3
A Muslim Minority

In one sense or another, the Dyula have been a Muslim minority since their arrival in northern Côte d'Ivoire in the seventeenth century. Such a statement can easily foster the illusion of historical continuity. The problem is that Muslim minorities can only be defined with reference to some other group or groups of people who constitute the non-Muslim majority. In the case of the Dyula, both the specific nature of the non-Muslim majority and the way in which the distinction between Muslim and non-Muslim is conceived have changed radically since the beginning of the twentieth century. This change in the context of what one might (awkwardly) call "minorityhood" has had far-reaching effect. The very notion of "being Muslim" no longer means the same thing in Korhogo as it did in the past.

It was the gold trade that first attracted Muslim traders from "Manden," the heartland of the empire of Mali along the Niger river, south toward the tropical forests. A major center of this gold trade was the town of Begho, in what is now northwestern Ghana, conveniently situated near the Lobi gold fields and on the way to the more important Akan deposits (Wilks 1962). This early (fifteenth- and sixteenth-century) trade in gold spawned the development of subsidiary trade in a variety of other commodities, creating demands in one region for products only found in another. In particular, kola nuts became a major export of the Akan forests.^[1] This north-south trade between the savanna and the forest attracted increasing numbers of Manding-speaking Muslim traders, who settled along the various routes throughout all of what is now northern Côte d'Ivoire. As it happened, the north-central part of Côte d'Ivoire, including

the region of Korhogo, was placed at a considerable disadvantage in the north-south trade as compared to either the northeast or the northwest. Due south of Korhogo, the savanna extends further, in a U-shaped depression; the kola-producing regions of the forest were thus either to the southwest or to the southeast. Even so, by the seventeenth century, the earliest Dyula settlements in the region, Dyendana and Faraninka, were established as outposts on one of the trade routes leading southwest. To the east of Korhogo, the town of Kong developed into a major trading center, but also into a centralized state.^[2] By the beginning of the eighteenth century, a Dyula, Seku Wattara, seized power in a coup d'état. Kong not only became the dominant power in the immediate area, but a force, both commercial and military, to be reckoned with well beyond its borders. At its apogee, it dispatched raiding parties as far away as the Niger river, and was able to hold off, if not to defeat, the armies of Ashanti.

Kong remained a major commercial center until its destruction at the end of the nineteenth century, but its military preeminence was relatively short-lived. Nonetheless, this burst of military activity was to have important consequences in the vicinity of Korhogo. Several chiefdoms trace their origins to bands of warriors setting out from Kong, if not always firmly under its control. These armies carved out little dominions among the acephalous, Sienar-speaking Senufo peoples to the west. In the early eighteenth century, one such band, under Jangarawuru Wattara, settled in the village of Pundya, along the route to the kola-producing forests of western Côte d'Ivoire. The Senufo villagers were rapidly driven off to a new site, a few kilometers away, as Jangarawuru's village, renamed Kadioha, established its hegemony over its immediate hinterland.^[3] Oral traditions indicate that the chiefdom of Korhogo was founded at about the same period and in a similar manner.^[4] However, Nanguin Soro, the warrior who founded Korhogo, was Senufo. The region saw the emergence of a number of small chiefdoms, some (but not all) tracing their origins to Kong, and most, with the notable exception of Kadioha, under Senufo rule. Kong,

at the height of its power, thus provided the impetus for the beginnings of political centralization of the Korhogo region, less than 150 kilometers away. Surprisingly enough, Kong never exercised any direct control over a region so close by. In any case, this process of centralization, either in Korhogo or among its neighbors, remained minimal until the mid nineteenth century, with the expansion of the nearby state of Kenedugu, whose capital, Sikasso, lay just north of what is now the border between Côte d'Ivoire and Mali. The reigning chief of Korhogo, Zwakonion Soro, chose to ally himself with his powerful northern neighbor as a means of consolidating his own power in the region.^[5] The village of Korhogo thus became the capital of a chiefdom that, though relatively small by absolute standards, was, if only by default, one of the major political centers in the region. As such, the village attracted Dyula settlers in increasing numbers.

By the end of the nineteenth century, both the chiefdom and the village of Korhogo were characterized by a matrix of ethnic categories and subcategories corresponding broadly to a system of division of labor and allocation of political authority (see table on p. 52). Most of the population consisted either of Sienar-speaking "Senufo" agriculturalists or of Manding-speaking Dyula traders. However, each of these categories was subdivided in significant ways, and there were several other groups who did not fit neatly into either category. In any case, the very name "Senufo" is a Dyula phrase, siena fo , "to speak Siena[r]." The majority of Sienar speakers, those whose primary occupation was agriculture, called themselves senambele. The senambele were in turn divided along linguistic lines: the Tiembara, speaking a northern dialect of Sienar, were the politically dominant group in the chiefdom and alone accounted for the majority of its population; the Fodonon, speaking a southern dialect of Sienar, were an encapsulated minority. Socially distinct from the senambele were the fijembele , members of hereditary (though not endogamous) groups traditionally associated with specific crafts: the fonombele , "blacksmiths"; the kulebele , "sculptors"; and the kpeembele , "brass casters." The kpeembele , I was told, once

spoke their own distinct language. According to Dolores Richter (1980: 15), "Kulebele and Fonombele support their oral tradition that they originated in Mali and migrated southward." Whatever the case, these craft groups now speak as a native language one dialect or another of Sienar, depending on where they are settled. In Korhogo, they all speak Tiembara. However, the Dieli, "leatherworkers," whom oral traditions identify as the oldest occupants of the site of Korhogo village, still have their own language distinct from either Sienar or Manding.^[6] Finally, the Milaga "blacksmiths" were native Manding speakers like the Dyula. The Dyula, in turn, were divided into two hereditary categories: tun tigi , "possessors of quivers"—that is to say, "warriors,"—and mory , "scholars." In principle, these hereditary categories were occupationally and linguistically defined. Reality, of course, was less static and more complex than such a neat system of divisions might suggest. Bilingualism was quite common, especially among members of linguistic minorities. Although Tiembara and Fodonon senambele might define themselves as "agriculturalists," agriculture was in fact an activity common to members of all groups. Various minority groups enjoyed monopolies, or quasi-monopolies, over certain occupations. Depending, in part, on the demand for their goods or services, members of such an occupational minority might practice their trade full time, seasonally, or not at all.

