All photos by Manu Sassoonian
The site of the necropolis of Bura-Asinda-Sikka in the lower Niger River valley in southwest Niger was discovered accidentally in 1975 by a young hunter who found two small terracotta heads. Three years later the discovery was reported to the Departement d'art et archaeologie (IRSH: l'Institut de Recherches en sciences humaines de l'Universite de Niamey). In 1983, Boube Gado (1944–2015)—head of the Department of Art and Archaeology, IRSH, Niamey—conducted a quick but rigorous archaeological excavation there, financed by the University of Niamey and the French Minister of Relations Exterieures. Among other things, Gado found tubular and oval-shaped anthropomorphic terracotta pots, some surmounted by heads or highly adorned equestrian warriors. The pots were placed upside down on the ground above burials and dated from the second to eleventh centuries ce. Gado wrote a short essay on these extraordinary terracotta vessels and his interpretation of their meaning for the exhibition catalog Vallees du Niger (Devisse 1993). The exhibit was held in Paris at the Musee National des Arts d'Afrique et d'Oceanie in 1993–1994 and travelled to museums in Europe, America, and West Africa throughout 1996. In the 575 page catalog, over forty scholars addressed aspects of the early history of the inland Niger Basin reconstructed through material objects and other sources. Gado's contribution is the only existing article to describe this Bura material. I was fascinated by these pots and surprised that so little had been published on them.
Gado (1993) tells us that, at the Bura necropolis, groups of pots covering several hundred square meters were found associated with burials: the archaeological excavation (25 × 20 meters in size) exposed 630 funerary urns placed upside down and close together. There were twice as many tubular urns as semi-ovoid ones (see maps and photo of site in Gado 1993: 365–66). The urns were placed on the ground several feet above the buried remains or they once covered or contained human skulls, teeth, bones, or offerings of cooked food for the dead.
Gado suggests that the vases with bodily features are “effigies” of those buried 1–1.2 meters below the surface and that some contained the head of a caretaker of the dead—perhaps a slave, servant, spouse, or kinsman sacrificed to serve the deceased in the afterworld. Gado maintains that the coffin-jars or anthropomorphic funerary urns from the Asinda-Sikka site are funerary statuettes. The terracotta pots vary in size but include tubular urns 70–80 cm in height and round or semi-ovoid urns, some of which are surmounted by standing figures, mounted horsemen, or heads (Fig. 1). Most urns, however, are decorated only with anthropomorphic features depicting eyes, nose, mouth, and/or coiffure. There are also representations of phalluses, breasts, and coiffures on the urns, which Gado says represent sexual differentiation. He interprets the iron arrowheads with hooked ends that were found beside every jar as remnants of a ritual sacrifice.1 The skeletons buried below the terracotta pots wore bracelets of brass or iron, brass nose rings, and sometimes strings of quartzite beads.
There are basically two shapes of pots found in the Bura necropolis: cylindrical and spherical. I suggest that most if not all of these represent genitalia. It is surely not an accident that the Hausa word bura means “penis” (Abraham 1962; Bargery 1934).2 The ancient stone megaliths at Tondidarou in Mali form a parallel example. To understand the sexually explicit forms on the Niger urns I consider them as a group, as a single category, and compare them with objects, myths, and rites performed by several other present-day peoples in nearby regions of West Africa.
The Bura terracotta pots that are reproduced in this essay were owned by New York collectors in the 1990s. They were purchased from American and European art dealers who in turn probably bought them from African runners and traders in the late 1980s.3 It is therefore an arbitrary selection. The collectors are anonymous.
