Masks Versus Cattle: The Ecology of an African Art Form

Abstract

Masquerades form an important part of indigenous African religions, and headpieces of masks are found all over the world as an iconic African art form. Contrary to expectation, however, the geographical distribution of masking is very restricted, occurring only in three areas of the continent, mainly situated in the forested zones. I explain this peculiar distribution by tracing the ecological and historical parameters of the societies that harbor mask rituals. The main ecological factors are the absence of cattle and the type of horticulture, while a crucial historical element is the long history of slave raiding and trading to which the continent has been subjected. These factors have led to societies in which both gender and local power arrangements show a marked indeterminacy, operating as internal arenas between men and women, and between different groups of men. Crucial in the explanation of masquerades is the fact that masking is completely absent from societies that practice cattle husbandry with the Bos indicus species (zebu). Since sleeping sickness prohibits cattle husbandry, wherever the tsetse fly thrives, masks appear.

The African Masking Configuration

My topic is the phenomenon of the African masquerade, a ritual type that is highly specific and has garnered much attention in ethnographic and art historical literature. I address this religious expression from the perspective of the general environmental turn in anthropology, reminiscent of classic cultural ecological explanations.¹ Since the distribution of mask rituals² is limited to just three zones on the continent, the central question is why this is the case; for this enquiry I use the lens of human ecology, by analyzing the ways in which historical, socio-ecological factors impact upon the local societies which house mask rituals.

Most people encounter African masks in museums or private collections, so the first association with the notion of “mask” is probably as an exotic headpiece, with a degree of strangeness and a strong visual appeal. The African definition is holistic: a “mask” is the entire costumed appearance, consisting of headpiece, costume, any adornment, and whatever is held in the hands; in some cases the same costume and headpiece may even become a different mask simply by changing the accessories. To this generalized African definition one has to add an invisible dancer, as well as a specific dance routine and music. The mask is part—even the center—of a mask ritual, and it is these rituals that form the subject matter of this study.³

Masks rituals form a recognizable and recurrent configuration that stands out in five ways. First, a mask ritual is highly orchestrated, is associated with a great sense of drama, and has layers of meaning; in more than one sense the mask is the ritual, so masquerades form their own category inside the local ritual repertoire. Second, masks figure within a restricted set of ritual types only. For instance, they only occasionally function in healing ceremonies, rarely in sacrifices, hardly ever in cyclical, agricultural rites, or rain rituals, and they never appear in marriage or birth rites. Rather, they are crucial in two rites of passage: initiation and death. In addition, power holders such as kings and secret societies use masquerades for the maintenance of law and order. All mask rituals are intensely gendered, functioning in male rituals such as boys’ initiations and funerals of men, but hardly ever in the female counterparts to those rites. Though masks form a separate category in most cultures, they have a strong association with the bush that surrounds the village, with ancestors, and with general wildness, a world deemed to be accessible to men, but not to women. One major taboo present throughout masking cultures is the threat that masks pose for the fertility of women, a pervasive notion that defines the world of masks as exclusively, even aggressively, masculine.

Third, masks form the focus of an ideology that hinges on the notion of the secret. The fact that masks are costumed men is deemed to be known only to those initiated into them; the uninitiated—mainly women—do not have the right to that knowledge. While this secret is as public as it is obvious, masking does raise a persistent virtual barrier between genders. Fourth, this ideology of secrecy not only pervades gender relations but, in many cases, undergirds an internal organization based on male initiation and age called a secret society; such a power association uses masquerades in its initiation ritual, cuts through familial connections and crosses village borders, and serves as an important player in the local political arena. Finally, masquerades are among the most spectacular rituals in Africa, and masking societies are performance-oriented, as well as highly appreciative of visual displays and art forms in their material culture: this is a ritual that people cherish, and by which they define their own culture and ethnic identity.

These five factors make up the mask configuration: an apparition set inside a captivating performance that forms a high point of ritual and social life, the core of a power organization, and the focus of an ideology that defines gender relations and transforms boys into men. Wherever masks are found, this whole configuration tends to be present (Fig. 1).

Fig. 1

Two “tree” masks in the Dogon village of Tireli. In Dogon, the notion of the tree carries a broad range of associations, all of which revolve around the notion of “bush,” the source of all power. Photo: van Beek (2008)

Methodology and the Map

Explaining the large-scale distribution of any cultural phenomenon calls for a broad comparative study of societies with and without masks in order to identify both commonalities and systematic differences. This requires substantial ethnographic data, both contemporary and historical, most of which is found in secondary sources, anthropological as well as art historical. The central question was triggered by my own field research in Africa among one group without mask rituals, and one in which they are central.⁴ By implication, the historical period covered depends on the ethnographic sources, the majority of which stem from the late colonial period up to the present, with an emphasis on the early post-colonial era. A broad comparison necessitates a generalizing approach to local conditions, so I mix generalization with “causal stories and histories that recur in different times and places” (McCay, 2008: 2; Vayda, 1995: 363) in order to discern patterns in historical and social formation⁵ and to identify systematic similarities and differences between masking and non-masking cultures; this research design calls for special attention to any exceptions in order to avoid confirmation bias.