The relationship between occupation and hereditary status was particularly complex among the Dyula, who represented a substantial proportion—perhaps as much as onesixth—of the chiefdom's population. Long-distance trade was effectively a Dyula monopoly, made possible by their links with other Manding-speaking trading communities outside the region. However, as we have seen, major northsouth trade routes tended to bypass Korhogo either to the east or to the west. Weaving was another Dyula monopoly; indeed weaving, rather than trade, was a ubiquitous occupation among the Dyula. Cloth was, for the most part, traded to the Senufo, as were imported luxury goods, whose local

distribution the Dyula controlled. Yet the Dyula themselves were further subdivided, as we have seen, into the hereditary categories of mory and tun tigi. These categories applied to entire kabilas , or clan wards. The domains of warfare and especially politics were the preserve of tun tigi kabilas , Islamic scholarship the domain of the mory. Unlike other hereditary distinctions in the region—between Dyula and Senufo, between senambele and fijembele— the division between mory and tun tigi was of greater ideological than economic significance. The real units of economic specialization among the Dyula were the individual kabilas. Only a few mory kabilas were devoted to advanced Islamic scholarship, and only those tun tigi kabilas with rights to chiefly office in a village or chiefdom were specifically concerned with politics. Aside from such specialized kabilas , the distinction between tun tigi and mory involved no monopolies; a mory individual might well take up arms as a profession, just as a tun tigi might devote himself to study. Rather, this division reflected the notion that warfare and politics on one hand, religion and scholarship on the other, were antithetical careers. Both, however, were compatible with trade and weaving, which were ultimately the economic mainstays of the Dyula community.

The distinctions among all of these various categories—Dyula and Senufo, senambele and fijembele, mory and tun tigi— were reflected in the religious domain. Within the village of Korhogo, where each of these categories was represented, every group—with the significant exception of the Dyula mory , the Islamic scholars—had its own separate initiation society, known as poro in Sienar, lo or jolo in Manding. Each society initiated its own members separately. Each had its own distinct set of masks and ritual paraphernalia (though a certain amount of borrowing could and did occur) associated with both funeral and initiation ceremonies. In other words, the religious practices of a significant proportion of the Dyula population—the tun tigi or "warriors"—did not differ significantly in many respects from those of their senambele or fijembele neighbors.

One might be tempted to conclude that precolonial Korhogo was characterized by the coexistence of two clearly distinct

religious systems: an initiation-society complex on one hand and Islam on the other. Other aspects of tun tigi behavior would lend credence to such a view: the offering of animal sacrifices to "fetishes" (jo ), the material embodiments of various spirits; drinking alcoholic beverages; nonobservance or irregular observance of Muslim prayers and of the fast at Ramadan. Nevertheless, the tun tigi were considered to be Muslims, not "pagans."^[7] In effect, the distinction between Muslims and non-Muslims, at least within the confines of the region, coincided exactly with the distinction between Dyula and non-Dyula. Non-Muslims in the region were collectively known as banmana , literally "refusers," that is to say, "unbelievers," "those who refuse Islam."^[8] In ordinary speech, the term banmana was generally used to refer to the Senufo; Sienar was called banmana-kan , "the unbelievers' language," as opposed to Dyula-kan , or Manding. However, language itself was not the determining criterion for religious identity; the Milaga blacksmiths, who spoke Manding, were classed as "unbelievers" rather than as Muslims, as were the Dieli leatherworkers. These distinctions were reflected in marriage patterns. Tun tigi and mory groups could give and receive wives to and from one another; they could also receive wives from "unbelievers," but not bestow their own women in marriage to them, since Muslim women ought only be given in marriage to Muslim men. The Muslim identity of the tun tigi was publicly expressed, not by their observance of the strictures of Islamic law, but rather by their participation in festivities associated with Muslim calendar holidays: donba (Arabic mawlud), tabaski (Arabic 'id al-kabir) and sun kalo (Ramadan).^[9]

In short, religion in precolonial Korhogo very neatly mirrored the local system of hereditary social categories. The distinction between Muslim and non-Muslim corresponded to that between Dyula and non-Dyula; among the Dyula, in turn, only the mory were obligated in principle to follow the strictures of Islamic law diligently. This correspondence between the categories "Dyula" and "Muslim" made perfect sense in terms of the local monopoly held by the Dyula in long-distance trade. The long-distance trade network of which Korhogo was a part was entirely in the hands of a

Muslim diaspora; one could not participate in such trade without being a Muslim.^[10] However, to demonstrate one's identity as a Muslim outside the region, one had to be reasonably familiar with the basics of Islamic ritual: at the very least, one had to know how and when to pray. Precisely because both the prayers and the instructions for their proper use were transmitted in Arabic written documents, the preservation of such knowledge implied the existence of a body of persons literate in Arabic. The maintenance of this tradition was the responsibility of the mory , although by no means all mory individuals had to be literate in Arabic. Relatively few could read and write Arabic with any degree of fluency; many more were capable of reciting a written text, although not necessarily of understanding its precise meaning.

