Vernacular architecture in eastern Senegal: Chaînes opératoires and technical choices

Introduction

Architecture is an important component of cultural identity, but surprisingly it has attracted limited attention in West Africa, and there is still a lack of references that address the dynamics of architectural techniques through space and time, and the factors involved. Moreover, there is a particular need for architectural studies that focus on local building materials and knowledge, as these are rapidly transforming.

Among the scarce available and relevant studies conducted in West Africa, an ethno-archaeological study of water settlements in southern Benin was exemplary in addressing the influence of environment, i.e. the water levels, on different forms of architecture, but it relates to a very specific type of architecture in a single cultural group (Pétrequin and Pétrequin, 1984). Studies have also been conducted by architects in the Dogon Country in Mali, focusing on the specificities of Dogon architecture in a rocky refuge zone (Lauber and Cissé, 1998). In Senegal, the most informative studies are the descriptions of architectural styles of various ethnic groups by an architect, four decades ago (Dujarric, 1986; Dujarric et al., 1976). Several ethnographic studies also provided interesting information on the architecture and ways of life of communities (e.g. Richard-Molard, 1953–1954; Thomas, 1964), but all emphasize a neat correspondence between an architectural style and an ethnic identity. Rare are the studies mentioning architectural variability, like that of Bourdier and Minh-Ha (1996) mentioning, for example, architectural differences within a village, linked to the presence of inhabitants of different identities. Even fewer question the transformations of the construction techniques of an ethnic group after its displacement into a new environment (Seignobos, 1982).

Numerous studies have questioned the definition of ethnicity in the post-colonial context (e.g. Amselle and M’Bokolo, 1985; Barth, 1969), as well as the links between material culture and ethnicity (e.g. Gosselain, 2000; Kasfir, 1984). A strict correspondence between ethnic identities and material culture cannot be assumed a priori, and borrowing mechanisms have been evidenced, notably in the domain of ceramic traditions (Gelbert, 2001). Nevertheless, there are mechanisms of transmissions of knowledge inside the ethno-linguistic groups, and sometimes a deliberate wish to be distinguished from neighbours, which can explain correspondences between material culture and cultural identity (Gallay et al., 1998; Hodder, 1982; Mayor, 2010, 2011). These potential correlations have to be questioned with sound data, and the field of architecture is especially interesting for discussing the history of techniques and population. However, no research has systematically compared architectural techniques of several cultural groups in different ecological zones, and explored the phenomenon of transfers of technique, as has been done notably for other categories of material culture, like ceramics.

Our ethno-archaeological study was conducted since 2015 in eastern Senegal, in the Faleme valley and the Bedik Country. Our survey includes 33 villages inhabited by Soninke, Malinke, Jakhanke, Fulani, Bedik and Jalonke people. Our aim was to precisely document current architectural techniques and their evolution during the past two to three centuries, across several ecological zones and several ethno-linguistic groups. The aim was also to analyse the factors that influence technical choices, either environmental, cultural or socio-economic. This study seeks to develop a database useful to archaeologists interested in the interpretation of past West African settlements, and an archive of a rapidly disappearing body of local knowledge, which is part of a regional cultural heritage and history. The overall methodology, precise descriptions of different mud-construction techniques present in different parts of the world and the attempt to understand the material culture dynamics in the domain of vernacular architecture should be a reference that arouses the interest of a wider audience.

The architecture of villages in eastern Senegal, at a first glance, gives an impression of monotony: houses are mostly circular and all have earthen walls and conical roofs covered with straw. Nonetheless, a closer observation of the chaînes opératoires (sometimes translated as ‘operational sequences’) of wall and roof construction reveals considerable variability. Following the description and mapping of different construction techniques, this study outlines the patterns observed, as well as the mechanisms that can explain them. This approach provides evidence of the borrowing of techniques in relation to the various historical contexts of the past few centuries in the studied area. It further shows that technical choices are influenced not only by cultural identity, but also by environmental constraints and matrimonial practices between groups, which play a crucial role in the adoption of new techniques.

This study is part of a doctoral thesis and was developed in the framework of the international and interdisciplinary research project Human Population and Palaeoenvironment in Africa, directed by E Huysecom and co-directed by A Mayor at the University of Geneva (Huysecom et al., 2016, 2017; Mayor et al., 2018, 2019). It was included in the global aim of the project, which was to explore technical dynamics during the last two millennia in the Faleme valley.