Using the available ethnography to map masking societies in Africa, I and my colleague Harrie Leyten examined 92 societies in or around three zones and identified 77 reported to host mask rituals (Fig. 2).⁶ Fifteen societies inside or close to these zones have no mask rituals, and neither do a dozen societies beyond the main zones we included to check our hypothesis. Plotting societies on such a map inevitably implies a “yes/no” answer to whether mask rituals are present, which tends to efface internal variation. For instance, mapping does not account for the relative importance assigned masks, since their role can vary widely. In practice, the relative importance of masks in rituals forms a continuum ranging from marginal masking, where a few rituals feature a single mask, to intense masking where elaborate rituals revolve around many. Moreover, the individual histories of specific masking types fall beyond the scope of the map. The same holds for the difference between fully functioning masquerades and extinct ones; in quite a few cases masks are on the verge of disappearance, or have now disappeared, while even in ethnic groups that do have mask rituals, not all villages have them. All of these are counted as masking groups.

Fig. 2

Distribution of ethnic groups with masks in Africa. ^© DeVink Mapdesign 2020

Mask rituals are found in what I call the “Mask Crescent,” a broad band running from Senegal eastwards over the whole width of the subcontinent, and then curling downward around the Bight of Biafra, as far as the eastern limits of the Congo Basin, plus a sprinkling in East Africa. The Crescent is not continuous and divides into three distinct zones: West Africa (Zone 1), West–Central Africa (Zone 2), and Central–East Africa (Zone 3). In the mask literature, the interruption in the Ghanaian region has been dubbed the “Akan Gap” (McNaughton, 1992). Not all ethnic groups in the three zones have rituals with masks, but many do, and very few cases of masking cultures occur elsewhere.

The Human Ecology of a Religious and Artistic Phenomenon

Masking is a high-investment and high-impact ritual and my quest is for the rationale behind such an effort. Masquerades have been the subject of numerous publications, mainly from the perspective of art studies, most of which focus on the many layers of signification and history of the headpieces. I look at the phenomenon of masking as such and argue that the peculiar distribution of mask rituals in Africa is determined by specific ecological, historical, and sociological factors. My focus is on the social fields or arenas that result from specific eco-historical contingencies that are conducive for masking, primarily gender and power (Fig. 3). This schema highlights the classic approach of human ecology, as it developed from Julian Steward and Marvin Harris, to Andrew Vayda and Tim Ingold (see Jochim, 1981; McCay, 2008; Moran, 2000: 59–64). Each social formation or local society is largely informed by two general factors: its ecological adaptation and its history; the enveloping semicircle highlights their status as important background factors. The ecological adaptation—the way a community has adapted to its physical environment in extracting resources and making a living—enables the choices and options the society has for acting within the context of its physical setting, and these choices shape many aspects of the sociocultural habitus of the group. Two fields within the social formation are hypothesized to be crucial for masking, namely gender and power. The bold circle indicates mask rituals as the central topic. In turn, the mask rituals inform local and ethnic dynamics of identity construction. The arrows point in both directions, since the relations are reciprocal, but—as indicated by their thickness—not always in equal measure.

Fig. 3

Theoretical schema

My basic thesis is two-tiered: first, given certain historical and ecological dynamics, societies tend to develop marked gender and power arenas; second, these arenas facilitate the creation of mask rituals, which in turn impact on both arenas. The ultimate causes reside in ecology and history, while the proximate causes of mask rituals lie within the local communities, with a crucial emphasis both on the relationship between the genders, and on the different forms of male bonding. I define as an arena any social field characterized by contestation, competition, or at least a noted indeterminacy, meaning that the relations among the various actors or groups in the social field are not fully crystallized, and are subject to negotiation and diverging definitions. The fields of gender and power are not necessarily arenas in all societies, but only in certain social formations and under specific conditions. I found that masquerades predominantly function in societies where one or both these fields have the characteristics of an arena—usually both, since they are related.

In local communities in Africa, gender means the social construction of being either male or female. Gender is an important social field anywhere but constitutes an arena if the relations between men and women are under negotiation, occasional redefinition, or contestation. That is not the case in all cultures, for in societies with either gender symmetry or rigorous gender asymmetry there is no indeterminacy. Masquerades appear when gender is an arena and show this clearly since the masks are consistently gendered themselves. Mask dancers are almost always male, for even if a mask portrays a woman, it is a man who dances with it. Not only are mask rituals a masculine domain, but masking rules exclude women, performances aim to frighten women, and, on the face of it, the ritual world of masks appears to be definitely unfriendly to women; numerous publications have stressed this male dominance in masquerades, even in studies highlighting the role of women inside the masking complex (e.g., the Senufo, Glaze, 1981, and the Diola, Langeveld, 2003). That does not mean that the women have no recourse to ritual ascendancy themselves, but it does point to the mask rituals as an important factor in the local gender balance.

