In 1929, French colonial officials in Mauritania began monitoring a young man named Yacouba Sylla, the leader of a religious revival in the town of Kaédi. A Sufi teacher (shaykh), Yacouba Sylla had incurred the hostility of local administrators and the disdain of Kaédi's elite for preaching radical reforms of social and religious practice and for claiming authority out of proportion to his age and his rather minimal formal education. He claimed to derive his authority instead from a controversial shaykh named Ahmed Hamallah, then in exile from his home in Nioro, French Soudan (now Mali).

French West Africa, c. 1930 (Don Pirius Cartographic Services)

Despite local opposition, mostly from the dominant Pulaar-speaking Tukulor population, Yacouba Sylla quickly gathered a following drawn largely from among Kaédi's minority Soninke population. His supporters came from a variety of backgrounds: merchants, scholars, slaves or former slaves, members of stigmatized occupational castes, the poor. In December of 1929, the French deported Yacouba from Kaédi, ultimately placing him in detention in Sassandra, in the colony of the Ivory Coast. In his absence, his followers continued to spread his ideas and the revival became more intense. By January, it involved over six hundred people and had come into frequent and increasingly violent conflict with other residents of Kaédi. Largely on the receiving end of the violence, Yacouba's followers saw their homes burned and their shops looted, and they were attacked in the streets. The crisis came to a head on 15 February 1930, when the local gardes de cercle shot and killed at least thirty of them, and in the days that followed imprisoned and exiled over one hundred to all corners of French West Africa.Footnote ¹

After this nadir, Yacouba Sylla and his followers experienced a reversal of fortunes. In the late 1930s, the administration began releasing the “Yacoubists,” as they called them, from detention, and were surprised when most of them decided to gather in the Ivory Coast rather than return to Kaédi or Nioro. Yacouba, who himself had been born near Nioro, moved to the Ivoirien town of Gagnoa in 1939 and turned his attention to commerce and plantation agriculture. He gathered his followers around him to form a new community, slowly accumulated great wealth and became involved in Ivoirien politics at the highest levels. During the political jostling after the Second World War, Yacouba and his followers mobilized Muslim voters in the southwest of the Ivory Coast to support Félix Houphouët-Boigny's Parti Démocratique de la Côte d'Ivoire (PDCI). During Houphouët-Boigny's subsequent presidency (1960–1992) the Yacoubists served as important spokespersons for the regime, bolstering its legitimacy among Muslims and in turn receiving informal patronage.

Over the last sixty years, the face that the Yacoubist community has presented to the outside has been a thoroughly masculine one, dominated by the employees of its transport companies and stores, by its crop brokers and credit lenders, its school children (girls were taught within the community's compounds while boys attended French schools), and its political figures and members of civil society. The historical sources reflect this masculine surface. French archives and internal traditions alike depict a religious organization founded by, fought over, and led with great success by men. Standing as a synecdoche for the community as a whole is Yacouba Sylla himself, whose followers called him “Ba Yaxuuba”: Father Yacouba.

In its earliest days, however, the Yacoubist “movement” was dominated, at least numerically, by women. A retrospective census made by the local French commandant in March of 1930 listed 350 total adult “followers” of Yacouba; 225 of these were women. Women were even more predominant during the early 1930s, when Yacouba's followers were making the transition from participants in a religious awakening to members of an organized spiritual community. Based on administration estimates, of the fewer than five hundred followers of Yacouba who remained free in March 1930, only twenty-six were adult men.Footnote ²

This essay places the actions of Yacouba Sylla's female followers within four different theoretical models and explores the narrative implications of each explanatory context. I argue that, beyond the need to create a minimal coherence and to account for available data, decisions about contextualization and thus of narrative shape are ideological rather than empirical.Footnote ³ I therefore propose that attention to narrative strategy rather than the production of social theory offers the best way of addressing the concerns of those drawn to questions of agency.

AGENCY, CONTEXT, AND EXPLANATORY NARRATIVES

Historical narratives cannot avoid explicit or implicit explanations of why people act the way they do, what it means to say they “decide” to act one way or another, and under what conditions they can affect the flow of events of which they are a part. Some of the most powerful approaches to such questions of “agency” have relied on the idea of hegemony developed by Antonio Gramsci. Under conditions of hegemony, subjectivity is constrained by what Fernand Braudel called the “limits of the possible.” But where Braudel defined these limits in terms of the slowly changing material conditions that lie outside of consciousness, Gramscians attribute them to the immediate power of a state or a class to shape consciousness in its own image. Ruling elites achieve this effect by saturating the conceptual tools that could be used to articulate dissent, leaving prospective rebels or revolutionaries with a fragmented or “contradictory” consciousness and thereby dramatically reducing their ability to envisage systematic alternatives to subordination.Footnote ⁴

In this frame, “autonomous” activity is described relative to the ruling apparatus, so that agency emerges in two basic ways: in the realm of intentionality through the appropriation of the symbols or signs of hegemonic discourses to cobble together dissonant (and thus “resistant”) discourses, and in the realm of action through the displacement of the projects of those seeking to secure domination. While such analysis provides a useful guide for examining the forms that resistance to domination takes and the ways in which its protagonists articulate their goals, it says little about how such re-conceptualizations are achieved. More troublingly, by equating that which is unintelligible in the language of domination with structural incoherence, and internal contradictions among dissidents with subjective disjuncture, it forecloses the possibility of coherent alternative visions; coherence can only be the hallmark of domination. The implications of this approach are often paradoxical when applied to the techniques of rule developed by colonial states. Colonial orders devalue the agency of their subjects in direct proportion to the authority they claim for themselves.Footnote ⁵ Recovering subaltern agency thus becomes a critical task for precisely those analysts most inclined to emphasize the difficulties surrounding the emergence of such agency. Efforts to resolve this paradox become increasingly fretful as the concepts involved are progressively redefined into abstraction.Footnote ⁶

The root of this problem is the tendency to collapse the difference between various phenomena that travel under the name of agency and to conflate processes operating at very different scales.Footnote ⁷ It is useful, in this regard, to observe that the question of agency can also be thought of in terms of the narrative strategies used to link an event and its context. Historians who emplot their analyses in narratives of change cannot avoid attributing agency, and these attributions imply modes of explanation that are easily confused with the descriptions of experience provided by historical actors themselves.Footnote ⁸ In terms of the dramatic qualities of these narratives, agency depends upon the consistency between the setting of an event and the action through which it takes place. Explanations that depict their subject as the point of convergence of causal forces assume a high degree of formal consistency between the setting and the agent.Footnote ⁹

Seen from this angle, the reasons for the binds produced by analyses of hegemony (and those of other approaches as well) become clear. Histories that place the subaltern and the metropolitan in the same interpretive framework—that is, which make colonialism itself their setting—typically adopt a narrative form that begins (and often ends) with the acts of the state itself. The “appropriation” of metropolitan signs or tools—a key event in many contemporary histories of colonialism—only reinforces the place of the colonial power as the setting of the story. Understanding the European “scene” remains an implicit prerequisite for understanding the action as a whole. Non-European actors may appear on stage, but non-European contexts do so only under the rubric of an authenticity that is at best naïve about the reach of hegemony or reactionary in its invocation of an essentially local subject.Footnote ¹⁰

