supposed newness of culturalist racial thought and in its depiction of the older forms, which in fact were neither invariably hierarchical nor invariably built around a core of biological theory.²⁰ Past social practices that are universally accepted as classic examples of "racism," including colonial racisms, were informed by a wide variety of ideologies, many of which had little to do with racial science. Far more influential than raciology, for example, was the anthropological concept of clearly bounded "cultural monads," a concept directly connected with contemporary culturalist thought. As Balibar recognizes, the idea of "racism without races" is far from revolutionary.²¹

In succumbing to the search for origins, scholars overlook a central theme of the historical literature on racial science itself, which charts the latter's own varied and multiple sources, including many that were neither "racial" (in the conventional sense) nor scientific. Perhaps most significant of these sources was the concept of "barbarism" and its foil, "civilization," from which modern race-thinkers inherited the project of comparing all humanity according to a single, universal standard.²² It should be axiomatic that the history of a phenomenon cannot be found solely by looking for its earliest manifestations as it is defined a priori; to paraphrase Friedrich Nietzsche, in defining a phenomenon we deny it a history.²³ Many scholars nevertheless insist on an absolute divide between racial thought and the concepts that contributed to it, and search doggedly for the precise moment that race emerged from (for example) the discourse of barbarism. As Stoler archly observes, they come up with widely divergent dates.²⁴

One of the implications for the study of the colonial world should be clear. If "race" is assumed to arise solely from scientific doctrines, then its presence in the non-Western world must be traced solely to the West. And that, in fact, has become a standard narrative. Building on a set of functionalist assumptions often associated with Immanuel Wallerstein, authors describe how Western expansion called racial thought into being as a way of structuring the worldwide division of labor between the subservient "periphery" and the ruling "core." In tying race so mechanistically to the structures of global capitalism, such analyses remove it from the realm of intellectual history. The result is the kind of conceptual inflation Miles warns against; indeed, several authors explicitly insist that race can be defined only in terms of the "inequalities ... inherent in a social structure" of Western conquest.²⁵ Again, such explanations are undermined by a historical literature on the multiple sources of Western racial thought, sources that included, for example, seminal debates about European difference.²⁶ Still, despite disagreements concerning the precise relationship between imperial expansion and the rise of racial thought, there is wide consensus that race was invented in the West and carried to the rest of the world in the toolbox of colonial rulers.²⁷

What happens if we abandon the fixation on scientific doctrines and instead recognize racial thought as a shifting field of discourse, a general set of assumptions that humankind is divided among constituent categories, each of which is distinguished by inherited traits and characteristics? ("Racism," a belief that racial qualities can be ranked according to moral status and other criteria, will then be seen as but a possible form of "racial thought.")²⁸ Raciology and other Western racisms then appear as historically specific manifestations of a much broader trend in Western thought—and in human thought generally. Relatively few of the intellectual currents that contributed to raciology were peculiar to the West: drawing boundaries between peoples or ethnicities and even ranking them according to universalizing registers of inferiority and superiority have been far from unusual in world-historical terms. Concepts convergent with the ideal of civilization,²⁹ ostensibly inclusive yet contributing to hierarchical beliefs and practices that look uncomfortably like "racism," have occurred in many non-Western intellectual traditions, Zanzibar's among them.³⁰ Indeed, Igor Kopytoff has argued that discourses of civilization and barbarism—including tropes marking barbarians as physically different—have been so common in sub-Saharan oral traditions that one might almost speak of them of as part of a continent-wide political culture.³¹

Still, the approaches we have been discussing all assume that race can be fundamentally distinguished from other ways of categorizing difference, such as ethnicity or xenophobia, by the central conceit that cultural identities, and hence cultural boundaries, are fixed in the body, or the "blood."³² The ubiquity of this distinction is surprising, since it is of recent vintage: it came into its own only after World War II, as social scientists strove to isolate raciology, recently discredited by the Nazis, from more general concepts of ethnic and national difference. ("Ethnicity," in fact, was something of a neologism.)³³ Yet the distinction is misleading, for it obscures the fact that all these modes of thought build on the same two core elements. First is the assumption that humanity consists of a discontinuous series of authentic cultural wholes, each internally homogeneous, the creation and property of a distinct "people."³⁴ Second is the metaphor of descent, that is, the general idea that the peoples who are the guardians of these discrete cultures are somehow linked by consanguinity. Such "blood ties" are imagined with greater or lesser degrees of vagueness. Language we call "race" places more explicit emphasis on the metaphor of descent—or, indeed, on the conviction that the "blood relationship" is more than mere metaphor—than does language we call "ethnicity."³⁵ The distinction between "race" and "ethnicity," then, is one of degree not kind; and, rather than regard them as qualitatively distinct, it is more useful to recognize them as modes of thought that fall toward opposing ends of a single continuum, the "aura of descent" hovering around them all.³⁶

I do not mean to deny the usefulness in historical analysis of distinguishing descriptively between doctrines or political ideologies based on explicitly racial concepts and those based on other kinds of ethnic criteria. (So, for example, contrary to the claims of neo-conservative critics, one must acknowledge the historical specificity of white supremacy and the unique forms of oppression it has produced.³⁷) But it should be recognized that at any particular historical juncture all such doctrines are part of a common discourse of difference that categorizes humanity via metaphors of descent. The "family resemblance"³⁸ among them must be grasped if we are to understand how virtually any form of ethnic thought carries implications that are capable of being elaborated in terms of qualities fixed in the "blood." Such elaboration—the process of racialization—is rarely the work of a unified cadre, inspired by a single coherent doctrine. Rather, it emerges from the debates of a diverse range of intellectuals, drawing on multiple and overlapping sources, united only by the general assumptions of racial/ethnic discourse.

Nationalism is one variant of ethnic thought that has proven especially susceptible to racialization.³⁹ By the mid-twentieth century, the politics of the nation-state had become a global "categorical order," a set of concepts taken for granted by leading political thinkers throughout the colonial world.⁴⁰ Of course, this did not automatically lead to unity. Nations have never been defined by any single set of criteria, and conflicts have always been rife among nationalist ideologues over who, precisely, belongs to the national community.⁴¹ It has become a truism that nations are defined as much by exclusion as inclusion, that is, in terms of who does not belong. When such exclusionary rhetoric is considered alongside the genealogical metaphors that underlie all ethnic discourse, one can understand how virtually any form of national thought—including, as we shall see, thought based on an ostensibly inclusive, universalizing language of civilization—might be interpreted in ways that denigrate certain categories of people by dint of their descent. There is no firm line between national thought and racial thought, and a racial paradigm of exclusion and dehumanization is implicit in virtually all nationalist projects, even the most liberal.⁴²

The misconceptions I have been highlighting are especially common in the literature on Africa, much of which continues to approach ethnicity, race, and nation as analytically distinct. There is a tendency to regard only the last of these as the product of ideologies consciously crafted by African thinkers. Ethnicity and race, on the other hand, are often inflated into "sociological facts," nowadays usually explained instrumentally. Some scholars, as we have seen, dismiss ethnic thought as little more than a European invention, accepted by Africans in order to secure access to resources controlled by missionaries and the colonial state.⁴³ Others, more persuasively, recognize the role played by African intellectuals, who cast ethnic appeals in ways calculated to resonate with concerns that had been shaped by labor migration, clientelism, and other processes associated with the construction of the colonial political economy. Yet even these authors assume that ethnic rhetoric can be traced to European origins; hence, as Leroy Vail argues in an often-cited overview, African intellectuals should be seen as "brokers." More fundamentally, these authors stress that the power of ethnic appeals can be explained only in terms of material need. In downplaying all other factors, which Vail dismisses as "non-rational" and therefore irrelevant, they minimize the impact of the content of ethnic thought, even while focusing on the intellectuals themselves.⁴⁴

The continuing assumption of a strict analytic divide between nationalism and other forms of ethnic thought is part of a more general failure to break fully with the nationalist paradigm that has shaped African historical studies for the past forty years. And this has prevented historians of Africa from addressing the kind of questions that have marked other literatures on ethnic politics. Those other literatures have examined how intellectuals crafted locally compelling languages of belonging and exclusion whose affective power rarely had much to do with utilitarian logics.⁴⁵ In contrast, the Africa literature has largely treated political thinkers "empirically," describing their roles in creating and leading parties and associations, but reducing the content of their thought to the formulae of a nationalist calculus: "nationalism" as inherently inclusive and liberatory, "ethnicity" as the divisive legacy of colonialism. The result is a history written from the perspective of the nationalists who took control of the state from the departing colonial powers.⁴⁶

Many of these shortcomings have been remedied in the past decade or so with the emergence of a new literature that seriously engages with the thought of African intellectuals who debated the public good in ethnically specific discourses that were once dismissed as "tribalism" or "sub-nationalism."⁴⁷ Rather than force ethnic thought onto a Procrustean bed of utilitarian "rationality," these studies reconstruct alternative rationalities that did not necessarily rest on logics of straightforward material advantage. By tracing the deep histories of these discourses, they demonstrate the limits of colonial-era "invention."⁴⁸ They also demonstrate that ethnic thought had a multiplicity of sources, not just the material demands created by the colonial state as interpreted by instrumentally minded political entrepreneurs, let alone the ideological constructs imposed by colonial rulers. Thus they avoid what Nancy Hunt calls the cliché of the colonial encounter, the nationalist paradigm that interprets modern African history in Manichean terms.⁴⁹

But studies of the racialization of African ethnic thought, and of the exterminatory violence it has at times engendered, have proven stubbornly resistant to this kind of analysis. Most studies of African ethnicity deal with vertical or regional divisions that were once commonly described as "tribes": ethnic categories that are imagined to exist side by side, each an "incipient whole society." Genocide and other forms of exterminatory violence, on the other hand, tend to be accompanied by rhetoric in which ethnic categories are imagined as hierarchical strata, linked to one another in relationships that structure the entire society; the violence itself is prompted either by the subordinate group's attempt to throw off those it sees as its oppressors or the dominant group's attempt to preempt such a revolt.⁵⁰ Such situations of "ranked" ethnic stratification parallel the Western experience of race; in fact, there is a sociological literature that distinguishes "race" from "ethnicity" precisely by the presence of such ranking.⁵¹ A recognition of those parallels is what has prompted several authors, including Mamdani, to write of "race" when describing clashes like those in Rwanda or Zanzibar. But, combined with the persistent assumption of an originary distinction between "race" and "ethnicity," as well as the assumption that race is a Western invention, the parallels have also prompted many of the same authors to assume that the primary force behind racialization was colonial indoctrination.

As a result, relatively little attention has been paid to African initiatives in the racialization of "ranked" ethnic thought.⁵² One suspects that many authors are immobilized by a set of political scruples borne of the assumption that all racial thought originates in white supremacy; in short, they don't want to appear to be blaming the victim.⁵³ That assumption and those scruples seem confirmed by the undoubted influence that Western teachings had on African racial thinkers, as well as by the fact that the antagonists in some of these conflicts, including the "Arabs" and "Africans" who fought for political advantage in Zanzibar, defined themselves in terms that converged with major categories of Western racial thought.⁵⁴

These characteristics are amply displayed in the Zanzibar literature, much of which portrays race as a peculiarly Western disease, introduced to the islands by the effects of colonial rule and even the deliberate machinations of European officials and educators.⁵⁵ Authors vary in their identification of the malady's local vectors. Many focus on immigrants from the African mainland, reputedly the ASP's most militant loyalists. Mainlanders presumably were more susceptible to European influences, including mission education, than were people rooted in the islands' Islamic communities, who in contrast continued to nurture more flexible local concepts of belonging. Other authors blame the ruling "Arabs," who championed British policies that propped them up as a racial elite.⁵⁶ Both variants reflect the political scruples already mentioned, borne of postcolonial politics. In their cruder forms, they constitute nationalist orthodoxies—one supportive of the revolutionary government, the other of the opposition—in which politicians who before independence took the lead in fomenting racial politics absolve themselves of all responsibility by simply blaming the colonial state and its stooges.⁵⁷ Whether crude or scholarly, such interpretations depict African intellectuals as having either too little autonomy or too much: either dupes whose thinking was easily molded by British mentors or steadfast anti-imperialists who clung to an authentic subaltern consciousness. Neither circle spoke to the other, and the anti-imperialists avoided all contamination from colonial thought.

Below, I will illustrate how these assumptions might be subverted by reconstructing the precise conversations and debates from which racial thought emerged, a task that requires abandoning images of authentic indigenous discourses and imported infections, and the "logic of the trial" that goes with them. Such a reconstruction reveals that indigenous intellectuals spoke to one another more than they addressed the colonial state or responded to its demands, and that their impact on the emergence of racial thought was arguably greater, and certainly more direct, than that of colonial educators. (As Anthony Appiah has observed, the colonizers were never as fully in control of intellectual life as the nationalist elite subsequently made them appear. ⁵⁸ ) I have suggested elsewhere that the Asp's racial rhetoric emerged within a discursive framework of nationalism whose terms had been set not by colonial officers but by members of Zanzibar's elite intelligentsia. ⁵⁹ Although the idea of the nation-state may ultimately be traced to the West, its pervasiveness in Zanzibar largely resulted from the influence of local intellectuals, who by the 1950s had persuaded thinkers across the spectrum that politics was about the rights of national groups. Zanzibaris disagreed only over how those groups should be defined.

Yet the intelligentsia's role in the racialization of political thought went still further: their own definition of the nation, despite its inclusiveness, directly contributed to the emergence of the Asp's more explicitly racial nationalism. The process was catalyzed by class factors. In its earliest manifestations, ZNP nationalism was liberal and inclusive, growing from a deep intellectual tradition that defined local communities of belonging in terms of an expansive "civilization" marked by the attributes of an Arabcentric urban high culture. But because the cultural criteria used to mark this civilization were largely class based (that is, beyond the reach of the poor), stressing them as marks of national identity risked dividing the very nation that the nationalists aspired to unify. Balibar has observed that this is why proponents of "civilizing" nationalisms often turn their constituents' attention to the threat of "false nationals" ("Jews, 'wogs,' immigrants") who threaten the nation from within. ⁶⁰ In Zanzibar, as elsewhere, the concept of civilization became an effective tool of nationalist mobilization when it was used to define barbarians to be purged.

Long before the Time of Politics, Zanzibar's elite intellectuals had devoted much energy to reflecting on the history of the Islamic civilization of the Swahili coast, of which they considered Zanzibar the exemplar. This civilization, they argued, distinguished the coast and islands from the rest of East Africa. In the 1950s, the implications of their historical narratives were noticed by the less cultured activists of the African Association/ASP, who accused the intelligentsia of seeking to exclude anyone who did not fully identify with urban high culture, particularly agricultural workers and the urban poor, many of whom traced their roots to the African mainland. To the intelligentsia's invocation of an Arab-driven history of civilization, African Association propagandists responded with historical narratives of Arab conquest and enslavement, and with an alternative definition of national identity based on race rather than civilization. By the late 1950s, ASP charges of racial blood guilt were routinely met by ZNP charges of innate barbarism. The resulting rancor spiraled into every corner of society, contributing directly to racial violence. Thus any attempt to trace the history of popular racial nationalism must begin, paradoxically, with the intelligentsia's rhetoric of history and civilization.