The second arena is power. Two meanings of the term are important here: the sociological concept that marks asymmetric relations among people, and the more religious one implying the force inherent in the otherworld—the power of gods, spirits, ritual, and magic. These two kinds of power are very closely related, and masking is one major way in which the sociological power system merges with the otherworldly force (see, e.g., Ellis & Ter Haar, 2004 for the close link between religion and politics). Power is a function of all social relations, but one main issue in African local communities is the manner of male bonding, both at the local level and in higher echelons. Again, the field of male power is considered an arena if the relations between men are under negotiation and periodic redefinition or contestation. Political systems with rigorous hierarchies, fixed positions, and clear inheritance rules appear much less prone to masking than polities in which different sources of power are in competition, and where power positions shift with personality, age, and wealth. Mask rituals are seen to operate in political configurations with a degree of indeterminacy—open power arenas—and with diffuse power systems, and that indeterminacy is one of the factors underlying the masking phenomenon. That raises three questions: What is the common denominator in the ecology and history of masking societies? Why do the fields of gender and power have the characteristics of arenas in these societies? And how do mask rituals interact with these arenas?⁷

Ecology and History

The three masking zones are in different ecological areas, but a number of ecological and historical factors combine to generate similar adaptations, so a generalization is possible. The forested areas of Africa form the core of the Crescent, with extensions into dryer ecotopes, where slash-and-burn horticulture or crop rotation and fallowing are practiced. Women are largely responsible for the production of food, particularly root crops such as cassava and yams, alongside some rice, sorghum, and millet. This subsistence economy is supplemented by a wide variety of cash crops such as cocoa, palm oil, nuts, or pineapple—or, in the drier areas, cotton, tobacco, and peanuts, generally the responsibility of men. Usually, the men own the land the women cultivate (Kaberry, 2005: 35). Our ethnographic sources mostly stem from the late colonial and early post-colonial eras, where in the masking zones the colonial system demanded the production of cash crops, in which male labor was largely engaged. In some parts of the Masking Crescent, the forcible conscription of men to labor on large-scale commercial farms owned by the colonial powers increased the need for women to care for staple crops (Guyer, 1984). All in all, in the Crescent women tend to feed the family.

The second and more long-term impact in the Crescent has been slavery.⁸ The Crescent includes most African areas scourged by the slave trade, but the three zones do not narrowly coincide with the historical areas of major slave trading. Zone 1 was not central in slave trading but the area between Zones 1 and 2 was a major slaving center; in Zone 2 Calabar and Douala were indeed major slaving ports. In Zone 3 the coast of Gabon is largely mangrove forest with few ports, and Angolan slaves were transported by the Portuguese to Brazil both from Cabinda and Luanda, but also from Benguela, far south of the Mask Crescent; anyway, the actual masking areas in Zone 3 are far from the coast.

Nevertheless, the influence of slavery on masking in all three zones has been profound, even if rather indirect. While the presence of Europeans on the coast as such did not lead to masks, it did create a general insecurity in a much larger area, an insecurity which generated societal types that would prove to be conducive to masking rituals. Under the influence of Atlantic slavery, centralized local polities—which included some masking societies—developed at the coast, which organized trade networks and slave raids far in land (Goody, 1971), thus taking the shockwaves of the enslavement way into the interior.

About the same time as the intensification of the Atlantic slave trade, in the northern savannah belt a string of jihads marked the establishment of Fulbe emirates throughout the West African Sahel until the end of the nineteenth century (Last, 1974). These cavalry-based states were organized by wide kinship networks and alliances and survived longer than the rebellion-prone states of the coast. However, they also depended on a “slave mode of production” (Lovejoy, 2012), leading to ever larger numbers of enslaved people on slave-run plantations. Slave raiding and trading was pervasive among these revivalist Muslim movements that persisted until well into the nineteenth century (Last, 1974; Lovejoy, 2012; Wolf, 1997).

This dual threat of enslavement from the Atlantic south and the Sahelian north held mainly for Zones 1 and 2. The wide swathe of Zone 3 with its great rivers, such as the Congo, formed another axis of slave raiding and trading, and the impact of the Atlantic trade depended on having access to these trade arteries and good ports (Vansina, 1990). Trade, mainly in ivory, began with the arrival of the Portuguese in the sixteenth century. With a growing demand for enslaved labor in the Americas, the extraction area moved further inland, by-passing those areas closer to the coast that were hard to reach. Owing to the large distances, an elaborate transport system emerged, creating a network that branched ever deeper into the interior.

The low population densities and migratory subsistence patterns in Zone 3 allowed people to move in order to avoid the threat of enslavement, opening new frontiers of forest and establishing isolated settlements with little centralization and few external contacts, developing secret societies and healing associations that furnished links among them. In the areas out of reach of the trading system, internal political dynamics prevailed. In Zones 1 and 2, with their much higher population densities, movement was not an option. Some communities could organize a military defense, but that demanded a cavalry and only happened occasionally (e.g. the Mundang of Cameroon, Adler, 1982). The general reaction to endemic insecurity was to fortify villages or to settle on inaccessible terrain, or to use other available means of defense, resulting in what Wolf (1997) calls the “shatter zones:” regions with many small ethnic groups huddled together. Mountains and hillsides form a prime location for such defensive settlements; flood plains are another option. But West Africa has few mountain ranges or major rivers, so many settlements had to rely on defensive architecture. These isolated defensive communities functioned as self-sufficient entities, without intensive links to other settlements: the very preconditions for intricate local power arenas.