The argument that to exhibit artifacts that lack archaeologically documented data and export papers, or even to publish their photographs, would contribute to looting and the loss of evidence needed to reconstruct the past has led to heated debates; African Arts has addressed the problem in one special issue devoted to ceramic arts (1989, vol. 22, no. 2, especially Berns 1989a, 1989b) and in another on “Protecting Mali's Cultural Heritage” (1995, vol. 28, no. 4; see also van Wyk 2001; Greenfield 1996 ; McIntosh 1989, 2016; Schmidt and McIntosh 1996; and also Dorfman 1998 on the archaeology of Classic Maya sites). The First Word and Dialogue in two recent issues of African Arts (2019, vol. 52, no. 1 by Roberts and Hersak, and vol. 52 no. 3 by Ndiaye and Monroe) address the now highly politicized and heated repatriation debate; Nayeri (2018) reported on the situation in France; and an October 2019 international symposium in New York at Columbia University on restitution may further impact the status of African art collections world-wide. It is estimated that between 50% and 90% of the sites in southwest Niger have been destroyed by looters and that looting of funerary art from sites in the area of Bura has been systematic since 1994.4 Roderick McIntosh says that art/artifacts from pillaged archaeological sites remain “forever in cruel chronological and cultural limbo” (McIntosh 2016: 60; McIntosh and McIntosh 1986: 51). I suggest that these works of art from Niger deserve greater study and indeed should be included as part of the data considered by archaeologists. Whatever new finds may be discovered, there is already a sufficient number of Bura objects in private collections to allow us to appreciate their beauty, to discern patterns in their formal variety, and to speculate about their meanings. I recognize the impassioned arguments against looting archaeological sites, but I believe that preventing research on “looted” objects impedes attempts to understand and interpret them.5
To set these spectacular urns in a broader context, I cast my ethnographic net farther afield than is customary to include twentieth century reports from areas in West Africa—Mali, Niger, Nigeria, Ghana, Cote d'Ivoire, Cameroon, Togo—and, by means of comparisons with this more recent ethnographic material, suggest possible deeper symbolic interpretations than have so far been provided for the archaeological data. While the horsemen and (later accidentally detached) heads that are placed atop some of the urns are the most commonly represented types of Bura sculpture in American and European collections, I will not dwell on them. Instead I concentrate on the overall form of the funerary pots themselves, both tubular and rounded, and on the anthropomorphic designs on their surface that represent heads, navels, and genitalia and argue that these are more than “effigies” of the deceased, as Gado proposed.
The Bura funerary urns are highly sexualized. Most of the terracotta urns from Niger that are in private American collections depict heads with faces or abstract coiffures with or without features (Figs. 2–5). There are also plain, semi-ovoid urns or jars, phallic-shaped urns (Figs. 6–10); and an oval one that is decorated on top with a bas-relief labia majora and what appear to be decorative beads as in a cache-sexe (Fig. 11).6Figure 12 appears to represent an umbilical hernia surrounded by decorative cicatrization, but it might equally be interpreted as a phallus or a stylized head. Figures 9 and 10 are phallic in overall form, but may also be read as heads with limbless bodies. Indeed, in these and in other tubular and oval urns (e.g. Figs. 2, 6–7) there is ambiguity as to whether one should read the decoration as a depiction of hair on the head or the glans (head) of the penis. While Figure 2 is generally phallic in form, a face—with ears, eyes, and a mouth—is depicted in low relief on the linear hair decorating the top of the head, and a vagina may be discerned at the center of the top. Figure 13 is similarly and ambiguously both male and female.
Apart from the representational decorations, the urns are basically of two shapes: elongated and cylindrical (Figs. 6–10) and spherical (Figs. 11–12, 14). It is possible that the forms are of equal or greater importance than the designs on them and that the elongated forms are male symbols and the rounded spherical forms are female ancestral markers. It is more likely, if one accepts that the coiffure represents the head of the penis, that most of the rounded forms are also male or even that these urns represent both male and female simultaneously. The simultaneous portrayal of head, navel, and genitalia on the vessels from Niger is striking, but it is not unique. The Ga'anda ceramic pots that represent the spirit guardian Mbir'thleng'nda and the culture hero Ngum-Ngumi are decorated with both male and female human characteristics: arms, breasts, navel, and often male genitalia and various male weapons.7
Funerary terracottas in various forms were once widely distributed in West Africa. I compare the Bura pots with anatomical features and globular pots on which human figures are seated or emerge from the top which are not necessarily funerary in nature. These pots are known from Nok, Katsina, and Sokoto in Nigera and date to 1000 bce. Aside from the well-known Middle Niger region of Mali (eleventh–thirteenth century)8 and the Ife terracottas from southwestern Nigeria (twelfth–fifteenth century)9, the tradition extended to the Koma in northern Ghana (eighth–sixteenth century),10 and later (seventeenth–early twentieth century) to the Asante in Ghana, who buried dead infants and those born with deformities in pots (McLeod 1984: 366–67; Rattray 1927: 61) and placed terracotta heads near cemeteries (Sieber 1972, 1973; Gilbert 1989; Boachie-Ansah 2000).