Of course, it would be misguided to neglect the impact of Western thought, including general Western concepts of ethnicity and nation. But such concepts were usually introduced indirectly and only after much reworking.⁶¹ The key actors were local intellectuals, for whom Western thought was but one of many influences, of many provenances; only they were capable of innovating versions of national thought that were locally compelling in ways that pallid imitations of Western discourse could not be. The potential complexities are well illustrated by Zanzibar's political intelligentsia, a vibrant and self-conscious intellectual community that was hardly susceptible to colonial control. The most prestigious intellectual circles were drawn from town-based elite families who considered themselves ethnically Arab. Their main idiom of discourse was Islamic, although, as we shall see, a new cadre of secular intelligentsia began to emerge early in the twentieth century, often from the same families that dominated the ranks of the ulamaa (religious scholars). The most learned knew Arabic, with which they read not only religious texts but also Cairo newspapers that championed Islamic modernism and Arab nationalism. Families were especially proud if they could send their children to study in Cairo. Intellectual accomplishment was an important component of family pride, and, given the kind of learning that was most valued, it is not surprising that such accomplishment was closely connected to a family's ability to claim elite status as Arabs.⁶²

Zanzibar's intelligentsia, then, were keenly cosmopolitan—more so, in many ways, than their European rulers and sometime teachers—and, though not isolated from colonial discourse, they had ample intellectual resources to be able to engage with the ideas of nationalism without merely parroting Western ideas. In fact, influence flowed in more than one direction: Zanzibari intellectuals had a marked impact on the thinking of British educators and administrators. Colonial historians acknowledged the influence of their local informants far more readily than is consistent with the image of a discrete and overpowering colonial discourse.⁶³ The lower echelons of the provincial administration were staffed almost entirely by members of local elite families, who were especially prominent as mudirs, the rough equivalent of district officers in other British colonies. These men were routinely commissioned to write reports on local customs and history, which were then circulated throughout the colonial bureaucracy.⁶⁴

In sum, it is just as misleading to speak of two discrete spheres of discourse—one colonial, the other indigenous—as it is to speak of the colonial state's domination of its subjects' consciousness. It is equally misleading to assume that popular discourse existed in isolation from that of the elite intelligentsia. The subalterns who would later support the ASP were not as lacking in political awareness in the interwar years as the standard sources assume;⁶⁵ they were listening to and arguing about many of the same issues that propelled the urban intelligentsia. Like any change in political culture, new ways of thinking of ethnic difference emerged from circuits of discourse in which diverse intellectuals spoke to one another—elite and popular, European and African—and in which their ideas were interpreted and debated by the population at large. Thus the history of the popular racial nationalism of the 1950s must be traced to the elite intellectual discourse of a generation before.

None of the above should be taken as dismissing the significance of instrumentalist or structural factors; in Zanzibar, as elsewhere, state formation was central to the context in which a discourse of race was created.⁶⁶ But the process of building a racial state did not begin with European rule, as one would expect from most of the instrumentalist literature. It can be traced at least to the actions of the Omani sultans who conquered Zanzibar in the nineteenth century. The sultans also sponsored the settlement of Arab planters and Indian financiers, as well as the import of large numbers of plantation slaves from the mainland. Thus the groundwork was laid for the major ethnic divisions of colonial Zanzibar: Arabs, Indians, indigenous islanders, and African mainlanders. British rule accentuated these divisions, in part through economic policies aimed at preserving the Omani Arabs as a landlord caste (Indian creditors, for example, were discouraged from foreclosing on mortgages) and administrative and educational policies that bolstered their position as a political caste. Although slavery was abolished, mainlanders continued to come to the islands to work as labor tenants ("squatters") on clove and coconut estates.

These ethnic divisions formed a widely recognized hierarchy, with Omani Arabs perceived to be at the top and, at the bottom, slaves and people of slave descent (real or presumed). In the middle were the bulk of the islands' indigenous inhabitants, who often made vague claims of descent from distant Middle Eastern ancestors. These lines of descent were usually fictive; in any case, they could rarely be traced through precise genealogies. Islanders claimed ancestral origins in a variety of exotic places, the most common being the Persian town of Shiraz, whence came the legendary founders of several of the ancient city-states of the East African coast. Such claims are recognizable as variants of a deep political tradition, widespread among speakers of Bantu languages, whereby leading families preserved myths linking them to exogenous conqueror-heroes who had introduced the technologies or forms of social organization by which civilized life was defined.⁶⁷ The more specific fashion of claiming Arab or Persian ancestry largely stemmed from coastal Muslims' desire to distinguish themselves from their non-Muslim neighbors and slaves. The fashion had long been present, but it became more pronounced under the rule of the Omani sultans.⁶⁸

The existence of this hierarchy in the minds of most Zanzibaris implied a widespread degree of Arab hegemony. Paradoxically, the persistence of the hierarchy was ensured by the permeability of ethnic boundaries: people of mainland origin (including descendants of slaves) might realistically aspire to become accepted as Shirazi, and Shirazi might realistically aspire to become accepted as Arab. Such claims constituted an affirmation of values by which the Islamic, urban-centered culture of the coast was esteemed as the embodiment of ustaarabu, "civilization," and as the antithesis of the non-Muslim culture of the interior, whence came people scorned as "barbarians" or washenzi (sing., mshenzi). These values were further encouraged by British rule, which was informed by the official policy that Zanzibar was a "protected Arab state," its head of state the sultan, its ruling caste Arabs. In fact, beginning in the 1920s, administrators noticed an acceleration in the adoption of the ethnonym "Shirazi," culminating in 1940 with the founding of the Shirazi Association in Pemba.⁶⁹

Yet no single definition of ustaarabu was accepted unquestioningly by all who aspired to embrace it. On the contrary, constant disputes over its precise meaning, including disputes over who should and should not be admitted to the status of the "civilized," are what ensured the persistence of the general idea. Central to claims of Shirazi identity was a deep-seated ambivalence over how to regard the Arab elite.⁷⁰ On the one hand, claims of Shirazi identity constituted a way to use Arabcentric notions of ustaarabu to distance oneself from mainlanders; fear and disdain of mainlanders was the primary motivation of the Pembans who founded the Shirazi Association. But, on the other hand, members of the older local families often used Shirazi identity to express status anxiety vis-à-vis the Omani ruling caste, whom they regarded as upstarts: more civilized than slaves and other mainlanders, of course, but parvenus when compared to the families who traced their roots to ancient Shirazi settlers.

During the Time of Politics, nationalists on both sides exploited this ambivalence in their attempts to win over the islands' Shirazi majority. The African Association used the rhetoric of racial solidarity to advocate the unity of Shirazi and mainlanders against a common Arab enemy. They eventually persuaded a portion of the Shirazi Association to merge with them to form the ASP. But many of the erstwhile members of the Shirazi Association allied themselves instead with the ZNP, persuaded by the intelligentsia's chauvinistic depiction of a nation founded, not on the solidarity of race, but on the shared values of ustaarabu, coastal exceptionalism, and disdain for mainland Africans.

Schoolteachers formed a disproportionate share of the secular intelligentsia that forged a nationalist discourse between the wars. This was not uncommon in colonial Africa, where schoolteachers were more numerous than other figures (say, journalists) defined by the institutions and discourse of national "modernity." Teachers were also more ubiquitous: most were posted to village communities, where they were in an excellent position to forge versions of nationalist ideology couched in local cultural idioms.⁷¹ But Zanzibar's schoolteachers were unusual in their belief that they were destined to lead not only by dint of their mastery of Western education but also by dint of intellectual attributes that they considered part of their distinct racial inheritance.

This situation was the outcome of several convergent trends. From its inception early in the century, Zanzibar's system of secular education was geared toward training an administrative cadre that could serve as the functionaries of a modern state. (Unlike other British colonies, where mission schools prevailed, virtually all secular education in Zanzibar was state-run.) Following general governing principles, that cadre was drawn principally from Arab families.⁷² But persistent Arab preponderance in government schools, especially in the higher grades, cannot be understood simply as the result of British policy. The racial vision was never pursued with the rigor that marked, say, Catholic education in colonial Rwanda; in fact, throughout the 1920s and 1930s, the government made repeated efforts to expand education for rural non-elites—efforts that foundered, as we shall see, on villagers' keen distrust of secular education. The urban gentry, by contrast, quickly overcame such suspicions,⁷³ no doubt influenced by modernist ideas emanating from the Middle East that stressed the benefits of a secular, government-sponsored education.⁷⁴

In 1923, the Education Department opened its first post-primary institution, the Teachers Training School (TTS). The new school was intended to do more than merely train teachers; classroom teaching was to serve as a forcing-bed for the cultivation of an administrative elite. That policy was endorsed by the islands' leading families, who, speaking via the Arab Association, had been demanding that government provide opportunities for sons who were being marginalized by the foundering of the plantation sector.⁷⁵ For the next thirty-five years, TTS and its successor institutions⁷⁶ would train a disproportionate share of the islands' administrative staff, many of whom began their careers in the classroom and finished them as prominent nationalists, particularly within the Arab Association and ZNP.

By 1930, these young schoolteachers had begun to cohere as a self-aware intelligentsia, imbued with a responsibility for building the "modern civilization" on which a nation could be based. Their cohesion was mediated in large part by the Education Department, in particular by the remarkable L. W. Hollingsworth, first director of TTS, whom many leading intellectuals later remembered as a formative influence.⁷⁷ In 1927, Hollingsworth launched the teachers' journal Mazungumzo ya Walimu, intending it as a forum in which teachers might continue the debates on social topics that he encouraged in the schools. Despite Hollingsworth's loose supervision during its initial decade, Mazungumzo was written and edited largely by Zanzibaris. Many of postwar Zanzibar's most prominent intellectuals and politicians first published there in the 1920s and 1930s.⁷⁸ The magazine thus served as the center of an inordinately influential circle of intellectuals, who no doubt propagated many of their ideas in the classroom before becoming directly involved in politics.⁷⁹

From its earliest numbers, Mazungumzo reveals the schoolteachers' confidence that they were a cultural elite: masters of the skills of modernity as well as of what they considered proper Islam, aloof from the rustic "natives"⁸⁰ whose children they were charged to teach but at the same time conscious of their obligations to exemplify Arab civilization in the villages where most were posted. Hollingsworth and his Zanzibari colleagues encouraged such ideals, although they frequently felt compelled to admonish Mazungumzo's readers not to yield to outright disdain. (Some teachers were given to abusive language in the classroom, calling their pupils washenzi, "barbarians.")⁸¹ The thinking of British educators reflected a mixture of considerations. They saw it as their long-term job to help lay the foundations of a modern nation (albeit one that would not assume self-government until well into the future), making their charges "better Africans" rather than a species of European. Their main strategy, encouraged by the Tuskegeeism prevalent in East African educational circles, involved cultivating an indigenous elite that would lead local communities in the building of modernity.⁸²

But British pedagogues were not the schoolteachers' only mentors. Many had older relatives active in the Arab Association, which had been founded in 1911 in a spirit of pan-Arab nationalism.⁸³ Many, too, were related to the islands' leading ulamaa, who since before the turn of the century had been reading and debating Islamic modernism. From those intertwined intellectual movements, perhaps more than from British teachers, they learned of their unique responsibility to lead in building a "modern civilization." Nineteenth-century intellectuals in Cairo and Damascus had originally borrowed this concept from Europe (although one should not ignore the rich intellectual legacy inherited from Arab thinkers like Ibn Khaldun), but by the interwar years it had been elaborated to include elements intended to challenge Western pretensions to represent the moral if not technological epitome of "civilization." Arabs must learn the history of their own civilization, argued the pan-Arabists, and in this crucial nation-building task schoolteachers were to play the leading role. The central figures in spreading this message were pedagogues based in Baghdad, but their influence was felt among educators throughout the Arab world, and it is no doubt significant that many of the first generation of instructors in Zanzibar government schools—that is, the new intelligentsia's own teachers—were recruited from Cairo, another center of interwar pan-Arabism.⁸⁴

These convergent pedagogies, then, all encouraged schoolteachers to think of themselves as a cultural and political elite. But none, at least in their original versions, had envisioned schoolteachers as ethnically distinct from the communities they were to lead. Such a distinction, of course, was in accord with Britain's unusual form of indirect rule over the "Arab sultanate." But it did not originate with the British; rather, like the policy of "Arab rule" itself, it was an outgrowth of precolonial concepts of ustaarabu and Arab hegemony. External intellectual currents were significant largely in how they influenced the intelligentsia in their elaboration of those inherited concepts.