The Arenas

The significant role played by women in food production, alongside historical insecurity, led to specific community characteristics conducive for masking as the fields of gender and power morphed into arenas. One crucial social element is matrilineal descent. Although many patrilineal cultures exist in the three zones, most matrilineal societies fall inside the Crescent—as do almost all known African societies with double descent (see Murdock, 1959). As Harris has argued (1979: 64), women’s responsibility for food production does not in itself lead to matrilineal descent. However, in combination with internal wars and surely with the continual threat of slave raiding, this type of descent reckoning becomes productive: the men are often absent for long periods and the women have responsibility for the household or extended family, and control over land resides in their brother. Thus slavery was probably a major factor in either generating or at least institutionalizing matrilineal descent (Vansina, 1990). Moreover, matriliny can undergird an array of political systems: all the zones, but especially Zone 3, show a wide range of centralized and decentralized political arrangements based on matrilineal descent, and this flexibility of political organization is crucial in coping with high physical insecurity.

Descent reckoning covers a wide array of kinship relations. As Forde noted: “Any principle of [kinship] recruitment can itself be a matter of degree” (1964: 38). Thus, matriliny forms the kinship logic underlying very different types of unilineal systems, ranging from small, nameless, and genealogically shallow mini-groupings to named, omnipresent matriclans with clear leadership structures (Holy, 1996). But whatever form matriliny takes, some dilemmas in this descent system persist.

The local power arena is closely linked to descent. In patrilineal societies a core of men share the same descent group, and thus easily form what Keesing (1975) calls the “board of directors”; they live with their paternal lineage members, their wives have left their parental homes, and the localized clans and lineages provide the social and political structure of the village. In matrilineal systems, while descent reckoning follows the mother’s line, it is the men who occupy positions of public representation and community power, with the women more confined to—and often dominant in—the domestic sphere. This so-called “matrilineal puzzle” forms a challenge for men to organize themselves, and the options for male bonding depend largely on residence rules, the place where a new couple will live. One logical option would be to take up residence with the husband’s mother’s brother, avunculocal, since that would unite men from the same matriline in one spot, but that calls for the couple moving to a location where neither of them grew up, so this residence rule is rare—and avunculocal societies tend not to have masks indeed.⁹ More common are matrilocal residence where a man lives with his in-laws, and patrilocal residence where he stays with his father; however, the latter belongs to his own mother’s lineage, while the man’s main lineage authority is his mother’s brother, who lives elsewhere. In neither case does the man live with closely related males, and the male groups tend to be small and short-lived, dependent on the success of the individual man. For instance, among the Bemba, a masking society on the border between the Democratic Republic of the Congo and Zambia, a man provides labor in the fields of his in-laws. If he works hard, begets children, and accrues wealth and prestige, he attempts to relocate his wife and children to a newly built settlement, his own “village” where he is boss, but has few male lineal relatives (Richards, 1982).

Variations abound and any system is dynamic. Whatever the rules of marital residence, the various forms of matriliny result in villages with a strong, independent position for women, a tight bond between a woman and her family, and a lack of self-evident bonding between males. Genealogically, the lineages tend to be shallow and the male power structures are weak; the villages often split or move with the slash-and-burn agriculture of the forest. This means that the power balance between husband and wife is skewed toward the wife and is under continuous pressure, since the wife retains a strong bond with her brother: after all, she furnishes his heirs. Male bonding can rely only to a small degree on close kinship between males, and is under continuous redefinition—in short, an arena.

Not all masking societies are matrilineal, for many groups in the Crescent are patrilineal and yet have masking rituals. The main factor is the same though: male political indeterminacy, showing as an open power arena, even if the reasons for this differ. These societies are part of an ecosystem in which female labor is important. Especially in Zones 1 and 2, these patrilineal villages result from defensive measures against insecurity, and are deeply rooted in small territories. They are politically self-sufficient and show a high village endogamy: most women marry within their natal village and so remain in easy contact with their kin.¹⁰ Although the husband holds more sway over his wife than in matrilineal cultures, her position is strong as she remains close to her support group.

The main capital in the societies of the Crescent is labor. Access to fields is usually quite open, with cultural norms of sharing and distribution that ensure that all able-bodied men and women dispose of cultivable land. Hence, the patrilineal descent groups do not control access to resources, and thus have to compete with other types of male bonding, such as voluntary associations, and groups based on age. The result is a mesh of different types of power positions, so that male bonding in these villages forms an internal arena in itself. Though in any local society descent groups are permanent, age shifts continuously, meaning that these societies show a structural tension not just between men and women but also between old and young men; even if the old hold the reins of power, they depend on the young for labor.

In many of the patrilineal societies of the Crescent, the strength of the lineage system is therefore curtailed by principles other than descent.¹¹ Such a village tends to be a rather closed universe, with descent groups operating only within the village limits: another village means another set of clans—the fragmentation by historic threats is still present. Although religious and economic ties may provide links between the villages, kinship seldom transcends the village border. Whenever power associations—also called secret societies—occupy this political vacuum, they serve as links between the autonomous settlements. These associations have a prime link with mask rituals: when secret societies arrive on the scene, masks can never be far behind. In matrilineal societies, too, with their weak male bonding, there is ample social space for such power associations as an alternative way of male organization, and these associations invariably command the masking rituals.

