ABSTRACT

Amongst the Kusasi, compound abandonment involves curation of usable material culture including shrines. Despite the fact that curation is a normative process, there are instances when compounds are abandoned without the curation of useable material culture. In these instances, shrines stand a good chance of being output into the archaeological record. This paper examines the likelihood of shrines being output into the archaeological record and provides suggestions about the strategies archaeologists can employ to discover shrines at abandoned compound sites.

Keywords: Abandonment, curation, compounds, shrines, ethnoarchaeology, Ghana

SHRINES AND COMPOUND ABANDONMENT

Anthropologists working in northern Ghana have long noted the importance of shrines to the order of society. Fortes (1949:46) remarks that amongst the Tallensi the stages of a compound head’s personal destiny, past, present, and future are perpetuated in the many shrines and other ritual objects that are housed within the compound. Other scholars have also touched upon how social relations and patterns of descent are mirrored in the variable placement and size of different shrine types and how shrines commemorate important events, people, and relationships from the past.¹ Most of the objects that anthropologists call shrines are made from durable, common materials such as: stones, ceramic pots, iron rods and bangles, and animal bones and horns. These are materials that stand a good chance of surviving should they get output into the archaeological record. Determining the presence or absence of shrines in the archaeological record could aid in drawing inferences concerning the size, composition, and developmental stages of residential groups. Identifying shrines in the archaeological record requires an understanding of the processes that condition their deposition or discard. The goal of this paper is to examine the sorts of factors that condition shrine deposition among the Kusasi of the Upper East Region, Ghana.

The Kusasi make their shrines from durable materials that stand a good chance of surviving once deposited into archaeological contexts. The frequency that the Kusasi output shrines from systemic contexts is, however, quite low. When a man dies, the younger members of his residential and descent groups inherit his shrines. When people abandon compounds, they take their shrines with them. In short, whether shrines make it into the archaeological record largely depends upon whether they are inherited or curated, and this is ultimately connected to the circumstances surrounding compound abandonment.

All archaeological sites are abandoned and in this sense it is not surprising abandonment has such an important role in shrine deposition. Circumstances of abandonment determine the constitution of de facto refuse at a site. De facto refuse consists of usable or repairable items of material culture left behind when abandonment occurs. The circumstances surrounding abandonment, whether it occurs rapidly or gradually, and with anticipation of return or not, will largely determine the make-up of de facto refuse at a site (Stevenson 1982; Tomka 1993).

Abandonment conditions curate behaviour. Curate behaviour “consists of removing objects from one site and transporting them elsewhere in anticipation of future use” (Schiffer 1976:56). If abandonment occurs gradually, people have the opportunity to plan if, and how, they will curate the objects and features they wish to have at the new location or area. Anticipation of return also affects whether objects or features are curated (Stevenson 1982; Tomka 1993). If they do not anticipate returning to the site, occupants will take away usable materials. If return to a compound is anticipated, and if the new location is close enough to the old, entire “structures may be dismantled and building materials transported” (Cameron 1993:5). Curate behaviour has the potential to deplete de facto refuse, particularly if curation is delayed. If abandonment occurs gradually and people anticipate returning to the abandoned site, curation may be drawn out with items being moved to the new area or locale, as they are needed.

Other processes can affect the make-up of de facto refuse at a site. These include lateral cycling, draw down, scavenging, and collecting and looting (Schiffer 1976; 1985). Lateral cycling “occurs when an object is transferred from one user to another. It includes the many processes by which used, but usable, objects circulate within a sociocultural system and persist in time” (Schiffer 1976:39). Draw down “refers to the tendency for people not to replace worn or broken items when they know they are about to move” (Lightfoot 1993:166). Scavenging refers to the depletion of de facto refuse in the form of neighbours taking away useful items from the abandoned area or site (Schiffer 1985). Collecting and looting refers to the removal of artefacts from de facto refuse by non-residents (Schiffer 1976:35–36). This paper examines compound abandonment and the effect of curation upon the production of de facto refuse, particularly shrines.

The paper begins by providing background information and describing the methods used in the present study. It describes compounds as long-term depositional events and highlights the various transformations of material culture from state to state in systemic contexts over time. Curation and abandonment are the two most important processes governing whether shrines are output from systemic contexts to archaeological contexts. Using observations made in the settlement of Zorse in the Upper East Region of Ghana, the paper will provide information about the preservation of shrines in abandoned compounds. These observations suggest that shrines appear in archaeological contexts as a result of events that are catastrophic from the perspective of the residential group. In conclusion, a methodology for future archaeological research in the area is proposed, paying attention to methods for identifying shrines in the archaeological record.