Pan-Arabist teachings about history, for example, reinforced local distinctions of civilization and barbarism in ways that accentuated their implicit racial undertones; pan-Arab nationalism was no different from any other in its compatibility with the idea of race.⁸⁵ Reeva Simon writes that pan-Arabist narratives were inspired by volkisch philosophies of "universal history," in which the origins of civilization were traced to a "primeval ancestor nation" that carried it to other parts of the globe during wanderings from an original homeland. Adapting such theories to their own ends, the pan-Arabists focused on ancient Semites rather than Aryans, invoking the American anthropologist James Henry Breasted to describe how "Semitic waves" emanating from Arabia and the Fertile Crescent had spread civilization to the rest of the world. Despite differences concerning the precise identity of the earliest Semites, all agreed that their crowning achievements, the Arabic language and Islam, were the gift of the Arabs.⁸⁶

Hollingsworth and his colleagues, Zanzibari and British, also urged the teaching of history, of much the same kind. Although the ideas of "universal history" pervade the entire colonialist historical literature, the schoolteachers encountered them most readily in school texts, including several prepared by Hollingsworth himself. Most accessible was Milango ya Historia (Gateways of History), a world history primer that he wrote in collaboration with his TTS assistant Abdulla Ahmed Seif and other members of the Mazungumzo editorial board, and that was used widely from its first appearance in the 1920s throughout the rest of the colonial period.⁸⁷ The primer paints its principal narrative, the progress from savagery to civilization, in broad strokes, emphasizing how relatively small nations such as the Greeks and Hebrews spread civilization to the rest of the world. Its definition of civilization must have seemed familiar to its adult readers: proper government (meaning obedience to a single monarch), literacy, and, implicitly, monotheism.⁸⁸

The implication was clear: the attributes of civilization had been introduced to East Africa by outsiders. And to make it clearer still, whenever a foil was needed to highlight the accomplishments of (say) the Greeks, the primer drew its example from the foolish customs of pagan Africans.⁸⁹ Aside from such stray comments, only one of Milango's thirty-eight chapters concerns Africa; the text flatly explains that is because black Africans are barbarians who have never done anything of historical significance. But the primer notes one exception: peoples living along Africa's coasts are worthy of study because they had come under the influence of foreigners. To demonstrate, it offers a condensed version of the story elaborated in Hollingsworth's other texts, telling how centuries of Arab presence had rendered East Africa's coastal population more civilized than their upcountry neighbors.⁹⁰

Texts like Hollingsworth's thus flattered the schoolteachers by giving the imprimatur of colonial modernity to their inherited ideas of ustaarabu and coastal exceptionalism. It is hardly surprising, then, that such ideas were prominent in the many essays on local history the schoolteachers wrote for Mazungumzo. The editors solicited these essays as part of a nation-building project in which colonial, Arabist, and inherited themes played out in an ironic counterpoint. The project was launched in 1930, when Abdulla Ahmed Seif came across an essay by the Gold Coast educator A. W. Norris that he decided to reprint in Mazungumzo. Norris's refrain was the need for teachers to cultivate "PRIDE OF RACE"; this phrase, the essay's title, recurs throughout, always in capital letters. Such pride, Norris observed, has been necessary to the success of "every known empire and nation"; without it, no nation can "rule itself or others." And the best strategy for cultivating national pride, he argued, was to cultivate historical memory, which might be learned by listening humbly to the tales of the elders.⁹¹

Norris's essay impressed Seif not only for its emphasis on history and nation-building but also for how it chastised teachers who thought themselves superior to those around them: as we have seen, Zanzibar educators harbored similar concerns. But the message originally crafted for readers in Gold Coast had the opposite effect in Zanzibar. Norris had urged teachers to seek pride not in their modern education but in the history of their entire "race," alluding specifically to "the Akan race" (or Akan "nation"), that is, the ethnic category that comprised the majority of the Gold Coast population. He thus addressed a contradiction that beset most British colonies, where the language of modernity threatened the "traditional" authority of chiefs and elders on which colonial officials relied: by encouraging political traditions built around metaphors of seniority and common descent, officials like Norris hoped to cultivate "tribal" (or "racial") solidarity between chiefs and commoners, including the educated "youngmen" who resented chiefly prerogatives. But Zanzibar's very different traditions of ustaarabu and Arab hegemony ultimately rested on metaphors of separate descent, thus shaping the schoolteachers' historical narratives in ways that emphasized not unity but difference. Hence most of the narratives published in Mazungumzo concerned the Arab or Persian figures whom local lore remembered as having built the city-states that made the coast distinct.⁹² The most prolific contributor of historical essays, M. Abdulrahman, wrote a series that depicted the nineteenth century as the story of Arab heroes who explored the mainland, contending with the depravity of African cannibals.⁹³

The authors of such narratives plainly drew from colonial texts; Abdulrahman acknowledged as much. But he also relied on oral informants, and it is evident that colonial sources only reinforced distinctions between Arab civilization and African barbarism whose roots were deep and multiple. Some of these roots were revealed in a 1938 debate about the history and meaning of the words ustaarabu (civilization) and mstaarabu (civilized person). The authors all agreed that the terms were derived from an Arabic form meaning to adopt Arab culture; in tracing the etymology, some displayed a sophisticated knowledge of Semitic prehistory that suggests the influence of the pan-Arabists. Their central dispute concerned whether an African who had been educated only in the ways of Western civilization might properly be described as mstaarabu or whether alternative words should be used. One suspects that semantics were not the only source of such misgivings: Zanzibaris were known to disdain upcountry people as barbarians even if the latter had attained a Western education—a disdain that British educators, who put more stock in Eurocentric ideals than Arabcentric ones, found amusing.⁹⁴ None of the authors imagined an African model of civilization, and at least one assumed that all the continent's ancient states and empires could be attributed to Arab influence.⁹⁵ Although the path to civilization was ostensibly open to anyone who wished to travel it, the mstaarabu who completed the journey could be recognized only in contrast to another left behind.

That "other," the African barbarian, was not, strictly speaking, defined racially; after all, an mstaarabu, in the word's original (Arabic) sense, was one who became Arab. But even though the concepts of Arab civilization conveyed in both colonial and pan-Arabist texts focused on culture rather than blood, they also relied, to a greater or lesser extent, on the metaphor of descent. The same is true of local notions of ustaarabu and Arab and Shirazi identity.⁹⁶ So, although Mazungumzo's historical narratives were multi-racial, in that they portrayed civilization being shared by diverse peoples, they were not non-racial; in their focus on waves of civilized peoples spreading to East Africa and elsewhere, sharing a common ancestry and identity, which were kept intact over centuries or even millennia, they reproduced many of the assumptions of racial thought. It is no surprise, then, that at least one of the contributors to the debate on ustaarabu, having made the small leap from civilization to descent, made the further leap from descent to the body—an association not universal to racial thought but not unique to raciology, either.⁹⁷ After describing how the Arabs' introduction of Islam had prompted "the natives of the coast to regard them as their leaders, and mimic them in every way," this author added that, as a mark of having become more civilized, the coast population's skin became clearer and brighter than that of "barbarians" (washenzi), "that is, the bush-people of the African interior."⁹⁸

A discourse of racial chauvinism, then, was beginning to emerge from the intertwined discourses of civilization, modernity, and nation-building. The interconnections can be seen in Mazungumzo's campaigns for the reform of popular culture, another instance in which the nation-building concerns of British educators converged with those of Zanzibari colleagues.⁹⁹ The schoolteachers published many condescending essays on village customs, particularly dance rituals, in which they stressed an obligation to reform popular culture along the lines of modernity and (Arab) civilization. The critiques were often harsh: dances were described as foolish, indecent, and unhealthy—a common adjective was kishenzi, barbarous—and at least one headmaster thrashed pupils he caught participating in them. Such condemnations must have aroused deep resentments, since festive dance had been central to how islanders defined their very communities.¹⁰⁰ Even more central was Islam, the rituals of which, more than any others, delineated the boundary between civilized folks and washenzi. Yet the intelligentsia made little effort to hide contempt for villagers' religious practices. In accord with reformist currents throughout the Islamic world, they were particularly aroused by the practice of melding religious ceremony with festive dance, a practice, they maintained, that smacked of pagan idolatry.¹⁰¹ At least one leading cleric argued that Islam forbade all festive dances, even those not combined with a religious rite.¹⁰² Although the secular intelligentsia rarely went that far, their general critiques conveyed the message that the villagers' very Islam was tainted by barbarism and needed to be corrected by Arabs.

Other objects of the intelligentsia's scorn were the village Koran schools, or madarasa, and once again their preoccupations, inspired by modernist Islam, converged with those of British educators. The latter regarded the madarasa as obstacles to the spread of secular education in the countryside. The town ulamaa, meanwhile, scorned the madarasa's pedagogical practices, particularly teaching children to recite Koranic verses in Arabic, "like parrots," without understanding their meaning. In the mid-1920s, a commission consisting of four eminent qadis and the director of education recommended that both problems be tackled by integrating religion into the curriculum of the government schools, where selected verses would be taught in Swahili. But villagers clung to the madarasa, resenting any attempt to limit their children's ability to recite in the language of the Prophet.¹⁰³ These tensions were still simmering ten years later, when they were captured in a widely circulated report by Mohammed Abeid el-Haj, a local official and former schoolteacher who would soon be among the most influential intellectuals in the colonial administration. El-Haj described the madarasa teachers as ignorant and avaricious, routinely mistreating their students and exploiting them for personal gain. But, he asserted, such abuses were characteristic of "African teachers" only. Arab teachers, in contrast, were expert and conscientious.¹⁰⁴

It is ironic that Islam, with its well-deserved reputation for condemning all ethnic distinctions, should have provided some of the language with which the intelligentsia constructed their chauvinist rhetoric; it is also a powerful index of the breadth of sources on which racial thought can draw. Yet the irony is not unheard of: like all universalizing creeds (including Enlightenment ideals of civilization and progress), Islamic ideologies have often been used to express difference.¹⁰⁵ In Zanzibar, locally inherited concepts of Arab hegemony were compounded by the teachings of Cairo-based modernists who made an exception for Arab ethnic solidarity, arguing that it alone was sanctioned and even encouraged by Islam as necessary for the well-being of the faith.¹⁰⁶ In the societies where they originated, such teachings gave religious support to calls for national unity. But in a place like Zanzibar, they were divisive. They told islanders that, in order to be moral beings according to the religious ideals they themselves held dear, they had to accept the leadership of an ethnically distinct Arab elite.

Because the intelligentsia's interwar teachings on Arab superiority also stressed the lesson of coastal exceptionalism, their impact on indigenous islanders was ambivalent. But the message conveyed to African mainlanders was straightforward. Mainlanders were described unreflectively as washenzi and hence automatically excluded from any community built on the values of ustaarabu. The Arab Association was particularly plainspoken. Its weekly paper Al-Falaq, with which the Mazungumzo circle had close connections, frequently stressed the inherent differences between islanders and mainlanders: whereas islanders had become "sufficiently Arabianized" to have lost most of the qualities of African barbarism, mainlanders are but "primitive natives ... whose culture and history are in process of formation only now."¹⁰⁷ These themes became pronounced during the association's first attempt at mass political mobilization in the late 1930s, when, locked in a bitter conflict with the Indian merchants who purchased most of the clove crop, its activists sought to enlist islanders' support for boycotts of all Indian businesses. They did so with a deeply chauvinist rhetoric that made purge categories of both Indians and mainland Africans.¹⁰⁸ Even though Indians were the activists' primary targets at the time, the dehumanizing rhetoric about mainlanders was to have a more potent afterlife in the racial divisions that racked the islands during the Time of Politics.

Hovering around all the paternalist rhetoric about how islanders had been civilized by Arab tutelage was the specter of slavery, a specter that would haunt some of the worst violence of the later years. In accordance with local sensibilities, slavery was rarely mentioned openly before the 1950s. Yet, in their historical essays, the intelligentsia sometimes referred delicately to the "help" slaves had given Arab settlers in building up the wealth of Zanzibar's "Great Arabic Empire,"¹⁰⁹ or indulged in the kind of apologies for "Arab slavery" that would later become standard fare in ZNP journalism. (In a prominent 1937 polemic, the young man who would later found the ZNP argued that "such was the happy state of slaves, that they loathed freedom."¹¹⁰) In 1938, Al-Falaq went even further, obliquely suggesting that slavery was the very institution that had done the most to civilize African "natives," imposing the control necessary to channel their anarchic energy into productive labor.¹¹¹

These ideas reflect with particular clarity how multiple intellectual strands became densely intertwined in Zanzibari discourses of difference. Bringing barbarians into the light of religion and civilization had once been the key justification of enslavement in Islamic doctrine. Al-Falaq's rhetoric shows that doctrine lingering in post-emancipation Zanzibar, complemented by European concepts of labor-discipline.¹¹² It also shows the influence of pan-Arabist journalism, in which the contrast between a supposedly benign "Arab slavery" and the cruelties of its Western counterpart was a key point in the defense of national honor.¹¹³ Such apologies grew out of two separate historical misunderstandings. First is the assumption that, despite its long and varied history, an institution like slavery could be characterized and labeled in fixed terms: even though slavery in the Arab Middle East (and Islamic Africa) was indeed often "benign" relative to New World forms, the East African experience shows that such relaxed relations of bondage could become transformed over time.¹¹⁴ So the label "Arab slavery" in fact corresponded to no single set of practices. Yet such labeling was central to nationalist understandings of history, in which institutions were said to reflect discrete national spirits. In this case, the history of Zanzibar slavery was said to demonstrate the humane paternalism of Arab civilization.

Second, whatever slavery looked like in Zanzibar at any given moment, there was nothing particularly "Arab" about it: contrary to conventional assumptions, many masters were Africans. This misunderstanding stemmed not from imported nationalist philosophies but from local usages, in which the claim of Arab status connoted descent from the planter elite and the absence of slave ancestry. Memories of slavery, in other words, were central to local understandings of Arab identity, and that placed the intelligentsia in the uncomfortable position of having to apologize for it. Given the bitterness with which ASP nationalists would later accuse them of the inherited sin of slavery, it is ironic that the intelligentsia chose to use slavery's history as a narrative tool in their construction of an Arabcentric national identity. But the choice was virtually forced on them by past practices in the construction of ethnic difference.

The acute racialization of political discourse was an aspect of the nationalist mobilization that only began in earnest during and after World War II. The intelligentsia were the pioneers of that mobilization, and their language of civilization, nation, and modernity had a more direct impact on it than did more esoteric notions of raciology. Yet this impact was profoundly contradictory. Most ambivalent were the responses elicited from the village non-elites who were widely regarded as indigenous to the islands. Arab Association activists invited indigenous islanders to participate in the nation-building project, telling them that they were destined to join ranks with Arabs because of their shared history of ustaarabu and Islam. But the invitation was unmistakably condescending, for, although the intelligentsia encouraged islanders to consider themselves superior to mainland barbarians, they also implied that Arabs alone were in full command of the civilizing arts. That condescension can be read obliquely in the tone of the few Swahili items that Al-Falaq began to carry in 1946. (Before then, the paper had published only in English and Arabic.)¹¹⁵ It can be read more directly in the paper's assumption that Arab leadership was the norm in all affairs, civil and political, and that, unless Arabs are in charge, "any institution established ... by Natives" was doomed to fail.¹¹⁶

Given such rhetoric, it is easy to understand why the hegemony of Arabcentric notions of ustaarabu did not translate into simple consent for Arab leadership but, on the contrary, sometimes encouraged racial resentments against Arabs. This was dramatically revealed by a wartime surge of Shirazi ethnic nationalism in southern and eastern Zanzibar Island, a development that, as the outcome of several converging intellectual trends, provides an excellent illustration of the diverse sources of racial thought.

Before the war, villagers in these agriculturally marginal areas had commonly called themselves "Hadimu," an ethnonym that reflected historical memories of vassalage to the Omani sultans. It was derived from an Arabic word for "servant" or "slave."¹¹⁷ During the nineteenth century, Hadimu cultivators were pushed out of the fertile central and western portions of the island by Arab settlers who established large estates of clove and coconut trees. By World War II, their villages had become reserves of seasonal plantation labor, in good years extruding the majority of their male population to assist in the clove harvest.¹¹⁸ Not surprisingly, the intelligentsia's historical narratives skirted the processes by which the expansion of the Arab-dominated plantation sector had relegated the Hadimu to the island's rockiest fringes. The most common ploy was simply to deny that tensions had ever existed and to explain the Hadimu's present-day poverty in terms of their innate fecklessness, typical of "natives."¹¹⁹

Within the Hadimu fringe itself, mounting generational conflicts in the 1930s and 1940s had given rise to acute anxieties about the cohesion of village institutions and to dreams of community renewal. Hadimu elders were convinced that the seasonal outmigration of younger villagers had caused a loss of respect for village traditions. In fact, much of the rancor arose when the migrants used their wages in efforts to intrude in village customs more forcefully than their elders deemed proper, paying bridewealth or sponsoring other festive rites by which authority and prestige were accrued. These rites were the very ones that the intelligentsia had condemned as feckless and uncivilized (Mazungumzo's harshest critiques were directed at Hadimu dances), yet for villagers they were central to how the local community was imagined. Late in the war, when new labor regulations backed by the Arab Association threatened migrants' ability to invest in these rites, they seemed to many villagers part of the elite's overall assaults on indigenous culture. So, when the young men who constituted the bulk of the Hadimu migrants launched a boycott of clove labor in 1944, they won broad support from the elders with whom they had only recently been at loggerheads. This newfound solidarity surprised many observers and made the boycott total in the southern district of Makunduchi, where it was concentrated.