Children are crucial in these gender and power equations, so fertility is important. This holds for other parts of Africa as well, but masking societies in the Crescent combine a pronatal attitude with a low bride wealth; often the groom works for his in-laws to procure a bride (bride service) (see Goody, 1973). Matrilineal systems transfer little bride wealth anyway, since few rights are actually transferred, but the patrilineal cultures of the Mask Crescent have a similarly limited flow of goods from the groom to his in-laws. In quite a number of Zone 2 societies, bride wealth is “a symbolic token in the form of services, goods or money, given to the bride’s parents by the prospective groom” (Njoku, 2020: 83). In some instances the idea of bride wealth is explicitly rejected—“a bride should not cost anything”—and just a ritual exchange of minor gifts with symbolic value accompanies the bride’s move to her groom’s compound.¹² Alternatively, bride wealth may comprise luxury goods and expensive consumables; capital goods are rare. Because people in the Crescent tend to marry within the village, both spouses can easily trace several kinship links and cross-cousin marriage is not uncommon.

The Masking Option

Within the two arenas of gender and power, it is masking that offers men the means to construct a local stronghold in the power balances. The most obvious effect is in gender relations. The masking configuration is an endeavor that provides the men with a significant ritual edge over the women, based on religious notions. Masks are deemed to stem from the bush, often as embodiment of bush spirits, and form a threatening presence to the women. Not only are the women not initiated into the masks, but they are even supposed not to know that there are men dancing in these costumes. It is difficult to imagine that the women really do not know this, since in African villages very little remains hidden, but the discourse of the secret is strong. When Ndembu boys enter the forest for initiation, the men insist to the women that “monsters had eaten their sons,” and they had a “knowing laugh” about the women’s gullibility (Turner, 1967: 224, 231). In fact, the women had to provide food for their sons and knew exactly for whom they were cooking. But the women play along for solid reasons. The powers inherent in the mask are taken to be inimical to their main asset, fertility, and any contact with even the smallest part of the mask—say the fibers of the costumes—is risky. In addition, they are subject to threats during the performances, for instance when the leading elders shout that the “good heads have to kill any woman they see” (Van Beek, 1991). Even if the secret is public, the fear is real: women are often genuinely frightened by mask performances even if they know that members of their own family are dancing (Binkley, 1987: 92) (Fig. 4).

Fig. 4

Rabbit masks at the 2008 mask ritual in the village of Tireli, Dogon country, Mali. Photo: Walter van Beek

So masquerades serve as a constructed male power base in a unsteady gender balance, and the power arena is closely linked with the gender one. The masking configuration often leads to a voluntary male organization that transcends village boundaries; indeed, secret societies abound in the Crescent. Such societies are constructed upon ranks which men can obtain by a series of initiations, usually with masks; in constructing ever more “secrets,” the public secret of the mask morphs into a generalized system of concealment. Initiation secrets tend to be cognitively empty, meaning that there is no semantic or encyclopedic content, just a regulated discourse on concealment and disclosure (Taussig, 1999). But the ownership of constructed secrets becomes a source of power for the men vis-à-vis the non-initiated, and changes a diffuse power system into a ranked one that is structured on limited access and controlled—even orchestrated—disclosure. Secret societies dichotomize the village into the initiated and non-initiated and divide the initiated into echelons of power through progressive “revelations.”

Not all women have to be afraid of masks. Postmenopausal women have little to fear from masked figures and are often quite familiar with them. But some younger women may also have a special relationship with masks: many African masking systems know a category of “sisters of the mask,” who provide the dancers with food and drinks; they show no fright whatsoever, since as partial initiates they have nothing to fear from the masks as to their fertility. Both exceptions prove the rule of the increased gender separation by masks. But the very mask ideology itself enhances the status of women in a positive way too, undergirding their pivotal role of mothers, as the ultimate progenitors. By its symbolic performance, the masquerade positions the masks as an alternative form of fertility (Van Beek, 1991), but for its efficacity it does rely on the consent of the women. Through their overt denial of women’s importance during the ritual times, masks ultimately stress the crucial value of fertility, a dialectical notion that shows to the full in the concept of female genital power (Grillo, 2018). Whenever men-and-masks transgress a boundary between the genders, women may use another means to balance the power between the sexes, with their genitals. For instance, the women of the town disrobe and together roam the streets naked, the postmenopausal women dominant among them. In the Crescent this is the ultimate show of basic power, that of female genitals, and men have no recourse whatsoever against it. This genital nudity is defined as the mask of the women, and women tend to use this as a last resort, as the countervailing threat against male abuse (Van Beek & Leyten, 2023: 206 ff.).

Finally, for men as well as women, the masking option is attractive since it offers the whole society a captivating performance of spectacular apparitions that holds the audience spellbound. Masks are unthinkable without a public, and for the young men dancing them the girls form their main target audience: “Without women, dancing has no sweetness” (Vogel, 1997: 180). In village life these dances are the absolute highpoint, and groups strongly identify with their masquerades, between villages and between ethnic groups, while some masks nowadays even serve as icons of a whole nation.

In short, the masking configuration provides the men with a power base against the women, and a structuring device among themselves, while it undergirds the position of the women with an ideology of fertility; for all concerned it is highly pleasurable and socially valued. Given the general characteristics of the Crescent societies, it is surely an attractive option for the men, and in many ways also for the women.

The Zebu Exclusion

Thus far I have shown why societies within the zones would be “masking-prone,” but why are masquerades absent in the rest of the continent? Three types of borders on our map call for an explanation: the internal ones of the Congo Basin and the Akan Gap, and the external one separating the Crescent from the bulk of Africa.