ETHNOGRAPHIC BACKGROUND

The first anthropologically related research on the Kusasi occurred in the 1930s with a major ethnographic survey of northern Ghana conducted by Rattray (1932), which provides details about Kusasi social organization, history, material culture, religion, and language. During the same decade, administrative reports appeared on Kusasi history and socio-political organization (Syme 1932) and Kusasi agriculture, land use, and management (Lynn 1937). The last seventy years of anthropological study has been dominated by recurrent developmentally oriented research on Kusasi agriculture and the physical environment (Blench 1998; Chilalah 1957; Cleveland 1980, 1986, 1989, 1991; Devereux 1989, 1993; Webber 1996a, 1996b; Whitehead 1988; Wiszniewski 1955).²

Observations for this study were made in Zorse, a densely populated rural settlement located near the Northwest outskirts of Bawku, the capital of Bawku East District, Upper East Region (Fig. 1). The area covered by Bawku East District and the neighbouring district of Bawku West is the home territory of the Kusasi. Approximately 3,000 square kilometres, Kusasi home territory is bordered by the Red Volta River, Togo, Burkina Faso, and the Gambaga escarpment. The climate here is arid and semi-tropical. Precipitation is largely limited to the wet season, which lasts from May to September. The dry season is marked by dust storms and Harmattan winds and lasts from November to April.

The Kusasi are politically acephalous sedentary horticulturalists (Blench 1998; Cleveland 1980; Lynn 1937; Webber 1996a). They have a patrilineal segmentary social organization and speak a Gur (Niger-Congo phylum) language (Naden 1988). Polygyny is common in Kusasi society and patrilocal post-marital residence is the norm. The basic unit of settlement is the yir, an earthen walled, multi-courtyard compound occupied by a residential group consisting of male agnates, their wives, and children. Traditionally, compounds were subject to abandonment as part of the farming cycle. As lands surrounding the compound became less fertile, residents abandoned compounds in favour of locales on more fertile lands. An abandoned compound is a dabog, a term that also refers to a lineage and a compound whose founder is deceased.

Follow for extended description

DATA COLLECTION AND METHODS

Observations used in the present study were made during nine months of fieldwork, from July 1996 to April 1997. The research team included my wife, Rebecca, and our interpreter, Cletus Anobiga. Our study was designed to provide archaeologists with a framework for identifying and interpreting shrines in the material remains of the archaeological record. Research goals included: (1) compiling a shrine typology; (2) deriving a sample of compound plans, including the locations and distributions of shrines; (3) examining abandoned compounds to determine whether shrines get output and survive in the archaeological record.

Information from 237 interviews with 153 compound heads was used to compile a shrine inventory. In total, 1,600 shrines belonging to roughly a hundred types were recorded and classified into five groups, as will be discussed in more detail below. Information was gathered concerning the creation of and patterns of inheritance for shrines, and whether people ever discard shrines. Physical characteristics of shrines such as size and form and raw materials used in construction were recorded. Details on domestic activities and activity areas in and around compounds were recorded along with observations of construction episodes at a half dozen compounds. Scale diagrams were made for twenty-seven compounds from the interview sample and show the locations and distributions of different shrines within compound space. Further interviews were conducted with the heads of these compounds in order to elicit reasons and rationales for the placement of shrines.

Subsequent to collecting data about occupied compounds, research focused on abandoned compounds. In a forest reserve directly to the south of Zorse, three compound sites abandoned during the 1930s were examined. These examinations consisted of pacing out the approximate dimensions of the compounds using elders’ memories and surface scatters as guides. Notes on surface materials were also compiled. Similar work was done on three compounds abandoned in the reserve within the last twenty years. Interviews with neighbours and relatives elucidated the reasons for abandonment. Detailed notes and sketch maps were made of compound remains. These examinations revealed whether or not shrines survive immediate abandonment and hence furthered understanding of the depositional processes acting on shrines in the archaeological record.

SHRINES

In the context of this study, a shrine is a physical object at which people make animal sacrifices and pour libations in honour of supernatural otherworldly agents including ancestors, spirits of the land, divining spirits, and nature spirits. Shrines are the foci of ritual practice because they are the dwelling places of the supernatural agents the people pay homage to. A shrine may be made up of a number of objects or it may consist of a single object. The most common objects are animal horns and tails, stones, trees, ceramic vessels, iron rods and bangles, plant roots, and calabashes.

Table 1: Frequency of Different Shrine Types in the Sample of 153 Compounds.

For analytical purposes, I placed the shrines I recorded into five categories: (1) land gods, (2) personal destiny shrines, (3) ancestor shrines, (4) shrines for spirits of the wild, and (5) other shrines (See Table 1). Analytical categories are based upon function and associated spirit more than physical form. This is largely because the physical form and objects used to make a shrine are not always correlated with the function of and spirits associated with the shrine.

The term land gods refers to the spirits of particular places and to the shrines that commemorate human relations with those spirits (see Mather 2003). The Kusal term for land gods is tengbana, which literally translates as “skin of the earth.” Most land gods are large stones set into the ground of forest groves, hills, ponds, and other natural features. People appeal to the land gods when there is drought, epidemics of disease, or outbreaks of blight on crops. Appeals are also made when witchcraft threatens the larger community or when it faces physical attack. Land gods guard the community much like parents safeguard their children. They hold spiritual and social authority over human communities.