More sophisticated imaginings of coloniality have organized their analyses not in terms of projects, displacements, and appropriations, but rather in terms of the “entanglements” that emerged as African systems of meaning and order were (often violently) taken apart and woven into new, syncretic structures. Such accounts can lend themselves to explanations more multi-faceted than the binary opposition between “European” and “local” knowledges or practices. The approach is, however, fundamentally synchronic; exploring the processes by which colonial knots came to be tied in the first place is eschewed in favor of “tracing” entangled objects and logics back and forth from one register to another. Change, insofar as it is present at all, is either attributed abstractly to conquest or to subsequent structural adjustments within the relationships among people and things. By shifting the scale to “micropolitics” and iterated daily practices, such studies fail to account for the purported necessary relationship between entanglement and coloniality in the first place. Colonialism becomes a setting detached from any specific set of actors but one that completely defines the actions that take place on its stage.Footnote ¹¹ Other kinds of narratives are possible, however, including ones that allow setting and act to be causally distinct while still linked in the formal production of meaning.

WOMEN AS FOLLOWERS OF YACOUBA: DECISION MAKING AND SOCIAL CHANGE

The stories told about Yacouba Sylla's followers by their present descendants provide little direct help with our topic. The main hagiographic and oral sources from within the community minimize the importance of women among their shaykh's early supporters. Those women who do appear in the accounts of the early years of the community are wives or mothers of the leading men, such as the long-suffering, extremely pious Mama Hawa Cissé, Yacouba's mother, or Nene Kaba, the wife of Yacouba's chief disciple and “mother” to the community in Kaédi during her husband's exile and detention from 1930 to 1933.Footnote ¹³ In 2001, BaThierno Marega, one of the oldest surviving companions of Yacouba, argued that men's contributions in the early history of the community were most important because it was men who were arrested and killed, who sacrificed themselves. If women followed Yacouba in greater numbers, this was because they are more sensitive to religious ideas and so quickly recognized his spiritual importance.Footnote ¹⁴

The earliest French commentaries on the matter employed similar logic. In August 1929, Kaédi's residents celebrated the mawlid nabi, the anniversary of the birth of the Prophet Muhammad. For many Soninke in Kaédi, the mawlid was a time for repentance for infractions committed against neighbors and for collective forgiveness and reconciliation. Judging the current practice superficial and insincere, Yacouba had urged his followers to use the occasion to make detailed, public confessions of their infractions. Whether by Yacouba's suggestion or their own initiative, several confessants revealed recent adulterous affairs, which startled many in the community and raised concerns within the administration.Footnote ¹⁵ According to Mauritanian Governor René Chazal, Yacouba's female followers took the lead in the public confession of adultery, creating irreparable breaks with their families. To account for their willingness to take such dramatic action Chazal first reached for stereotypes about Soninke women being fickle, but he also offered a more complex explanation: Soninke women were moved by Yacouba's preaching because they were “more emotive and more sensitive to mystical revelations.”Footnote ¹⁶

Both BaThierno's invocation of women's “sensitivity to ideas” in general and Chazal's reference to Soninke women's “sensitivity to mystical revelations” shared an “attentiveness” to women's spirituality, even as they dismissed its broader historical relevance. The differences between their accounts merely reflected the differences in the ways the two men evaluated Yacouba's teachings: BaThierno praised women for recognizing Yacouba's significance while Chazal belittled them for being duped by his charlatanism.

It is not hard, though, to link such stereotypes to more “respectable” explanatory frameworks. In a groundbreaking study of women's relationships with Sufi orders (tarîqas) and saints (walîs) in Morocco, Daisy Hilse Dwyer suggested that women can have a specifically gendered relation to their social setting that allows them to act upon possibilities presented by that setting that are unavailable to men. Women who experienced crises were much more likely than were men to shift their allegiances from a Sufi tarîqa or walî, whose attachment they had “inherited” from their father or husband, to one that proffered more effective spiritual aid. Women thus tended to privilege flexibility over stability, while it was harder for men to re-affiliate because their religious associations were more thoroughly integrated with lineage identities and other social networks. Men also tended to be more insulated from the kinds of crises that provoked women's search for additional spiritual resources.Footnote ¹⁷

Dwyer's model suggests we could set the early days of the Yacoubist revival against two intersecting backdrops: the wider context of competing religious identities, and those specific social changes or stresses that were experienced asymmetrically by men and women. This is, it turns out, quite easy to do. Yacouba's preaching in Kaédi in 1929 can be seen as simply the latest round in a struggle between different Sufi constituencies that had been going on since the turn of the century. Yacouba Sylla's teacher, Shaykh Hamallah, had become famous in West Africa for preaching a reform of the central liturgy (wird) of the dominant Tijani Sufi order. Hamallah argued that one of the wird's key prayers, the jawharat al-kamâl, the “Pearl of Perfection,” should be recited eleven times rather than the customary twelve.Footnote ¹⁸ These recitations were counted by fingering prayer beads, and so Hamallah's followers (the “Hamawis”) came to be colloquially called “eleven-bead” Tijanis, as distinguished from the majority “twelve-bead” branch. The eleven-bead/twelve-bead question had been hotly controversial in Kaédi since 1900, when Hamallah's own teacher, Shaykh Sidi Muhammad al-Akhdar, had first visited the town and preached the eleven-bead reform.Footnote ¹⁹ In the intervening years, debates over the political implications of the eleven-bead reform and the networks of Sufis centering on Hamallah had figured heavily in French Islamic policy in the region. Intellectuals who had worked with French officials to construct what David Robinson has called “relations of accommodation” provided intellectual critiques of the eleven-bead position, while figures like Seydou Nourou Tall, who served as France's central Muslim intermediary in the 1930s and 1940s, worked to discipline Hamallah's followers and bring them back into networks of religious patronage led by approved, twelve-bead shaykhs.Footnote ²⁰

Thus when Yacouba Sylla presented himself as a disciple of Shaykh Hamallah, he was immediately identified with the eleven-bead movement and his activities aroused the hostility of many of those who were affiliated with the twelve-bead Tijaniyya (including most of the town's political elite) or who simply hoped the whole controversy would die out.Footnote ²¹ French observers understood the entire conflict between Yacouba's followers and the other residents of Kaédi to center around the competition between the eleven- and twelve-bead groups. In fact, many eleven-bead Tijanis rejected Yacouba's leadership, and so the situation as a whole was characterized by a proliferation of claims to religious authority and a fragmentation of spiritual affiliations.