For the Congo Basin the main factor is demography. In both the rainforest and the forested savannah, subsistence is based upon a wide spectrum of food production, with foraging and fishing as prime activities, and a very flexible division of labor. Mobility acts as a defensive strategy against slave raiders or other external threats. These societies consist of small groups, hamlets, and bands that coalesce, fission, and fuse again in a constant dynamic, and any leader depends to a large extent on personal abilities.¹³ The crucial arenas that mask rituals address are not present in such a widely dispersed population, which precludes such rituals. In this basin the population density remains well below four per square kilometer; it is only beyond that “historical watershed” (Vansina, 1990: 99), that processes of village formation and political centralization appear. Below that limit, the gender and power arenas evaporate between the trees, and masks with them.

The Akan-speaking groups in southern Ghana are clearly matrilineal, herd no cattle, and surely have excellent sculptors (Cole & Ross, 1977), but nevertheless have no masks. Again, residence rules provide a clue for this absence. Southern Ghana shows a peculiar rule of marital residence called duolocality (Abu, 1983), in which husband and wife do not cohabit but continue to live apart, each with his or her matrilineal relatives. Usually, the spouses live at close quarters, and the wife can visit her husband at night, an arrangement that coincides well with individual agency, and mobility for the wife when doing business—in which her husband should invest; it also makes divorce easy (Hagan, 1983). Worldwide, duolocal residence is rare, and occurs only in matrilineal settings, where it has the advantage that brothers do not split up, and neither do the sisters; brothers work together on their tasks, and women club together over their own chores. This feeds into a strong matriline, since all matrilineal-related males live together in one household. This solution of the matrilineal puzzle comes at the cost of what most cultures would define as the core of the marital union: a nuclear family that lives together. The same holds in Akan: the classic tension in matrilineal systems—that between husband and brother over who controls the woman—is absent here, because the husband simply does not live with his wife; her brother does. Since the latter is dependent on his sister to have inheritors, the position of the woman inside these matrilineages is fully embedded and extremely strong (Kwame, 1983). Significantly, in their position as mothers the women wield considerable power; they have their own place in the palace, and their own ritual seats—the famous stools—while some even earn a reputation in warfare.

The political structure among the Ashanti—the central Akan group—is clear and unambiguous as well: in the palace, in the kingdom, in the chiefdoms, down to the individual households, men and women hold undisputed positions within the political system. This stability is based upon the structure of the matrilineage mentioned above and upon the many rituals surrounding the throne. What is also remarkable is the absence of secret societies. Ashanti men were organized in military companies, usually joining that of their father, and had to perform military service, entirely within the state system. Such a hierarchical, matrilineal, and duolocal society lacks the arenas addressed by mask rituals, because the indeterminacies that feed masks are simply not present. Domain demarcation between the genders is institutionalized, the power configuration is not up for debate, and the authority of the king remains unchallenged. In such a system, mask rituals or secret societies would be subversive.

As for the external border of the Crescent, the systematic and large scale difference between masking and non-masking societies lies in the absence of cattle from the three masking zones, and the main ecological factor is the absence of sleeping sickness. The core of the Crescent is made up of the wetter parts of Africa, which house the tsetse fly, Glossina palpalis or G. fuscipes, the hosts for Trypanosoma brucei, the vector for sleeping sickness, an affliction that attacks wild animals, but also cattle and humans. Since these flies need shade and therefore thrive in dense vegetation, this illness made cattle husbandry impossible in the broad forest band in which the Crescent is situated, at least before the tsetse eradication campaigns. The combined distribution of G. palpalis and fuscipes is almost isomorph with the mask zones (Fig. 5). It is not the threat to humans that is relevant here but that to cattle, and the contours of the Masking Crescent are for a large part defined by the absence of Bos taurus indicus, the zebu, the bovine subspecies that is dominant in African cattle husbandry, and lacks resistance against the vector.¹⁴

Fig. 5

Distribution of Glossina palpalis, the “riverine species” of the tsetse fly. The second variant, Glossina fuscipes, has a slightly larger distribution, extending towards the eastern side of Africa, as far as the Kenyan and Tanzanian coast

Masking cultures do not herd cattle; and wherever cows appear, masks tend to be absent. Why? The main reason is that cattle are capital, an asset that can be accumulated and that requires management, as well as organized transfer between generations: cattle form the epitome of inheritable wealth. This has crucial social implications. Pastoral cultures are usually patrilineal, with wide-ranging clan and lineage systems, deep and segmented genealogies, and kin-based management systems that include vast tracts of land and a multitude of settlements and villages. Thus, cattle-keeping leads to segmentary societies, with cattle playing a significant role not only as capital but also as the focus of a pervading cultural fascination which leaves little room for mask rituals.¹⁵

The clearest example of this exclusion is the sharp eastern border of Zone 3, for the masks never seem to cross the western branch of the Great Rift, the mountain ranges that cradle the string of Great Lakes. This precise border is quite revealing. The eastern rim of the Congo Basin houses cultures that narrowly fit my analysis, while the countries of Uganda, Rwanda, Burundi, and Tanzania have a completely different cultural habitus. East of the Rift one is in cattle country, with societies that both in ecology and in social organization heavily depend on cattle. In these societies the zebu is all-important as the main source of livelihood, and as icon of wealth. Cattle herding trumps horticulture, and the animals are milked; men herd, women milk. These people drink cow’s milk, whereas people in most of the Mask Crescent cannot digest it well.¹⁶ Herding means mobility and, even if full nomadism is rare, cattle husbandry demands movement for pasture over large spaces, since the African savannah is not nutrient-rich. Such a way of life generates societies with large lineages of considerable depth, and complex kinship systems that extend over vast distances, and thus cultures that are in constant interaction with surrounding groups—almost the opposite of the societies in the Crescent.