Personal destiny (win) is a guiding light that shines upon the individual and provides good fortune, health, well-being, and fertility. Each individual has a win, a unique destiny that comes from the high god Na-Win. Consequently, personal destiny shrines are the most commonly encountered shrines, accounting for 28 per cent of the shrines recorded in the present study. A personal destiny shrine is a plastered, rounded mound of earth. In most cases an individual’s personal destiny shrine is located directly outside that individual’s room. Exceptions to this rule are found in the case of the personal destiny shrines of a compound head or of a deceased compound head. In these cases the shrine is generally located either in the public yard, or saman, in the front of the compound, or in the main courtyard of the compound. The personal destiny shrine of a deceased compound head is maintained by the dead man’s senior son, who refers to the shrine as ba-win (father’s personal destiny).

Spirits of the wild (kinkiriis) are also called “bush spirits” and “fairies.” Spirits of the wild are of the “wild” not necessarily because they originated there but because they are without a house; they have no connections to living persons or social groups and hence no source for libations and sacrifices. Many of the spirits in the wild once belonged to houses, but this was so long ago that they have become forgotten and fallen out of the pool of ancestors that receive sacrifices from the living members of the house. In order to be re-enshrined, the kinkiriis take one of two courses. They appear to one of their agnatic or cognatic descendants in the form of natural phenomena or they reincarnate in newborn children. The former situation usually leads to the creation of a chameleon (dendet) shrine, while the latter situation requires the erection of a birth shrine.

Spirits of the chameleon shrine come to an individual, or call to an individual by appearing as a pair of chameleons mating in the bush. Chameleon shrines make up 8 per cent of the total number of shrines recorded in the present study. The consequence of seeing the chameleons is blindness. To prevent blindness one has to kill the chameleons, remove their heads and make a protective charm from them, and erect a second shrine at one’s personal destiny shrine. The shrines usually consist of a small ceramic vessel, often a water pot, next to a personal destiny shrine. Commonly, people mould a stylized chameleon above the water pot – sometimes in the mound that forms the destiny shrine; other times on a wall. In some cases, people will build and plaster the mound around the water pot , leaving the neck and the mouth of the vessel accessible.

Birth shrines house spirits that cause unusual pregnancies or births. They make up almost the same percentage of the total number of shrines, just under 8 per cent, as chameleon shrines. Women who experience unusual pregnancies or births place the shrines outside their rooms. When the child or children mature and occupy rooms of their own, they relocate the shrines outside their rooms next to their personal destiny shrines. In some cases, individuals locate their birth shrines in the front yard, often at the foot of or built onto the wall of a granary. The most frequently encountered shrine of this type is for breach births (tula), and it protects the granary from thieves. There is only one way to enter a granary, through the roof. As a thief enters the granary, legs dangling outside, he or she becomes trapped and caught by the spirit of the shrine. Most birth shrines are undecorated ceramic vessels with a spout on their shoulders that spirits use as an entrance. In the case of twins, the pot has two spouts, one for each spirit.

Ancestor shrines include objects that enshrine maternal and paternal ancestors. Paternal ancestor shrines consist of plum-sized stones set into a mound built onto the outside wall of the compound just north of the front gate. The stones enshrine the spirits of former members of the residential group’s patriline, including father, father’s wives, grandfathers, grandfather’s wives, and great-grandfathers and their wives. Maternal ancestors are enshrined in animal horns (goat, cattle, and wild animals) that are often associated with other objects, including ceramic vessels, stones, animal tails, and paraphernalia used for divining. A third type of ancestor shrine is the chief shrine. Chief shrines house the spirits of maternal or paternal ancestors. They enshrine the latent power within families of chiefly lines and the status of those who fall within the line of succession is reflected by possession of the shrine.

The shrines that belong to the ‘other’ category are made from a variety of objects, and most of them include more than one object. The seven most frequently encountered objects include animal horns, animal tails, ceramic vessels, stones, calabashes, iron rods and bangles, and trees. The shrines within this category perform one or more of three functions: (1) harnessing or controlling natural forces, (2) launching and protecting against magical attacks, and (3) diagnosing and treating illnesses. The shrines are found throughout the domestic space of the compound, though entranceways are the preferred placements.

Examinations of the spatial distribution of shrines within and around twenty-seven compounds provide some indication of the relative predictability of shrine placement (see Table 2). Several trends are readily seen. Just over half of the shrines within the sample are located in the courtyard while nearly 37 per cent of the shrines in the sample are located in the front yard. Paternal ancestor shrines are only found in the front yard on the exterior wall of the zong, the compound entrance hut. Maternal ancestor shrines are the predominant types of shrine found within the zong. Personal destiny shrines, paternal ancestor shrines, and chief shrines are never found in personal rooms. Only shrines belonging to the ‘other’ category are found in animal yards. Drawing upon these observations, it may be possible for archaeologists to determine the type of shrines they find in the archaeological record if they can determine the approximate layout of the abandoned compound from which the shrines were excavated. The rest of this paper will address the likelihood of finding shrines at abandoned compound sites.

