5
The Ritual Arena

In 1972, when I first came to Korhogo, the controversy between the Wahhabis and their opponents was raging. The issue most hotly discussed was, ought one pray with arms crossed or with arms outstretched? My initial reaction, I must confess, was one of puzzlement. Why should this particular detail matter so very much? Obviously, the manner of praying is a symbol of other differences; however, as academics, we must be especially wary of the temptation to conclude that it is only a symbol of deeper differences. Controversies, after all, are familiar enough to academics, as long as they are controversies about ideas. Had Wahhabis and anti-Wahhabis contended about doctrine, I doubt that I would have found their arguments very perplexing. Of course, there really are differences between the ideas of the Wahhabis and the ideas of their opponents, differences that parties to either side of the debate are quite capable of articulating. One might conclude that each manner of praying symbolizes a set of ideas about Islam and about the world. The problem with such a conclusion is that the ideas were discussed much more dispassionately than the issue of prayer; the symbols aroused deeper emotions than the entities they seemed to symbolize.

Symbols pose a special problem for "intellectualists" like Robin Horton (1971, 1975a, 1975b), who rather self-consciously avoids confronting the problem of symbols and of the rituals in which such symbols are regularly embodied. There are, it must be said, sound reasons for such reticence; the exegesis of ritual symbols—a fortiori those of another culture—is a perilous exercise at best. Provided that, in one way or another, the symbols "express" the cosmology, they do not present an entirely intractable puzzle. But what if the symbols

have no relation to cosmological ideas? In what way, for example, does praying one way rather than another explain, predict, or control anything at all?

In fact, religious controversies among the Dyula, at least in the past fifty years or so, have been in one way or another about ritual, much more explicitly than about cosmology. The lo societies were abolished on the grounds that it was improper for Muslims to offer blood sacrifices to local spirits and to participate in initiation rituals. Few, if any, questioned the existence, the nature, perhaps even the power of the spirits. "Fetishism," for modern-day Dyula, consists, not in believing in spirits, but rather in worshiping them. Religious debate is concerned, not with what Muslims should believe, but with how Muslims should behave. The issues are essentially moral, not cosmological.

Like morality, and unlike cosmology, ritual is concerned with the actions of individuals, with what people should or should not, must or must not do under specific circumstances. Moreover, ritual obligations are in and of themselves moral obligations, whereas it is quite unclear exactly what one is morally obliged to believe as a Muslim, aside from the fact that there is no God but God and that Muhammad is his messenger. Debates about ritual are thus both literally and metaphorically debates about morality; deliberately praying the "wrong" way is immoral, both in and of itself, and because of what such "wrong" prayer expresses. It is precisely for this reason that ritual issues are the ones that incite most passion. The ideas of one's opponents are simply ignorant or silly; their behavior, on the other hand, is objectionable, if not frankly evil.

Of course, there is nothing new about the idea that ritual "expresses" moral notions; since Durkheim and Robertson Smith, such a statement has been a commonplace in anthropology. The problem with the Durkheimian formulation and with most of its successors is that it sees ritual as the expression of the relationship between "society" and the individual; ritual symbols are simply the "collective representations" of this relationship, taken as given. "Society" is taken

for granted, as if it were always clear to which society any given individual or group belongs. Conflict, in the religious domain, thus appears to be anomalous, a symptom, for Durkheim, of "anomie." However, there is nothing at all anomalous about conflict and disagreement, at least in the history of Islam—or, for that matter, Christianity. It would be extremely rash to dismiss differences of opinion about ritual among the Dyula as evidence of "anomie."

The question remains, how do such differences of opinion express different moralities? What, in other words, is at stake besides ritual itself? The answer, I would suggest, is that rituals express, among other things, the existence and nature of moral communities.^[1] Moral communities, however, are by no means identical with "society." "Society," as such, is an abstraction of sociologists and anthropologists; no one is conscious of belonging to a "society" per se, but individuals are quite aware of their allegiance to specific, named moral communities. As we shall see, there are different kinds of moral communities, and individuals and groups may simultaneously belong to moral communities of different orders.

It is important to bear in mind that differences of opinion among the Dyula are not about ritual per se, but about the relationship of ritual to Islam. People are specifically concerned with the question of what Muslim ritual is. In other words, what kinds of ritual are associated with, appropriate to, or required of Muslims, and what kinds of rituals are not? The question is, of course, highly ambiguous. For example, is a ritual "Muslim" if only Muslims perform it, even if it is not required on religious grounds? Such questions, however ambiguous, remain fundamental precisely because Islam provides legitimacy to the existence of the community whose very nature is expressed by ritual; it is Islam, in other words, that makes a community a moral community.

Broadly speaking, Muslims belong to two contrasting kinds of moral community: on one hand, global and universalistic; on the other hand, local and particularistic. In the first instance, there is the umma , the global community of

Muslims, a community that transcends regional, national, political, ethnic, and linguistic boundaries. Within this community, all believers are in fundamental ways morally equivalent. In principle, a single standard of conduct applies to all individuals, a single set of rules by which believers may find favor with God, or else call down His ire. Admittedly, this moral uniformity is not absolute, and Islam gives explicit recognition to certain distinctions, notably between males and females and between adults and children. Even these differences are relativized in the moral domain; ideally, the piety of women is not supposed to differ radically from the piety of men.

There is only one umma , one global community of Muslims, but, of course, there are many Muslim communities. The morality of the umma is of necessity impersonal; only an abstract code of conduct can govern relationships between Muslims throughout the entire world. The local community, on the contrary, is a highly personal realm, governed by regular face-to-face interaction. Within such communities, the differences between members, rather than their equality before God, are of paramount importance. Those distinctions that, if recognized, are at least relativized at the global level—between male and female, adult and child, slave and free^[2] —are of primary importance within the local community. To these, each local community may add a host of distinctions that the global community of Islam in principle ignores entirely: ethnicity, kin-group affiliation, types of occupations, generation, order of birth, and the like. In a fundamental sense, the global community postulates the essential equality of all humans, the local community their essential difference.