As recently as the 1960s, among the Lobi (Birifoh section) in northern Ghana and Burkina Faso, the deceased were buried in a sitting position inside very large clay pots that were made by women. The pots were simply decorated by means of a rope rolled on a stick and a few dots made with a pebble; they were only half-dried in the sun so they would not crack. A pot for a woman took four days to mold and dry, men's pots were thinner, and dried faster—in three days. Women were mourned for four days and men's funerals lasted three days: the numbers three and four are associated with male and female. After burial in a pit dug in the ground, a small broken pot was placed on top and covered with soil. In other cases, when the deceased was the last in the mother's line, a pot-shaped grave was dug, the corpse was placed lying down, and the top of the grave was covered with a big pot upside down. A small hole was made in the pot, which some believe was to allow the spirit of the deceased to pass out to the spirit world.11
When talking about terracottas, it is significant to know who makes them. The literature on potters and pottery making in Africa is vast; see especially Herbert (1993) on potters, smelters, and smiths; Barley (1984) on Dowayo potters and transformation; and Berns (1993) on women and clay. Most but not all pottery making in West Africa is done by women, and pottery making itself is closely identified with fertility and creativity—Herbert (1993: 216), for instance, discusses the role of the potters who transform raw food into meals and raw clay into finished forms through the medium of heat and who are often midwives (see also David, Sterner, and Gavua 1988: 366). Pots and people, furthermore, share certain anatomical features: both have “mouths,” “necks,” and “bodies,” and both may be decorated and adorned.
The shapes of the Bura fired urns resemble unfired granaries in present-day Niger, in the oasis of Fachi in the Tenere desert and in Songhrai Niger (see Gardi 1973: 166, 167, 169) and in Affala Niger.12 These granaries are built by the coil method, a pottery-making technique used on a large scale (Bourgeois and Pelos 1996: photo 51). They also resemble the huge jars kept inside houses in the village of Nigari (Tomboko) that hold water or grain (Desplagnes 1907: 166, fig. 108, plate LVI) and the round clay ovens in Tillaberi, Niger, and along the rivers in Ghana that are used to smoke fish.13 In the Niger-Benue valley of northern Cameroon, the Dowayo men build mud granaries and women decorate them with designs of breasts, hands, navels, faces, and lizards (Barley 1994). The Batammaliba of northern Togo decorate their adobe house facades with incised cicatrization marks and thereby turn their houses into architectural self-images; the granary is believed to be the stomach of the house and is offered food before its human inhabitants (Blier 1987). According to Smith (1989), the Gurunsi in northern Ghana view houses, pots, granaries, and graves as merely different kinds of earthen containers—their funeral pots and houses share similar decorative patterns. The Akan of Ghana equate the womb and the pot and associate a mother's nurturing role with the preparation of food. The Dowayo associate heads, bellies, and pots, especially water pots, in all their rituals (Barley 1984: 98).
Comparative ethnographic examples are useful to help us understand the archaeological ones. Similarities between granaries and terracotta funerary urns and vessels suggest the idea of transformation and rebirth. Granaries are storehouses for seeds that are used for food and for grains that will germinate. Bura pots, similarly, are like storehouses for the deceased, who will be born again into the world of the dead or whose souls or spirits will return to God. In this comparison, pots are like storehouses for seeds of the future.