Table 2: Distribution of shrines within the sample of twenty-seven compounds.

COMPOUNDS AND DEPOSITIONAL PROCESSES

One day, while watching a friend fix a roof to a granary in his front yard, I noticed a few pigs grubbing around for food. They found scraps in two large potsherds on the ground near the front gate. A grinding stone with a deeply grooved surface held water from which the pigs drank. My friend stated that his father used the stone to grind tobacco for snuff. Unlike his father, my friend was a smoker and did not need a stone to grind tobacco. The stone had lain in the front yard, where it served as a trough, since his father’s death.

The sherds and grinding stone exemplify the recycling and reuse of material culture that takes place within each compound. Tools and objects are modified throughout their life cycle. Broken pots and sherds, for example, are used as lids on other vessels, to give water and food to animals, to support cooking pots, and to make termite traps. When sherds are broken into smaller pieces, people discard them on a heap outside the compound where daily trash burning, walking, and trampling break them down even further. Ground surfaces are littered with tiny pieces of ceramic that have passed through various stages or use-lives. The recycling of objects and tools is a predepositional process, a cultural transformation that affects the condition of objects before they are discarded. Recycling is an example of what Schiffer (1976:37–38) calls “S-S processes,” transformations of material culture from state to state within systemic contexts.

Several natural transformations affect Kusasi material culture. Plaster erodes off compound walls under heavy rains, scorching heat, and powerful winds. Without repair, walls decay and rooms become uninhabitable. The same natural phenomena and the recurrent footfall of occupants and visitors wear floors. Many shrines are architectural features built from the same materials and subject to the same erosional forces affecting walls and floors. The mounds built for shrines commemorating paternal ancestors and personal destiny spirits, for example, are slowly worn down by exposure to the elements.

The Kusasi repair their architecture seasonally. Walls and floors are replastered at the end of the wet season. Shrines are also repaired. Severely damaged architectural features are destroyed and built anew. New rooms and courtyards are added to existing ones. Remodelling occurs whenever the natural elements have taken their toll or when the residential group grows in size because of the introduction of a new wife or the birth of children. No matter why renovations are needed, rebuilding episodes also exemplify S-S processes; material culture is altered though still used in the same ongoing behavioural system.

A compound is a long-term depositional event. Some compounds in Zorse have been occupied for several generations, experiencing more than a dozen serious episodes of remodelling. Discarded material culture, primarily ceramic sherds, can be found in the front yards of these compounds, mixed and scattered throughout the ground surface. Generally, this detritus forms the bulk of the material assemblage that will ultimately be output from systemic context to archaeological context.

Deposition from systemic to archaeological contexts does not occur until a compound is finally abandoned. Rate and mode of abandonment vary as to circumstances (Cameron 1993:3). Historically, the most common reason for abandoning compounds was the depletion of soil fertility of both compound farms and bush farms. Abandoned compounds remained within the settlement system as sites for farming or as the future locale for a land god shrine for commemorating apical male ancestors and founders. Over the last hundred years, abandonment has become less frequent due to major increases in population density and decreases in available lands for resettlement. In Zorse, because of shortages of unused land for resettlement, compounds are rarely abandoned. Abandonment still occurs, however, in those cases where a compound head dies without having male relatives to inherit his position. This can occur if the compound head has no brothers or sons, or if his brothers have compounds of their own and his sons are too young to assume his role. People call these deaths ‘untimely,’ or ‘coming too early’ in the sense that there is no one to continue leading the residential group.

The circumstances surrounding abandonment affect whether and how material culture assemblages will be curated (Kent 1993; Stevenson 1982; Tomka 1993). When a family abandons one compound and founds or moves into another compound, they delay curation until they need their material possessions in their new compound. Constraints upon curation include size and weight of the items involved and the distance to travel between the old and new residential sites (Schiffer 1976:33–34).

People curate shrines along with other useful items. Shrine curation involves libations and sacrifice. Should a compound head have several shrines, curation can be a costly procedure, with the potential that each shrine requires its own sacrifice. Curation also involves divination to learn whether enshrined spirits will allow abandonment without provoking misfortune and whether the spirits are willing to live at the new compound.

People also pass shrines from systemic to systemic contexts by way of ‘lateral cycling.’ When a compound head dies, the shrines he managed as a senior member of the descent group will be inherited by the man directly after him in the line of succession. The deceased’s senior son will inherit the shrines that belong to the man’s residential group and those he got on his own volition (e.g., medicine shrines). In the former case, the deceased’s successor moves the shrines from the deceased’s compound to his own compound. In the latter case, the senior son takes possession of the shrines and, if he no longer lives with his father, he will relocate the shrines to his own compound.