Ritual provides a ready-made medium for the expression of such principles of moral equivalence or difference. The moral axioms of ritual action are, as it were, childishly simple: equivalent categories of persons perform identical actions; distinct categories of persons perform different actions. In ritual, even more than in everyday life, there is a

fundamental interrelationship between the theatrical and the sociological notion of "role." It follows that the two kinds of moral communities, global and local, are associated with two different kinds of ritual. In ritual that expresses the importance of the umma as a moral community, all participants should act identically; on the other hand, ritual associated with local communities as moral communities will involve different categories of persons in explicitly different capacities. These two kinds of ritual can justifiably be contrasted as "universalistic" as opposed to "particularistic."

It is tempting to conceptualize particularism and universalism as opposing poles of a single continuum. The implications of such a continuum would be that a community's emphasis on universalistic ritual is in inverse proportion to its emphasis on particularistic ritual. Such a conceptualization is ultimately teleological: as societies evolve from small-scale microcosms to large-scale macrocosms, religions evolve from polytheistic particularism to monotheistic universalism. In the long run, the nature and direction of change should be predictable. The crucial assumption on which such a unidimensional model is built is that microcosm and macrocosm, particularism and universalism, are polar opposites. There is, of course, another possibility, which is that they might constitute separate dimensions. It is not necessarily a contradiction to assert that all Muslims are alike, and that they are different. It is precisely as Muslims that they are alike, and in terms of other social identities that they may be different.

It is through ritual that groups and individuals express the kinds of social identities they hold to be salient and morally acceptable, if not enjoined. In a fundamental way, attitudes toward ritual have as much to do with religious sensibilities as with doctrine; rituals may seem meaningless or empty as well as morally objectionable. Over the years, religious sensibilities in Koko have clearly changed, though not always in obvious and easily predictable ways; moreover, it is certainly not now, if it has ever been, the case that such sensibilities are uniform. The differences are differences of emphasis, of

universalism as opposed to particularism—of what kind of universalism, what kind of particularism, and of their interrelationship.

Islamic Particularism

Until the middle of the twentieth century, particularistic ritual of one kind or another pervaded religious practice among Dyula Muslims. It served, in other words, to express and ultimately to legitimate social distinctions between categories of groups and persons within the local community. This is not to say that universalistic ritual was absent, or even unimportant. Islam is, of course, a universalistic religion. The local Muslim community must, of necessity, be part of the umma , of the global community of Muslims. If not, what sense does it make to call oneself Muslim? Of course, if being Muslim is to have more than a purely local meaning, it must have some tangible, universally recognizable, signs; these are, quite obviously, prayer, fasting, abstention from alcohol and ritually unclean meats—in short, the various ritual prescriptions of the sunna. All Muslims—whether or not they perform these actions punctiliously—recognize them as quintessentially "Muslim." All of these signs, all of these ways of behaving, were familiar to the Dyula in and around Korhogo. As members of a Muslim trading diaspora, they were, in any case, involved in a supra-local economic as well as religious system. Yet even those very actions that, in a supra-local context, expressed a "universalistic" content served, in a purely local context, to distinguish between social categories of Muslims, and not only to differentiate Muslims from unbelievers. As we have seen, only the mory , the "scholars," were bound to observe this ritual code scrupulously. The tun tigi , the "warriors," might observe the code only sporadically.

The lo societies into which all male tun tigi were initiated were even more expressly particularistic. The term lo referred, not only to the initiation society as a whole, but to specific named spirits, as well as to the particular masks representing them. Each spirit, each mask, was ultimately the

property of a specific clan ward, rather than of the initiation society as a whole. As a result, each local society (and there might be more than one in a large village or quarter) took the form of a fortuitous constellation of the spirits and masks owned by its constituents units. The particularism that characterized the lo societies, as well as the poro societies of their Senufo neighbors, was expressed in the idiom of ownership. This idiom operated at a plurality of levels. Within a single lo society, a mask was the property of a single ward. Within the village as a whole, it was the property of one particular lo society; within the region, it was the property of a specific village. Ultimately, masks were ethnic properties, Dyula masks as opposed to Dieli or Senufo masks.^[3] Some masks were unique; others were commonly found in many Dyula communities. Some masks were harmless, and acted the buffoon in public to amuse the crowds; others, the sight of which might even be forbidden to all non-initiates, had the power to kill. Not surprisingly, the most powerful masks tended to be the unique ones, as rights over their exclusive ownership were more jealously guarded. Ownership of such powerful masks was a matter of considerable local pride, a pride the mory could share with their tun tigi neighbors even though they did not participate directly in the initiation ritual.

The ritual practices of the mory and of the tun tigi were, in a sense, complementary. The ritual observances of the mory provided a link between the local Dyula community and the Islamic world in general; the initiation societies of the tun tigi furnished visible signs, in the form of masks, of the differences between one community and another, indeed between one clan ward and another. Each segment of society, mory and tun tigi , participated to some extent in the symbolism of the other. After all, tun tigi were Muslims, who could and sometimes did pray; mory , as members of specific villages or quarters, also, if only indirectly, "owned" masks. Even so, the local community, as a Muslim community, was characterized by two very distinct patterns of ritual. Adherence to the sunna , while it might be a global sign of Muslim identity, was associated locally with a specifically mory identity. Only

some third category of ritual could adequately signify a Muslim identity in a local context, a category of ritual in which mory and tun tigi participated on an equal footing. This category consisted, in the first place, in the celebration of Muslim calendar holidays: sun kalo , "the month of fasting"—that is, Ramadan; tabaski , the annual slaughter of sheep during the season of pilgrimage to Mecca; donba , commemorating the birth of the Prophet. Along with this celebration of the Muslim year was the celebration of a Muslim life, the rituals of birth (den sereli ), marriage (furu ), and death (su ko). In the broadest sense, a Muslim was one who celebrated the annual Muslim holidays and who was named, married, and buried in a Muslim fashion. Participation in these rituals distinguished silama from banmana , Muslim from unbeliever, far more effectively than adherence to the sunna.