The boycott's transformation from a labor action into a movement of communal revival was completed by activists from the Shirazi Assocation (SA). The intellectuals of the SA's Zanzibar Island branch were decidedly different from the sophisticates of the Arab Association and Mazungumzo; their president, Ameri Tajo, was a madarasa teacher from Makunduchi who had little command of English.¹²⁰ More to the point, they did not consider themselves members of the Arab elite, and most of their activities, which focused on extending to indigenous islanders the prerogatives enjoyed by Arabs, were vociferously opposed by the Arab Association.¹²¹ Nevertheless, much of their rhetoric was derived directly from the Arab intelligentsia's own discourse, particularly the historical narratives of Arab conquest. In 1944, as Al-Falaq thundered against the Hadimu clove pickers for their laziness and disloyalty, the SA told them that the cause of their troubles was not the labor reform per se but the entire history of Arab dispossession. They reminded villagers of the servile origins of the ethnonym "Hadimu" and, cleverly making use of some of the Arab Association's own propaganda, argued that the reforms were but the latest move in the Arabs' efforts to enslave them. In that spirit, they urged villagers to reject "Hadimu" identity and adopt instead that of "Shirazi." Their success was remarkable. Virtually overnight, villagers throughout the region embraced the SA's language, saying to the Arab elites, in effect: We are not slaves or servants, not "Hadimu," but civilized people, "Shirazi," with an inheritance of ustaarabu as deep as your own.¹²²

The discourse of Shirazi ethnic revival, then, arose from an amalgam of sources, some strictly local, such as the dreams of community revival, others borrowed from the subaltern intellectuals of the SA, whose rhetoric in turn was derived from their opponents among the intelligentsia. Yet, despite this diversity of origins, and despite the anti-Arab resentments expressed by SA supporters in the Hadimu fringe, Shirazi identity was founded on the same distinction between civilization and barbarism that had been elaborated by the secular intelligentsia. And that meant that the category most commonly excluded from the SA's vision of the nation were not Arabs but mainlanders. This was most apparent in Pemba, the birthplace and stronghold of the SA. Because ownership of clove estates was more evenly distributed there than in Zanzibar Island, Pemba was not plagued by the economic tensions that plagued the Hadimu fringe. Hence in Pemba, enmity toward Arabs never overtook the hostility to mainlanders that had prompted creation of the Shirazi Association in the first place.¹²³

By the end of the war, then, the intelligentsia had set many of the basic terms of political discourse, including discourse used to express resentment of elite Arabs themselves. Even the subaltern intellectuals of the African Association, who claimed to speak for mainlanders, accepted the intelligentsia's teachings that the nation must be built on the values of civilization and modernity, that those values had been introduced to Africa from abroad, and that upcountry Africans had received them late, from Europeans rather than Arabs.¹²⁴ Historical narratives written by African Association journalists differed from the intelligentsia's mainly in their abolitionist perspective—their emphasis that the British had brought Africans not only enlightenment but also redemption from Arab oppression. But even this was not simply a matter of parroting British teachings. Colonial historiography, in fact, more closely resembled the intelligentsia's whitewashing of "Islamic slavery" and glorification of the civilizing effects of Arab rule. In contrast, the racial nationalists of the African Association told the story of Arab rule almost purely as one of conquest and enslavement, which tied mainlanders together in a history of shared victimhood.¹²⁵

African Association propagandists thus seized on elite historical narratives and made their racial undertones explicit, complementing them with notions borrowed in part from the rhetoric of pan-Africanism. (In a sad irony, though, instead of following the pan-Africanists' quest for the African roots of "civilization," they accepted a reactionary Eurocentric vision.)¹²⁶ This move enabled them to craft more straightforward appeals to anti-Arab sentiments than the SA's ambivalent rhetoric of ustaarabu. With its emphasis on skin color and "blood purity," their propaganda advocated a hardening of the flexible boundaries between ethnic categories, arguing that God himself had decreed that the races and nations be kept separate. An interpretation inflected by the nationalist paradigm might emphasize that this move revealed the influence of Western racial philosophies. That was no doubt true for pan-Africanism itself. But few of the African Association propagandists were well read, and most derived their pan-Africanism second or third-hand. And whatever the ultimate derivation of pan-Africanist ideas, local intellectuals were the ones who did the job of elaborating them in a local idiom and applying them to the circumstances of local politics.

The African Association's rhetoric of racial solidarity appealed most directly to mainlanders, but, with the introduction of electoral politics, its activists also used it in efforts to win support from indigenous islanders. The contest between the two rival visions of the nation—one based on ustaarabu, the other on race—was most hotly fought in the Hadimu fringe, owing to the intense ambivalence there about Shirazi identity. In 1957, Ameri Tajo and other Makunduchi activists led a portion of the Shirazi Association to unite with the African Association, thus forming the ASP. Tajo embraced his new partners' stress on racial unity, and together they excoriated Arabs as oppressive aliens.¹²⁷ The partnership proved brittle, however, and two years later Tajo and his associates left the ASP to form the anti-mainlander Zanzibar and Pemba People's Party (ZPPP), which allied itself to the ZNP. In marked contrast to his former party, Tajo's new one vilified mainlanders as barbarians, and at rallies he shared the stage with members of the intelligentsia who used history lessons to persuade listeners that by origin and essence Shirazi were the same as Arabs.¹²⁸ In the 1961 and 1963 elections, the ZPPP and ASP battled fiercely and violently for Makunduchi's votes.¹²⁹

As the pace of political mobilization accelerated after 1957, members of the intelligentsia who in their youth had written high-minded essays on civilization and uplift took to blaming mainlanders for a wide array of ills. Such rhetoric appeared even in Mwongozi, postwar Zanzibar's most sophisticated political weekly. Published by two former Hollingsworth students and closely affiliated with the ZNP, Mwongozi had been a voice of Islamic modernism, ecumenical tolerance, and anti-racialism. Yet its essays on history and culture were imbued with all the assumptions of coastal exceptionalism, increasingly expressed as contempt for mainlander savagery. So, for example, adapting to its purposes a deeply rooted popular discourse about criminality and difference,¹³⁰ Mwongozi blamed crime on immigration policies that admitted too many mainlanders (and ASP supporters), "foreigners who have been allowed to pour into our islands as if this were a rubbish-heap for every type of filth," and castigated its political rivals for wanting to "leave the door open to ... thieves, cannibals, and naked people." By the eve of independence, the paper was regularly vilifying mainlanders as enemies of Islam and warning that under a ZNP government their "beastialities" (sic) would not be tolerated. ¹³¹

Thus both sides engaged in the spiral of reciprocal dehumanization that culminated in bloodshed.¹³² Of the two, only the subaltern intellectuals of the ASP espoused a politics that was "racial" in the conventional sense, focusing on the divinely ordained solidarities of common origin manifested by bodily markers; ZNP propagandists, in contrast, remained more preoccupied with the intellectual traditions that had shaped their youthful essays for Mazungumzo, stressing the lessons of Islamic universalism and the distinctions between civilization and barbarism. But in terms of their dehumanizing impact, it is difficult to discern between these two rhetorics. And, no matter who was most to blame for the rising intensity of racial politics, it is clear that the discourses of civilization and race informed and fed off one another.

Tracing the precise etiology of racial thought therefore suggests the pitfalls of posing a categorical distinction between race and other forms of ethnic thought. Ultimately, the question is semantic: at what point do we want to use the word "race" to describe a given current of ethnic thought? Few would object to using the word to describe the rhetoric of the postwar African Association; such a description might in fact be useful for historical analysis, since the association's propaganda conveyed more straightforward messages of expulsion and extermination than were initially apparent in the intelligentsia's ambiguous rhetoric of civilization.¹³³ But to draw an absolute distinction between these two discourses would blind us to the borrowings between them and the ways in which they grew from one another. The subaltern ideologues operated within the same general intellectual milieu as the intelligentsia; indeed, the intelligentsia taught them many of the fundamental lessons of exclusionary ethnic nationalism. The debates and interconnections between these very different groups of political thinkers were central to the intellectual work involved in the creation of racial thought. Such work was not merely ancillary to the rise of racial identities but was constitutive of them, and to ignore it is to accept that racial identities reflect inherent differences—it is to accept, that is, that "racial groups" existed prior to their mobilization by ideologues, who merely forged political allegiances from the facts of their existence.

In focusing on the elite intelligentsia who came of age between the wars, I have told only a small part of the story, slighting the ASP demagoguery that contributed most directly to the bloodshed of the early 1960s and omitting entirely the processes, far from inevitable, by which racial thought was transformed into popular conceptions capable of motivating transgressive violence.¹³⁴ But because the intelligentsia exerted such an overwhelming influence on Zanzibari political culture, this focus is an essential starting point. It is pertinent in other ways as well. Of all Zanzibaris, the intelligentsia experienced the most direct and sustained influence of colonial education. Reconstructing the development of their thought thus offers a pointed demonstration of the limits of colonial indoctrination. It also demonstrates how even the most liberal and mainstream of nationalist rhetorics might become racialized, not by the sinister workings of colonial mentors (throughout the postwar years, in fact, British officials strove to damp down ethnic conflicts, not fan them), but by political actors responding to political opportunities. Neither the convergence of African and European idioms of ranked difference nor the interplay between them, as in the rich intellectual cross-fertilization that took place between Teachers Training School graduates and their British mentors, is sufficient to justify the assumption that African concepts of racial politics originated solely in the European imagination.¹³⁵

This emphasis on the limits of European indoctrination should not be taken to imply that Zanzibar's racial clashes were manifestations of some ancient, inbred enmity. That, in effect, is the false choice proposed in much of the instrumentalist literature on African ethnicity: since it is plainly ahistorical to view tribalism as primordial, one must therefore see it as a consequence of colonialism.¹³⁶ Yet, even though the links between colonial rule and contemporary ethnic politics are unmistakable, they alone are not sufficient for explaining the often profound resonance of ethnic demagoguery, especially its ability to evoke ethnic violence. For that, one needs a historical perspective that is both deeper and broader than a simple focus on the colonial state. The racialization of Zanzibari ethnic thought was indeed a modern process, accomplished by modern thinkers, Zanzibari as well as British. But the modes of thought subjected to the process were neither invented from whole cloth nor imported anew; many had been inherited from precolonial intellectual traditions. At the same time (and contrary to primordialist assumptions), those traditions had always been given to innovation, adaptation, and change. So, for example, although distinctions between Shirazi and other islanders were rooted in precolonial thought, their meanings had changed dramatically in the nineteenth century, with the rise of the plantation sector and Omani rule, and again in the twentieth, with the imposition of the colonial state and the rise of nationalist politics. At the same time, memories of historical events placed limits on innovators' ability to invent new political traditions: this appears with particular clarity in the dilemma faced by elite intellectuals who found themselves having to apologize for "Arab slavery."

As in other parts of the colonial world, racial thought in Zanzibar was derived from a multiplicity of sources, foreign and domestic, and the innovators who rethought and combined them came from many walks of life. Such diversity implies that we cannot universalize any one path toward racial thought. Yet many studies of the non-Western world still reflect a tendency to universalize the experience of the West, and, in particular, the United States.¹³⁷ Hence the continued assumption that ethnicity and race had separate origins, the latter arising from imported European doctrines. Although it is undeniable that the divide between "race" and "ethnicity" is fundamental to contemporary American political discourse, that discourse is the product of a long and complex history, which we cannot assume to have been replicated anywhere in identical terms. Abandoning such self-centered assumptions is a necessary first step toward a precise understanding of how racial thought developed elsewhere in the world. And that in turn may help us relinquish the last vestiges of the racial thought in which we ourselves have been schooled: the assumption that racial boundaries have any basis in phenomena that exist apart from the subjective perceptions shaped by our several histories.

Research and writing were made possible by a Fulbright grant and fellowships from the Social Science Research Council and American Council of Learned Societies, the Institute for Advanced Study, the John Simon Guggenheim Memorial Foundation, and Northwestern University. Thanks also to Nicola Beisel, Gary Burgess, Frederick Cooper, Steven Hahn, James Kern, Pier Larson, John Lonsdale, Nancy Maclean, Dylan Penningroth, David Schoenbrun, Thomas Spear, and participants in the Johns Hopkins University History Seminar for salutary criticisms of earlier drafts.

Jonathon Glassman, an associate professor of history at Northwestern University, studied with Steven Feierman and Jan Vansina at the University of Wisconsin. This article is drawn from a book-length project on the racialization of political thought in colonial Zanzibar, which will include an examination of how racial identities were reproduced by the acts of violence that ended the sultanate in the mid-twentieth century. In many ways, the project builds on Glassman's first book, Feasts and Riot: Revelry, Rebellion and Popular Consciousness on the Swahili Coast, 1856–1888, which traces how debates over the structures of power and community in a provincial region of the Zanzibar Sultanate yielded discourses of ethnic and racial identity in the pre-colonial century. The book won the 1996 Herskovits Prize of the African Studies Association, awarded annually for the best book on Africa in any discipline.

Notes

¹� A. W. Norris, "Pride of Race," Normal Magazine 4, no. 4 (April 1930): 30–33. The magazine was later renamed Mazungumzo ya Walimu, hereafter Maz.

²�Maz., n.s. 1, no. 3 (May 1957): 14.

³� Zanzibar consists of two main islands, Zanzibar (or Unguja) and Pemba. From 1890 to December 1963, it was a British protectorate, its official head of state chosen from the Omani Arab dynasty that had established the sultanate early in the nineteenth century. Shortly after independence, the sultan was overthrown by forces loyal to the Afro-Shirazi Party. In April 1964, Zanzibar and the mainland nation of Tanganyika united to form the United Republic of Tanzania.