What are the chances of finding shrines in archaeological contexts? Curate behaviour eliminates shrines from material assemblages, making it highly unlikely that shrines will form part of the de facto refuse at an abandoned compound site (Schiffer 1985:26). There are, however, reasons to suspect that shrines will be deposited into archaeological context whenever abandonment occurs because of an untimely death. Observations at three abandoned compound sites support this hypothesis.

THREE ABANDONED COMPOUNDS

The three compounds chosen for analysis differ as to the number of years that have passed since they were abandoned. The first compound was being abandoned during fieldwork for this study, while the second and third compounds were abandoned seven and twenty years ago respectively. Theoretically, each compound represents a different stage of abandonment. They are ordered, by way of presentation, into a sequence. Compound 1 represents the first stages of abandonment, while compound 3 shows us how an abandoned compound site appears following the post-depositional disturbance caused by farming. Compound 2 illustrates an intermediate stage of abandonment – this structure is no longer used as a residence, but it has yet to be knocked down to make way for farming.

Compound 1

Observations of compound 1 were made on March 3, 1997 (Fig. 2). The compound head died in 1992, his final funeral was held in 1994, and the compound was finally abandoned in the middle of December in 1996. The widow and her children continued to live in the compound after the compound head died. They finally abandoned the compound because the widow married one of the compound head’s brothers (a man who had the same grandfather).

The wife and her children have curated many of the items of material culture from the compound, including the roofs of buildings. Notwithstanding this, a large quantity of useable items and materials remain in the compound. As she continues to settle into her new home, the widow will take the articles she needs for her new compound. The deceased compound head’s brothers, men from the same father, have rights to all of his personal belongings, including the shrines, and they removed most of these items after the final funeral rites for the deceased. Consequently, most of the materials left in the compound are the property of the widow.

Follow for extended description

The process of abandonment will not be complete until the deceased’s relatives perform the appropriate rites to transfer the paternal ancestor shrines that he was responsible for. At this time, the paternal ancestor shrines are still in place on the wall of the zong. Once the appropriate rites are conducted, the men in line to inherit the deceased’s property (i.e., the senior living descendants of the ancestors in question) will relocate the shrines to their respective compounds. Apart from the paternal ancestor shrines, there is no evidence for shrines in other parts of the compound. The wife has already moved her personal destiny shrine and birth shrines. There is no evidence left of these shrines. This is unexpected because the shrines of this type are frequently built into mounds and the remnants of the mounds would be visible even after the shrines were removed. Likewise, there is no evidence for other shrines in the compound despite the fact that, according to the widow, her husband had controlled other shrines.

Compound 2

Compound 2 was abandoned in 1992 after the sudden death of the compound head (Fig. 3). Apparently, the compound head had no living relatives within the settlement. The deceased’s only child, a son, was too young to take control of the house. Since there were no relatives to take responsibility for the child, he moved with his mother to his mother’s natal residential group.

Several residents of the sub-settlement wanted to farm where the compound stood, but they could not agree on who could claim the land. The walls of the compound were left standing. Neighbours took roofs from buildings right after the compound was abandoned, while the widow moved her personal belongings to her father’s house. Since abandonment, the compound has served as a pig barn.

The compound head’s grave is in the front yard about four metres southwest of the compound entrance. A water pot, likely once used as a burial shrine, sits atop the grave. Approximately four metres west of the grave, there is a pile of household refuse. Directly west of the compound entrance, around six metres away, there is a granary. A birth shrine rests at the foot of the granary. Midway between the compound entrance and granary there is a plastered mound, severely eroded, that once served as the compound head’s personal destiny shrine. The bathing area between the wife’s room and son’s room includes the plastered mound and stone that served as the wife’s personal destiny shrine. A large potsherd covers the mound. Just south of the personal destiny shrine, there is a small water pot on top of an upside-down metal bowl with a rusted-out bottom. Ethnographic observation suggests that these latter objects once served as a chameleon shrine. In the son’s room, partially buried in the floor near the entrance at the north of the room, there is a flat, black, circular stone on top of a medium-sized water pot. The water pot is upside down, placed within a metal ring (most likely the neck of a metal basin), and its bottom is broken. The metal ring was lodged into the top of a circular hole, with the ceramic vessel fit snugly into the ring and resting directly in the hole. Drawing from ethnographic examples, these objects were used as a shrine prior to abandonment.

Follow for extended description

Compound 3

Abandoned compound 3 is located approximately forty metres north of compound 2 (Fig. 4). It was abandoned in 1979 following the death of the compound head and his wife. The couple’s only child, a son who worked as a police officer in the south, died shortly after his parents. The son left no descendants.

A few years after abandonment, members of the compound head’s descent group knocked down the compound walls and began using the site for farming. Three graves, housing the compound head, his wife, and son, lie in the former front yard (Fig. 4). Near the graves a water pot rests at the foot of a small tree. Little remains of the compound, save for a large roughly rectangular area of lightly coloured soil, about eighteen metres long east to west and nine metres wide north to south. Concentrations of charred ceramic sherds and cement are scattered around this area. Whatever de facto refuse remained after abandonment is now destroyed and thoroughly mixed with soil and farming detritus. The only indications a compound existed on the site are the shade tree and graves, and a myriad of remnants from household features and activities, seriously disturbed and damaged.