These celebrations of the communal Muslim identity of the Dyula were, by and large, festive occasions,^[4] characterized by drumming and dancing of all sorts. Lo society masks were by no means excluded from these Muslim holidays, but would emerge in particular to celebrate the end of Ramadan.^[5] Characteristically, the name the Dyula use for the Prophet's birthday, donba , literally means "the big dance." These occasions were marked by decidely particularistic manifestations, emphasizing social distinctions within the Muslim community at the same time as they celebrated its overall Muslim identity.

The donba celebrations in the village of Kadioha, which I witnessed in 1985, are a good example of "particularistic" Muslim ritual. Unlike donba celebrations in neighboring Dyula communities, those of Kadioha have a particularly militaristic flavor. Kadioha was the seat of one of the few Dyula-ruled chiefdoms in the Korhogo region,^[6] a fact its peculiar donba celebration seems to flaunt. However, the atypicality of Kadioha's celebrations is, in a larger sense, quite typical. Each village—in some cases, each clan ward—takes pride in its own particular way of celebrating calendar holidays, marriages, and so forth. The differences are in some cases minute, but they are symbolic markers of an esprit de minaret ,

a means of asserting the village's or ward's distinct identity—its distinctiveness, as it were. To return to donba in Kadioha, the festivities begin after midday prayer, when bands of adolescent boys, grouped clan ward by clan ward, roam dancing throughout the village, girdled with a cloth, their faces daubed white, sporting bows and arrows (mostly the rubber-tipped variety in 1985), with which they gesture menacingly. This was the only instance in which I witnessed mock warfare in Dyula ritual of any sort or in any place; even in Kadioha, it is strikingly unusual, restricted to this one day each year. At the same time, adolescent girls, again grouped ward by ward, also dance throughout the village, separately from the boys. After the evening prayer, the military glory of the chiefdom is once again celebrated. One by one, young men representing each ward dance, brandishing a spear, before an assembly of elders. Each, in turn, commemorates the military exploits of his ancestors. After every ward has had its turn, there is a procession to the hilltop that overlooks the village, which houses jinns. These particular jinns, it should be noted, are Muslims, and they are credited with repulsing a "pagan" Senufo attack on the village; so it is not, after all, inappropriate to honor them as part of the festivities of a Muslim holiday. After the procession returns, elder men and Koranic students adjourn to different places in the village to chant liturgical texts in Arabic. They gather, in fact, at the ceremonial meeting places of the various makafus ; in other words, one's kin-group membership determines where one will adjourn. The chanting, which continues until daybreak, if not even later, partakes of the festive tone of the entire occasion. Groups of Koranic students and younger men may jovially compete with one another to see who can chant the loudest or most melodically. The chanting is punctuated by the periodic distribution of refreshments, as lavish as possible, such as grilled meats or sweet drinks. Like the chanting, there is a competitive edge to the catering, with prosperous families and individuals vying to provide the most expensive and elaborate food in sufficient quantity so that everyone may be served.

Clearly, these festivities serve to mark the existence and importance of a variety of social identities and distinctions situated at different levels: Muslims—the village as a whole, but also its nearby jinns, its protective spirits—in contrast to their "pagan" Senufo neighbors; the village, as opposed to other nearby Dyula villages, in terms of its glorious (or at least glorified) military past, as the seat of a chiefdom; the various clan wards of the village, who dance separately, and who compete with one another in providing the most lavish refreshments during the sessions of chanting. Within the village as well as within each ward, gender and generational differences are also given ritual expression. Adolescent males and females dance in separate groups. Juniors dance, elders are danced to. Younger men are, at least symbolically, warriors, while elders preside over, if they do not participate in, liturgical chanting. (In this, as in other contexts, young men relatively advanced in Qur'anic studies are partly assimilated to elders; they, too, have special knowledge that gives them privileged access to God.) Thus the festivities as a whole recapitulate many of the major divisions of the social universe: Dyula Muslims versus Senufo "pagans"; village and clan ward affiliation; males versus females; elders versus juniors; even the relatively wealthy versus the relatively poor, in terms of who provides refreshments for whom.

Neotraditionalism

As long as the local Dyula community and the local Muslim community were one and the same, celebrations of Muslim calendar holidays such as the donba festivities in Kadioha could bear a plural symbolic weight, simultaneously expressing the allegiance of the entire community to Islam and the internal division of this community into its salient components: male and female, elder and junior, slave and free, members of one kin group or another. By the early 1950s, this was certainly no longer the case in Korhogo, which attracted a sizable community of Muslim immigrants, many from outside Côte d'Ivoire, and where a growing number of Senufo

were converting to Islam. Of course, such Muslims observed the same Muslim calendar holidays, but they did not, by and large, participate in the festivities that marked these occasions for the Dyula. As a result, such celebrations could no longer symbolize a Muslim identity per se. Rather, they were typical of the Dyula, but "Dyula" and "Muslim" were no longer synonymous in a local context. In a real sense, the festivities ceased to be religious and became part of the local folklore, like Mardi Gras in certain Catholic communities.