⁴� To describe the revolution in this way is to take a stand on an issue that has aroused much contention. The extent of racial violence in 1964 is undeniable, yet several authors have argued that "so-called ethnic divisions" merely masked the fundamental fact that the coup was "not an ethnic, but a social revolution." This distinction between "ethnic" and "social" phenomena is, of course, false. I fully endorse these authors' contention that racial identities are mere "images people have of themselves and others." But the events of 1957–1964 are eloquent testimony to the impact such images can have on people's thoughts and actions, and it is imperative that the historian try to account for how they were constructed. Quotes are from L. Rey, "The Revolution in Zanzibar," in Lionel Cliffe and John Saul, eds., Socialism in Tanzania (Dar es Salaam, 1972), 1: 30; and (for "images") Abdul Sheriff, "A Materialist Approach to Zanzibar's History," in A. Sheriff and Ed Ferguson, eds., Zanzibar under Colonial Rule (London, 1991), 7.

⁵� For a forceful statement, see Michael Lofchie, "Zanzibar," in James Coleman and Carl Rosberg, eds., Political Parties and National Integration in Tropical Africa (Berkeley, Calif., 1964), 506. The quoted phrase is from Clifford Geertz, "The Integrative Revolution: Primordial Sentiments and Civil Politics in the New States," in Geertz, ed., Old Societies and New States (New York, 1963). Prominent examples of Western perceptions of the Zanzibar violence include V. S. Naipaul's 1979 novel, A Bend in the River, and Gualtiero Jacopetti's 1966 film, Africa Addio.

⁶� I will consider the African literature below. The classic argument for South Asia is Gyanendra Pandey, The Construction of Communalism in Colonial North India (Delhi, 1990); for broad critiques, see Sudhir Kakar, The Colors of Violence: Cultural Identities, Religion, and Conflict (Chicago, 1996), 12–24; C. A. Bayly, Origins of Nationality in South Asia: Patriotism and Ethical Government in the Making of Modern India (Delhi, 1998); Rosalind O'Hanlon, "Historical Approaches to Communalism: Perspectives from Western India," in Peter Robb, ed., Society and Ideology: Essays in South Asian History (Delhi, 1993), 247–66.

⁷� See the suggestive comments in Michael Chege, "Africa's Murderous Professors," The National Interest 46 (Winter 1996); also Gyanendra Pandey, "In Defense of the Fragment: Writing about Hindu-Muslim Riots in India Today," Representations 37 (1992): 27–55.

⁸� Allowing for the oversimplification necessary in summing up a complex and nuanced work, this stands as a fair representation of the argument in Michael Lofchie, Zanzibar: Background to Revolution (Princeton, N.J., 1965). Anthony Clayton criticizes Lofchie for underestimating the divisiveness of ZNP rhetoric; his own account of the Time of Politics serves as a useful corrective: The Zanzibar Revolution and Its Aftermath (London, 1981), 37–49. Lofchie's study remains the standard account of the islands' political history, although subsequent authors have emphasized more the colonial derivation of Zanzibari racial thought: for example, Alamin Mazrui and Ibrahim Noor Shariff, The Swahili: Idiom and Identity of an African People (Trenton, N.J., 1994); Laura Fair, Pastimes and Politics: Culture, Community and Identity in Post-Abolition Urban Zanzibar (Athens, Ohio, 2001). For a fuller discussion of the ASP's racial rhetoric, see Jonathon Glassman, "Sorting Out the Tribes: The Creation of Racial Identities in Colonial Zanzibar's Newspaper Wars," Journal of African History 41 (2000): 395–428.

⁹� Ann Laura Stoler, "Racial Histories and Their Regimes of Truth," Political Power and Social Theory 11 (1997): 183–206, quotes from 185.

¹⁰� Loïc Wacquant, "For an Analytic of Racial Domination," Political Power and Social Theory 11 (1997): 221–34, quote from 222.

¹¹� René Lemarchand, "Revolutionary Phenomena in Stratified Societies: Rwanda and Zanzibar," Civilisations 5, no. 1 (1968); Catharine Newbury, "Colonialism, Ethnicity and Rural Political Protest: Rwanda and Zanzibar in Comparative Perspective," Comparative Politics 15 (1983): 253–80.

¹²� The concept is often associated with the sociologist Michael Banton, The Idea of Race (London, 1977); Jean-Pierre Chrétien makes use of it in his studies of Rwandan intellectual history, some of which are cited below.

¹³� Villia Jefremovas observes that most authors neglect the role of elite indigenous intellectuals: "Treacherous Waters: The Politics of History and the Politics of Genocide in Rwanda and Burundi," Africa 70, no. 2 (2000): 298–308.

¹⁴� Mahmood Mamdani, When Victims Become Killers: Colonialism, Nativism and the Genocide in Rwanda (Princeton, N.J., 2001). This aspect of Mamdani's interpretation contradicts his avowed determination to go beyond analyses that focus on state agency. (It also contradicts his condemnation of "the search for origins" as "the original and persistent sin of Western history writing" about Africa [50].) Yet some of the literature that Mamdani claims to transcend demonstrates the important role played by Rwandan intellectuals in crafting racial myths. Examples include Jan Vansina, L'évolution du royaume Rwanda des origines à 1900 (1962; rpt. edn., Brussels, 2000); Claudine Vidal, Sociologie des passions: Rwanda, Côte d'Ivoire (Paris, 1991), 45–61; Jean-Pierre Chrétien, "Hutu et Tutsi au Rwanda et au Burundi," in Jean-Loup Amselle and Elikia M'Bokolo, eds., Au coeur de l'ethnie: Ethnies, tribalisme et état en Afrique (Paris, 1985), esp. 146–50; René Lemarchand, Ethnicity as Myth: The View from Central Africa (Copenhagen, 1999). This literature renders implausible any suggestion that such intellectuals merely parroted their European teachers. Although racialization was indeed largely a product of the colonial era, the historical record prompts a leading authority to "reject decisively the judgment of those who attribute the distinction between Tutsi and Hutu, and their mutual hostility, to the ideas and actions of colonial masters." Jan Vansina, Le Rwanda ancien: Le royaume nyiginya (Paris, 2001), 177.

¹⁵� Robert Miles, Racism (London, 1989). For a critique of the continued tendency to inflate race into a sociological category, see Mara Loveman, "Is 'Race' Essential?" American Sociological Review 64, no. 6 (1999): 861–98. Understanding race as a mode of thought has become all the more prevalent as scientists have demolished lingering notions that racial boundaries have any biological significance, yet it long predates those advances. It informed Hannah Arendt's influential concept of "race-thinking" (The Origins of Totalitarianism [new edn., New York, 1973; orig. 1951]), and can be found in Max Weber, who did not consider race or ethnicity to be sociological categories. Weber, Economy and Society, Guenther Roth and Claus Wittich, eds. (New York, 1968), 1: 385–98; Ernst Moritz Manasse, "Max Weber on Race," Social Research 14 (1947): 191–221.

¹⁶� For a consensus that race be regarded as "doctrine," see Michael Banton, "The Concept of Racism," in Sami Zubaida, ed., Race and Racialism (London, 1970), 17–34. Miles, Racism, prefers the word "ideology." Arendt's idea of "race-thinking" is more flexible, but she, too, treated it as specific to Western thought.

¹⁷� Nevertheless, folk racisms are usually regarded as distinct from the real thing, some authors simply refusing to consider them manifestations of racial thought, others assuming that popular notions arose as pale reflections of ideas whose origins lay in more erudite circles. Examples of the first approach include Banton, "Concept of Racism"; of the second, Eric Wolf, "Perilous Ideas: Race, Culture, People," Current Anthropology 35, no. 1 (1994); and K. Anthony Appiah, "Race, Culture, Identity: Misunderstood Connections," in Appiah and Amy Gutmann, Color Conscious: The Political Morality of Race (Princeton, N.J., 1996), 30–105. For an incisive discussion of "the continual barter between folk and analytical notions," see Wacquant, "For an Analytic of Racial Domination."

¹⁸� Etienne Balibar, "Is There a 'Neo-Racism'?" in Balibar and Immanuel Wallerstein, Race, Nation, Class: Ambiguous Identities, Chris Turner, trans. (London, 1991), 21. This literature first arose in response to the arguments of Banton and other British sociologists that the anti-immigrant rhetoric of Tory politicians in the postcolonial UK could not properly be deemed "racism." Banton, Idea of Race, chap. 9; Martin Barker, The New Racism: Conservatives and the Ideology of the Tribe (London, 1981).

¹⁹� Stoler, "Racial Histories."

²⁰� Raciology was more about explaining difference than about ranking, and some raciologists considered themselves anti-racists, that is, politically opposed to ranking. For examples, E. W. Count, ed., This Is Race (New York, 1950); Benoit Massin, "From Virchow to Fischer: Physical Anthropology and 'Modern Race Theories' in Wilhelmine Germany," in George W. Stocking, ed., Volksgeist as Method and Ethic: Essays on Boasian Ethnography and the German Anthropological Tradition (Madison, Wis., 1996), 79–154.

²¹� Balibar, "Is There a 'Neo-Racism'?" 23. The term "cultural monads" is from Wolf, "Perilous Ideas."

²²� Classic accounts include Philip D. Curtin, The Image of Africa: British Ideas and Action, 1780–1850 (Madison, Wis., 1964); George W. Stocking, Race, Culture and Evolution: Essays in the History of Anthropology (Chicago, 1982).

²³� Friedrich Nietzsche, Toward a Genealogy of Morals, section 13.

²⁴� Stoler, "Racial Histories." An example is Ivan Hannaford, Race: The History of an Idea in the West (Baltimore, 1996). The first half of Hannaford's book is an erudite exercise in correcting for the opposite error, that is, the assumption through backward induction that an act of apparent racism from the pre-modern past was the product of raciological doctrines.

²⁵� John Rex, "The Concept of Race in Sociological Theory," in Zubaida, Race and Racialism, 39. For similar arguments, see Immanuel Wallerstein, "The Construction of Peoplehood," in Balibar and Wallerstein, Race, Nation, Class, 71–85; Roger Sanjek, "The Enduring Inequalities of Race," in Steven Gregory and R. Sanjek, eds., Race (New Brunswick, N.J., 1994), 1–17; Eduardo Bonilla-Silva, "The Essential Social Fact of Race," American Sociological Review 64, no. 6 (1999): 899–906; Howard Winant, The World Is a Ghetto: Race and Democracy since World War II (New York, 2001).

²⁶� Banton, Idea of Race.

²⁷� Such is the view of Banton, Idea of Race, despite his skepticism that Western expansion caused the emergence of racial thought.

²⁸� These distinctions are developed further by Anthony Appiah, from whom I have derived some of my language: In My Father's House: Africa in the Philosophy of Culture (Oxford, 1992); and "Racisms," in David Theo Goldberg, ed., Anatomy of Racism (Minneapolis, 1990), 3–17. (Appiah writes of "racialism" rather than "racial thought.")

²⁹� These comments do not refer to the French Enlightenment concept of civilisation, which, strictly speaking, included the embrace of universal (that is, Western) reason. But the eighteenth-century French neologism was coined to express more general concepts that long predated (and outlived) it, few of which were peculiar to the modern West. Lucien Febvre, "Civilisation: Evolution of a Word and a Group of Ideas," in A New Kind of History, Peter Burke, ed. (London, 1973), 219–57.

³⁰� China is the best-known example, for which see Frank Dikötter, "Group Definition and the Idea of Race in Modern China," Ethnic and Racial Studies 13, no. 3 (1990). Many of the concepts that were to become racialized in colonial Rwanda had their origins in earlier distinctions of boorishness and urbanity: Vansina, Rwanda ancien, 172–77; David Newbury, "'Bunyabungo': The Western Rwandan Frontier, c. 1750–1850," in Igor Kopytoff, ed., The African Frontier (Bloomington, Ind., 1987), 164–92. Such distinctions were of ancient standing in Bantu-speaking central Africa: for example, John Thornton, "Mbanza Kongo/Sao Salvador: Kongo's Holy City," in David M. Anderson and Richard Rathbone, eds., Africa's Urban Past (Portsmouth, N.H., 2000), 67; Jan Vansina, The Children of Woot: A History of the Kuba Peoples (Madison, Wis., 1978). Examples from southern Africa include the contrast between cannibalism and order, and the harsh attitudes harbored toward Khoi and Basarwa hunters and gatherers by their neighbors: Carolyn Hamilton, Terrific Majesty: The Powers of Shaka Zulu and the Limits of Historical Invention (Cambridge, Mass., 1998), 212–13; Robert J. Gordon, The Bushman Myth: The Making of a Namibian Underclass (Boulder, Colo., 1992), 31–32; John Comaroff, "Of Totemism and Ethnicity," in John and Jean Comaroff, Ethnography and the Historical Imagination (Boulder, Colo., 1992), 49–67. Even those whom state dwellers scorned as stateless "barbarians" could harbor their own values of civilization, which might become hegemonic over their immediate neighbors; this irony is illustrated by the pastoralist Maa-speaking "sections" that came to dominate East Africa's Central Rift Valley in the eighteenth and nineteenth centuries: Thomas Spear and Richard Waller, eds., Being Maasai: Ethnicity and Identity in East Africa (London, 1993). A striking parallel to the Zanzibar case involves the so-called originaires of French West Africa, whose disdain for certain of their fellow Africans was rooted in concepts of civilization and boorishness that played both on precolonial, Islamic standards and those adopted from French rulers: Rebecca Shereikis, "From Law to Custom: The Shifting Legal Status of Muslim originaires in Kayes and Medine," Journal of African History 42 (2001): 261–83.

³¹� Igor Kopytoff, "The Internal African Frontier: The Making of African Political Culture," in Kopytoff, African Frontier, esp. 49–50, 56–57. Kopytoff notes that discourses of civilization and barbarism shaped the competing claims of peoples who met at the continent's many internal frontiers.

³²� A narrower understanding of "race," as a form of thought that categorizes humanity according to objectively existing somatic traits such as skin color, is historically and conceptually untenable: John Szwed, "Race and the Embodiment of Culture," Ethnicity 2 (1975): 19–33; Appiah, "Race, Culture, Identity"; Donald L. Horowitz, Ethnic Groups in Conflict (Berkeley, Calif., 1985). Again, these views were anticipated by Weber, who stressed the degree to which perceptions of group physical differences are culturally determined and therefore as subjective as other perceptions of common ethnic descent. Economy and Society, 2: 392.

³³� Miles, Racism, 42–44; Saul Dubow, "Ethnic Euphemisms and Racial Echoes," Journal of Southern African Studies 20 (1994): 355–70; Elazar Barkan, The Retreat of Scientific Racism: Changing Concepts of Race in Britain and the United States between the World Wars (Cambridge, 1992), esp. 296–302. As Barkan observes, the trend began between the wars. Similar trends can be found in the history of the French word ethnie, which, for example, was first applied to Rwanda only in the early 1960s. Dominique Franche, "Généalogie du génocide rwandais: Hutu et Tutsi; Gaulois et Francs?" Les temps modernes 582 (1995): 10; Claudine Vidal, "Le génocide des Rwandais tutsi et l'usage public de l'histoire," Cahiers d'études africaines 38, no. 150–52 (1998): 660.