Follow for extended description

DISCUSSION

The Kusasi use the term dabog to refer to an abandoned compound. Dabog is also reserved for a descent group whose founder is deceased and to a compound whose founder is deceased. Abandonment is almost synonymous with the death of individuals who hold offices of authority, whether they are compound heads, senior members of descent groups, or chiefs.

In the majority of cases, the death of an office holder does not lead to abandonment of territory or residences. When a compound head dies, for example, the surviving senior member of his residential group, a brother or son, will assume his role. If the deceased was also the compound founder, the compound will thereafter be referred to as a dabog (a social and spatial entity whose founder is deceased) and a yir (a physical structure occupied by the founder’s descendants) (Mather 1999:139).

The three compounds examined above were abandoned under uncommon circumstances; a compound head died and no one in his residential group could assume his role. In each case, a residential group ceased to exist and the physical shelter it occupied was abandoned. These are favourable events from the vantage of archaeology and shrines because, if there is no one to inherit the compound head’s office, it is likely there is no one to inherit and curate the compound head’s personal shrines. When the compound is abandoned, the shrines are effectively discarded from systemic contexts.

Compound 1 differs from compounds 2 and 3 because the compound head’s shrines were or will be curated. His personal shrines are now located in the compound where his children and other relatives live. When his sons are old enough, the senior son will assume control of the shrines as head of the residential group and descent group. The shrines will be held as the common property of all of the deceased founder’s sons. The compound founded by the compound head has been abandoned but the residential group he founded still exists. As his sons mature into adulthood and marry, the group will grow and fission, and the founder’s personal shrines will be curated down through the generations. The shrines that the compound head controlled as the senior member of the descent group founded by his father are or will be located in his successor’s compound. The paternal ancestor shrines, for example, will be relocated once the appropriate rituals are performed at the deceased’s successor’s compound (Mather 1999:89–90).

Compounds 2 and 3 represent catastrophic abandonments because the residential groups have ceased to exist. The bloodlines started by the compound heads are ended and the shrines the men controlled are now refuse. Compound 2 likely contains six shrines: the compound head’s personal destiny shrine, the wife’s personal destiny shrine and chameleon shrine, an unknown shrine resting on top of one of the graves in the front yard, a birth shrine, and an unknown shrine in the floor of the son’s room. If and when the compound walls and floors are broken down to make way for farming, it is likely that at least three of the shrines, the wife’s personal destiny shrine and her chameleon shrine, and the unknown shrine located in the sons’ room, will be destroyed. In addition, the compound head’s personal destiny shrine will continue to erode away under exposure to the elements. This leaves two shrines that stand a chance of surviving post-abandonment disturbances.

Unlike compound 2, compound 3 has suffered serious post-abandonment disturbance. Nevertheless, an object that could be a shrine, consisting of a water pot at the foot of a tree, survives in what was once the front yard of the compound. There is no guarantee the shrine will be output into archaeological contexts, though it should be noted it has lasted in its present state since 1979.

In summary, abandonment can lead to the discard of shrines if it also involves the demise of a residential group. Despite the fact that shrines can be discarded, post-abandonment processes disturb and further deplete de facto refuse. Given ideal circumstances, such as with compound 2 where post-abandonment farming does not take place, shrines may well remain intact and in situ. It is more likely, however, that post-abandonment farming will destroy discarded shrines along with the rest of the compound. Consequently, to find evidence for shrines one will have to look towards the detritus that forms the bulk of the impoverished material assemblages found in archaeological contexts.

CONCLUSIONS: LOOKING FOR SHRINES

What should one look for to find evidence of human occupations from the recent past? To identify compound sites one can focus on several features. Trees are an excellent sign of human occupation. Compounds occupied for more than a single generation typically have at least one large tree shading the front yard with several graves located nearby. Trees and graves characterize abandoned compound sites, and as the description of compound 2 shows, they are often associated with shrines. Ethnographic observations indicate that shrines are placed in predictable locations in occupied compounds. Archaeologists can expect that shrine remains will be in close proximity to where the shrines they once belonged to were placed in the compound when it was occupied. The front yard and courtyard are the primary areas of interest, while personal rooms and the zong are secondary areas of interest.

After an abandoned compound site has been discovered, excavations should focus on covering as much area of the site as possible in order to discover whether shrines were present when the compound was occupied. Archaeologists interested in finding shrines should focus on horizontal rather than vertical excavations. Particular attention should be given to discerning entranceways to the compound, internal courtyards, and rooms. Given that compounds only have one entrance and that this entrance is always located on the western side of the compound, archaeologists can move eastwards to find the inside of the compound and hence the internal courtyards and westwards to locate the front yard.