There is, indeed, a decided ambivalence concerning this whole domain of ritual. In an ethnically plural Muslim community, these celebrations seem to distinguish Dyula Muslims from other Muslims, rather than, as in the past, Dyula Muslims from Senufo unbelievers. As folklore, as the self-conscious expression of custom, of tradition, of one's own "roots" as opposed to the roots of one's neighbors, they are more frankly and flagrantly particularistic than ever. The question is no longer whether they are part of Islam, but whether it is proper for Muslims to participate in them. These were the very grounds for abolishing the lo societies among the Dyula tun tigi , as well as, somewhat later, excision ceremonies for adolescent girls; such practices, it was argued, were "pagan" and not Muslim. While there were, of course, additional extra-religious motives for abolishing initiation and excision ceremonies, the principle that they were religiously improper was upheld and vindicated. Singing, drumming, and dancing, particularly on occasions such as donba or during the month of Ramadan, have been called into question for similar reasons. The scholars of Koko are quasiunanimous in voicing their disapproval. Muslim holidays, they argue, are serious occasions, and such frivolities only distract participants from true religious concerns. Such objections are, for the most part, unchallenged and ignored. The scholars, while they have failed to persuade many people to cease the festivities, have indeed convinced them that the singing and dancing have nothing to do with religion.

Perhaps the most paradoxical outcome of this process of the "folklorization" of formerly religious ritual has been the

resurgence in recent years of lo masquerading. When I was first in the field, in 1973, I concluded that such masquerading had almost completely ceased. Exceptions were negligible, as when Dyula youths in Kadioha borrowed the costumes of their Senufo neighbors to run around the village at the end of Ramadan, playing the role of bogeymen to smaller children and clowns for adults. In 1984, on my return to the field, I began hearing reports of much more elaborate masquerading activity. At first, I assumed—much to my chagrin as an ethnographer—that such activity had escaped my attention a decade before. The truth turned out to be more surprising. During the preceding decade, there had been a revival of lo masquerading in certain Dyula communities. This revival involved a collaboration between two generations, the old men and the adolescent youths. The old men were the last generation to have been initiated, and, as such, some were nostalgic about the lo societies. The middle-aged men, the generation that followed, consisted precisely of those who had revolted against initiation, and who were responsible for the demise of the lo societies. The fascination of adolescents seems, at first sight, to be evidence of a pendulum swing of opinion, generation by generation. Yet the reality is more complex. Lo masquerading represents something entirely different for adolescents and for their elders. The adolescents have never really experienced the lo societies in operation, even at a distance. On the other hand, the masquerades of the poro societies of their Senufo neighbors are quite spectacular. Such masquerading, rather than the whole process of initiation, fascinates the adolescents. The lo masks have been revived, but not the lo societies. In many respects, lo masquerading, like the dancing during Ramadan and donba , has lost its religious content and become "folklore." What was, for a whole generation in revolt against initiation, a symbol of unacceptable "paganism" has become, for a new generation, a symbol rather of "tradition," of local pride. Not surprisingly, attitudes toward this kind of masquerading are highly ambivalent. There is no question, for example, of reviving masks in Koko; those communities where masquerading

has resumed are all villages or small towns. Dyula in Koko sometimes comment superciliously that their cousins in the bush are somewhat backward in religious instruction, and that if they only knew better, they would abandon masquerading. Clearly, such behavior, even more than dancing during Ramadan, is not quite proper for good Muslims. On the other hand, Dyula in Koko are not only in constant contact with groups and individuals involved in lo masquerading, but they regularly attend events where such masquerading takes place. Not even the scholars publicly condemn the masquerading.

One way or the other, neither the Muslim holiday festivities nor lo masquerading symbolize the religious identity of participants any longer. No one is obliged to participate; they are largely, if not purely, for enjoyment. They are neither Muslim nor pagan, but rather Dyula, and often typical of a specific local community or even a single clan ward. They remain ritual expressions of particularism, but they have been desacralized. Such expressions are no longer legitimated by the religious idiom; on the contrary, from a religious point of view, their propriety can be called into question, particularly as they distinguish the Dyula from their Muslim, rather than from their "pagan," neighbors.

It follows that the primary ritual symbols of religious identity must now be symbols the Dyula share with these new Muslim, non-Dyula neighbors. These are, of course, the ritual expressions of Muslim universalism, of the umma , the global Muslim community, the symbols that once, locally at least, typified the mory rather than all Muslims per se. The most compelling symbol of this universal Muslim identity is prayer (seri ), the regular performance of the five daily prayers. The symbolic equivalence of prayer and Islam is embodied in ordinary conversation. One does not generally ask about a stranger, "Is he a Muslim?" ("Silama lo wa?"), but rather, simply, "Does he pray?" ("A be seri ke wa?"). Prayer is by no means the only such symbol: abstaining from alcoholic beverages and from the worship of "fetishes" (jo ) through blood sacrifices, as well as fasting during the month

of Ramadan, have a similar function. Still, of all these symbols, prayer is far and away the dominant one. This is because prayer is a positive and public action, which can readily be observed. One can observe someone drinking alcohol; one cannot literally observe someone abstaining. The importance of visible signs of compliance explains why Dyula spit so frequently in public during the daytime at Ramadan. The rationale is that a person who swallows his own saliva is breaking the fast, and while one cannot observe someone fasting, one can observe him spitting. Spittle is in this way turned into a tangible (sometimes all too tangible) sign of the refusal to eat or drink during the daytime.

Prayer, then, is the visible sign of the proper observance of the code of the sunna , and thus of true inclusion in the umma , the abstract community of believers. But the sunna itself is a complex code, embodied in texts written in Arabic, a foreign language, accessible only to scholars, whose interpretation is indeed subject to scholarly dispute over legal subtleties. Unfortunately, it holds true in Islam that ignorance of the law is no excuse. For the ordinary believer, partly if at all literate in Arabic, this possibility of ignorance can lead to a sort of existential dilemma. In principle, good faith is not enough; one must constantly strive to know as precisely as possible the details of the code one is morally bound to follow, details that are not accessible firsthand but only through the medium of scholars (karamogos ) who can translate them into the vernacular. This anxiety extends to the very domain of prayer itself. Prayers are not all equal. Particular prayers may be more or less efficacious, depending in the first instance on the time and circumstances in which they are performed. More important, prayers may be valid or invalid; a lapse, even involuntary, in ritual detail may "spoil" a prayer. These preoccupations are reflected in the sermons that have relatively recently become an integral part of Dyula funeral ritual. Scholars frequently stress the importance of prayer in general, but equally the necessity of praying correctly: the correct pronunciation of prayers in Arabic, the correct techniques for performing ablutions, indeed what may or may

not render the water used for ablutions (seri ji ) ritually impure. Nor may one conclude that these details reflect the concerns of scholars as opposed to their (often captive) audiences. Bilingual manuals in Arabic and French, replete with illustrations, explaining the details of prayer and ablutions and transcribing Arabic prayers more or less phonetically into French, are readily available in the marketplace and are eagerly studied by those pious Dyula whose literacy in French surpasses their knowledge of Arabic.