There were reasons why some women in Kaédi might have seen this state of affairs as opportune. In the 1920s, the Middle Senegal River Valley, where Kaédi was located, had witnessed a series of socioeconomic transformations that unsettled gender and age dynamics. Male elders had blamed increases in labor migration, the withdrawal of state support for slaveholding, and the imposition of new legal codes for weakening control over both women and young men. Some women had benefited from the increased income and independence that came with male migration, but in many cases it had also led to a rapid increase in divorces, disputes between husbands and wives, and problems with marriage engagements. Rapid monetization of social relations and the increasing role of the colonial administration in channeling wealth through patronage networks seem to have increased inequalities among Kaédi's families. Women who came from slave or casted backgrounds encountered serious obstacles to establishing their own households and accessing the material, social, and cultural capital needed to take advantage of the changing economic landscape. They thus increasingly found themselves sliding even further down the social ladder.Footnote ²²

Such conditions would have affected many of the revivalists, since the best evidence indicates that most of Yacouba's early followers were socially marginal. In terms of social relations, a person's status as a slave, former-slave, or member of a caste was signaled and reproduced through control over access to education and bridewealth, and the regulation of the reproduction of households.Footnote ²³ As a result, associated practices make good proxies for measures of wealth and status. In the aftermath of the deaths on 15 February 1930, the administration investigated the condition sociale of twenty-two of the dead. Of the fifteen for whom status could be determined, all but one were classified as anciens captifs, “former slaves.”Footnote ²⁴ Observers on all sides noted the relative poverty of many of Yacouba's early followers, their inability to control the reproduction of their own households, and their engagement in cultural practices normally associated with slave or caste status, and movement detractors put this evidence to rhetorical use. The “orthodox” Tijani leadership of Kaédi viewed the revival with great condescension and claimed that Yacouba's reforms threatened the established social order. French officials, too, rapidly adopted this position.Footnote ²⁵

Much of the oral evidence for early poverty and marginality is clouded by the spiritual significance that suffering later took on for Yacouba's followers. Yacoubist memory is structured by what Benjamin Soares has called an “economy of martyrdom,” in which a Sufi community combines mass affiliation, a collective sense of persecution, and leadership through management of networks of redistribution.Footnote ²⁶ Such institutions can create incentives to exaggerate, or at least highlight, past injuries.Footnote ²⁷ Be that as it may, those elements of the hagiography that directly index the status of Yacouba's followers before the revival support the reports left by French observers in recording the injuries of social inequality rather than direct religious repression.

Traditionists and historians within the community address the question of the social origins of these early disciples indirectly but unambiguously.Footnote ²⁸ Yacouba Sylla himself was taken as representative of the background of his followers; Yacouba's mother and others of her generation are said to have repeatedly testified that Yacouba's family had been the poorest in the vicinity of Nioro. Other forms of social inferiority were visible through their subsequent transcendence. The most important example was the conscious orchestration by Yacouba and his deputies of marriages between followers from casted families and those from “noble” families and between descendants of slaves and “free” persons. Community leaders claimed to me that this policy had been so successful that few preserved any memory of a given individual's past social position. Nonetheless, a collective past of deprivation could be readily acknowledged.Footnote ²⁹

Some of Yacouba's followers, including the Kaba family of merchants, did have wealth, though the prominence given in the official memory to these few figures supports the impression of their uniqueness. Furthermore, even they held little influence. Soninke residents of Kaédi were on the whole subordinate to the majority Tukulor population, except for a few select intermediaries who distributed patronage from French officials and Tukulor landlords. These Soninke political elites, such as the family of Gattaga's chef de village, Dieydi Diagana, were openly hostile to Yacouba's revival.Footnote ³⁰

Dwyer's model suggests that socially marginal women would be more likely than would men to see the arrival of new sources of spiritual power as providing potential resources for responding to these growing pressures. This then would explain why more women from eleven-bead families joined Yacouba Sylla's revival than did men. Furthermore, given that women outnumbered men in the movement by a ratio of two-to-one, that regionally polygyny was generally associated with substantial wealth, and that many of Yacouba's followers were unmarried, it is unlikely that many of these women were from polygynous households where the husband was a follower. In all probability, then, most of his female followers were unmarried, divorced, or had husbands who did not participate (at least publicly) in the revival. French observers even claimed that some women had “crossed over” from the twelve-beads to Yacouba's party.Footnote ³¹ By contrast, no men were reported as having made such a move, even temporarily.

Furthermore, if the rigidity of local religious identities had the result of raising the “decision threshold” for women who sought to change the direction of their spiritual orientation, and thus encouraged or necessitated more radical breaks from past behavior once the decision to change had been made, it may account for the apparent “radicalism” of Yacouba's female followers who were willing to leave their husbands (though not their children) behind. For women whose family relations reproduced their dependent status, or whose husbands had been absent for an extended period of time in migrant labors or in the colonial service, the option that Yacouba's arrival provided to use religious allegiances to dissolve marital ones may have been particularly welcome.

However, the implication that these decisions can be explained by reference to some feature of women qua women assumes that Soninke women in Kaédi were a relatively homogenous group, at least insofar as their attitudes toward such social stresses were concerned. Yet, at least some of Yacouba's female followers were from affluent families, such as the Kaba family, and these in particular seem to have joined with their male relatives, rather than to free themselves from them. Furthermore, the framework would seem to suggest that some distinct factor must have been involved in the decision of approximately 125 men to affiliate with Yacouba. Here, too, one might introduce factors of status and age, and suggest that young, unmarried, poor men were in positions of dependency analogous to those of non-elite women, and thus would have faced similar incentives to re-affiliate.Footnote ³² But this is no more successful in explaining the attraction of wealthy or older men to Yacouba than it is that of well-off women. Finally, the entire framework, which approaches religion from the perspective of how affiliations define corporate groups, evades something that was explicit in the claims of both BaThierno Marega and Governor Chazal: that there was something specific about Yacouba's “ideas” or “mystical revelations” that made his revival attractive to women.

Thus, this story has its holes, both in the sense of offering a somewhat unsatisfactory (“two-dimensional”) depiction of its characters' motivations, and in its only partial ability to account for all the available evidence. Still, the framework does provide an acceptable explanation for the fact that so many of Yacouba's followers were unmarried or otherwise independent women, and we may not want to fault it for failing to answer other interesting but distinct questions. The story does “work” to some extent, both in the sense of attaining plausibility and in effecting a change in our understanding.

GENDER, “MORAL PANIC,” AND CHANGING BOUNDARIES

This is not, however, the only possible story, and we can solve some of its problems by shifting from a socioeconomic setting to an ideational or symbolic one. Actors can make choices based on relative costs and benefits, but they can also define the terms of the choices available to them—that is, they can be consistent with their setting but also act upon it. One narrative that allows for such activity is that of the “moral panic.” Moral panic stories demonstrate how, in Sonya Rose's terms, “disturbances within the social order are dealt with, in part, symbolically and in a manner determined by … a logic belonging to the structural properties of cultural symbols.” Such disturbances, in turn, unsettle carefully constructed representations of group and individual identities and thus become “represented as sources of subversion and disorder,” drawing on tropes deeply embedded in discursive traditions.Footnote ³³ Rather than requiring a causal factor unique to women's affiliation preferences, the moral panic story's emphasis on representations of femaleness shifts from an analysis of “women” as social agents to an analysis of gender as social discourse.