If the practice of Islam distinguished one category of Dyula from the other, allegiance to Islam, more than anything else, served as the defining criterion of Dyula identity. Only Islam distinguished Dyula unambiguously from all of their neighbors. After all, some Dyula participated in initiation-society activities. Like the fijembele , the Dyula enjoyed a hereditary monopoly over craft production—namely, weaving. Even language was hardly an unambiguous marker of identity, for the Dyula shared a native language with "pagan" Milaga blacksmiths (admittedly not a very numerous group).^[11] In any case, Manding-speaking banmana were to be found in large numbers in regions to the north and to the west of Korhogo. Religion, rather than language, linked the Dyula of Korhogo with the whole community of Muslim traders in the Western Sudan. Indeed, a Hausa quarter existed as close by as Kong (Binger 1892: 297), and several families of Koko's Dyula community in Korhogo trace their origins to immigrant Hausa traders who were able to assimilate without difficulty.

In short, by distinguishing the Dyula from their immediate neighbors and linking them with Muslims outside the region, Islam was crucial to their maintenance of a monopoly over long-distance trade. Since trade and weaving were locally associated, it served equally to preserve the even more

crucial monopoly over the production of cloth. It follows that the Dyula had no particular interest in converting their neighbors to Islam; on the contrary, in order to protect their monopoly, they had, if anything, an interest in discouraging such conversions. One could argue that the Dyula scholars, whose Arabic literacy furnished them with a lucrative monopoly on written magical charms,^[12] did not even have an interest in reforming the religious practice of their tun tigi coreligionists. An early colonial administrator in Côte d'Ivoire related an anecdote that, while it pertains to a community outside the Korhogo region, equally reflects the attitude of Korhogo's Dyula scholars.^[13] A "pagan" chief had decided to convert to Islam, and learned the prayers and the rudiments of Arabic, as well as observing the tenets of Islamic law with all the enthusiasm of a neophyte. The local scholars apparently found this rigor disquieting, if not embarrassing. When epidemics decimated both the cattle and the human population of the village, they reportedly addressed the chief in the following terms: "Do not be astonished if the village is subjected to Allah's wrath. He has done well by putting everyone in his place. You have transgressed His orders, and have abandoned the faith of your fathers, whom God manifestly commanded to drink palm wine, to eat unclean meat, and to worship only stones, mountains, and trees. Do not be astonished if your ancestors, and God himself, are punishing your village. This is what God has revealed to us." The story is perhaps apocryphal, a caricature, but not necessarily a radical misrepresentation, of what was the attitude of some, if not all, scholars. There is no evidence that the scholars of Korhogo actively discouraged conversions among the "unbelievers," or reform of the religious practices of their tun tigi cousins—nor, for that matter, is there evidence that they had occasion to do so. Rather, they not only tolerated but fully accepted the status quo. In fact, they apparently winked at the limited participation of mory youths in the initiation-society rituals of their tun tigi kin. Of course, such youths were neither formally initiated nor involved in any blatantly unorthodox activities, such as the offering of sacrifices or drinking

beer. However, some of my older mory friends in Koko confidentially—but with unabashed nostalgia—recounted to me how they had entered the sacred forests of their tun tigi relatives, stripping off their Muslim robes and dancing with the initiates during various masquerades. Scholars must have been aware of such goings on; that they condoned such behavior, even among mory , is a sure indication that they were not characterized by any particular zeal to reform the system.

If Dyula scholars so readily accepted such a dual standard of religious practice, it is also because the hereditary division into mory and tun tigi reflected a deep-seated principle, the separation of religion from politics. This principle was a cornerstone of Dyula policy, and a reflection of their status as a Muslim minority. In keeping with this policy, it was the responsibility of religious leaders openly to proclaim their support for whatever regime happened to be in power, provided that the regime tolerated the free exercise of Islam as a religion. Such moral support included the provision of charms and prayers, when solicited, to ensure the success of the regime's enterprises. It is easy to misrepresent such a stance as quietist (at best) or collaborationist and opportunistic (at worse). On the contrary, such behavior was a moral obligation on the part of religious leaders, but for that very reason was not binding on the Dyula community at large. Ordinary Dyula were free to support one faction or party or another, in or out of power, while scholars were expected to endorse whomever happened to be in power at the time. The regime in place, in other words, was not to be opposed on religious grounds—which is not to say that it was not to be opposed by anyone under any circumstances. The principle placed both religious leaders and Islam in general outside the political arena.

The arrival of the French did not in any way alter this principle of the separation of "religion" from "politics." Colonial rule came late to northern Côte d'Ivoire. It was only in 1895 that the French established a post on the Bandama River, in the vicinity of the region; in 1903, they chose to move the district headquarters to the village of Korhogo, which was to

remain the "capital" of the north. The last decade of the nineteenth century had been a turbulent one, as three rival empires vied for control of the region: Kenedugu, whose capital, Sikasso, lay north of Korhogo; the empire of Samory, a conqueror, whose home base had been to the west, in Guinea, but whose armies, driven eastward by a revolt, subjected vast areas; and finally, of course, the French. Each of these armies, in turn, exercised their dominion over Korhogo. In the early days of French rule, administrators reported that their subjects fully expected their dominion to be as shortlived as those of their predecessors, and that the French, like the others before them, would soon go away or be driven away by some, yet more powerful, army. In a sense, these rumors were not entirely false. The reign of Gbon Coulibaly, chief of Korhogo, who had been installed by the king of Kenedugu, and who had submitted to Samory and later the French, outlasted the entire period of French rule. He died in 1963, three years after Côte d'Ivoire became independent. Colonial rule was but a long episode in Gbon's reign.

It became the responsibility of Muslim clerics to declare open and enthusiastic support for the French, just as they had declared their loyalty to previous African rulers. Indeed, an early colonial study of Islam in Côte d'Ivoire (Marty 1922) cites in its entirety an address from the "Muslims of Korhogo to the people of Mecca," written during World War I, which demonstrates the lengths to which such a principle could be taken:

The present document, written by Mamadou Soumare, son of Brahima Soumare, of Korhogo, is destined to be sent to the Muslims of Mecca through the intermediation of the governor of Côte d'Ivoire, so that they understand that we, the Muslims of the post of Korhogo, are very happy to hear that they are now allied with the French to fight the Germans.