Herding societies are strictly patrilineal and patrilocal, and their local patrilineages not only manage important capital, but also dominate marital transactions. Large bride wealth is required for marriage, implying the transfer of a substantial number of cattle, since each marriage involves a serious exchange of wealth for the rights in the bride and her offspring.¹⁷ In these cultures the gender relations are more skewed and much less a matter of negotiation, and the dominion of men over women is considerably greater; many of the transactions between men concern payment and restitution of bride wealth, and court cases are full of discussions on paternity. In fact, inherited wealth and accumulated capital make masking pointless; these patrilineal structures have boards of directors that are simply too well structured. In such systems, masquerades as an independent force would be a highly disturbing factor, threatening the social fabric of male power.

Nevertheless, there is a sprinkling of masking societies in East Africa—isolated cases in Malawi and on the border between Mozambique and Tanzania—that are probably the remains of an older and larger spread of masking. The region between the western and eastern branches of the Great Rift historically served as a corridor through which cattle herders migrated from the north to the south (Mwai et al., 2015), sometimes in conquests of neighboring societies. During their migration, these cattle herders intersected with the forested zone of the matrilineal belt, which ran all the way to the Indian Ocean (Murdock, 1959), thus forcing a patrilineal wedge between the Congo Basin and the matrilineal areas close to the Indian Ocean, which were cut off from the rest of Zone 3.

In this area another historic lethal epidemic has been crucial as well. Until the end of the nineteenth century, East and Southern Africa were free of tsetse, but in 1887 a virus from central Asia hit the area, rinderpest.¹⁸ The epidemic spread rapidly and devastated the continent: it killed over 90% of all cattle. Thus, by the end of the nineteenth century the land was emptied of its cattle and its keepers, and the grasslands turned wild quickly, transforming the closely cropped grass into woody grassland and thorn bush, the so-called mopane bush. Wild mammal populations increased rapidly, accompanied by the fuscipes variant of the tsetse fly. It is estimated that most of the present large game areas of East and Southern Africa have been the result of this combination of rinderpest and varieties in tsetse.¹⁹ The East African enclaves of masking have survived thanks to the reprieve through the rinderpest epidemic, plus an internal dynamic which rested on the persistence of their secret societies.

At the southern limit of the Crescent, the line of masking societies closely follows the limits of the open rainforest in its transition to dry forest—the type with less tsetse. The latter area is where cattle can be kept, and in fact are kept. It is an area with a long history of internal wars, of political dominance of one group over another, each with their own specific rules of descent and political centralization. In Angola, the virtual border curves northwards with the elevation of the Angolan highlands; the higher regions are more conducive to cattle husbandry, and so less so to masking cultures.

Due to tsetse cattle are absent in the coastal areas of West Africa as well, yet Zones 1 and 2 extend well into the savannah, and at their northern rim people do have masks in a region where cattle thrive. But here the mixture of agriculture and pastoralism is different from that in East Africa, owing to an ethnic division of labor. In the northern parts of Zones 1 and 2 sedentary groups practice horticulture, while specialized cattle holders have a different ethnic affiliation; the latter are dispersed over the whole area—in fact, throughout the Sahelian zone—and they are predominantly Fulbe. The sedentary masking societies cultivate, and the Fulbe herd cattle and, after the harvest, let their cattle graze on the stubble of the sorghum and millet, fertilizing the fields with dung—in principle, an ecologically sound ecosystem. Some cultivators may own some cattle and have the Fulbe herd them. Though the horticulturalists closely interact with the Fulbe in the economic sphere and their activities complement each other, their cultures do not mesh: they form two completely different types of societies on both sides of the ecological coin; while one has masquerades, the other definitely does not (De Bruijn & Van Dijk, 1995).

Socially and politically this joint ecology is vulnerable and may lead to tensions, a precarious balance of power that may easily tip either way; at present (2023) the problems in the Sahel testify to this ecology’s volatility. As for masking, such an ethnic specialization means that, even though far removed from forested areas, the local horticulturalists fit the pattern described above: independent patrilineal villages with some age organization, a high village endogamy, low bride wealth, and little inheritable wealth. In fact, some of the most exuberant masking traditions are found precisely in this Sudanic–Sahelian belt. North of the masking border—which in Zone 1 follows the Niger Bend—rain-fed grain cultivation is no longer feasible and one is in the real Sahel, where the only ecological option is pastoralism—so no masks. The dip in the northern border of Zone 1 is formed by the Futa Highlands, where the Fulbe stem from, a cradle of pastoralism and a major pathway for Islam, for which masks are anathema.