The social changes underway in the region around Kaédi do seem to have coincided with an intensified debate over appropriate gender roles, and we can understand Yacouba's preaching as part of that debate, both reflecting and transforming a tradition of representational practices. In addition to championing Hamallah's eleven-bead reform, Yacouba preached the equality of all before God, attacked certain women's clothing and dance styles as immoral, and banned the wearing of gold. When many of his followers complained they were too poor to marry, Yacouba called for bridewealth levels to be lowered; tradition holds that forty-two marriages were celebrated the next day. Yacouba particularly encouraged followers from occupational castes and ex-slave lineages to marry others who were from “noble” or “free” families, in violation of norms of endogamy.Footnote ³⁴

These reforms all centered on the rejection of symbols of marginality that operated by marking someone as a subordinate, “immature” member of society. Yacouba's teachings thus identified pathways by which his followers could assert their right to be accepted as free, adult members of the community regardless of their social background. Yet, since adulthood was an explicitly gendered concept, Yacouba's reforms and even his own identity were predicated on particular notions of masculinity and femininity that reflected, in turn, gendered experiences of marginal social status.

In many parts of Muslim West Africa, including Soninke communities, the performance of certain dances and the wearing of revealing clothing were ways of signaling the supposed sexual availability of female slaves to men of noble rank. They contrasted visibly with the disciplined bodily comportment that Islamic reform movements had made a centerpiece of Muslim identity in the previous century.Footnote ³⁵ Yacouba Sylla's followers argue that their denunciations of women's dances and clothing were attempts to make the women of the eleven-bead community “respectable,” enabling and compelling them to play the role appropriate to an adult woman in an ideal Muslim community.Footnote ³⁶ In banning both transparent clothing and “slave” dances, Yacouba was not only furthering a major thrust of religious reform, he was also providing a means by which female ex-slaves could recover their symbolic honor.Footnote ³⁷

Reforms of dancing and attire responded to the perceived moral crisis in a way that served the interests of Yacouba's male followers, who could now not only marry but marry “respectable” Muslim women. The same could be said of Yacouba's ban on gold jewelry. Gold was not just a symbol of wealth and status in society; it had a particular role in gender relations. Married women used the money they earned from their portion of bridewealth and from labor on their personal lands to purchase gold. Gold so bought became part of private wealth that their husbands could not appropriate.Footnote ³⁸ Yacouba's ban on gold may thus have severely limited married women's autonomy. But this would have been most noticeable for women who had access to such income in the first place. For those too poor or marginal to marry and collect their own wealth to store as gold, or for those whose bridewealth had been very small to begin with, abandoning this abstract right might have been a small price to pay for the tangible, immediate benefit of increased status and the freedom to establish their own households.

This moves, then, in almost the opposite direction of the Dwyer-inspired story: rather than fleeing family affiliations, women who joined Yacouba's community may have been inspired to do so out of a desire to establish new families, ones of their own design. As Luise White argued for participants in Kenya's Mau Mau crisis, these may have been less “the social rearrangements of women and men in intense crisis” than the creative responses “of women and men whose definitions of gender were in intense crisis.”Footnote ³⁹ Even though each particular symbol may have meant something different for different women when viewed through social lenses, Yacouba's reforms involved a complex package of gender transformations whose significance derived from both the entire system of shifts and individual women's specific circumstances. For some, escape from one form of dependency may have been only a precondition for entering into another, less onerous form of dependency. Similarly, for women who had seen their lack of “noble” comportment reinforce their social marginality, a shift in their “cultural capital” might have made a real impact on their status. For wealthy women, the ban on gold may have provided an opportunity to convert material status into religious and cultural respectability.

The revival may have also provided an opportunity for some women to make an intervention in religious norms, to alter prevailing spiritual practices to better suit their perceived needs. The surprising decision of women to seize upon Yacouba's expanded definition of confession and their choice to specifically foreground adultery in public pronouncements suggests an attempt to alter the practices of the men in the community. Confession may have been attractive not because of any particular propensity of women towards such displays, but because it, like changes in attire or women's participation in Sufi recitation (dhikr) ceremonies, forced men into uncomfortable corners where they either had to accede to women's demands, admit their own culpability in irreligious behavior, or denounce the revival as a whole.Footnote ⁴⁰ There is also evidence that some of the practices introduced by Yacouba's followers made use of Islamic spiritual power to provide therapies of a specifically gendered nature. Although the practice is controversial even within the community and has provided the grounds for denunciations by rival Hamawis, at least some contemporary gatherings of Yacoubist women center around the use of Yacouba's baraka (grace) or haidara (presence) to cure problems with fertility or family health.Footnote ⁴¹

Taken together, these interventions constituted a system of transformations made coherent by the fact that their understood meaning was decisively religious. Thus that Yacouba's criticisms of women's attire are said to have started first with the distinctly noble female members of his host family linked his redefinition of gender roles to his attacks on wealth, status, and endogamy.Footnote ⁴² By creating a common identity within which a whole set of symbols could be rejected as a package, he leveled the social distinctions among his followers. Defining the community in terms of the rejection of slave and casted pasts made those pasts a collective one, so that no stigma attached to a particular individual; they became a religiously meaningful memory of generalized suffering and oppression, as opposed to one merely reflecting (and thus reinforcing) the ignominies of inequality.

If shifts in gender relations were constantly mediated by religious beliefs and practices, the reverse was also true. The most radical of Yacouba's reforms in the eyes of many of his opponents was his opening of Sufi dhikr ceremonies to the women of Kaédi.Footnote ⁴³ In this, Yacouba moved beyond creating new routes to adulthood for his followers and actually altered the meaning of adult womanhood, at least insofar as it touched on religious life. This change was felt as a threat to many free men's masculinity, causing Yacouba to be widely denounced as a cause of the very moral degeneracy his followers saw him as fighting. Yet just as poor women may have been little troubled by the ban on gold jewelry, so too a new, expanded role for women may have been a low price for socially marginal men to pay in exchange for their own free-adult masculinity.

Taking the content of Yacouba's teachings as central to our explanatory framework, the tradeoffs his followers faced appear to have involved both altered social identities and modifications to strategies of acquisition. Distinctions between what we would identify as economic and symbolic values collapse. Not surprisingly, this makes Yacouba's preaching about bridewealth—a practice that practically defines the intersection of economic and social interaction in anthropological theory—appear to be a decisive event of the community's early history.

The costs of “bridewealth”—a vague category that included a wide range of giftsFootnote ⁴⁴—seem to have gone up dramatically with expanded opportunities for wage labor in the early decades of the century as older men struggled to control the timing of the establishment of new households.Footnote ⁴⁵ In 1929, Yacouba declared that high bridewealth had reduced women to the status of commodities and he asked the imams of Kaédi to set a maximum payment. His calls linked recent social changes with widely accepted moral norms in a way that created the conditions for broad-based support. There were, for example, several ways for male elders in Kaédi in the 1920s to present a reduction in bridewealth for their daughters as a pious act, especially if the prospective son-in-law was a religious figure.Footnote ⁴⁶ Women themselves controlled only a portion of the bridewealth payments that supposedly belonged to them; for poor women in particular, inflation and the expansion of the gifts destined for the fathers of brides had emptied dowry proper of anything more than symbolic value.Footnote ⁴⁷ Claims that reductions in bridewealth could both restore the piety of the community and protect the rights of all men and women to contract marriages were likely to be popular; indeed, some of the imams apparently complied and persuaded a number of male elders to lower their demands.Footnote ⁴⁸

On the other hand, when Yacouba pressured the imams to establish a maximum bridewealth for all families it was likely seen as an unprecedented incursion of religious authority into family affairs. It was also not justifiable by reference to any Islamic texts popular in West Africa. Bridewealth did a large amount of social and cultural work and certain members of the town must have been reluctant to forgo its use. The very feature that made it frustrating to poorer men made it attractive to the wealthy: by enabling families to make marriage an agonistic pursuit, it facilitated the emergence of hierarchy, status, and dependency.