Moreover, we are very unhappy to see the Germans make war on the French.

Whoever does not wish to see the French in our colony [Côte d'Ivoire] is also held in contempt by us Muslims, since our prosperity depends entirely on the arrival of these latter in our colonies.

It is moreover thanks to the French that we are spared the ravages and pillages of Samory, slavery, and wars between one village and another.

At present, we are free, we can live, work in peace, and perform our prayers in tranquility [emphasis added].

Since we have learned that all the Muslims of the Holy City have allied themselves with the French against the Germans, our concern has been to pray to God that the Allies annihilate Germany.

Let Germany be annihilated.

Let the Germans (poor, rich, princes, kings) surrender and submit themselves to the French.

We note that the French race is better than all the others.

When you give a chicken to a Frenchman, he pays you the price of a sheep; one day of work for him is paid by a month's salary (unlike other races).

Which explains why we consider His Excellency the governor as our father and mother….

Needless to say, these words are not of our own invention, but follow qur'anic rules and the point of view of Islam.

Moreover, Muhammad (our master), during his existence in the Holy City, was very fond of the French race. This explains why we Muslims should continue to love the French as he did

(Marty 1922: 486–87; my translation)

This document, originally written in Arabic, is nevertheless reproduced by Marty in its African author's French translation, replete with mistakes in French usage. Its inclusion as such in the book is intended quite obviously as a rather supercilious testimony to the loyalty of Korhogo's Muslim community to France. To modern readers, it seems either contemptibly and embarrassingly sycophantic or else crassly hypocritical, in either case a pure product of "collaboration" in the worst sense of the word. Such an interpretation, however, rests on a fundamental misunderstanding. The author of this missive was in fact acting precisely in the way Muslim clerics were (and are still) expected to act. The letter stresses the loyalty of Korhogo's Muslim community to the regime in place, and goes so far as to imply that this loyalty is a religious obligation. Its inclusion, in its entirety, in Marty's volume demonstrates that certain colonial administrators ac-

cepted such declarations at face value—an implicit, if unintentional, testimony to the efficacy of the policy. But the letter is hypocritical only if its rhetorical excesses are accepted too literally, although of course such a literal interpretation served Muslim interests all the better. Rather, such declarations served the more fundamental purpose of reassuring the French that the Dyula, as Muslims, were not opposed to the regime on religious grounds, precisely, as the letter indicates in no uncertain terms, because the French permitted Muslims to pray in all tranquility.^[14] Such a stance did not prevent individuals, as long as they were not Muslim clerics, from opposing the regime on other, more purely "political" grounds, provided that the issue of "religion" remained above the fray. Indeed, a number of Dyula from Koko quarter in Korhogo were active militants in the early years of the independence movement. By diligently separating "religion" from "politics" and by expecting religious leaders to affirm loyalty to those in political power, the Dyula, as a Muslim minority, successfully avoided any political repression of the religious community. In effect, they affirmed that there could be no "Muslim opposition" as such. On the other hand, ordinary Dyula were free to join forces with the opposition as long as it was defined in explicitly "political" rather than "religious" terms. This, too, was to the advantage of the Dyula community, should the regime be overthrown. In other words, the continued toleration of the Dyula presence as a Muslim minority involved their continued loyalty, in principle, to "the" political regime, without requiring individuals to commit themselves firmly to any particular regime. This ambiguity was only possible as long as "religion" and "politics" remained conceptually separate domains.

Neither colonial rule nor the postcolonial regime have changed these fundamental attitudes about the relationship of religion and politics among the Dyula. However, colonial rule was to bring other changes, which deeply affected Dyula conceptions about who they were and about the relationship between "being Dyula" and "being Muslim." In the first place, the pax colonia irreversibly altered trade routes by

opening up the forest to Muslim traders who had previously had to obtain the kola produced there from frontier markets. Muslim traders and would-be traders rapidly emigrated to the south of Côte d'Ivoire, and the Dyula from Korhogo were no exception. Here, the Dyula found themselves lumped together as "immigrant strangers" along with Muslims from other parts of Côte d'Ivoire and from neighboring countries, and were more systematically exposed to different conceptions of Islam. At the same time, manufactured goods were increasingly coming into competition with locally produced items, to the detriment of most traditional craftsmen, including Dyula weavers.

Early in the colonial period, partly as a result of some of these changes, groups and individuals in Korhogo began converting to Islam. The most spectacular conversion, about 1920, was that of the Senufo chief of Korhogo, Gbon Coulibaly. If we are to judge by Gbon's political career, this conversion is unlikely to have stemmed from deep moral convictions; throughout his entire reign, Gbon managed to convince those in power—Samory, the French, the RDA independence movement—of his loyalty, never switching allegiance too early or too late. Whatever his motives for conversion, it is clear that Gbon, much to his own convenience, was able to model his observance of Islam on the practices of the Dyula tun tigi. In other words, as French observers noted (Marty 1922: 173), he prayed irregularly and inaccurately, drank beer, and, most important of all, continued to participate in poro activity. Gbon's conversion was apparently welcomed by the Dyula scholars of Koko, his neighbors, with whom he had always been on good terms, and who now served, formally or informally, as his religious advisors. At the time, the conversion may have seemed essentially a symbolic gesture, one that did little to change the status quo. Yet it set a precedent that was to rock the foundations of the Dyula conception of their place in the local social universe.