The orientation of the urns (“top” vs. “bottom,” right-side-up vs. upside-down) is unexpected. The terracotta pots are placed “upside down”: the opening faces the ground. The anthropomorphic features—eyes, nose, etc.—are read in this position; they are added so that the “bottom” of the vessel becomes the “top” of the terracotta sculpture (see Figs. 2–7, 9, 11–15). Indeed, most of the terracotta urns can only stand upright in this “reversed” direction, taking advantage of the flaring lip of the “top” of the pot, which now becomes its “foot”. This would appear to render the object (qua container or “pot”) nonfunctional—although Gado tells us that the Bura vessels that were placed three to four feet above the main burial did contain bones of a deceased or of a caretaker to the deceased. In a similar manner, granaries full of seeds in Niger are also reversed in form, as are the big storage pots overturned on the ground atop Lobi burials and Ga'anda graves.14 This reversal is visually powerful and ambiguous.
Markings on the “torsos” of the Bura terracotta figures and on the middle sections of the urns resemble body scarifications with designs that radiate from and thus emphasize the navel (see Figs. 6–10, 14–15). The projecting or herniated umbilicus itself draws the viewer's eye and suggests its former attachment to the womb. This undoubtedly evokes the typical notion of rebirth such as in the Dogon creation myth, in which the mythical ancestor and creator Nommo made a placenta from which emerged all living beings.15 Cicatrization or scarification marks applied to the area around the navel to beautify and eroticize are often made at key moments in a woman's life to mark her fertility (at puberty, marriage, giving birth), e.g. among the Ga'anda, Senufo, Bamileke, Bangwa, and further afield as well (see Berns 1988 and forthcoming, Glaze 1981, Malaquais 1994, and Brain 1980: plate 11.2).
Men's bodies are often used as palates for scarification (see Roberts 1988: 48). Scarification does not necessarily signify ethnic identity, and styles may change from one generation to another, as among the Tiv of Nigeria (Bohannan 1988). Among the Akan of Ghana, cicatrization is never for decorative purposes, but rather is evidence of cuts for medicine and protection from a deity. The cicatrization decorating the “torso” of one of the Bura pots (Fig. 15) resembles Islamic designs seen on present-day embroidery around the neck opening of Fulani and Hausa robes (grand boubou) and may represent protective magic squares (on which see Prussin 1986: 92–94; 1995).16 Similar geometric designs are also placed on the outside of houses in the region of Zinder, Niger (see Mester de Parajd 1992: 38, photos 59–62).
Many of the urns from Niger were presumably intended to hold the bones of the deceased, or possibly of a sacrificial victim (Gado 1993). Many of the larger ones that I examined had a small hole near the bottom or on the side at mid-height (see Figs. 2, 7, 15–16). Openings of this kind, Gado writes, regarding the vessels that he studied, were made after firing or by removing a disk or plug in the course of a burial rite. It is possible that some of those urns without such a hole were restored. One may speculate that the small hole allowed the release of fluid due to decomposition of the deceased. Considering its placement and comparative material, I suggest that such a hole was made in the body of the pot to allow for the release of the soul, as among the Lobi.17 The release of the soul is from the pot on the surface probably belonging to the caretaker and the main burial is below ground. This is symbolically acceptable though seemingly contradictory. Ga'anda women in Nigeria decorate the surface of their burial pots with marks that resemble their own cicatrization designs to ensure that the male or female soul contained within is properly cared for on its journey to the spiritual realm. Ga'anda men paint these designs on granary surfaces to ensure a fertile crop (Berns 1988). So Ga'anda women and wombs are associated with burial pots and the soul, and men are associated with granaries and germination. I suggest both symbolically imply rebirth.