³⁴� Nowadays, a belief in such "cultural monads" is typically associated with functionalist anthropology and is often criticized as such. But it also formed a central component of classic racial thought and was among the targets of Franz Boas's famous critiques of the latter. Stocking, Race, Culture; Stocking, Volksgeist; Wolf, "Perilous Ideas."

³⁵� An explicit emphasis on the blood relationship is not incompatible with an awareness that the relationship is a metaphor for something else. Sati al-Husri, a nationalist intellectual active in Baghdad and Cairo and a major influence throughout the Arab world, including among the Zanzibari intellectuals discussed below, recognized that common descent was merely a metaphor, yet he insisted that it was a useful metaphor for forging national unity. Sati al-Husri, "The Historical Factor in the Formation of Nationalism," in Kemal Karpat, ed., Political and Social Thought in the Contemporary Middle East, rev. edn. (New York, 1982), 39–43.

³⁶� This abbreviated discussion describes views held by a variety of authors, many of which were anticipated by Weber. The phrase "aura of descent" is from Ronald Cohen, "Ethnicity: Problem and Focus in Anthropology," Annual Review of Anthropology 7 (1978): 379–403. Discussions relevant to the metaphor of descent also include Charles Keyes, "Towards a New Formulation of the Concept of Ethnic Group," Ethnicity 3 (1976): 202–13; Eric Voegelin, "The Growth of the Race Idea," Review of Politics 2 (1940): 283–317; and various sources already cited. For an insightful discussion of the ambiguity of distinctions between race, nation, and ethnicity, see Elizabeth Tonkin, Maryon McDonald, and Malcolm Chapman, History and Ethnicity (London, 1989), 1–21. Religious ethnicities would seem at first glance to be an exception to this argument. Yet in fact the notion of religious identity as an act of choice was "a delayed result of the Reformation and a direct result of the Enlightenment ... Outside the West, religion remained an ascriptive affiliation" (Horowitz, Ethnic Groups in Conflict, 50).

³⁷� For a discussion of the neo-conservative position, see Sanjek, "Enduring Inequalities," 8–9.

³⁸� Horowitz, Ethnic Groups in Conflict, uses this phrase to describe similarities between idioms of kinship and ethnicity.

³⁹� We may define nationalism as any political philosophy based on the assumption that each of the mutually exclusive ethnic groups into which humanity is presumably divided ought "naturally" to control its own state. This gloss is derived chiefly from Weber, "The Nation," in From Max Weber, H. Gerth and C. Wright Mills, eds. (London, 1948). Tonkin et al. observe that "it is no more than a tautology to say that nations have ethnic origins" (History and Ethnicity, 18)—provided, of course, that one remember that ethnic communities are no less "imagined" than nations.

⁴⁰� Liisa H. Malkki, Purity and Exile: Violence, Memory, and National Cosmology among Hutu Refugees in Tanzania (Chicago, 1995); A. M. Alonso, "The Politics of Space, Time and Substance: State Formation, Nationalism and Ethnicity," Annual Review of Anthropology 23 (1994): 379–405; Benedict Anderson, Imagined Communities, rev. edn. (London, 1991). This is not to say that national thought was taken for granted at all levels of colonial society, as has been made clear by historians working on popular politics in Africa and South Asia.

⁴¹� As Eric Hobsbawm has written, the criteria are "fuzzy": Nations and Nationalism since 1789: Programme, Myth, Reality (1990; Cambridge, 1992), 6. For an unusually strong rejection of the search for a consistent definition, see Valery Tishkov, "Forget the 'Nation': Post-Nationalist Understandings of Nationalism," Ethnic and Racial Studies 23 (2000): 625–50.

⁴²� Authors who have recognized this include Etienne Balibar, "Racism and Nationalism," in Balibar and Wallerstein, Race, Nation, Class, 37–67; Zygmunt Bauman, "Soil, Blood and Identity," Sociological Review 40 (1992): 675–701; Carole Nagengast, "Violence, Terror and the Crisis of the State," Annual Review of Anthropology 23 (1994): 109–36; George M. Frederickson, The Comparative Imagination: On the History of Racism, Nationalism, and Social Movements (Berkeley, Calif., 1997), 77–97. For a splendid historical analysis that makes use of this insight, see James R. Brennan, "Nation, Race and Urbanization in Dar es Salaam, Tanzania, 1916–1976" (PhD dissertation, Northwestern University, 2002).

⁴³� See the works on Rwanda already cited. Other examples include Patrick Harries, "Exclusion, Classification and Internal Colonialism: The Emergence of Ethnicity among the Tsonga-Speakers of South Africa," in Leroy Vail, ed., The Creation of Tribalism in Southern Africa (Berkeley, Calif., 1991), 82–117; Fair, Pastimes and Politics; Mazrui and Shariff, The Swahili; Colin Leys, Underdevelopment in Kenya (Berkeley, 1974), 198–206. An influential statement dismissed ethnic thought as "false consciousness": Archie Mafeje, "The Ideology of 'Tribalism,'" Journal of Modern African Studies 9, no. 2 (1971): 253–61. For an "internal critique" of this position, see Terence Ranger, "The Invention of Tradition Revisited: The Case of Colonial Africa," in Ranger and Olufemi Vaughan, eds., Legitimacy and the State in Twentieth Century Africa (London, 1993).

⁴⁴� Leroy Vail, "Introduction: Ethnicity in Southern African History," in Vail, Creation of Tribalism, 1–19. Vail's essay has been taken as something of a manifesto by subsequent scholars who call their approach "constructivist," thus seeking to distinguish themselves both from those who interpret ethnic identities as the residue of some "primordial" essence and from more straightforward "instrumentalists" such as those mentioned in the preceding footnote. Their insistence that ethnic identities are created historically should be axiomatic, and makes sense only in contrast to a caricature of scholars they deem "primordialists," few of whom, in fact, argued that ethnicity was beyond history. When it comes to charting the historical processes that gave rise to ethnic identities, most "constructivists" remain instrumentalist in their stress on the demands of the colonial political economy and their neglect of inherited intellectual content. They also maintain a strict (and undefined) distinction between ethnicity and nation. These points emerge from a recent statement of the "constructivist" consensus: Paris Yeros, ed., Ethnicity and Nationalism in Africa: Constructivist Reflections and Contemporary Politics (London, 1999), and from a broad reading of the literature on African ethnicity.

⁴⁵� Walker Connor, "Beyond Reason: The Nature of the Ethnonational Bond," Ethnic and Racial Studies 16, no. 3 (1993): 373–88.

⁴⁶� Some of my language, and much of my argument, is derived from M. Crawford Young, "Nationalism, Ethnicity, and Class in Africa: A Retrospective," Cahiers d'études africaines 26, no. 103 (1986): 421–95. Young's comments were directed specifically toward literature on nationalist intellectuals, but they are applicable more generally.

⁴⁷� Examples include Steven Feierman, Peasant Intellectuals: Anthropology and History in Tanzania (Madison, Wis., 1990); John Lonsdale, "The Moral Economy of Mau Mau," in Lonsdale and Bruce Berman, Unhappy Valley: Conflict in Kenya and Africa; Book Two: Violence and Ethnicity (London, 1992), 265–504; Paul La Hausse, Restless Identities: Signatures of Nationalism, Zulu Ethnicity, and History in the Lives of Petros Lamula (c. 1881–1948) and Lymon Maling (1889–c. 1936) (Pietermaritzburg, 2000); David Lan, Guns and Rain: Guerrillas and Spirit Mediums in Zimbabwe (Berkeley, Calif., 1985); John Peel, "The Cultural Work of Yoruba Ethnogenesis," in Tonkin, History and Ethnicity, 198–215. See the discussion in Ranger, "Invention of Tradition Revisited."

⁴⁸� The phrase was coined by Hamilton, Terrific Majesty, though her book, too, overprivileges the role of Europeans. For critiques along these lines, see Ronald Atkinson, "The (Re)construction of Ethnicity in Africa: Extending the Chronology, Conceptualization and Discourse," in Yeros, Ethnicity and Nationalism, 15–44; Thomas Spear, "Neo-Traditionalism and the Limits of Invention in British Colonial Africa," Journal of African History 44 (2003): 3–27.

⁴⁹� Despite Young's plea of fifteen years ago, this cliché has been abandoned less quickly in studies of ethnic and nationalist politics than in studies of social history. Nancy Rose Hunt, A Colonial Lexicon of Birth Ritual, Medicalization, and Mobility in the Congo (Durham, N.C., 1999); Frederick Cooper and Ann Laura Stoler, "Between Metropole and Colony: Rethinking a Research Agenda," in Cooper and Stoler, eds., Tensions of Empire: Colonial Cultures in a Bourgeois World (Berkeley, Calif., 1997), 1–56.

⁵⁰� Horowitz (Ethnic Groups, chap. 1) is among those who have discussed the contrast between these two ways of imagining ethnic divisions. He usefully refers to them as "unranked" and "ranked" ethnicities; the phrase "incipient whole societies" is also his. René Lemarchand ("Revolutionary Phenomena") makes a similar distinction, highlighting the presence of ethnic "stratification" to explain the "revolutionary" upheavals in Zanzibar and Rwanda. John Comaroff ("Of Totemism") proposes the terms "totemism" and "ethnicity" to describe the distinction.

⁵¹� See the works already cited by Banton, Rex, Wallerstein, and Sanjek. A critique of such an approach is implicit in my earlier discussion of literature that approaches race as social structure, and that fails to distinguish between racism and racial thought.

⁵²� As observed by Mamdani, When Victims Become Killers, 15, although, as we have seen, he does little better.

⁵³� See Chege, "Africa's Murderous Professors," for suggestive comments.

⁵⁴� The latter factor helps explain why the term "race" has been less commonly applied to Rwanda's ethnic conflicts than to similar conflicts in Zanzibar: Tuutsi and Hutu no longer correspond to Western racial categories (but they once did, Tuutsi being imagined as "Hamites").

⁵⁵� Hence the leading authority on Zanzibari history, in language similar to that used by many authors, explains ethnic divisions by invoking British "preferences" informed by policies of "divide and rule": Abdul Sheriff, "Race and Class in the Politics of Zanzibar," Afrika Spectrum 36, no. 3 (2001): 307–08.

⁵⁶� For a pronounced version of the first interpretation, see Mazrui and Shariff, The Swahili; for the second, B. D. Bowles, "The Struggle for Independence," in Sheriff and Ferguson, Zanzibar under Colonial Rule, esp. 86, 92; B. F. Mrina and W. T. Mattoke, Mapambano ya Ukombozi Zanzibar (Dar es Salaam, 1980).

⁵⁷� Non-Zanzibari examples of similar practices are described in René Lemarchand, Burundi: Ethnic Conflict and Genocide (Cambridge, 1994); and Dikötter, "Group Definition." The official and oppositional Zanzibari histories differ in which groups are identified as British stooges. Official versions include Omar Mapuri, Zanzibar: The 1964 Revolution; Achievements and Prospects (Dar es Salaam, 1996); R. K. Mwanjisi, Abeid Amani Karume (Nairobi, 1967). Opposition versions include Ali Muhsin al-Barwani, Conflicts and Harmony in Zanzibar (Dubai, 1997); Zanzibar Center of Human and Democratic Rights, Zanzibar Dola Taifa na Nchi Huru (Copenhagen, 1994). A curious hybrid version can be found in the writings of Juma Aley, a onetime ZNP government minister who was imprisoned after the revolution and who since his release has published praise of his jailers. Aley repeats the ZNP charge that the British chief secretary personally published the most inflammatory ASP journals, and thereby absolves the ASP itself of any responsibility. Juma Aley, Zanzibar, in the Context (Delhi, 1988), 86–87, 94.

⁵⁸� Appiah, In My Father's House, 7. Appiah's comment stands in marked contrast to much of the literature on colonial education: see the critical review in Sybille Küster, Neither Cultural Imperialism nor Precious Gift of Civilization: African Education in Colonial Zimbabwe (Münster, 1994).

⁵⁹� Glassman, "Sorting Out the Tribes."

⁶⁰� Balibar, "Racism and Nationalism," 60.

⁶¹� It is by now a commonplace that anti-colonial nationalism was derived from European discourse: Elie Kedourie, Nationalism in Asia and Africa (New York, 1970); B. Anderson, Imagined Communities; Basil Davidson, Black Man's Burden (New York, 1992). Of course, this does not mean that nationalist thinkers in the colonial world had "nothing left to imagine": Partha Chatterjee, The Nation and Its Fragments (Princeton, N.J., 1993).

⁶²� The intelligentsia linked the transmission of religious expertise with the inheritance of Arab status: see José Kagabo, "Réseaux d'ulama 'swahili' et liens de parenté," in F. Le Guennec-Coppens and P. Caplan, eds., Les Swahili entre Afrique et Arabie (Paris, 1991), 59–72. For descriptions of nineteenth and early twentieth-century Zanzibari intellectual life, see Abdallah Salih Farsy, The Shaf'i Ulama of East Africa, Randall Pouwels, trans. and ed. (Madison, Wis., 1989); Anne K. Bang, "Sufis and Scholars of the Sea: The Sufi Family Networks of Ahmad ibn Sumayt and the Tariqa 'Alawiyya in East Africa, c. 1860–1925" (Dr.Art. thesis, University of Bergen, 2000); Randall Pouwels, "Sh. Al-Amin b. Ali Mazrui and Islamic Modernism in East Africa, 1875–1947," International Journal of Middle East Studies 13 (1981): 329–45. The elite character of these intellectual circles is implicit throughout the descriptions in Aley, Zanzibar, in the Context, who states that "Shirazi" and "African" circles "were less spectacular" than those dominated by Arabs and Comorians (58). Also see Muhsin, Conflicts and Harmony; Shaaban Saleh Farsi, Zanzibar: Historical Accounts, 2d edn. (s.l., 1995); Norman R. Bennett, A History of the Arab State of Zanzibar (London, 1978).

⁶³� For example, W. H. Ingrams, Zanzibar: Its History and Its People (London, 1931), 125, 189–90. Another historian, L. W. Hollingsworth, will figure prominently below.

⁶⁴� For the mudirs' reports, see I. H. D. Rolleston, Annual Report on the District of Zanzibar, 1935, BA 30/5, Zanzibar National Archives (hereafter, ZNA). These and similar reports continued to be written, and British officials often referred to them.

⁶⁵� Fair, Pastimes and Politics; compare with Lofchie, Zanzibar.

⁶⁶� Alonso, "Politics of Space."

⁶⁷� Kopytoff, "Internal African Frontier." East African examples are described in Jean-Pierre Chrétien, The Great Lakes of Africa: Two Thousand Years of History (New York, 2003); Steven Feierman, The Shambaa Kingdom (Madison, Wis., 1974).