At the same time, Islam was attracting quite another sort of convert in Korhogo. These were members of those hereditary craft groups who occupied a sort of liminal position in

local society, those who did not speak Sienar as a native language—that is to say, the Milaga blacksmiths and the Dieli leatherworkers—and were consequently neither "Dyula" nor "Senufo." As long as the crafts with which these groups were associated remained a principal basis of their livelihood, such a liminal status was, if anything, an advantage. However, the colonial economy, by freeing the movement of people and goods both within Côte d'Ivoire and across its borders, opened up new economic opportunities by undermining many traditional monopolies. A new Muslim Mande diaspora was forming communities in the south of Côte d'Ivoire. The Milaga were, after all, native Manding speakers. As for the Dieli, since no one else could understand their language, virtually all of them acquired a native fluency in either Manding or Sienar, and often in both, early in life. In order to assimilate themselves into these new diaspora communities, Dieli and Milaga needed only convert to Islam. Not surprisingly, they began to do so, assimilating themselves to their Dyula neighbors in the process. Unlike their Senufo chief, such individuals did not have the luxury of announcing their conversions with considerable fanfare while adopting a tun tigi model of Islamic practice. Rather, in order to mark their transition to Dyula status, they had to pray regularly, abstain from alcohol and impure meats, and in general imitate their more orthodox mory neighbors. Indeed, such standards of piety—obligatory hallmarks of Muslim identity in the new diaspora communities in the south—were gaining ground among the tun tigi as well.

As mory standards of piety concerning prayer, fasting, alcoholic beverages, and the like continued to gain general currency within Korhogo's Muslim community, the continued participation of tun tigi , Milaga, and Dieli in lo initiation societies was bound to become a matter of contention. Matters came to a head in the mid 1940s. Significantly, pressure to abolish these societies came, not from the scholars, but from young prospective initiates. There were a variety of reasons, not all of them religious, why these youths objected to the initiation societies. Initiates were subjected to physical and

psychological hardships; they were obliged to pay various fees and fines and to work in the fields of initiation-society elders. One purpose of the initiation societies was to reinforce the physical, psychological, and economic subjection of youths to their elders. However, the economic changes that accompanied French rule tended to make such youths more and more economically independent of their elders, as new opportunities opened up to replace declining traditional crafts such as weaving. Young men objected to the initiation societies, not only on ideological grounds, but also because the time, effort, and expense involved distracted them from their new careers, seemingly for no purpose. It is perhaps symptomatic that the ringleader of the resistance to the initiation societies was a Milaga "blacksmith" rather than a Dyula, a young peddler-turned-tailor from Korhogo named Seydou Fofana. Here, in other words, was a young man following an entirely new craft occupation—sewing machines had only just been introduced into Korhogo; moreover, he belonged to a group who had just recently converted to Islam and adopted a new mode of religious behavior, involving regular prayer, fasting, and other such signs of "orthodox" Islam. Seydou and his peers had a vested interest, of course, in abolishing initiation: why should they bind themselves to their elders when the means of independence were increasingly available? It would be unnecessarily cynical, however, to attribute exclusively economic motives to the young rebels. Some—Milaga and Dieli—were recent converts to Islam; others, the tun tigi , were nevertheless adopting new rules of religious observance. In either case, it might easily seem that the initiation societies were incompatible with "orthodox" Islam; their mory neighbors provided an obvious example. Indeed, the sincerity of Seydou and his followers was quickly put to the test. When they refused to bow to the threats of their own elders, they were summoned before the (now Muslim) chief or Korhogo, Gbon Coulibaly, who took an equally dim view of such rebelliousness, especially as his own authority over his Senufo subjects still rested in large measure on his control over the initiation societies. The youths were

severely beaten. When they continued to resist, their elders predicted that those who persisted would all be dead within a year. Such mystical death was commonly threatened as a punishment for those who flaunted the rules of the initiation societies. Under ordinary circumstances, the occasional recalcitrant individual might be poisoned or otherwise done away with in order to maintain the credibility of the threat, but the elimination of an entire cohort was out of the question. The abolition of forced labor in the French colonies in 1946 strengthened the hand of the rebellious youths. Until then, Gbon Coulibaly, as chief of Korhogo, had been able to send off young "troublemakers" or political opponents to the labor camps. One way or the other, the rebels held their ground, refusing to give in to threats or beatings, and demonstrating, by their very survival, their triumph over the mystical powers of the lo societies. This example sparked off a movement that led in a decade to the abolition of virtually all the Dyula initiation societies in the region. Henceforth, the religious practices of all Muslims were in principle uniform. The distinction between mory and tun tigi lost all salience whatsoever. Behavior that had been admissible, and indeed appropriate, for certain categories of Muslims in the past, was now considered distinctly "non-Muslim."

In short, to be Dyula, one had to be Muslim, and to be Muslim, one had to observe what were once only mory standards of piety. One could, however, be identified as "Dyula," not only in Korhogo, but also in the new and fast-growing communities in the south of Côte d'Ivoire, where many Dyula from Korhogo had emigrated during the first half of the twentieth century. In the south, all Manding-speaking Muslim immigrants were known as "Dyula," whether they were from Côte d'Ivoire or from neighboring countries such as Guinea, Mali, or Upper Volta. In short, the "Dyula" were a highly heterogeneous community of "outsiders" linked by a common language and a common religion. Originally, of course, the Dyula community of the Korhogo region had constituted itself in the same way, over the centuries incorporating Manding-speaking immigrants from a wide variety of

origins. However, by the twentieth century, the Dyula of Korhogo defined themselves as an indigenous, rather than immigrant, community. The assimilation of such groups as the Milaga and the Dieli had, for a while, preserved the division of the socio-religious universe of Korhogo into two categories: Manding-speaking Muslim Dyula and Sienarspeaking "pagan" Senufo. Indeed, the Dyula community of Koko could continue to assimilate outsiders, at least in small numbers, who would attach themselves as "strangers" to one local kabila or another.^[15] Even the presence of Senufo converts to Islam did not necessarily threaten the equation of religion and ethnicity, provided these converts followed the example of their chief, praying irregularly and continuing to participate in initiation societies. Under such circumstances, Senufo Muslims, behaving as the tun tigi used to do, constituted an intermediate category between Dyula Muslims and Senufo "pagans." Religious practice, rather than the profession of Islam as such, constituted a defining criterion of being "Dyula," whether in Korhogo or in the new immigrant communities of southern Côte d'Ivoire.