At first glance, it might seem that the Dyula have neatly compartmentalized ritual into two mutually exclusive domains: "religion," characterized by a preoccupation with universalistic ritual, notably prayer; and "tradition" (or, as the Dyula would call it, lantan , "custom"), the exuberant celebration of particularism through singing and dancing on calendar holidays. However, this compartmentalization breaks down during the life-crisis rituals associated with birth, marriage, and death, where it is not always so easy to separate "custom" from "religion." Of course, life crises are marked, not by isolated rituals, but by whole sequences of ritual events, some of which are unambiguously "customary," others unambiguously "religious." Thus the singing and dancing that accompany births and weddings, but also the funerals of elders, are clearly in the "customary" domain. On the other hand, Islamic law prescribes quite specific burial procedures for Muslims, and their public observance obviously falls within the religious domain. Still, aside from such obviously "customary" or "religious" ritual manifestations, all life crises involve the elaborate public distribution of prestations of various kinds, often on more than one occasion. The gatherings where these prestations are distributed are inevitably formal, and their sobriety contrasts with the exuberance of the singing and dancing, which are often going on at the same time. More important, one essential category of prestations involved on these occasions—indeed the central category involved in funeral ceremonies—is labeled saraka , derived from the Arabic term sadaqa. For example, as part of the fortieth-day ceremonies at a funeral, saraka is distributed

to the close kin of the deceased, particularly those not directly involved in the distribution of prestations (e.g., widows); the members of the senior generation of the deceased's kabila , in strict order of seniority; Islamic scholars present during the ceremony, in order that they may afterward recite blessings (duau ) for the deceased; representatives of all the other kabilas in the officiating makafu ; representatives of all the other makafus (if not kabilas ) in the village or Dyula quarter; representatives of all other villages or quarters present at the ceremony. The number of groups represented is often surprisingly large, if only because persons invited in their individual capacities as friends, kinsmen, or affines of the deceased or of members of the deceased's entourage are nevertheless treated as "representatives" of one social group or another for the purposes of the ceremony. In addition, various categories of joking partners of the deceased will snatch the portions destined to them, in ritualized mock theft: senanku , members of patronymic groups who joke with one another; "grandchildren," classificatory and real; and (in the case of a horon , a free individual) worossos , slaves "born in the compound," as opposed to purchased or captured slaves.

It is significant that all segments of the village or Dyula quarter must be represented. In this way, the very identity of the local community as a whole, in terms of its component parts—makafus and kabilas— is not only given symbolic expression but is in a fuller sense legitimated. Paradoxically, the moral unity of the community is expressed precisely in terms of the distinctions between different categories of members: members of different kabilas , of different generations, of free or slave descent. At the same time, the ritual concretely acknowledges the individual network of ties—of kinship and affinity, but also of friendship, clientship, and so on—linking the deceased and members of his immediate entourage to other individuals in the same kabila , to members of other kabilas in the same village or quarter, and finally to members of other local communities. A rigorous protocol governs the order in which prestations are publicly presented, as well as the relative amount distributed. In most

instances, these are token prestations: small change, a plate of cooked food, or (in town, where time and fuel are increasingly valuable) a handful of grain; however, given the large number of groups and individuals entitled to a share, the total amount distributed is never negligible. These prestations, while they are not, strictly speaking, a religious obligation, are nevertheless construed as a religious act, a gift of charity. Not only does the giver thus accumulate merit, but the ostensible purpose of the whole ritual is to obtain the blessings of the entire community, so that God may grant a long and happy life to a newborn infant, harmony and children to a marriage, or that He take mercy on the soul of the deceased in the afterlife. For these reasons, the ostentatious distribution of prestations constitutes an essential part of what is perceived locally as a "Muslim" wedding or funeral.

These distributions of saraka during life crises, like the festivities during calendar holidays, define individual participants in terms of their salient attributes: gender, generation, age, free or slave status, membership of specific kin groups and villages. Formerly, both kinds of ritual were unambiguously perceived as "Muslim" and served to mark a Muslim identity locally. Now, however, their "Muslimness" is perceived somewhat differently. Singing and dancing on calendar holidays are "Muslim" only because of the time when they take place. The occasion is intrinsically Muslim, the activities are not. By way of contrast, the occasions on which saraka is publicly distributed—births, weddings, and deaths—are not intrinsically Muslim. Rather, it is because certain of the prestations are defined as saraka that the ritual assumes a specifically Muslim character. To give saraka is a religious act, an act of piety, but, equally crucially, it is not an obligatory act, from a strictly religious point of view. Only Muslims give saraka (at least during life-crisis rituals),^[7] but one can, in principle, be Muslim without being obliged to give saraka on any particular occasion.