If the suggested associations between burial pot and soul and between burial pot and granary are accepted, then we may proceed to think again about those terracotta urns in which the “body” or “head” clearly resembles a phallus (Figs. 6–10, 12–13) and to consider ejaculation and saliva and their association with fertility and the creation of new life. The head and phallus are identified, not contradictory. This accords well with my own findings among the Akwapim Akan in Ghana. The word for “head” (ti) refers to the whole person (ti ade, lit. “head thing”) and to life, as in the expression wo tiri nkwa (“may you live long”; lit. “your head [should obtain] life”): This is a salutation to a newborn infant or a person escaping danger and recovering from sickness. The reply is me ti da ase (“[my head] thank you”). The Akan also believe that a man's ntoro (“spirit,” or sunsum, which is also used as a word for semen) “makes a child” and, after conception, “builds up the embryo in the womb.”18 The ntoro (also called agya-bosom, “father's deity” in Akwapim) is associated with the totemic group of the father; in the outdooring rite of a newborn the father transfers his spirit (sunsum or ntoro) by spitting into the infant's mouth and then bestows a name on the infant who is thus integrated into the world of the living. The infant brings to life the generic deceased of the group of patrifiliation. The link between death and sex is also implied ritually when, during a funeral, sexual license is not only condoned, but adjured. The women sing bawdy songs and mime sexual intercourse in dance (Gilbert n.d.).
It is likely that some of the many terracotta heads without bodies (Figs. 17–18) were once inserted into the tops of large funerary urns. In some urns (Figs. 16, 19) the heads are slightly loose. The heads appear to have been made at the same time as the pots, but separately; presumably once something was inserted inside the pot, the head was attached to the pot with clay that dried out over time. If this were so, then it might help to explain why so many small and large heads have been found (and collected). These pots and those with heads on top that are not removable (e.g., Fig. 13) both suggest a link, formally (and in some cases perhaps functionally), with later Akan terracotta heads and abusua kurowa (family pots). The latter are an elaboration on domestic pottery fashioned with a lid to represent the deceased or the funerary entourage. These covered pots held the hair of matrilineal relatives of the deceased19 and were left at a place distant from the actual graves in a place called asensie (“place of the pots”) at the completion of the funeral rites20 together with cooking pots and food for the deceased on his voyage beyond the grave.
The significance of the terracotta funerary urns unearthed from Bura may be connected to the ancient 1.2–1.5 meter-tall phallic-shaped stone megaliths in Mali from Tondidarou, also in the valley of the Niger River. The penis is the whole body of these large Tondidarou megaliths, but some are carved mid-height with an unmistakable umbilical hernia. The megaliths have mostly nonfigurative surface decoration, some of which resembles scarification patterns. Dembele and Person (1993: 446–47, figs. 6C and D) propose that triangular decoration on the body of the stone represents female genitals and the megalith form is phallic. The megaliths thus are a combination of both male and female at once.
In conclusion, archaeological excavations that have been conducted in sites from the middle Niger valley up to the frontier of Niger and Mali by the Institut de Recherches en Sciences Humaines (IRSH) of the University of Niamey will doubtless provide more information about the lives of the people who produced the artifacts discussed here. I have tried to show the value of ethnographic comparisons to understand these objects of the past. There is little that is new in the association of death and rebirth or in the combination of male and female genitalia in a single object. I believe we need to recognize that these African funerary urns with representations of male and female genitalia have not been dealt with adequately in part because of the heated current controversy over the looting of archaeological sites. I nevertheless have indicated a new direction that interpretation might follow.
A number of people have read earlier versions of this paper and thoughtfully raised critical issues: these include Marla Berns, Labelle Prussin, Barbara Bianco, Dominique Malaquais, Peter Mark, Bernard Gardi, and two challenging anonymous readers. I am grateful for their suggestions, some of which pointed in different directions. I am especially grateful for the intellectually rigorous critique and kind encouragement of Gillian Gillison and the generous, thoughtful, and patient advice of Marla Berns. I hope I have managed to address some of the points that they raised. I thank Bossman Murey, University of Ghana, Legon, for his willing aid. My article could not have been written without the generosity of the anonymous collectors who showed me their terracotta urns. Manu Sassoonian kindly agreed to photograph the urns. The friendship and discerning eye of the late Joseph Knopfelmacher will always be an inspiration to me.
One must accept this carefully as it might equally suggest that they were hunters or warriors, or even that the intent was protection from avenging spirits.
It is likely that the site called Bura in Niger has the same spelling as the Hausa word for penis, but not the same pronunciation, which would make it a homograph or homonym (R. Gaudio, personal communication).