In the 1930s, the Yacoubists made the even more radical decision to abolish bridewealth entirely, signaling the development of a new attitude toward outside society as a whole. If bridewealth symbolizes the notional freedom to enter into alliance with any family that shares the same “currency,” the contracting of marriages as pure, prestation-less exchanges narrows the scope of possible social interaction. The insistence that no bridewealth be given in exchange for women who married into the community would have acted as a strong deterrent for intermarriage with any family that did not become a member; few households would want to “give” a wife into the community without bridewealth and without assurance that it could obtain one similarly in the future. Indeed, one of the most important things the community guaranteed its male members was the ability to marry and establish their own families.Footnote ⁴⁹

The decision to create such a boundary between Yacouba's followers and the outside society responded to a different kind of moral crisis than had the calls for reduction in bridewealth in 1929. At a time when most adult male Yacoubists were dead or in exile, the group found itself in a dangerous situation but also in a moment of opportunity. If arguing that bridewealth was excessive had helped Yacouba articulate his critique of social inequality in a language that appealed to the pious even as it facilitated a shift in social status by his followers, the abolition of bridewealth transformed the isolation of Yacouba's followers into the defining feature of their shared identity.

Indeed, we could say that it was at this moment that Yacouba's followers became a community. If adhering to Yacouba's teachings or taking him as one's shaykh made one a follower of Yacouba, then adhering to the rules abolishing bridewealth made one a member of the Yacoubist community. This in turn brought with it all the other obvious elements of community identity. For example, it seems to have been around this time that the members of the community began holding all property in common. This tightened the control of the leadership even as it reinforced the uselessness of bridewealth: what could it mean to merely transfer goods from one part of a collective to another?

This decision to draw Yacouba's followers together into an endogamous community sharply distinguished from the outside world marked, in a sense, the final closure of the moral panic initiated by the socioeconomic changes of the early twentieth century. The transformations in fundamental gender and social norms suggested by Yacouba's teachings proved too radical for most residents of Kaédi, but too appealing to his followers for them to return to the status quo ante. Thus the only outcome that could end the panic—that could bring the story to an appropriate point of closure—was to redraw the boundaries of the community itself. Affiliating with Yacouba was more than just a way to access a new source of spiritual power; it was an opportunity to change the rules by which spiritual power operated.

At the same time, the story takes us quite far from the original question of women's participation in Yacouba's revival, and it reveals the limits to what can be achieved in this particular setting. Indeed, men seem to dominate the active roles in this narrative almost as much as they do in the official stories established by colonial and hagiographic records. This, too, seems to derive from the explanatory context, as the shift from women to gender, from the social to the symbolic, navigates a path back to the centers of power and the generative sources of discourse in the community. Insofar as our sources elucidate those centers and sources, they mark them as distinctively male. Furthermore, the story's setting has become so abstract, so deep-structural, that its actors have little capacity to influence it. Each break with local norms was only intelligible to its performers and audiences by reference to those norms themselves. If the setting was the logic governing the symbolic responses to “disturbances within the social order” (and the explanatory power of the anthropological models designed to translate that logic), it had, by the end of the story, undergone only a change of scenery.

SPIRITUAL PIONEERS: YEWTI KABA'S STORY

Such long-term continuities may be troubling for theorists of agency, but the problem they pose is not one that the Yacoubist community finds particularly interesting. When, at a haidara ceremony in 2001, I remarked on the way the community's use of Soninke songs and dances memorializing the events of their past had helped them preserve their distinctive culture, the presiding elder quickly corrected me: the Soninke language was only a vehicle for religious teachings, not something important in and of itself; the community was engaged in an act of spiritual, not cultural, conservation.Footnote ⁵⁰

The community's rhetoric takes little note of the “customs” like bridewealth and dancing that played such crucial roles in our moral panic story. In some ways, this avoidance corresponds to the logic of the hagiography's broader elision of women's participation in the community's early development: their participation and the ways they mark the group's distinctiveness draw too much attention to the social and conventional and thus take away from the miraculous and officially Islamic. More appropriate to the serious task of establishing the community's status as Muslim power-brokers in postcolonial Côte d'Ivoire are stories about the actions of individual religious reformers like Yacouba and his right-hand man, N'Paly Kaba, both of whom were emboldened and empowered by the spiritual grace (baraka) of Shaykh Hamallah as they confronted French oppression.

The hagiographies, like the French documents they counter-balance, make full use of the power of narratives built around such unique characters. By placing agency in a named person rather than a social category or system of symbols, more radical transformations of the setting are made possible, contingencies are multiplied, and the drama in general is increased. Contexts still exist, of course, but only to provide opportunities for the “heroes” of the story to demonstrate their ability to transform them. We might expect that our sources—records of barbarism, as Walter Benjamin reminds us—would conspire to put such stories out of the reach of those two hundred-plus anonymous women who joined Yacouba's movement in 1929. But one woman, at least, does have her own story.

In September and October of 1929, after Yacouba Sylla's expulsion from Kaédi but before the outbreak of serious violence in the town, tensions between Yacouba's followers and the French administration came to a minor crisis surrounding one of the most prominent female members of the religious revival. Sirandou Kaba (known within the Yacoubist community as Yewti Kaba) had been married to one Ibrahim Galledou, an employee of the colonial administration who was absent from Kaédi and, apparently, not a participant in the religious awakening. For reasons that remain unclear, the local administration had taken it upon itself to reunite Sirandou Kaba with her husband, posted at the time to Mederdra.Footnote ⁵¹

The young woman's father was Amadi Gata Kaba, an important local merchant and brother of N'Paly Kaba, Yacouba's closest disciple. It appears that Amadi Gata initially agreed with the plan to send Sirandou to her husband. Yet N'Paly Kaba, rather than Amadi Gata, took the lead in representing the family to the administration in the matter and, whether in response to Sirandou Kaba's own desires or out of strategic calculation, refused to cooperate. Commandant Charbonnier responded by trying to use physical force to place Sirandou Kaba on a boat leaving for Mederdra. In the dramatic confrontation that ensued, N'Paly Kaba slapped Charbonnier and was immediately arrested.Footnote ⁵²

That same day, Charbonnier left for a new post (on the very boat that was to have taken Sirandou Kaba), and Commandant Quegneaux replaced him. Quegneaux quickly investigated the affaire and determined there were “reasons that argue[ed] in favor of keeping [Sirandou] with her family in Kaédi.” But the change of heart did not come in time to avert an uprising by Yacouba's followers, outraged by this “unusual” interference in family politics and the apparent threat to their religious autonomy. Several of Yacouba's followers were arrested, weapons were seized, and two days later N'Paly Kaba was sentenced to four months in prison. Eleven others, including Amadi Gata and members of the influential Kaba and Tandia families, also received prison sentences ranging from one to eight months and were fined for their involvement.Footnote ⁵³