In short, the attitude toward ritual to which the majority of Dyula in Koko subscribe, an attitude that might be labeled

"neotraditionalist," admits of two tiers of religious ritual. The first, universalistic tier is typified by prayer. It is prayer, along with the observance of the sunna , which prayer expresses, that distinguishes Muslims from unbelievers, both locally and globally. The Dyula are Muslims because they are part of the umma , and not simply, as in the past, because they are Dyula. However, a second, particularistic tier, exemplified during life-crisis rituals, legitimates the identities of individual Dyula as members of local communities, and, as such, of kabilas , of generations, and so on, in religious terms. Islam still legitimates the local community, both as a whole and in terms of its salient internal distinctions. But the local community is now conceived as only a part, rather than as a microcosm, of the umma , the global community of believers.

Reform

In a sense, neotraditionalism seeks to have its cake and eat it too, to function at the most particular and most abstract level, by positing the coexistence of two conceptually distinct moral communities: the community of believers and the local community, the ritual practices of the one emphasizing the intrinsic sameness of all Muslims; those of the other, their difference. The possible logical contradictions implicit in such a stance are not necessarily felt as such; in different contexts, one or the other register takes precedence. However, it is also possible to attempt to point out and to refuse to tolerate perceived contradictions, by asserting that only one of these registers—the universalistic—is "truly" Islamic, that Islam recognizes the local community only as part of the umma as a whole and not in and of itself. Such a critique excludes all particularistic ritual from the realm of Islam. As we have seen, such a process has already taken place with respect to many particularistic rituals, which have been relegated to the status of "custom," lantan. In its weaker variant, the reformist critique simply carries the process to its logical conclusion, banishing all ritual manifestations of particularism to the

limbo of "local tradition," where Islam is, in the last analysis, irrelevant. In its stronger variant, however, such rituals are labeled anti-Islamic, a breach of both the spirit and the letter of the law.

The reformist stance has typified the various challenges to the Suwarian tradition of scholarship, a tradition that at present is solidly "neotraditionalist." Indeed, it was in its most radical form, exemplified by the Wahhabi movement, that reformism first made its appearance in Korhogo. The Wahhabis inveighed against all forms of particularistic ritual: the distribution of saraka during life-crisis rituals; Sufi orders; saint worship. Their opposition to such forms of ritual was perfectly consistent with their stance that hereditary distinctions between believers, in terms of ethnic origin, "caste," and free or slave status, had no place in Islam. Not surprisingly, the Wahhabis attracted converts among groups who were, in one way or another, hereditarily stigmatized by other Muslims as slaves, "casted" individuals, or "pagans."^[8] On the other hand, Wahhabism was also associated with well-to-do merchants. The combination is not really surprising. The pax colonia , by undermining many precolonial monopolies, opened up new avenues of opportunity, which such stigmatized groups were often quick to seize upon. Particularistic ritual, by emphasizing the salience of hereditary identities, condemns such groups and individuals to secondclass status as Muslims in spite of their prosperity. The Wahhabi ideology lends legitimacy to this newfound wealth, while at the same time strongly condemning ostentation, not only in the form of conspicuous ceremonial largesse but also, for example, in clothing. This refusal of ostentation is only apparently paradoxical, for ostentation legitimates rather than creates one's social position—in order to maintain one's standing, one may be obliged to spend more than one can afford—while it also sanctions a certain ethic of reciprocity, where prestige lies in giving rather than in possessing. To the Wahhabis, particularistic ritual is wasteful and vain, a sacrilegious exercise in self-aggrandizement. To their opponents, the Wahhabis seem stingy, while their conspicuous auster-

ity—white flowing robes for men and, increasingly, veils for women—seems simply a transparently disguised form of "reverse" ostentation.

Wahhabi ritual focuses exclusively on the universalism of the sunna , and consequently prayer is the focal symbol for defining the global community of believers. Yet if both Wahhabis and neotraditionalists stress the all-importance of prayer, what is to distinguish them, save that the Wahhabis abstain from particularistic ritual? The solution they have chosen is to pray differently. Neotraditionalists, following the Maliki rite overwhelmingly predominant in West Africa, pray with arms outstretched. Consequently, the Wahhabis pray with their arms crossed, in the fashion of the other three Sunni schools. Not surprisingly, neotraditionalists have focused their vehement disapproval of the Wahhabis on this manner of prayer, for it is in this way that the Wahhabis implicitly define themselves as belonging to a community of believers that excludes the neotraditionalists.^[9] Thus, their critics maintain that the universalistic ideology of the Wahhabis translates itself in practice into a form of (thinly) disguised exclusivism. Whereas they reject the claims of the majority of Muslims to belong to the true community of believers, their opponents characterize—or rather caricature—them as wealthy parvenus jealous to preserve their newfound prerogatives. If Wahhabis reject those religious rituals that legitimate the local community as it is "traditionally" constituted, it is—so their detractors claim—in order to form another sort of community where newfound wealth erases any stigma associated with one's origins.

Although the Wahhabis have successfully established a foothold in Korhogo, they have made few inroads among the Dyula of Koko. Nevertheless, a growing number of Muslims among the Dyula have come, if not to reject outright, at least to disparage particularistic ritual in Islam. This is especially true of younger individuals with Western-style schooling past the primary level, the individuals attracted, as we have seen, to the AEEMCI. The semi-official status of the organization ensures that it avoids the kind of controversy that the