Until the end of World War II, the fact that Korhogo was the administrative center of northern Côte d'Ivoire seems to have made relatively little difference to most of its residents. All of a sudden, after the war, just as the lo societies were being abandoned, the town began to mushroom. The town attracted a growing number of immigrants, Dyula and Senufo alike, from the surrounding countryside, but also from further afield, from other regions and indeed other countries, whose borders, after all, were not far away to the north. Many of these immigrants were Manding-speaking Muslims like the Dyula, and in smaller numbers they would have been assimilated into Korhogo's Dyula community in Koko. But they came in droves and settled, not in Koko, but in new neighborhoods, "across the stream." At the same time, the new Senufo immigrants to Korhogo began to convert to Islam, the first step toward a new career in trade, an occupation no longer monopolized by the Dyula of Koko. The neighborhood of Koko, the site of Korhogo's main mosque but also

of its largest sacred initiation grove, retained as its nucleus the descendants of Korhogo's various precolonial inhabitants. The Koko Dyula retained their rights over the imamate of the main mosque, but they lost all their monopolies—over trade, over Islam—in Korhogo. No longer were they the Muslim minority; they were only one among several.

The consciousness that they were no longer the only yardstick by which Islam was to be measured in Korhogo accentuated the pressures for a uniform standard of Islamic conduct in Koko's Dyula community begun by the abolition of the tun tigi , Dieli, and Milaga initiation societies. If it remained true that "being Dyula" also meant "being Muslim," the converse was hardly the case any longer. As a result, the question "What does it mean to be a Muslim?" acquired new salience.

The answer to this question lay, in part, with the Muslim scholars in the community, the karamogos. They were the ones with specialized knowledge about what Muslims ought or ought not do. Now that such knowledge was of concern to the entire community, new means of transmitting such information were necessary. In particular, the role of the scholar was broadened from that of teacher to preacher. In the past, this knowledge had been communicated largely on an individual basis, from teacher to pupil or from a scholar to an individual seeking to elucidate a specific problem. Now, public sermons by Muslim scholars began to supplement these older means of transmission as a way of reaching the community as a whole.^[16] Sermons were originally introduced as a part of funeral ritual, marking the fortieth day after death. Formerly, this had been the occasion for a particular dance known as majo. During the majo dancing, young men who were advanced students would deliver brief exhortations to the audience about the principles of proper Muslim behavior. This was now expanded into a full-scale "recitation" (kalan ), generally consisting of a long, extemporized sermon in Manding. The majo dancing, once the climax of the funeral ceremonies, was abandoned, except at the funerals of important elders. Nowadays, even when the dances are held, they

are much attenuated and take place separately from the sermons. More important, the sermons are no longer delivered by students but by established scholars, a sign that both the scholars and their audiences treat them more seriously.

Since their introduction, sermons have, in the space of only a few decades, come to replace dancing as the primary means of celebrating Islamic calendar holidays, as well as funerals. This is not, it should be stressed, necessarily a symptom of a wave of Islamic puritanism among the Dyula. Admittedly, the scholars, in a somewhat puritanical vein, disapprove of dancing altogether these days, though their exhortations, in this case, have fallen on deaf ears. More important, though, the sermons are festive occasions, and many people enjoy attending. Both sermons and dances are communal celebrations, and like all such celebrations, they transmit a variety of messages, depending on who organizes them, who attends, on which occasions they are held. In the past, dancing had been an appropriate means of affirming the Islamic identity of the Dyula community; as Muslims, they performed special Muslim dances on special Muslim holidays. Once the Dyula of Koko ceased to be the only, or even the principal, Muslim community of Korhogo, their dances, even if performed on Muslim holidays, were no longer unambiguously identified as Muslim dances; on the contrary, they were Dyula dances. Their performance was not the occasion for all local Muslims to celebrate together, but instead a means of distinguishing Dyula Muslims from other Muslims. Sermons, unlike dances, were a relatively unambiguous symbol of the collective Muslim identity of the Dyula. Such sermons were not peculiar to the Dyula of Korhogo. They were regularly held in the new immigrant Muslim communities of southern Côte d'Ivoire, and became part of a national (indeed an international) Muslim culture in the country. By adopting the sermons, the Dyula of Koko signaled that they were part of the national Muslim community of Côte d'Ivoire, and not simply the local Muslim minority of Korhogo.

If the sermons themselves proclaim the Muslim identity of the Dyula, their content serves to clarify the ramifications of this identity. The sermons stress, above all, the obligations of believers, in other words, what it means, in concrete terms, to be a Muslim.^[17] By their form and content, the sermons stress the fact that the Dyula of Koko are part of the universal community of believers of Islam. However, ceremonies have the power to transmit different, and even contradictory, messages at different levels. While the sermon is in progress, the scholar and his message may well be the center of attention, but the organizers and the audience are just as much a part of the proceedings. The sermons are public, and in principle anyone is welcome to come, but they are also one part of an elaborate cycle of funeral rituals; it makes a difference who is or is not invited, who does or does not attend. Funerals, along with weddings—the other major life-crisis ritual among the Dyula—transmit all sorts of messages about social relationships. One such message has to do with the very identity of the Koko Dyula community itself. Each wedding, each funeral, reaffirms—or redefines—the separate identity of the community as such. At crucial moments of the wedding or funeral, representatives of each constituent section of the community must attend. In this manner, the "we" who are holding the ceremony can be extended to include, not only a group of kin and indeed an entire kabila , but ultimately the whole community. But, like all blanket principles, the injunction that each segment of the community must be represented is more ambiguous than it might seem. Who is the community? What constitutes a segment? Who, in other words, are "we"?