Yewti Kaba herself figures not at all in the written hagiographies. Indeed, the Fondation Cheikh Yacouba Sylla's “official” history, Le sens d'un combat, includes no event at all between Yacouba's expulsion from Kaédi in August 1929 and the events of February 1930. But those oral accounts that focus on these months consistently frame the story in terms of the struggle of a group of men—N'Paly Kaba, Commandant Charbonnier, Amadi Gata, and Ibrahim Galledou—to control the movements of “Amadi Gata's daughter.”Footnote ⁵⁴ Other sources, though, are more suggestive. One of Adama Gnokane's informants describes her as having been “an object of veneration of the Yacoubist women.”Footnote ⁵⁵ The French documents generally share the male-centered focus of the oral traditions, but also hint that Yewti Kaba's own actions were the source of the agitation. Governor Chazal relayed an early report that her father, Amadi Gata, “had informed [the administration] that the aggressive attitude of his daughter, an adept agitated by the subversive theories of Yacoub Cilla, had obliged him before long to intervene disciplinarily and that in order to avoid such humiliation to his family, he agreed to send her away from Kaédi.”Footnote ⁵⁶ If, in fact, Yewti Kaba ran afoul of the administration and Kaédi's elites because of her role in stirring up a religious revival among the town's women, the complaints that Yacouba was a source of moral disorder may have had less to do with his critiques of bridewealth and social inequality than with the very visible actions of his female followers. Women may have featured so prominently in his movement because they were, at least in part, leading it.

Similarly, the decision to abolish bridewealth in the community looks quite different when seen as part of the same story as Yewti Kaba's failed deportation. If we return to the perilous moment in early 1930 when the community's men were mostly dead or in detention, we can read in another way the decision to make the group of followers endogamous. Recognizing the community's weakness at that moment, the administration seems to have sought to take the opportunity to disperse the group once and for all. Commandant Quegneaux encouraged those Soninke families who had remained “untouched” by Yacouba's preaching to adopt the Hamawi women and children left behind. Some of these were estranged from their non-Yacoubist husbands, others were unmarried, widows, or had husbands in exile. Quegneaux assumed they would be eager to incorporate into households, and that their attachment to Yacouba and to those in prison would thereby be gradually dissolved. However, much to the administration's concern, the other families in Kaédi, “even those who were united to the onze by lines of parentage,” refused, saying that the Yacoubist women were “too quarrelsome.”Footnote ⁵⁷

It is impossible to say whether the failure of Quegneaux's project had more to do with the prejudices of non-Yacoubists against Yacouba's followers, or with the refusal of Yacouba's female followers to be assimilated into twelve-bead homes, though we might suspect some of both. In any case, it is possible to see the decision taken by the male leaders of the Yacoubist community to forbid the “widows and orphans” of 15 February to marry anyone outside the group, along with their efforts to marry widows to followers using proxies, as attempts to normalize and institutionalize what was already happening on the ground.Footnote ⁵⁸ A majority of that amorphous group of women who had responded to Yacouba's preaching and who had participated in various aspects of the revival and confrontations that ensued had already decided to cast their lot with what must have seemed a rather risky undertaking. Only a minority had apparently decided to do their best to subdue their “quarrelsomeness” and seek re-assimilation into Kaédi's other households.Footnote ⁵⁹ The price paid by those who held their ground? Three or four years in Kaédi surrounded by rivals who feared them and without men to provide for them economically. The rewards? The same.

While there is little doubt that without the unprecedented changes in women's religious roles, Yacouba's teachings would not have so scandalized many of Kaédi's elites, it also seems likely that without a large number of women followers Yacouba would have had much less of an impact. In this sense, it seems appropriate to speak of women's roles in shaping Yacouba's early teachings and even in determining the course of the community's history. It is even possible to link women's early participation in the revival with their efforts to conserve the social and spiritual powers they had gained with the specific content of the Yacoubist doctrine. It is easy, when adopting some of the ethnographic language, to conflate transfers of bridewealth with weddings themselves. Barbara Cooper has pointed out that, in addition to their functional and rhetorical properties, bridewealth prestations also have a performative aspect, and wedding ceremonies and the various rituals that surround them often provide a crucial opportunity for women to establish their “worth” in society, to activate potential social relationships, and to create, reproduce, or challenge hierarchies. They are, in short, important sites of women's action within and upon society.Footnote ⁶⁰

If many of Yacouba's followers had indeed been prevented from marrying because of excessive bridewealth demands, then many of his female followers would have lacked an opportunity to assert their value through these ceremonies. If we set aside male-centered, highly abstract models that seek merely to calculate the debts and obligations that men incur from the exchange of women, and instead see marriage as an activity in which women were able to exercise considerable initiative, then the decision of the male elders in Kaédi to block marriages as a way of controlling young men's wealth would have been equally objectionable to young women. Marriage within the Yacoubist community, by contrast, provided women with entrance to an entire symbolic language designed to enable them to affirm their worth. All of the signs of free, adult status that were symbolized by such practices as modest vesture and refined dancing could be put forcefully on display through the commitment to the community that marriage within it implied.

Not the least of these symbolic practices were the nuptials themselves, which affirmed the ability of bride and groom to constitute themselves as givers, and thus as moral, pious, free persons. For as in the case of marriage to a religious scholar, the forgoing of bridewealth could itself become a gift. Furthermore, marriage was a gift that everyone involved gave simultaneously to everyone else: the bride and groom gave themselves to each other, their families gave them up, and Yacouba gave them all the spiritually meaningful context and security in which to give these gifts. And—a point in which community members take great pride—everyone gave their consent.

THE POWER OF A SPIRIT(UAL TRADITION): THE VISION OF FATIMA

This story of freely consenting disciples, motivated by their own individual spiritual (and not so spiritual) desires would seem to represent the zero-ground of agency, the frame of reference in which women's status as actors is most fully “restored,” with little hegemony or contradictory consciousness in sight. We can, however, subvert even this contextualization, for the idea that religious beliefs are freely chosen is, at best, in tension with other modes of explanation that are less liberal. Certain pieces of evidence can only be woven into the story if we take an even more expansive view of Yacouba Sylla's teachings, and place them in an explicitly Islamic framework—one in which our choice of dramatis personae is between a religious tradition and its supernatural forces.