Wahhabis openly seek. Rather than condemning particularistic ritual, it ignores it, stressing exclusively the practice of the sunna. This is a stance local scholars do not—indeed cannot—find objectionable, particularly as the association has been very diplomatic about recruiting their support. Younger members of the Western-trained elite, stationed for the most part away from Korhogo, still retain close ties with their families. Yet their social position, their place in a hierarchy of privilege—modest as it may often be—depends on a bureaucracy, on a system of values totally separate from and out of the control of the local community per se. This is not to say that their community of origin is of no importance to them. Strong affective bonds link them to kinsmen and friends. Moreover, there are individuals in or from Korhogo, indeed from Koko, with varying degrees of political connections and influence in the world beyond the local community, and who can, under certain circumstances, give a boost to their careers. For these young men, Koko is the hub of a network of friends, of kin, of useful connections. It is not, on the other hand, primarily a mosaic of kabilas , of elders and juniors, slaves and freemen. In terms of such traditional criteria of identity, they remain very junior, and as such their voices carry little weight in the affairs of the local community. They have appreciably more wealth and access to politically influential individuals than most members of their home communities, but not enough of either to merit treatment as quasielders. In any case, they are largely unconcerned by the very system of values that particularistic ritual seeks to underscore and dramatize. Neither their careers nor their daily lives depend very much on the prestige or influence they enjoy in Koko. For such individuals a more purely universalistic Islam suits their needs, particularly in a country where Muslims are a minority; the bonds of prayer constitute in and of themselves a link with their home communities that sets them apart from many of their fellow citizens from other regions. Yet their place in life is determined in the national, not the strictly local, arena. The sums that they find themselves obliged to spend for the marriages and funerals of kin are

both necessary—they are still a precondition for effective membership in one's "home" community—and onerous, if not wasteful, in that they serve to articulate the position of groups and individuals within a system of values to which they no longer subscribe without reservations.

Secularism

So far we have been able to distinguish between various attitudes toward ritual in terms of their relative emphases on universalistic ritual, stressing the membership of participants in the global community of believers, and on particularistic ritual, stressing the place of individuals within local, face-to-face communities. In fact, for the majority of Koko's Muslims—those I have labeled "neotraditionalists,"—both kinds of ritual are crucial. But is it possible, conversely, to identify a set of attitudes which de-emphasizes both kinds of ritual? For example, if one can be a "practicing Muslim," one can, by implication, be a "nonpracticing Muslim": one who drinks alcoholic beverages, who does not pray regularly or observe the fast of Ramadan, and so forth. Superficially, the behavior of a modern-day "nonpracticing" Muslim would seem very much like that of the tun tigi in the past. The resemblance, however, is only superficial. In the first place, the tun tigi were members of a hereditary category of Muslims; in the second place, ritual—calendar holiday festivities and life-crisis rituals, if not the prescriptions of the sunna— defined the individual as a member of the community of Islam.

"Nonpracticing Muslim," however, is not an entirely adequate characterization, placing, as it does, primary emphasis on behavior, rather than focusing on the significance of behavior within an overall system of values. An individual who fails to pray regularly may simply be a lax Muslim who accepts without question that prayer is a fundamental expression of "being Muslim." On the other hand, one can consider prayer as, at best, an act of personal faith; at worst, a social obligation; not, at any rate, as the core of Muslim identity. From this perspective, a Muslim is anyone who identifies as

such; whether and how this identification is translated into ritual terms is one's personal business. In this sense, the category "Muslim" functions much like an ethnic label: just as forms of speech, dress, and so forth, may symbolize ethnic identity but are not equivalent to it, so Islamic ritual may symbolize Muslim identity without defining it. Whether Islam is conceived as a faith in the full sense of the word—a question of interior commitment—or as an identity label, it is either more than or less than—in any case other than—a moral community.

This attitude bears close resemblance to a Westernized secularist mentality and one would expect to find it, if at all, among Western-trained members of the Dyula community. Public expression of such an attitude in Koko is still inconceivable. Public declarations of sympathy for the Wahhabis are unpopular, but they are not unknown. However, Wahhabism is at its very root a public commitment; secularism, to the extent that it conceives of Islam as a private faith and a public identity, is perfectly consistent with a pattern of behavior that might uncharitably be characterized as hypocritical. But, if no one flaunts such an attitude in Koko, the fact remains that certain categories of individuals are commonly suspected of practicing Islam only in public contexts—that is to say, of not praying when they can get away with it unnoticed. Suspicion falls first of all on the young—not all of them, by any means, for some have a reputation for being serious and others of being frivolous, if not good for nothing. In and of itself, the laxity of the young does not call into question the Islamic values of the majority. On the contrary, particularistic ritual emphasizes the difference between the young and elders (even incipient elders). The young are expected to behave differently until they begin to be socialized as elders, until their voice has at least minimal weight, at the moment they enter the lowest rung of the hierarchy of eldership by marrying and having children. The young may be sermonized ad nauseam about their immoral behavior, but to a certain extent such behavior is tolerated and even expected, provided they grow out of it.

But another category of individuals represents a far greater threat: those with high-paying employment outside the region. The better off they are, the more likely they are to fall under suspicion. These are the individuals who have not only received a Western education but have reaped its full benefits. They live in expensive villas and own modern appliances and fancy cars. More important, their very success is measured by their belonging to a social milieu comprised, in Côte d'Ivoire, mostly of non-Muslims, and one that aggressively sports a Western lifestyle. This success is admired in Koko, and its outward signs—villas, cars, refrigerators, videocassette recorders, and so forth—are exhibited whenever possible by individuals who, by the neotraditionalist standards of the majority, rate as unambiguously pious. But this very admiration is, in a sense, only a passive acknowledgement of success in an arena outside the local community's control. Success is not measured in terms of the values of the local community, as expressed in particularistic ritual. On the contrary, the values of the local community must adjust to forms of success outside its ability to define, in and of itself. These individuals are members of an elite—economic, bureaucratic, and political—as far removed from the abstract community of Muslim believers as from the local community. Religion, ethnicity, and local community of origin are by no means irrelevant; recruitment may depend on locally anchored networks of patronage, or alternatively may reflect attempts to achieve some sort of representative balance at the national level. Nevertheless, the elite is constituted according to a system of values that is both national (if not international) and secular. This is not to say, by any means, that members of this elite who happen to be Dyula necessarily reject a universalistic standard of Islamic values in favor of a purely secular one. However, more than anyone else, they are in a position to do so if they please.

Paradoxically, certain members of this elite participate quite actively in particularistic ritual, as systematically as (and perhaps more enthusiastically than) in prayer. Such participation is essential for anyone with political aspirations.