In 1972, when I first came to Koko, the answers seemed fairly straightforward. The Dyula community consisted of about twenty kabilas , grouped into larger units called makafus , linking one "host" kabila to a number of "strangers." For important ceremonies, each Dyula kabila in Koko was expected to send at least one representative. The Dieli and the Milaga were included as obligatory participants, and were thus

symbolically incorporated into the Dyula community.^[18] Although, by this time, many of Koko's Senufo residents were Muslim, they were for the most part absent from the proceedings, attending, if at all, in individual capacities, as friends, colleagues or affines of the deceased or his close kin. In other words, the ceremonies stressed the separateness of Koko's Dyula community from Koko's Senufo community, as well as from other communities, Manding-speakers or Sienar-speakers, Muslims or unbelievers, across the stream.

By 1985, both the rules and the practice had been modified. It was no longer obligatory for each kabila to send a representative; a delegate from each makafu was sufficient. More important, Muslim representatives from the various Senufo groups living in Koko (Tiembara, Fodonon, Fono, Kpeem, and Kule) were almost invariably present on such occasions. Their attendance, though it remained optional in principle, was considered absolutely normal. This change was initiated by the Dyula themselves, and not the Senufo: one does not attend such rituals without being invited. The Senufo presence concretizes a growing consciousness of the existence of a multi-ethnic Koko Muslim community as such. The phenomenally rapid growth of the town as a whole—it virtually doubled its size in ten years—has intensified the sense on the part of Koko's residents that they are an "old" community, divided and subdivided into ethnic and clan wards, as opposed to the amorphous settlements "across the stream." Outnumbered by the immigrants, Manding speakers and Sienar speakers alike, residents of Koko are united precisely by their "native" status. Paradoxically, most of Korhogo's residents now consider themselves Muslims. In town, at least, the Dyula of Koko are not a Muslim minority as such, but rather part of a Muslim majority. Their minorityhood is experienced in different terms; they are seen as autochthones as opposed to immigrants. Koko's mosque neatly symbolizes this dual affiliation. On Fridays and other major holidays, it serves the whole town; on ordinary days, it is the local mosque of Koko's residents. The site of the mosque and the fact that the office of imam is reserved for a member of spe-

cific kabilas of Koko are reminders that this was the first, and once the only, Muslim neighborhood of Korhogo, and that the residents of Koko, as the town's autochthonous Muslims, have certain symbolic prerogatives.

The Dyula of Koko are thus part of variously defined ethnic, neighborhood, and town Muslim communities, as well as, in a very real sense, of a national Muslim community within Côte d'Ivoire as a whole, not to mention the global community of Muslim believers. These communities are not mere analytical abstractions; the sense of community can always be given a symbolic expression, even though the symbols are often polyvalent and ambiguous. Yet this sense of community depends, in each instance, on a sense of its boundaries; in other words, one defines who one is in terms of who one is not. In the past, Islam had served unambiguously to distinguish the Dyula from their "pagan" neighbors as well as identifying them as part of the great West African Muslim trade network. Nowadays, the Dyula can define themselves only in terms of a series of distinctions—Dyula versus Senufo, autochthones versus immigrants, Muslims versus unbelievers—that crosscut one another. Indeed, religion no longer distinguishes the Dyula of Koko from most of their neighbors. Rather, it is because they are part of a national community of Muslims in Côte d'Ivoire that they are still part of a Muslim minority.

The emergence of a nationwide Muslim community has been a very recent phenomenon. During the colonial period, a more uniform Muslim culture had developed in the immigrant communities in the south of Côte d'Ivoire. There was no sense, however, that this culture was in any way national. On the contrary, it transcended national boundaries, involving Malian and Voltaic as well as Ivoirian Muslims. The creation of an Ivoirian Muslim community, conscious of itself as such, was a function of political rather than cultural change. Independence, and the creation of the postcolonial state, marked the turning point. Pluralism has, of necessity, marked the political style of Côte d'Ivoire, a country characterized by extreme ethnic, cultural, and linguistic diversity

for its size. Approximately sixty different languages are spoken,^[19] but no single group, however constituted, forms a majority. Consequently, the government must avoid appearing to favor any particular group too blatantly; ethnic and regional lobbies are continually clamoring for their fair share. Religions are in a similar situation: neither Catholics, nor Protestants, nor Muslims, nor "animists" —as they are designated in the official statistics—constitute a majority of the population. Religious communities can, like any other community, lobby for government resources. Only the "animists" are excluded from playing at this game and receive no government recognition. This is hardly owing to an anti-"animist" bias on the government's part, but rather to the fact that "animists" are a residual category, rather than a religious community per se. This leaves Muslims in competition with Catholics and Protestants for government resources. They may, for example, seek government funds for the erection of new mosques and churches. The new mosque in Koko is a product of such government largesse, intended as tangible evidence of the government's concern for the Muslims of Korhogo and of the nation at large. Of course, funds for such buildings are not available as a matter of course; religious communities, either at the national or local level, have to lobby for them. The government also allots television air time to different religious groups. Muslims, for example, have an hourly program called "Allahu Akbar" on Thursday evenings, just as the Catholics and Protestants have religious broadcasts on Sundays.