Much of what we know about the community's origins comes from those of Yacouba's opponents who sought to brand him as a heretic and rabble-rouser. It was they who passed information along to the French officials who created our archives. In addition to describing Yacouba's outrageous ideas and the scandalous actions of his followers, they also claimed that he had attributed the source of his reforms touching on women's dignity and the institution of marriage to a conversation with the Prophet's daughter Fatima.Footnote ⁶¹ Summarizing the events of 1929 and 1930, Governor-General Carde wrote this claim into what would stand as the official version of Yacouba's teachings for nearly seventy years:

The Fatima story has been a staple of commentary on the community ever since, though no one has ever offered any explanation or interpretation of it. Documents assembled by the scholars at the Centre des Haute Études d'Administration Musulmane (CHEAM) picked the story up from the early administrative reports and it passed into J. C. Froelich's oft-cited Les musulmans d'Afrique noire in 1962. The story was introduced into Anglophone scholarship by Jamal M. Abun-Nasr's The Tijaniyya, which drew on both Froelich and the CHEAM files. Finally, in 1980 Adama Gnokane, the first scholar to return to the original reports, simply repeated the Fatima story, again without comment.Footnote ⁶³

Those who codified this story refrained from explaining it, as if it provided its own context. To some extent, it did—for them. For Fatima was a fairly well defined figure in early twentieth-century French Orientalist thought. Through the 1920s and 1930s the canonical work on Fatima was Henri Lammens' Fatima et les filles de Mahomet. A critical study with a debunking, even condescending tone, Fatima dismissed the historical significance of the Prophet's daughter in early Islam, arguing that her veneration was a later development, orchestrated by first ^cAlid and then Shi^ci political activists and bolstered by naive mysticism. Lammens believed the cult had survived in Sunni Islam because of its usefulness in diverting the high regard most Muslims had for the Prophet's descendants (the shurafâ', all of whom descended from Fatima) away from Fatima's husband ^cAli ibn Abi Talib and their sons al-Hassan and al-Hussayn, the central figures in Shi^ci challenges to Sunni legitimacy. Lammens also associated Fatima with Islamic patriarchy, and argued that her status as a model of the ideal woman served to reinforce women's subordination. The main features of Fatima's hagiography that Lammens made available to Islamologists—events and symbols that turned around the themes of weddings, poverty, women's clothing, and bridewealth—may have resonated with the early reports about the specifics of Yacouba's moral reforms.Footnote ⁶⁴

These symbols aside, the story may have seemed important simply because of its considerable uniqueness; to my knowledge no other stories of revelatory visions of Fatima have ever appeared in the vast literature on Sufism or Sunni Islam. As such, it helped to identify Yacouba's preaching as part of a discrete religious movement, separate from either the Tijaniyya or various mahdist movements. At a time when reformist impulses had led most Sufi shaykhs to claim Muhammad as the source of their authority, the French may have thought that Yacouba's association with Fatima represented a kind of defective, if not heretical, reformism.Footnote ⁶⁵ In any case, the Fatima story served to emphasize the movement's heterodoxy by associating it with Shi^cism, to symbolize its unusual gender dynamics, and to emphasize the credulity of those who would follow such teachings.

In doing so, it intersected with another line of speculation about Yacouba that also seems to have had its origin in the Fatima story. At the end of the 1930s, Captain J.-L. Montezer's handbook on Islam for incoming colonial officers dismissively reported that Yacouba had decided to embark on his reforms “following a ‘dream,’” without mentioning Fatima. Montezer in turn was the direct source for a respected 1952 study by Alphonse Gouilly, who also referred to Yacouba's “dream.” Montezer's text also included the phrase “the absolute independence of the child within the family and of the individual within society,” lifted verbatim from Carde, and set the foundation for the elaboration of a mythology about sexual deviancy that ran parallel to the Fatima story. A 1941 CHEAM report by Lieutenant d'Arbaumont, for example, elaborated Charbonnier's fears about the social implications of Yacouba's acceptance of women at dhikr ceremonies into a litany of sexual fantasies about an orgiastic dance that supposedly followed the dhikr, fantasies that Jamal Abun-Nasr subsequently echoed uncritically in his authoritative monograph.Footnote ⁶⁶

By the time it reached Abun-Nasr's book, the Fatima story had traveled a great distance from those who first told it to Charbonnier. Where it came from remains a mystery. The uniqueness of the story in the Sufi tradition means there is no “evident” set of internal traditions that could help us interpret it, nor any rich ethnographic text that might ground it in some particular local inflections of Islam or a specific religious pedagogy. Moreover, the leadership of the Yacoubist communities in Gagnoa and Abidjan strongly rejects the story of Yacouba's vision of Fatima, so internal sources offer no clarification of what it may have signified. Nonetheless, even if the story was fabricated by Yacouba's rivals, they had some reason for presenting this particular rumor. Our only hope is to start by sketching the parameters of what those reasons might have been by examining Muslim attitudes towards Fatima in general, to present a kind of counter-narrative to Lammens and Montezer.

The same themes in the Fatima hagiography that might have caught the eye of the French may have also been known to the more-educated Muslims in Kaédi, and perhaps to the general public. In the story, Muhammad rejects Fatima's first suitors—including the future Caliphs ^cUmar and Abu Bakr—when they flaunt their great wealth, and he instead selects ^cAli despite his poverty. This may have resonated with Yacouba's criticisms of excessive bridewealth and the wearing of luxurious jewelry. Local debates over bridewealth reform may have reminded people that Muhammad gave Fatima, as her dowry, the power to intercede with God on the part of all Muslims on the day of resurrection—an attribute frequently attributed to Sufi saints. Finally, the “widespread” legend reported by Massignon that Fatima was without hayd (menstrual bleeding) or nifâs (puerperal bleeding), and thus legally able to pray and fast without any sexually determined interruptions, might have been a useful precedent in legitimating the presence of Yacouba's female followers in Sufi gatherings and at dhikr recitations.Footnote ⁶⁷ If so, then the distortions and elaborations of the story and of the gendered nature of Yacouba's community that d'Arbaumont or Abun-Nasr presented were not inventions in the sense of outright fabrications, bearing no relationship to lived experience. Rather, they were a result of the misleading re-contextualization of rhetorical tropes that had originally been mobilized for very different purposes.

Unfortunately, this line of thinking must remain speculative in the absence of more information about what the figure of Fatima meant to the residents of Kaédi. What is clear is that the generally pejorative connotations the story seems to have taken on in the French sources, and then in the derivative literature, would almost certainly not have obtained among either Yacouba's followers or the informants who first told the story to Charbonnier. Fatima would have been held in high esteem if for no other reason than her close proximity to Muhammad, ^cAli, Hassan, and Husayn, and her status as the mother of all shurafâ'.

If, as I have suggested, Yacouba's early religious revival was predominantly a women's affair, and women like Yewti Kaba may have even taken a lead in sustaining and radicalizing the movement, then the story that Yacouba's teachings had been inspired by a vision of Fatima might have been absolutely central to this early history. All the other reforms—of women's clothing, dancing, confessions of adultery, and bridewealth—may have presented themselves as a solution to women's anxieties about their worth in society and their control over their religious lives, especially for women who were casted or former slaves, and this solution may have been symbolized by the image of Fatima herself leading the revival through Yacouba.

Yet here the story branches, for the vision of Fatima does not emerge out of the “fabric” of Soninke society in the 1920s in any material, social, or discursive registers. As an explanatory device, it has a context, but one that is as ahistorical as are French stories about sexual deviancy or the oral tradition's remarks on women's sensitivity. Nothing indicates that anyone in Kaédi ever heard any of the stories about Fatima mentioned above, no sense that such transformations were part of the cultural “grammar” that allowed Yacouba and his followers to criticize the social order, no evidence that they were part of the “symbolic repertoire” available for negotiating their position in a changing society.