Establishing one's local identity in this way is a precondition for developing a base of support in one's home community. Active displays of generosity on such occasions translate success in the national arena into local terms. This double accumulation of prestige is a luxury the wealthiest and most powerful can easily afford. However, this is not the case for those who, in national terms, are simply well off but not really rich. Such individuals are faced with three conflicting sets of demands on their resources: maintaining a lifestyle, in terms of consumer goods, appropriate to their station; aiding close relations (friends, kin, affines) in their networks, particularly in their home communities; and distributing gifts and money generously on appropriate occasions in local particularistic ritual. It is far harder for the moderately prosperous, unlike the very wealthy, to have their cake and eat it too by satisfying all three sets of demands simultaneously. Yet, even if they participate actively and enthusiastically at marriages and funerals, wealthy bureaucrats are in a very different situation from ordinary members of the local community. Their situation is defined by values and standards independent of the local community; they have the freedom to choose whether or not to translate this prestige into local terms. Local standards of values remain in any case secondary, and their prestige in local terms depends on their position in the national arena. Their participation in local particularistic ritual is instrumental, a demonstration that they can impinge upon, rather than that they are subjected to, the local hierarchy of prestige.

Conclusions

As we have seen, each type of ritual stresses an individual's commitment to a specific kind of moral community, and consequently to a specific set of values. At stake are three different ways of defining the individual: as a member of a specific kin group within a specific village or quarter, as male or female, elder or junior, free or slave; as a "Muslim," more or less pious depending on the rigor with which he or she ob-

serves the ritual and moral code of the sunna ; and, finally, in terms of type of occupation, amount of wealth, degree of Western education, and political clout—in other words, for want of a better word, "class."

The first system of definition, in terms of village and kin-group affiliation, generation, gender, and so on, only retains its salience within the "microcosm," the close-knit community where everyone knows everyone else, where any given individual's position in the local hierarchy depends on a combination of ascribed characteristics and "reputation" together determining the weight his or her voice will carry in local affairs. The third system, in terms of the attributes of "class," transects not only the local but also the global "Muslim" community. An individual's position is expressed in terms of signs that have no place in religious ritual, either particularistic or universalistic; such signs include the amount and nature of goods they own and consume, the way they speak French, and ultimately the various means by which other individuals with established positions in the hierarchy mark them as peers, superiors, or inferiors. In other words, both the close-knit local community and the national system of "class" are essentially hierarchical. Grosso modo, to the extent that individuals are in a position to choose the system of values by which they wish to evaluate themselves and be evaluated by others, their stake in either system depends in large measure on the rank they can plausibly claim or, at least, reasonably aspire to. These two systems cannot, however, be placed on an equal footing. As we have seen, a high position in the national arena can be translated and transferred into a local idiom. The reverse is obviously not the case.

This leaves the second system of definition, in terms of a global "Muslim" identity. To the extent that one can meaningfully speak of a principle of hierarchy here, it is a hierarchy of piety, apparently divorced from such criteria as birth or class. The great in this world will not necessarily find favor in the eyes of God in the next. But this is not to say that the last shall be first—the Sermon on the Mount is not part of

Islam's debt to Christianity—or even that the global moral community of Islam is in some deep sense egalitarian. The moral hierarchy of piety may supersede one's rank in this world, however reckoned; it does not deny it. It is significant that the categories of individuals most readily attracted to varieties of reformism occupy positions that are, in certain respects, anomalous in terms of the local hierarchy of influence and the national hierarchy of class: those whose wealth or Western education places them "above" their station in the local hierarchy without justifying their inclusion in the upper echelons in the national hierarchy. This applies to the Wahhabi merchants, who may be wealthy but lack Western education, and who are often (but not inevitably) members of low-ranking hereditary categories in the "traditional" local scheme of values. But it applies equally well to younger secondary school and university students and graduates, who are educated but whose current prospects of achieving considerable wealth or political influence are severely limited, and whose very youth, in the absence of real wealth or clout, places them for the time being toward the bottom of the local hierarchy of respect and influence. By observing the sunna , as they interpret it, more rigorously than anyone else, they assert their moral superiority both over traditional elders, guilty in their eyes of mingling Islam with particularistic ritual, and the Westernized elite, suspected of a tepid commitment to Islamic strictures. In this respect, reformists do not so much deny as relativize the modern national values of "class." Wealth or education are not denigrated as such, but must be combined with the "proper" observance of Islam in order to determine the individual's ultimate worth.

Among neotraditionalists, the relationship between standards of piety and the locally defined hierarchy are expressed in somewhat different terms. Elders are expected to be more pious than juniors; men more pious than women; scholars more pious than ordinary believers. The universalistic values of Islam thus, in a way, legitimate the position of individuals in the local hierarchy of influence, while at the same time serving as a bridge between the purely local realm and the

wider community of Islam with which individuals are at any rate involved on a day-to-day basis.

In short, the notion of a global moral community of Islam defines a social arena in Côte d'Ivoire that lies in some respects between and in other respects apart from the close-knit local community, where an individual's position is determined in terms of age, gender, generation, free or slave status, and so on, as well as personal reputation; and the national arena, where wealth, education, and political influence reign supreme, and where symbols of identity are universally recognized. Each kind of identity—local, Muslim, and "class"—has its own quasi-autonomous system of symbolic articulation. At another level, these systems can be combined in various ways to give rise to different varieties of Islamic ideology, expressed in ritual terms. From this point of view, religious change cannot simply be reduced to shifts along a continuum from particularism to universalism. Instead, we are faced with a multiplicity of schemas, each purporting to define the identity of individuals but also the nature of the larger social entity within which this identity is situated, expressed in symbols embodied not only in religious ritual but also in everyday life. By the values they place on these different kinds of symbols, individuals indicate the terms in which they choose or refuse to identify themselves, to evaluate and be evaluated by others around them.