The government, in turn, needs to be wary of the ways in which it awards largesse to any religious community. For example, the construction of a mosque or church in any particular locality may be interpreted as an act of favoritism, not only of one religion over another, but equally of one town or region over others. Consequently, the government has favored the emergence of a national Muslim association, the AEEMCI (Association des élèves et étudiants musulmans de la Côte d'Ivoire, the association of Muslim students of Côte d'Ivoire). By recognizing and sponsoring such an association, by making it the unofficial—or perhaps quasi-official—

voice of the national Muslim community, the government can better channel the lobbying efforts of Muslims. Thus, the AEEMCI produces the weekly Islamic television program. It also holds national conferences for Muslim youths, such as the one held in Korhogo in 1985. Such conferences can only be held with government sponsorship and rely on the government for money, housing, and space. Again, considerable lobbying was necessary on the part of Korhogo's Muslims (including prominent members of the Koko Dyula community, but also, for example, the Lebanese Shiites of Korhogo) both to obtain such funds and to ensure that Korhogo would be chosen as that year's site.

For obvious reasons, government largesse to religious communities takes highly visible forms: television programs, public buildings (the bigger the better), national conferences. In this way, Muslims are made aware of what the government is doing for their religion. They are, by the same token, aware of what is being done for other religious communities. Seen in this light, Catholicism and Protestantism are direct competitors of Islam. A religious community's success in competing for these government resources depends, at least in part, on the size of its membership. Consequently, each religion has a vested political interest in recruitment. So-called "animists" represent a potential pool of converts. The Korhogo region, from this point of view, is particularly promising; though most townsmen claim to be Muslims, the majority of Senufo in the surrounding countryside are not affiliated to any of the three religious groups and remain unregenerate "animists" in the eyes of census takers. There are Catholic and Protestant missions both in town and in the countryside. These missions were established well before independence, and it would be ridiculous to suggest that their current activities are in any sense politically motivated. The Protestant mission of Korhogo is the most conspicuously engaged in a race for converts, but, for the missionaries, this is a race against the Devil for souls, and not a race against other religions for members. Since Vatican II, the Catholic missionaries have abandoned the image of a race in favor of a more ecumenical vision. Catholicism is impressive in terms of the

elaborateness of its hierarchy and the strength of its organization. Korhogo, after all, is a diocese in its own right, with its own bishop.

Compared to Catholics and Protestants, the Muslims of Korhogo have no structured missionary organization whatsoever. There are no professional Muslim missionaries, and consequently every Muslim is a potential missionary. It is through the network of contacts between believers and "unbelievers" that new converts are brought into the fold. For this very reason, I suspect that ordinary Muslims are more sharply aware of the competitive aspect of the relationship between religions in Côte d'Ivoire than are ordinary Catholics, and perhaps even Protestants. In Christianity, the responsibility for recruiting converts falls on the clergy; in Islam, it falls on the congregation. It is the responsibility of the Muslim community as a whole to set an example "unbelievers" would choose to emulate. For many of Koko's Muslims, Catholicism and Protestantism now constitute yardsticks, a means by which they can consciously—indeed self-consciously—assess their own performance. Such comparisons are not necessarily fair-minded; for example, the ideal behavior of Muslims may be compared to a stereotyped and caricatural vision of Christian behavior. Even so, such comparisons reveal a fundamental shift in the way Koko's Muslim community conceives of its place in the world. Formerly, the world of Koko was divided into silama and banmana , into Muslims and "pagan" unbelievers. The "pagans," it should be mentioned, have not disappeared from Koko. The poro society grove in the old Tiembara quarter is still active. Yet the Muslims of Koko are increasingly likely to compare their behavior to that of Catholics or Protestants "across the stream" rather than of their "pagan" neighbors next door. This comparison, however biased, is not always reassuring. One evening, at the end of a funeral sermon, a prominent member of Koko's Dyula community, a man who had personally been very active in lobbying both for the construction of Korhogo's mosque and for the AEEMCI conference, began to harangue the audience. He complained about the

beggars who clustered in front of the mosque during Friday services, aggressively clamoring for alms. The services at Catholic and Protestant churches on Sundays, he continued, were never the site of such sorry spectacles. "A be silamaya dogoya," he concluded: "This belittles Islam."

To conclude, the way in which the relationships between Muslims and non-Muslims has been conceived in Korhogo has changed radically in the twentieth century. Before the colonial period, Islam, like long-distance trade and weaving, was essentially an ethnic monopoly, part of a broader system of hereditary social categories. These hereditary categories accounted for differences in religious practice among Muslims, and not simply between Muslims and "unbelievers," and condoned behavior among certain categories of Muslims that, from a strictly legalistic stance, fell short of orthodoxy. Moreover, the system as a whole actually discouraged conversion to Islam. The relationship between Muslims and non-Muslims was essentially a complementary one, where each category had its theoretical place within the total system. During the colonial period, particularly after World War II, this system simply became obsolete. Changes in the economy had opened up new opportunities while rendering certain hereditary specializations increasingly unprofitable or open to outside competition. Islam simply ceased to be an ethnic monopoly, and the distinction between Muslims and non-Muslims had to be rephrased in an idiom other than ethnicity. The outcome was an increasingly legalistic conception of Islamic identity, with a specific focus on prayer. In a local context, Muslims were those who prayed, non-Muslims were (for the most part) those who continued to initiate themselves in sacred forests, to drink, and to sacrifice to "fetishes." The relationship was an asymmetrical one; "unbelievers" represented a pool of potential converts to Islam, but not vice versa. At the national level, however, non-Muslims were prototypically Catholics or Protestants, competing with Islam for resources and for converts.

By definition, Muslim minorities do not live in a vacuum. The existence of salient distinctions between Muslims and

non-Muslims raises not only the question, Who is to be classed as a "Muslim"? but also, What is the prototype of "non-Muslims"? The way in which each category is defined—and they can only be defined with reference to each other—determines in the first place the nature of their relationship. In Korhogo, we have seen that this has ranged from complementarity to competition. But, at the same time, the relationship bears on the central question, What does it mean to be a Muslim? The very nature of Islam itself as a religion—at least as it is conceived and practiced by particular people in a specific context—depends as much on "non-Muslims" as it does on Muslims themselves.