⁶⁸� For this and much of the following two paragraphs, see my book Feasts and Riot: Revelry, Rebellion and Popular Consciousness on the Swahili Coast, 1856–1888 (Portsmouth, N.H., 1995). Abdul Sheriff and Chizuko Tominaga believe that there was sometimes a kernel of truth to claims of Shirazi descent: see "The Shirazi in the History and Politics of Zanzibar," paper presented at the International Conference on the History and Culture of Zanzibar (Zanzibar, December 1992). For "Arabization" before the Omani period, see Randall L. Pouwels, Horn and Crescent: Cultural Change and Traditional Islam on the East African Coast, 800–1900 (Cambridge, 1987).

⁶⁹� Instrumentalist interpretations of the turn to Shirazi identity stress the material benefits that the colonial regime supposedly bestowed on "non-natives": see, for instance, Fair, Pastimes and Politics, 51. These interpretations are not always convincing: see Glassman, "Sorting Out the Tribes," 403–04. Furthermore, they make too much of the specific ethnonym "Shirazi," neglecting to recognize that admiration of Middle Eastern status could exist in the absence of that particular ethnonym. Evidence collected before the accelerated turn to Shirazi identity in the 1920s, for example, indicates that although indigenous islanders at that time called themselves "Tumbatu" and "Hadimu," they nevertheless already considered themselves descendants of Persians and as such more civilized than the barbarians of the mainland. F. B. Pearce, Zanzibar: The Island Metropolis of Eastern Africa (London, 1920), 248–52.

⁷⁰� This ambivalence is described by Ariel Crozon, "Les Arabes à Zanzibar: Haine et fascination," in Le Guennec-Coppens and Caplan, Les Swahili, 179–93.

⁷¹� For a suggestive account of the role of Kenyan schoolteachers in crafting a local cultural nationalism, see David Sandgren, Christianity and the Kikuyu: Religious Divisions and Social Conflict (New York, 1989).

⁷²� For an explicit statement, see the comments by W. Hendry, the director of education, on agricultural education, July 21, 1924, and June 8, 1925, AB 1/365, ZNA. A more general statement regarding the place of "Arab" leadership in educational policy can be found in W. Hendry, "Some Aspects of Education in Zanzibar," Journal of the African Association 27, 108 (1928): 351.

⁷³� For a firsthand explanation, see Zam Ali Abbas, "Yaliopita huzungumzwa (maendeleo ya skuli)," Maz., n.s. 1, no. 2 (January 1957): 15–18.

⁷⁴� Indians were also disproportionately represented in the islands' classrooms. For this and the next two paragraphs, see Bennett, Arab State, 194–96, 222–34, 245; Farsi, Zanzibar: Historical Accounts, 20; O. W. Furley and T. Watson, A History of Education in East Africa (New York, 1978), chaps. 6 and 10. Furley and Watson assume, erroneously, that because Zanzibari schools catered to all races, educational policies were non-racial; Muhsin also defends the colonial schools against charges of racial bias, although his memoirs are inconsistent on the issue (Conflicts and Harmony, 30–31). Compare Hendry's comments, cited above, which suggest a race-based policy of vocational "tracking." At any rate, the enrollment figures cited by Bennett and Furley and Watson confirm that, regardless of official intent, educational opportunities were skewed by race. Additional evidence of the discrepancy is overwhelming: see, for example, figures in CO 618/44/15, Public Records Office (hereafter, PRO), Kew, United Kingdom.

⁷⁵� Zanzibar Protectorate, Annual Report of the Education Department for the Year 1927 (Zanzibar, 1928), esp. 6–7; W. Hendry, February 27, 1934, and "Memorandum of the Arab Association," February 5, 1934, both in CO 618/60/15, PRO. Figures from 1937 indicate that teaching school was the most likely salaried post available in the administration for "African" and "Arab" school-leavers: AB 1/184, ZNA.

⁷⁶� In 1935, as part of empire-wide educational reforms, TTS was closed and its teacher-training functions divided between two new elite schools, Dole Rural Middle School and Government Secondary School.

⁷⁷� Hollingsworth later became director of the elite Government Secondary School, where his influence was, if anything, greater. The already-cited memoirs of Zam Ali Abbas, Ali Muhsin, Juma Aley, and Shaaban Saleh Farsi explicitly acknowledge how the experience of classroom teaching helped foster a sense of belonging to a leading intelligentsia; the latter three discuss Hollingsworth. Also see "Dr. Hollingsworth and Uhuru," Mwongozi, September 20, 1963; and Acting Dir. Educ., June 12, 1933, AB 1/76, ZNA.

⁷⁸� They include Abdullah Saleh Farsy, one of East Africa's most prominent Islamic scholars and public intellectuals; Yahya Alawi, who would become one of the highest-ranking Zanzibaris in the colonial administration (among his posts was information officer in charge of broadcasting); Zam Ali Abbas, founder of the Zanzibar Association, the islands' first explicitly nationalist organization; Ahmed Seif Kharusi, founder of the influential weekly Mwongozi; Juma Aley and Mohammed Salim Hilal Barwani, leading figures in the ZNP; Mohammed Shamte, chief minister of the ZNP/Zanzibar and Pemba People's Party government that would be overthrown in 1964; and Aboud Jumbe, second president of the revolutionary government.

⁷⁹� Given Hollingsworth's encouragement of discussion of topical issues as a way to hone expressive skills (and his sympathy for the goals of Arab-led cultural nationalism), there is little doubt that the schoolteachers inculcated some of their ideas in the classroom. Muhsin describes the classroom influence of A. M. al-Hadhrami, the leading figure on Mazungumzo's editorial board in the 1930s (Conflicts and Harmony, 64).

⁸⁰� This term, with all its pejorative implications, was used in the English-language journalism of the Mazungumzo circle. There was eventually a small debate over its appropriateness, sparked, ironically, by the objections of a British educator; Arab Association journalists defended the usage ("Native or African?" Al-Falaq, November 8, 1941). Compare the latter-day memories of Juma Aley, a leading member of the Mazungumzo circle who as a member of the ZNP government would become notorious for his disdain of Africans (Clayton, Zanzibar Revolution, 63); Aley claims that the British imposed use of the word "native" as part of their policy of stirring up racial hatred. Zanzibar, in the Context, 40.

⁸¹� This assessment and much of what follows is based in part on a broad reading of Mazungumzo from the 1920s and 1930s, as well as Arab Association journalism and files in the Zanzibar National Archives. Fuller documentation can be found in Chapter 3 of my book manuscript under preparation on the rise of racial thought in colonial Zanzibar.

⁸²� The version of Tuskegeeism introduced to East Africa in the 1920s stressed schoolteachers' roles as community leaders in "community development." See esp. Kenneth James King, Pan-Africanism and Education: A Study of Race Philanthropy and Education in the Southern States of America and East Africa (Oxford, 1971); also Lene Buchert, Education in the Development of Tanzania, 1919–90 (London, 1994). Zanzibar's longtime acting director of education, G. B. Johnson, was an internationally prominent advocate of Tuskegeeism, and his Swahili-language adaptation of Booker T. Washington's Up from Slavery (1901) was a standard reader in Zanzibar schools: Maisha ya Booker T. Washington, Mtu Mweusi Maarufu (London, 1937).

⁸³� For the founding of the Arab Association, see "Mafveraky: Who Is Who?" Al-Falaq, December 21, 1946; the context is described in AC 1/151, ZNA (thanks to Philip Sadgrove for directing my attention to the latter source). Previous works have mistaken both date and context: see, for instance, Alison Smith, "The End of the Arab Sultanate: Zanzibar 1945–1964," in D. A. Low and A. Smith, eds., History of East Africa, Vol. 3 (Oxford, 1976), 199.

⁸⁴� Such themes can be found in the writings of pan-Arabists and Islamic modernists at least as early as Jamal al-Din Afghani and the Egyptian nationalist Rifa'a Badawi al-Tahtawi: see Albert Hourani, Arabic Thought in the Liberal Age (Cambridge, 1983). For interwar currents, see Reeva Simon, "The Teaching of History in Iraq before the Rashid Ali Coup of 1941," Middle Eastern Studies 22 (1986): 37–51; Simon, "The Imposition of Nationalism on a Non-Nation State: The Case of Iraq during the Interwar Period, 1921–41," and Israel Gershoni, "Rethinking the Formation of Arab Nationalism in the Middle East, 1920–1945," both in J. Jankowski and Gershoni, eds., Rethinking Nationalism in the Arab Middle East (New York, 1997), 3–25, 87–104; Gershoni, "The Emergence of Pan-Nationalism in Egypt: Pan-Islamism and Pan-Arabism in the 1930s," Asian and African Studies 16 (1982): 59–94; C. Ernest Dawn, "The Origins of Arab Nationalism," in Rashid Khalidi, et al., eds., The Origins of Arab Nationalism (New York, 1991), 3–30.

⁸⁵� Arab nationalists had often stressed explicitly racial themes in their definitions of the Arab nation: see the passages quoted from Rifa'a al-Tahtawi and Lutfi al-Sayyid in Hourani, Arabic Thought, 79, 173; Eve Troutt Powell, A Different Shade of Colonialism: Egypt, Britain, and the Mastery of the Sudan (Berkeley, Calif., 2003); Powell, "Brothers along the Nile: Egyptian Concepts of Race and Ethnicity, 1895–1910," in Hagai Erlich and I. Gershoni, eds., The Nile: Histories, Cultures, Myths (Boulder, Colo., 2000).

⁸⁶� Simon, "Teaching of History," and "Imposition of Nationalism"; C. Ernest Dawn, "The Formation of Pan-Arab Ideology in the Interwar Years," International Journal of Middle East Studies 20 (1988): 67–91. The pan-Arabists, who were first motivated by opposition to Ottoman rule, adopted Breasted's theories in reaction to Ottoman authors who had celebrated the civilizing influence of ancient "Aryan Turks."

⁸⁷� L. W. Hollingsworth, Milango ya Historia, 3 vols. (London, 1925–31). It was promoted shamelessly in Mazungumzo and reprinted many times over four decades; the 1965 edition was approved for classroom use by the Tanzania Department of Education. Hollingsworth acknowledges the assistance of A. A. Seif, A. M. al-Hadhrami, and Mohamed Salim Hilal al-Barwani. Seif appears as translator on the title page of Vol. 1; however, in a list of titles on the back cover of the Swahili edition of another of Hollingsworth's works published by the same company (Macmillan), Seif is credited as translator of all three volumes: Historia Fupi ya Pwani ya Afrika ya Mashariki (London, 1966). Milango ya Historia was at first used as a reader in Standards 4 through 8, and also urged as a text with which teachers should prepare for teaching other classes; pupils in the upper standards were also assigned Hollingsworth's Short History of the East Coast of Africa (London, 1929). G. B. Johnson, "Report on Text-Books in the Swahili Language," Annual Report of the Education Department for the Year 1927, 27–31, CO 618/44/15, PRO; W. Hendry, "Memorandum," February 27, 1934, CO 618/60/15, PRO; "Syllabus, 1944," AD 1/80, ZNA. Geography primers told similar stories of civilization and barbarism, often in racial terms: B. M. Hart, Bara Afrika: Chanzo cha Jiografia ya Afrika (Nairobi, 1948); E. C. Francis, Afrika (London, 1952; first pub. in English, 1933); G. W. Broomfield and D. V. Perrott, Habari za Walimwengu, Book 3, Masimulizi ya Juma juu ya Waingereza (London, 1954).

⁸⁸� Explicit statements of these themes can be found in Milango ya Historia, 1: 15–27 (monarchy and literate learning) and 1: 42–48 (the ancient Jews and monotheism). Given the exemplary importance of the Greeks in Vol. 1, Hollingsworth could hardly insist on monotheism as part of his definition of civilization (as he does the other attributes), but the entire work, particularly the last two volumes, emphasizes the civilizing power of Judeo-Christian-Muslim values. Broomfield and Perrott's primer was less equivocal in proclaiming religion "the foundation of all true civilization": Habari za Walimwengu, Book 3: 101. Similar views of "civilization" were common in mainstream social science at the time: see Charles Ellwood, The Psychology of Human Society: An Introduction to Sociological Theory (New York, 1925); and other sources cited by Thomas Spira, Nationalism and Ethnicity Terminologies, Vol. 1 (Gulf Breeze, Fla., 1999), 69.

⁸⁹� For example, Milango ya Historia, 1: 93–94.

⁹⁰�Milango ya Historia, 3: 87–94. Also see Hollingsworth, Short History; and the work of his close colleague and collaborator, W. H. Ingrams, Zanzibar. These aspects of East African history figured prominently in the Middle School exams in 1933 and the TTS exams in 1957: AB 1/184 and AD 1/213, ZNA.

⁹¹� Norris, "Pride of Race." Seif explains the circumstances of the essay's publication in an editorial note introducing it. It originally appeared in the Accra Teacher's Journal 6 (October 1929), a publication that had been inaugurated one year later than Mazungumzo. Thanks to Cati Coe for sharing the latter information.

⁹²� Examples include M. Abdulrahman, "Safari ya Kilwa," serialized in six parts, beginning in Maz. 12, no. 2 (February 1938): 24–26; Muhammad Othman, "Siku Kuu ya Mwaka," Normal Magazine 4, no. 9 (September 1930): 112–15.

⁹³� M. Abdurrahaman [Abdulrahman], "Maisha ya Watu Wengine," Maz. 12, no. 11 (November 1938): 162–63; Maz. 12, no. 12 (December 1938); and following issues.

⁹⁴� "Mshenzi, n. wa- a barbarian, savage, one of the aborigines, a person untouched by civilization. Often used contemptuously by the coast native of those who come from the interior, although they are frequently more cultured and refined than the coast native!" Frederick Johnson, Standard Swahili-English Dictionary (Oxford, 1939), 419.

⁹⁵� Muhammad Othman, "Asili ya Neno 'Ustaarabu na Mstaarabu,'" Maz. 12, no. 4 (April 1938): 59–61; A. M. el-Hadhrami, "'Ustaarabu,'" Maz. 12, no. 4 (April 1938): 61–62; Ali Said el-Kharusy, "Asili ya Waarabu," Maz. 12, nos. 6–7 (June–July 1938): 81–83, 99–101. Parts of this debate were anticipated in the Dar es Salaam paper Mambo Leo of February and March 1923 (thanks to Katrin Bromber for this reference).

⁹⁶� The importance of the descent metaphor should be clear from all that has been said about Shirazi identity; it is also apparent in all the standard ethnographic studies. For an explicit statement, see Mohammed Shamte's careful consideration of how best to reckon ethnic identity, in "Mpemba," Maz. 11, no. 4 (April 1937): 52–54; Shamte opts for a strict reckoning based on descent.