Once an abandoned compound site is identified and excavated, one is still left with the task of determining whether the materials one uncovers are the remnants of mundane objects and behaviours or whether they are the remains of shrines. Unfortunately, shrines are unlikely to be deposited intact, a somewhat discouraging fact. How is one to discern whether the stone, potsherd, or hoe blade one finds in the archaeological record was a shrine and not just another mundane object?

Various activities and materials create and leave physical traces upon the objects forming shrines. Pots used to prepare medicines are charred black on their bottoms and sides from direct exposure to fire. Sacrifices leave ample amounts of animal blood on the surfaces of shrines and upon the ground surfaces around the shrines. Libations of millet beer and water mixed with millet flour are regularly poured upon shrines. Inside surfaces of the pots that contain medicines absorb residues from the different plants soaking within the pots.

Morphological, histological, and chemical criteria may be used to identify remnants of plants and plant foods on artefact surfaces (Fullagar et al. 1996; Hillman et al. 1993). Blood proteins can be identified in residues using various immunologic techniques including ouchterlony (OCH), radio-immuno-assay (RIA), gold immunoassay (GIA), crossover-immuno-electrophoresis (CIEP), and enzyme-linked-immuno-absorbant-assay (ELISA) (Downs and Lowenstein 1995; Eisle et al. 1995; Fiedel 1996; Kooyman et al. 1992). Lacking target compounds, use of these techniques will require compiling a reference collection by sampling contemporary plants and food plants from the study area and compiling a comprehensive list of animals commonly sacrificed at shrines.

In conclusion, focusing on horizontal excavations and using residue analysis, archaeologists stand a reasonable chance of uncovering shrines in the archaeological record. The likelihood of finding intact shrines is remote. Notwithstanding this, drawing upon ethnographic examples, it might be possible to infer shrine type from spatial context (e.g., personal destiny shrines are generally found near personal rooms). If we can identify and type shrines in the archaeological record, we can extrapolate to the make-up and social dynamic of residential groups represented by material culture assemblages from the past that are comparable to Kusasi material culture.

ACKNOWLEDGMENTS

Research would not have been possible without the encouragement, understanding, and participation of the residents of Zorse, the late Zorsenaba, the Tengindana, and my interpreter, Cletus Anobiga. The Ghana Museums and Monuments Board kindly extended a research permit for the fieldwork. Dr. Nicholas C. David provided funding support through the Social Sciences and Humanities Research Council of Canada (Grant No. 410-92-28). This research was conducted as part of the Mandara Archaeological Project (Ghana Phase) directed by Nicholas C. David.

NOTES

1. Rattray (1932) provides details on shrines amongst northern populations in general, Kirby (1986) amongst the Anufo, Goody (1962) amongst the LoDagaba, and Kroger (1982) amongst the Bulsa.
2. Work has also been carried out on the phonology and syntax of Kusal, the language spoken by the Kusasi (Spratt and Spratt 1968, 1972), on Kusasi vernacular architecture (Bourdier and Min-Ha 1985), Kusasi history (Hilton 1962), and Kusasi marital exchange and prestations (Awedoba 1989a,b, 1990).

REFERENCES

Awedoba, A.K. 1989a. “Notes on Matrimonial Goods Among the Atoende Kusasi, Part 1.” Research Review 5, no. 1:37–53.

———. 1989b. “Matrimonial Goods among the Atoende Kusasi Contingent Prestations, Part 2.” Research Review 5, no. 2:1–17.

———. 1990. “Matrimonial Goods among the Atoende Kusasi: Matrimonial Prestations and Exploitation, Part 3.” Research Review 6, no. 1:49–56.

Blench, R.M. 1998. Developing Common Property Resource Management: Upper East and Northern Regions, Ghana. London: Overseas Development Institute.

Bourdier, J-P., and T.T. Min-Ha. 1985. African Spaces: Designs for Living in Upper Volta. New York: Africana Publishing.

Cameron, Catherine M. 1993. “Abandonment and Archaeological Interpretation.” In Abandonment of Settlements and Regions: Ethnoarchaeological and Archaeological Approaches, ed. Catherine M. Cameron and Steve A. Tomka, 3–7. Cambridge: Cambridge University Press.

Chilalah, G.C. 1957. “Advances in Agriculture in Kusasi, Northern Ghana.” Ghana Farmer 1:198–201.

Cleveland, David A. 1980. The Population Dynamics of Subsistence Agriculture in the West African Savanna: A Village in Northeast Ghana. PhD dissertation. Tucson: University of Arizona.

———. 1986. “The Political Economy of Fertility Regulation: The Kusasi of Savanna West Africa (Ghana).” In Culture and Reproduction: An Anthropological Critique of Demographic Transition Theory, ed. H. Pennewerker, 263–93. Boulder, CO: Westview Press.

———. 1989. “Developmental Stage Age Groups and African Populations Structure: The Kusasi of the West African Savannah.” American Anthropologist 91, no. 2:401–13.

———. 1991. “Migration in West Africa: A Savanna Village Perspective.” Africa 61:222–46.