They were eagerly collected in the mid-1990s, after which time looting was better controlled through the efforts of local government and national heritage institutions to interdict objects leaving the country and bilateral protection accords that further restricted importation of archaeological material into the United States. The illicit market for Bura urns was further shriveled by the appearance of highly restored pastiches, some of which may be found in West African tourist markets today.
1999 statistics from “The Looted Heritage,” Illicit Antiquities Research Centre, Cambridge University. https//paul-barford.blogspot.com/2012/10/looting-some-statistics
An anonymous reviewer was of the opinion that it would be “unethical” for African Arts to publish this Research Note because the objects discussed come from anonymous collections and were “looted.” The reviewer said that in order to write about them I should travel to Niger and examine the objects excavated by Gado myself. I am an anthropologist who has worked in one Ghanaian kingdom for over forty years. My scholarship has always been based on comparison. When I first saw these Bura ceramics in New York in the 1990s I determined to know more about them and to try to understand them. In my view, even objects without provenance deserve serious scholarly attention. The objects that Gado excavated at the site of Bura-Asinda-Sikka have been in Niamey, Niger, IRSH since 1985; some of these heads, figures, and urns are illustrated in the Vallee du Niger exhibition catalog (1993: 360–64, 371, 373, 550–57). A few other Bura objects in private collections are illustrated in Schaedler 1997: 70–72. They have been largely ignored in the academic literature.
I do not know how many pots at Bura were adorned solely with female genitalia, nor do I know if it indicates a female burial. This is the only one I have seen in a private collection or in photographs.
According to Berns, the gender ambiguity in Ngum-Ngumi iconography may be related to supernatural powers (1989: 52; 2011: 517, 525). Bisexual/hermaphroditic sculpture is fairly common in Dogon wood sculpture (see Nooter 1993: 31, fig. 6; 209, fig. 83).
Cf. head on a pot from the Olokun sanctuary now in the National Museum of Ile-Ife (Vallees du Niger 1993: 314)
See Anquandah and van Ham (1986) on Koma funerary terracottas. More recent excavations suggest these terracottas were not associated with burials and had other ritual uses (Bossman Murey, personal communication).
Bossman Murey, personal communication.
See Mester de Parajd for large round granaries (4.50 m with a thickness of 10 cm) that have an opening on the top and side from the region of Madaoua, Niger (1992: 56, 60, photos 113, 115); round granaries with a circular design in a more modest size (2.5 m by 3 m in height) from along the river towards Tillaberi, Niger (1992: 56, photo 110); and granaries that seem to rest on the soil but have a raised interior base in cylindrical form from the region of Tahoua, Niger (1992: 58, photo 112, and 60, photo 114). Also see Prussin (1986: plate 1) for granaries outside Tera, Niger.
See Mester de Parajd (1992: 63 photo 127) for a round oven with two openings from Ayorou near Tillaberi, Niger. Round clay ovens are also used for smoking fish on the coast and along all the rivers in Ghana.
Marla Berns (personal communication). They serve as grave markers and also cover up small Ga'anda spirit pots.
Griaule and Dieterlen (1954–57). See Blom (2010: 79f.) for fifteenth–sixteenth century examples of male and female wood figures with umbilical hernias from Tintam, Bondum.
Islam was introduced to this part of Africa early: The great empires of Ghana, Mali, and Songhay were at their height from the ninth–sixteenth century. Scarification is prohibited in the Islamic context, but this need not be considered a contradiction.
Among the Akan, burial of individuals varies according to their status: the soul (kra) is believed to leave the body upon death and return to the creator god called Nyame; the ancestral-ghost (asaman) may linger for a while before going to the land of the ancestors.
The spirit (ntoro) of the man is passed into the woman during the act of coition to mingle with her blood (mogya) to form the child (Rattray 1923: 50 and 1927: 51 on the Asante). In Akwapim, sunsum means the “soul or spirit of a man”; out of a sense of social decorum, many say there is no Akan word for semen and call it obarima hu nsu (“water from a man”).
To assure the living family is united with the deceased.