There is, however, another type of bovine in Africa, Bos taurus taurus, a more ancient resident on the continent, usually called the West African shorthorn or savannah shorthorn (Fig. 6). In the Masking Crescent these cattle occur in small numbers in isolated places, as remnants of older populations. These smaller, shorthorn, flat-backed cattle support forest conditions rather well, resist tsetse better than zebu, and can subsist on a wider range of foods and fodders. Living in isolated pockets in the mountain regions, most of these taurine populations are in fact dwindling.²⁰ Their largest extant population is just north of Zone 2, in northeast Nigeria and north Cameroon (Seignobos & Thys, 1998: 12).

Fig. 6

Shorthorn taurin. ^© C. Seignobos and reproduced with permission

The place of the shorthorn in West African societies demonstrates that it is not the cow as such that precludes masking, but the way in which people interact with it. As Christian Seignobos and Éric Thys show in their overview of taurins in West Africa (Seignobos & Thys, 1998: 9–14), these cattle interact with the horticultural societies of the Crescent in various ways, but always with a limited role in the ecosystem, while their numbers have never been large enough to warrant a pastoral lifestyle—which their habitat would not allow in any case. These cattle remain marginal in the adaptation to the habitat; in these societies horticulture comes first. Taurins are neither milked nor used for traction, but serve as an investment in an economy that offers few other options to invest surplus; they also function as offerings in major sacrificial events. Usually, these cattle are primarily a status symbol, wealth to be spent in the owner’s lifetime, especially at his funeral. They are transient wealth, symbolic riches, and seldom figure in bride wealth transactions. Thus, taurine keeping does not lead to strong and corporate lineages, and societies with masking rituals may raise taurines and vice versa. For instance, among the Bamileke and Bamum of west Cameroon (both masking cultures), shorthorns must have been present for a long time, though they have disappeared now (Blench, 1998). They were left to roam free, considered not to be fully domesticated, and their main purpose was to serve in rituals, sacrificed for prestige and social well-being. Shorthorns do serve as crucial symbols in the cultures that have them and tend to be ascribed with symbolic meanings that resemble those of masks (Fardon, 2007; Van Beek, 1998), but that does not lead to any systematic exclusion between mask and taurine.

Conclusion

To date, masks and masquerades have largely been the province of art historical and religious studies. I have shown how an ecological approach offers new insights into the adaptive regularities that govern the incidence of this form of African art. The map of the locations of masquerades (Fig. 2) illustrates a specific distribution of this iconic feature of African expressive culture. The existing literature on African masks focusses on individual cases—formerly just the head coverings and more recently the performances—and I used these case descriptions and analyses to explore a chain of stochastic causality. My research design sought patterns, based on the assumption that such high-profile and high-investment rituals have to be grounded inside their host societies, and that systematic, large-scale differences between societies are best explained by the material conditions of ecology in tandem with history. I looked for the eco-historical communalities between the Crescent societies, and investigated the generalized fundamental differences from non-Crescent societies.²¹ I started with forest ecologies and the presence or absence of the zebu, which proved to be a main factor of difference with non-forest adaptations, with an endemic illness as a limiting factor. I identified the proximate causes for masking inside the local communities in the ways that the arenas of gender and power were organized.

The results of this endeavor have led to surprisingly reliable predictors of whether a particular society is open for masking. Based upon these observations, it is clear that masking rituals will be absent in the following three cases: societies with segmentary patrilineal systems; societies with zebu husbandry; and societies with strong political lineage and clan systems; the three often combine. Masks will similarly be highly unlikely in societies with very low population densities and societies with other kinds of inheritable wealth or capital. On the other hand one can expect with some confidence mask in societies where the women’s share in food production is large, and mask rituals to be at home in self-sufficient villages with either matrilineal or patrilineal descent plus age groupings, with a low bride wealth and labor as the only capital.²²

I have not explained the presence of masks, but rather the societal proclivity to adopt masking rituals, and not all societies where masks would fit in actually do have them. The proximate causes still need choices made by actors, triggered by specific events that induce people, usually men, to adopt masquerades, either to invent them or, more commonly, to borrow them from neighboring communities. Some cases are reported where people have bought masks from neighbors, even stolen them. Furthermore, my explanation does not predict what type of head cover, costume, or dance will prevail in any specific group: that part of human creativity escapes this explanatory pathway. But neither the wide range of meanings nor the artistic freedom in carving mask tops “precludes them from being subject to causal inquiry” (Vayda, 2008: 320) as far as the conditions for their adoption are concerned.

In the end, each social formation has its own contradictions, its own potential conflicts of interest among categories of people; all solutions to the imperative of gaining a livelihood and living together come at a cost, at least for some. In the Crescent societies cattle cannot be kept, and men depend in large part on women for food, thus facing the challenge of how to control their women, how to organize themselves, and how to link their progeny to themselves. Within the enabling constraints of their ecologically and historically generated societies, the men use masking to construct a virtual stronghold, and to gain an, admittedly fleeting, ascendancy over women. The ultimate causes are ecological, the proximate causes are social, and the mask ideology itself again has an ecological referent. In most masking cultures the masks are deemed to stem from the bush, representing spirits or animals from the wilderness; more generally, they are considered emanations of the power of the bush. Thus, one pervading interpretation is that in masquerades the bush enters the village to revitalize the community. In its symbolism, masking shows that people live in and from their environment and in doing so shape their surroundings (Ingold, 2000) not just as a result of their ecological history but also in the performative construction of their religious art.