⁹⁷� The literature still suffers from embarrassed silence about the degree to which Arab-centered notions of skin color as status marker were and remain widespread on the coast. A good indication of this is the furor raised among Lamu intellectuals by the publication of Abdul Hamid M. el Zein, The Sacred Meadows: A Structural Analysis of Religious Symbolism in an East African Town (Evanston, Ill., 1974), which describes such notions. Also see Fair, Pastimes and Politics, 94–96; and, for nineteenth-century material, Fred Morton, Children of Ham: Freed Slaves and Fugitive Slaves on the Kenya Coast, 1873 to 1907 (Boulder, Colo., 1990). For the Arab and Islamic world more generally, see John Hunwick, "Islamic Law and Polemics over Race and Slavery in North and West Africa," Princeton Papers: Interdisciplinary Journal of Middle Eastern Studies 7 (1999): 43–68; Bernard Lewis, Race and Slavery in the Middle East: An Historical Enquiry (New York, 1990); Powell, Different Shade.

⁹⁸� Othman, "Asili ya Neno." He reminds readers of the standard etymology, which traces mshenzi to the Persian word zinj, meaning black, and the Swahili word he uses to describe coast peoples' supposedly lighter skin color is far from neutral; the color is cleaner, purified, brightened (takata). Othman cites a book entitled History of Zanzibar, probably Ingrams and Hollingsworth's 1925 school text. But his essay reflects a variety of influences, including local understandings of Islam.

⁹⁹� For cultural nation-building in the Tuskegeeist principles favored by British educators, including the preservation of "tribal customs," see King, Pan-Africanism, 167–70, 263.

¹⁰⁰� Examples include Zam Ali Abbas, "Kwenda kwingi kuona mambo," Normal Magazine 3, no. 6 (July 1929); Abdul Rahman Muhammad, "Sikukuu ya Mwaka," Normal Magazine 3, nos. 10–11 (November–December 1929): 148–49, 162–63; Saleh Muhammad, "Ubaya wa somo juu ya mwari wake," Normal Magazine 4, no. 12 (December 1930): 147–48. The essays represent a debate between those who wanted to reform and preserve such dances as symbols of national pride and others who wanted simply to suppress them. By the 1930s, the reformist position had won the day, at least in the pages of the teachers' journal. For festive dance and community identity in the previous century, see Glassman, Feasts and Riot.

¹⁰¹� In theological terms, their concerns were with the dangers of bidaa or "innovation," forms of worship sanctioned neither by God nor His prophets, and shirk, mixing worship of God with pagan idolatry. For such concerns among East African ulamaa, see Pouwels, Horn and Crescent; Farsy, Shaf'i Ulama; and Bang, "Sufis and Scholars of the Sea"; and, for Islamic modernists more generally, Hourani, Arabic Thought, 225, 231–32. (The latter source refers to Rashid Rida, whose paper al-Manar was read in Zanzibar.) These concerns meshed with British officials' desire to discourage festive rites they considered economically wasteful. But the ulamaa's independent theological motives were well established; Islamic modernists had been criticizing such rituals for over a generation. See documents from 1935–1936 in AB 30/22, ZNA. The ulamaa continued to condemn such customs well after the campaigns: Muhammad Saleh Abdulla Farsy, Ada za Harusi katika Unguja (Nairobi, 1956). For the ulamaa's leadership in similar campaigns on the mainland coast, see Sarah Mirza and Margaret Strobel, eds. and trans., Three Swahili Women: Life Histories from Mombasa, Kenya (Bloomington, Ind., 1989); Ali bin Hemedi el-Buhuriy, "Habari za Mrima," Mambo Leo, nos. 141–47 (1934–35).

¹⁰²� Sheikh Tahir [Abubakr el-Amawi] (senior qadi), March 21, 1936, AB 30/22, ZNA. (The qadi of the minority Ibadhi sect appended a postscript expressing his agreement.) For other ulamaa who took similarly puritanical views, see Saidi Musa, Maisha ya al-Imam Sheikh Abdulla Saleh Farsy katika Ulimwengu wa Kiislamu (Dar es Salaam, 1986), esp. 65 and following. But Musa's account must be treated with caution, especially when he attributes such views to Abdullah Saleh Farsy, the subject of his hagiography.

¹⁰³� Bennett, Arab State, 228–29; Furley and Watson, History of Education, 122–23; Teaching of Koran and Arabic in Government Schools, 1924–1957, AB 1/390, ZNA. The comparison to parrots was made by one of the qadis on the commission (Ingrams, Zanzibar, 230).

¹⁰⁴� M. A. el-Haj (Mudir, Koani), "The Koran Schools," 1936, AK 33/294, ZNA. For the report's circulation within the Education Department, see AB 1/82, AB 1/390, ZNA.

¹⁰⁵� Louis Brenner, "Muslim Representations of Unity and Difference in African Discourse," in Brenner, ed., Muslim Identity and Social Change in Sub-Saharan Africa (Bloomington, Ind., 1993), 1–20. Also see J. Glassman, "Stolen Knowledge: Struggles for Popular Islam on the Swahili Coast, 1870–1963," in Biancamaria Scarcia Amoretti, ed., Islam in East Africa: New Sources (Rome, 2001), 209–25.

¹⁰⁶� Hourani, Arabic Thought, 299–301 (for Rashid Rida); Gershoni, "Emergence of Pan-Nationalism," 76–77 (for Hasan al-Bana).

¹⁰⁷� "Our New Educational Chief," Al-Falaq, May 20, 1939; "Native Trusteeship (1)," Al-Falaq, July 29, 1939; "Memorandum by the Arab Association," February 5, 1934, CO 618/60/15, PRO; "Report on Zanzibar Education (II)," Al-Falaq, October 1, 1938.

¹⁰⁸� Lofchie, Zanzibar, 119–26; Zanzibar District Monthly and Annual Reports, BA 30/5–7, ZNA; Al-Falaq, 1938, passim. A leading member of Mazungumzo's editorial board, Mohammed Salim Hilal el-Barwani, was among those most active in this campaign: Secret Police Bulletin, July 4, 1938, AB 12/114, ZNA. I borrow the phrase "purge category" from Brennan, "Nation, Race and Urbanization."

¹⁰⁹�Al-Falaq, June 15, 1940; M. A. el-Haj, "Hadimu Land Tenure," May 1940, AK 33/294, ZNA.

¹¹⁰� Ali Muhsin, "Slavery as It Used to Be Practiced in Zanzibar," Makerere College Magazine 1, no. 4 (August 1937): 111. Muhsin's essay was celebrated in Al-Falaq; for a full discussion, see Glassman, "Sorting Out the Tribes." Six decades later, Muhsin continued to issue similar apologies: Conflicts and Harmony, 177–86.

¹¹¹� "The So-Called Native Lethargy," Al-Falaq, September 3, 1938. The article was published as part of an ongoing campaign to persuade the authorities to create the legal apparatus by which estate owners might more effectively control agricultural labor.

¹¹²� Frederick Cooper, From Slaves to Squatters: Plantation Labor and Agriculture in Zanzibar and Coastal Kenya, 1890–1925 (New Haven, Conn., 1980), for British efforts to convince Zanzibari planters of the merits of industrial work-discipline.

¹¹³� Eve Troutt-Powell and John Hunwick, papers presented to the Workshop on Slavery and the African Diaspora in the Lands of Islam, Northwestern University, April/May 1999.

¹¹⁴� Frederick Cooper, Plantation Slavery on the East Coast of Africa (New Haven, Conn., 1977); Glassman, Feasts and Riot, 79–114; and, for a comparative example, see Peter Wood, Black Majority: Negroes in Colonial South Carolina from 1670 through the Stono Rebellion (New York, 1974).

¹¹⁵� In a regular column, "Father Zanzibar" addressed his "children" in a patronizing tone unlike anything in the English section ("Risala ya Unguja," Al-Falaq, 1946). Otherwise, most Swahili-language items consisted of advice on cooking, child-rearing, decorating, and other non-political matters directed specially at women; again, such items did not appear in the English language section. By 1951, the paper was publishing Swahili translations of leaders from its English and Arabic sections, but Swahili continued to be its tertiary language.

¹¹⁶� "Itihad-el-Watani," Al-Falaq, February 11, 1939.

¹¹⁷� More precisely, it reflected the villages' subjection to the island's most powerful indigenous potentate, the Mwinyi Mkuu, who in turn became the sultans' direct vassal. John Gray, History of Zanzibar from the Middle Ages to 1856 (London, 1962); Pearce, Zanzibar: The Island Metropolis; Mohammed Abeid el-Haj, "Hadimu Land Tenure," May 1, 1940, AK 33/294, ZNA; Hemed Jabir el-Farsy, "The Mwinyi Mkuu," Appendix C, Zanzibar District Annual Report, 1935, BA 30/5, ZNA.

¹¹⁸� Charles Sacleux, Dictionnaire Swahili-Français, 2 vols. (Paris, 1939–41), 2: 622; R. H. W. Pakenham, Land Tenure among the Wahadimu at Chwaka, Zanzibar Island (Zanzibar, 1947); John Middleton, Land Tenure in Zanzibar (London, 1961); Abdul Sheriff, "The Peasantry under Imperialism," in Sheriff and Ferguson, Zanzibar under Colonial Rule.

¹¹⁹� Among the most telling examples is M. A. el-Haj, "Hadimu Land Tenure"; we have already encountered the influential el-Haj's disdain for village madarasa teachers.

¹²⁰� Lofchie, Zanzibar, 169. For Tajo's lack of English, see CO 822/1376, PRO; for the Shirazi Association generally, AB 12/2, ZNA.

¹²¹� This included the SA's advocacy for indigenous islanders in the ethnically structured rationing schemes being discussed in the opening years of the war, as well as their demand that a non-Arab be appointed mudir in the southern district of Makunduchi. See J. O'Brien, October 4, 1944, plus other documents in AB 4/39, ZNA; R. Pakenham, Zanzibar District Monthly Report, December 1943, and other documents in BA 30/7, ZNA; "Shirazi Association," AB 12/2, ZNA.

¹²²� I spell out this interpretation in a section of a forthcoming book; its sources include various documents in AB 4/39, ZNA; District Reports in BA 30/5–8, ZNA; Timothy Welliver, "The Clove Factor in Colonial Zanzibar" (PhD dissertation, Northwestern University, 1990), 380–84; Pakenham, Land Tenure; Pascal Bacuez, "Une ethnographie dans son contexte: Administration coloniale et formation identitaire," Cahiers d'études africaines 38, no. 149 (1998): 103–33. British officials overestimated the SA's role in instigating the boycott; in fact, the evidence suggests that the SA largely took advantage of a situation that had been initiated in the villages.

¹²³� My interpretation contrasts with others that read Shirazi ethnic revivalism only in terms of a nationalist master narrative: either a divisive identity imposed by colonial rule or a unifying one crafted to defy the British. The first of these interpretations, which accords with the official ASP view, is echoed in Fair, Pastimes and Politics; for the second, see Sheriff and Tominaga, "The Shirazi." See Glassman, "Sorting Out the Tribes," 404–05 n.

¹²⁴� Even the fiercest anti-Arab nationalists believed, incorrectly, that Africans had never developed their own systems of writing, and that they had always depended on Arabs and Europeans for enlightenment and "true religion." This is apparent from a perusal of the African Association weekly Afrika Kwetu; an explicit example is "Kale hata leo," January 13, 1955.

¹²⁵� The discussion of African Association rhetoric in this and the next two paragraphs is based mostly on Glassman, "Sorting Out the Tribes." For the British defense of "Islamic slavery" generally, see Paul E. Lovejoy, Transformations in Slavery: A History of Slavery in Africa (Cambridge, 1983), 261–68; and Morton, Children of Ham; the views of several Zanzibari officials can be found in Sir Arthur Henry Hardinge, A Diplomatist in the East (London, 1928).

¹²⁶� Jonathon Glassman, "Between Two Worlds: Diasporic Visions and Racial Politics in Colonial Zanzibar," a paper presented at the Workshop on the Western Indian Ocean since 1800, Yale University, November 3–5, 2000.

¹²⁷� See the description of the Makunduchi rally at which the ASP was launched: Afrika Kwetu, February 14, 1957.

¹²⁸� Anti-mainlander sentiments can be found in the ZPPP newssheet Sauti ya Wananchi, among other sources; see AK 20/1, ZNA; also Saud A. Busaidy (DC Urban), March 6, 1963, AK 31/15, ZNA. In a particularly telling moment, Tajo shared the stage with Zam Ali Abbas, long a figure on Mazungumzo's editorial board; their speeches are summarized in "Umma Hay!" Mwongozi, April 21, 1961.

¹²⁹� The tensions at Makunduchi are well reflected in "Makunduchi," and "Mtumwa mwenye ari si mungwana asiekuwa na ari," Afrika Kwetu, October 29 and November 5, 1959; also Security Report, July–August 1961, CO 822/2046, PRO. The political events of the Time of Politics are best recounted by Lofchie, Zanzibar, although his assertion that ZPPP never posed a threat in Makunduchi is belied by his own data, which show that in 1963 the party lost to ASP there by a mere sixteen votes.

¹³⁰� I spell out the history of this discourse in Chapter 6 of my forthcoming book; for a summary, see "Gangsters, Thieves, and the Construction of Race in Colonial Zanzibar," a paper presented to the African Studies Seminar, University of Wisconsin, Madison, April 10, 2002.

¹³¹� Quotes from "Barua," by "Mzanzibari," Mwongozi, May 3, 1957; "Yepi Yaliyowaleta Pamoja Wananchi na Wazalendo," Mwongozi, March 3, 1961; "Declaration from Lamu," Mwongozi, September 13, 1963. Similar rhetoric can be found in an official ZNP statement from October 1957, CO 822/1377, PRO.

¹³²� I paraphrase Comaroff, "Of Totemism."

¹³³� Glassman, "Sorting Out the Tribes."

¹³⁴� The latter theme is explored in two chapters of my forthcoming book that draw on comparative studies of contemporary ethnic violence in South Asia and Central Africa.

¹³⁵� Frank Dikötter draws similar conclusions in his critique of a different literature: "Racial Discourse in China, Continuities and Permutations," in Dikötter, ed., The Construction of Racial Identities in China and Japan (London, 1997), 12–33.

¹³⁶� For a prominent example of such an argument, see Bill Berkeley, The Graves Are Not Yet Full: Race, Tribe and Power in the Heart of Africa (New York, 2001). I hasten to add that, despite his emphasis on colonialism's historical role, Berkeley refuses to absolve African politicians of responsibility for pogroms and massacres; compare with the "academic sycophants" denounced by Lemarchand, Burundi. And Berkeley fully recognizes the parallels between tribalism and race.

¹³⁷� This tendency seems most pronounced in the sociological literature, perhaps because that literature has been dominated by American scholars for most of the past century. Howard Winant, "Race and Race Theory," Annual Review of Sociology, 2000.

Content in the History Cooperative database is intended for personal, noncommercial use only. You may not reproduce, publish, distribute, transmit, participate in the transfer or sale of, modify, create derivative works from, display, or in any way exploit the History Cooperative database in whole or in part without the written permission of the copyright holder.