Devereux, S. 1989. Food Security, Seasonality and Resource Allocation in Northeastern Ghana (ESCOR R4481). London: Overseas Development Institute.

———. 1993. “Observers are Worried: Learning the Language and Counting the People in Northeast Ghana.” In Fieldwork in Developing Countries, ed. S. Devereux and J. Hoddinott, 43–56. Boulder, CO: Lynne Reiner.

Downs, E.F., and J.M. Lowenstein. 1995. “Identification of Archaeological Blood Proteins: A Cautionary Note.” Journal of Archaeological Science 22:11–16.

Eisle, J.A., D.D. Fowler, G. Haynes, and R.A. Lewis. 1995. “Survival and Detection of Blood Residues on Stone Tools,” Antiquity 69:36–46.

Fiedel, S.J. 1996. “Blood from Stones? Some Methodological and Interpretive Problems in Blood Residue Analysis.” Journal of Archaeological Science 23:139–47.

Fortes, Meyer. 1949. The Web of Kinship among the Tallensi. London: Oxford University Press.

Fullagar, R., J. Furby, and Bruce Hardy. 1996. “Residues on Stone Artefact: State of a Scientific Art.” Antiquity 70:740–45.

Goody, Jack. 1962. Death, Property and the Ancestors: A Study of the Mortuary Customs of the LoDagaa of West Africa. Stanford: Stanford University Press.

Hillman, Gordon, Sue Wales, Frances McLaren, John Evans, and Ann Butler. 1993. “Identifying Problematic Remains of Ancient Plant Foods: A Comparison of the Role of Chemical, Histological, and Morphological Criteria.” World Archaeology 25, no. 1:94–121.

Hilton, T.E. 1962.“Notes on the History of Kusasi.” Transactions of the Historical Society of Ghana 6:79–96.

Kent, Susan. 1993. “Models of Abandonment and Material Culture Frequencies.” In Abandonment of Settlements and Regions: Ethnoarchaeological and Archaeological Approaches, ed. Catherine M. Cameron and Steve A. Tomka, 54–74. Cambridge: Cambridge University Press.

Kirby, Jon P. 1986. God, Shrines, and Problem-Solving among the Anufo of Northern Ghana. Berlin: Dietrich Reimer.

Kooyman, B., M.E. Newman, and H. Cerl. 1992. “Verifying the Reliability of Blood Residue Analysis on Archaeological Tools.” Journal of Archaeological Science 19:265–69.

Kroger, F. 1982. Ancestor Worship among the Bulsa of Northern Ghana: Religious, Social and Economic Aspects. Munich: Klaus Renner.

Lynn, C.W. 1937. “Agriculture in North Mamprusi.” Ghana Department of Agriculture Bulletin 34.

Mather, Charles M. 1999. An Ethnoarchaeology of Kusasi Shrines, Upper East Region, Ghana. PhD dissertation, Calgary: University of Calgary.

———. 2003. “Shrines and the Domestication of Landscape.” Journal of Anthropological Research 59, no. 1:23–45.

Naden, T. 1988. “The Gur languages.” In The Languages of Ghana, ed. M.E.K. Dakubu, 12–49. London: Kegan Paul.

Rattray, R. S. 1932. The Tribes of the Ashanti Hinterland. Oxford: Clarendon.

Schiffer, M. 1976. Behavioural Archaeology. New York: Academic Press.

———. 1985. “Is there a ‘Pompeii Premise’ in archaeology?” Journal of Anthropological Research 41:18–41.

Spratt, D., and N. Spratt. 1968. Collected Field Reports on the Phonology of Kusasi. Legon: University of Ghana Institute of African Studies.

———. 1972. Kusal Syntax. Legon: University of Ghana Institute of African Studies.

Stevenson, M. G. 1982. “Towards an Understanding of Site Abandonment Behaviour: Evidence from Historic Mining Camps in the Southwest Yukon.” Journal of Anthropological Archaeology 1:237–65.

Syme, J.K.G. 1932. The Kusasis: A Short History (mimeo).

Tomka, Steve A. 1993. “Site Abandonment Behaviour among Transhumant Agro-Pastoralists: The Effects of Delayed Curation on Assemblage Composition.” In Abandonment of Settlements and Regions: Ethnoarchaeological and Archaeological Approaches, ed. Catherine M. Cameron and Steve A. Tomka, 11–24. Cambridge: Cambridge University Press.

Webber, Paul. 1996a. “Agrarian change in Kusasi, North-East Ghana.” Africa 66, no. 3:437–57.

———. 1996b. “News from the Village: Agrarian Change in Ghana.” Geography Review 9, no. 3:25–30.

Whitehead, A. 1988. Distributional Effect of Cash Crop Innovation: The Peripherally Commercialised Farmers of North East Ghana. Sussex: Institute of Development Studies.

Wiszniewski, T. M. 1955. “Dry Season Vegetable Gardening in Kusasi.” Gold Coast Department of Agriculture Quarterly Newsletter 9, no. 1:3–5.