The source of this vision raises troubling questions. Did it form part of a “lost” tradition, evident to people in Kaédi in 1929 but not described by any source? Or was it a unique, contingent act of genius on the part of Yacouba or some other revivalist who summoned the vision from his or her mind? For a theory of agency, the difference is significant. It is not so clear that it matters from the perspective of narrative. The distinction between having no context at all—either because it was a creative eruption or because the vision was a genuine supernatural phenomenon—and having an unknown context, one that unfolds off stage, is not, in the end, a meaningful one from the perspective of agency as a property of dramatics. This is a limitation of our method, but also, as we shall see, its most liberating feature.

CONCLUSIONS: EMPLOTTING POWER AND AGGREGATING STORIES

Professional anxiety over the Fatima story comes from a very different place than those forces that marginalized it and its contexts in the historical record. Supernatural agency is not only compatible with the Yacoubists' interpretation of their own history, it is demanded by it. What makes it quite difficult for most of them to embrace the story of the vision of Fatima is the firm rejection by the community's leadership of any suggestion that women took a leading role in these formative events. However, the community's own narratives cannot account for the vast disparities observed between the number of women and men who openly declared their support for Yacouba. Nor can they account for what is, for other Hamawis, the most scandalous feature of Yacouba's community: the active participation of women in dhikr and haidara.

Thus, the explanations presented here rely on manifestly “intrusive” readings of the sources. They do not recuperate authentic subaltern historical visions encoded in the oral sources and suppressed by colonial hegemony. But neither do they depend on Rashomon-style appeals to the power of perspective, or argue for the “undecidability” of knowledge. For the “found” narratives do not contradict the “made” narratives on any of the basic facts about women's roles; rather they simply dismiss women's relevance by dismissing their explanatory significance in the broader context of the story they wish to tell. We make our interventions and transform the dramatic agency in the story by changing setting, demonstrating that some different context reveals a hidden meaning in the evidence. Historians are dependent on re-contextualization if they want to do anything other than reproduce the arguments embedded in their sources, and such re-contextualization, in turn, grounds and imposes the historian's interpretive authority.

What historians actually do when they read sources “against the grain” in order to “give voice to the voiceless” or “make the subaltern speak” ultimately has little to do with theorizing about constraints, mentalités, or “the limits of the possible.” Historical figures act as if they had agency when we construct our stories so that they might. To ask if they actually had this agency outside of our stories is to misunderstand what historical agency is, to take it for a property of lived experience rather than of experience explained. If we formulate questions about our own practice in terms of distinctions between events and contexts rather than a dichotomy between structure and agency, we enter a terrain that we are much better equipped to negotiate.

This is not to say that we can separate techniques of narration and emplotment from more general statements about power. It is precisely a commitment to identifying the operation of power on different scales of experience that leads us to examine the dramatics of narrating human action. What we have generally misunderstood, however, is that what power means for humanistic historians is not necessarily the same as what it means in the social sciences. As a practical matter, power affects agency in our narratives less by constraining the actions of persons within them—for historical narratives should always strive to protect contingency—and more by limiting the contexts available to emplot those actions.

One consequence of this is that the choice of which re-contextualization should structure our narratives is not one reducible to empirics, and while alternatives can be concatenated they cannot be aggregated and still retain the dramatic effect of attributed agency. Any one of these stories could stand “on its own” as an explanation of women's participation in the Yacoubist revival, but none can do anything other than thematize the others, subsume them into its own narrative and, therefore, its own particular locus of decisive action.Footnote ⁶⁸ Whether the women's actions emerge, following Dwyer, out of the context of changing social obligations and needs; out of the opportunity for self-fashioning provided by momentary weaknesses in the institutions that reproduced gender hierarchies, as the moral panic literature might suggest; out of a commitment to a project of spiritual reform mobilizing a wide range of practices and norms to fashion a new form of Islamic piety; or as contextless carriers of a discontinuous impulse of moral and cultural transformation; in each case their agency is a function of the way the narrative stages the interplay between their actions and these specific contexts. What we choose among, then, are not so much explanations or forms of agency, but unique settings.

Some of the most difficult challenges are posed by limits on the ways that these settings, and therefore our stories, accumulate, and on how our historical actors fit into the broader pictures fashioned for the consumption of non-specialists and even non-professionals. As Steven Feierman has pointed out, histories that are always cautious to frame subaltern agency within the constraints and discourses of domination—and indeed, which deem it the height of agency to “displace” or “appropriate” those constraints and discourses—can reinforce the false universalism according to which only stories centered in the metropole, which employ explanatory contexts grounded in knowledges taken implicitly to be shared by historians and their audience, can be articulated in professionally acceptable languages.Footnote ⁶⁹

One might construct a story in which Yacoubists responded to the changing socioeconomic conditions brought by colonialism, to French discourses of liberty and equality, and to administrative policies that granted religious freedom while simultaneously equating heterodoxy with subversion, and fashioned a hybrid religion out of these forces and their own “traditions.” This would be both perfectly plausible and in keeping with current orthodoxies about the capacities and limits of domination. It would also reproduce dominant patterns in the literature on African Muslim communities in which women are virtually ignored, in the literature on African Islam in which local beliefs and practices are explained by reference to a reified and theologically normative “tradition,” and in the new colonial history that interprets such movements in reference to Western discourses of freedom, rights, modernization, or democratization.

Instead, each version of the origins of the Yacoubist community presented here becomes progressively less reliant on Europe's implicit knowledges and the stories it has aggregated to itself. The setting of each moves farther from the colonial state in its guises as motor of social change (e.g., dissolving old forms of dependency), or purveyor of symbols of modernity (e.g., “free labor” or “equality”). The result has not, however, been to generate those current shibboleths: authenticity, essentialism, or autonomous action. Rather, it suggests that the most powerful form of agency is that which displaces not the structures of dominance, but rather the historian's own expectations and ingrained suppositions.

Some of the most powerful recent attempts to confront disciplinary strangleholds on interpretation have sought to disrupt hegemonic narratives by opening up historical explanation to inhuman agency. By “technologizing” and denaturalizing agency, historians like Timothy Mitchell and Dipesh Chakrabarty and sociologists like Bruno Latour have reminded us of the ways our accounts can serve to reproduce a liberal, euro-centric notion of the subject.Footnote ⁷⁰ But the costs of such maneuvers are high. They achieve their analytic power by sacrificing the humanity of their actors, and their critiques of rationalism easily turn to critiques of reason itself. The approach I suggest here constitutes an assertion that such moves are unnecessary and, from a humanistic perspective, undesirable. For if there is a potential foundation of a loosely bounded humanism with which historical practice can be in deepest sympathy, it is the way that meaning is generated through storytelling and explanation through loose contextualization. As Ranajit Guha recognized over twenty years ago, the contradictions involved in seeing the subaltern as the subjects of their own history cannot be resolved by those who distance themselves from “the prose of counter-insurgency only by a declaration of sentiment.”Footnote ⁷¹