Geographic and historical context

Climatic–environmental conditions in eastern Senegal are oriented along a north–south gradient (Ben Yahmed and Ba, 2007). The southern zone, in the region of Kedougou, is characterized by precipitation of more than 1000 mm per year and marked relief (Figure 1). This Soudano–Guinean bio-geographical zone is dominated by wooded savanna and dry forest. The vegetation includes a high proportion of African mahogany (Khaya senegalensis), barwood (Pterocarpus erinaceus), African locust bean (Parkia biglobosa) and African fan palm (Borassus aethiopium) in some areas, as well as African lowland bamboo (Oxytenanthera abyssinica). The central Sudanian savanna, which receives between 800 and 1000 mm of rainfall per year, is characterized by a less marked relief and variable vegetation. While the northern part is wooded with a high proportion of bamboo and fan palm, its southern portion, while more humid, is distinctly less wooded, with bamboo almost completely absent. According to the traveller André Rançon (1894), this area contained bamboo stands before 1890, but they were destroyed by disease shortly before his arrival. The northern Sudanian zone, which is dryer, receives only 500 to 800 mm of rain per year, and is characterized by Sudanian vegetation, with mostly wooded savanna including baobab trees and occasional gallery forests.

OPEN IN VIEWER

Figure 1.

Map showing villages and ethnic groups studied. © T. Pelmoine and A. Mayor.

With regard to the cultural context, eastern Senegal is characterized by several ethnic groups with distinct languages (Albenque, 1967; Dupire, 1963; Gomila and Ferry, 1966; N’Diaye, 1970; Smith, 1965a, 1965b). They fall under two main large linguistic categories: the Manding and the Fulbe. By ‘Manding’, we refer to all people speaking a language of the Mande family, including the Soninke, Malinke, Jakhanke and Jalonke. Under ‘Fulbe’ we place all Fulbe-speaking groups of diverse historical trajectories and of variable episodes of inter-mixing, including the Fulbe of Bundu, Tamgué and Bandemba. In addition, there are the Bedik people, speaking a language, like the Fulbe language, belonging to the west-Atlantic family.

Today, in all geographical zones, Manding and Fulbe groups are sedentary agriculturalists practising polygyny and patrilocal residence. Nevertheless, endogamous marriages are still very much respected in the south, a trait that explains the prevalence of contrasting cultural identities, whereas the social mixing of northern groups has driven their cultural ‘homogenization’. All these groups are currently Muslim, with the exception of the Bedik, an animist population converted for the most part to Christianity.

After the decline of the Empire of Mali during the 16th century, eastern Senegal has witnessed the rise of numerous local powers between the 17th and 19th centuries. The Fulbe kingdom of Bundu established itself in the north, where it imposed its authority and Muslim religion on the resident Manding populations (Clark, 1996). In contrast, the central part of this area saw the development of several small warrior Malinke kingdoms, called Dantila, Sirimana and Beledougou (Aubert, 1923). The southern zone, located at the northern margin of the Jalonke kingdoms of Tembafou and Sangalan, and of the Fulbe kingdom of Futa-Jalon, acted as a refuge zone, notably for the Bedik, who withdrew into the hills during moments of insecurity and expanded again in the plain during times of peace. This kind of situation finds parallels in the Dogon Country in Mali (Gallay, 1994).

Material and methods

In our aim to develop a reference database on vernacular architecture able to fuel the debate on cultural transitions and the adoption of techniques, we draw on methods developed in ethnoarchaeology and on comparative technology approaches (Gallay, 2011; Gosselain, 2017; Mayor, 2017). To document and understand the observed diversity of technical choices, we used an extensive research strategy that considers numerous villages established in variable environmental contexts and inhabited by distinct cultural groups.

The scale of our research area is vast and required a significant amount of fieldwork. During 10 months of fieldwork, conducted between January 2015 and March 2018, we studied 60 compounds in 33 villages (see Table 1 and Figure 2). By compound, we refer to a collection of buildings and spaces belonging to an extended family and placed under the authority of a household chief. Across all compounds, we recorded the construction chaînes opératoires of 355 dwellings. By dwelling, we designate the architectural structures used primarily for sleeping and the safeguarding of personal items. Other than dwellings, compounds are typically composed of kitchens and granaries, which will not be considered in this article. As a general rule, we selected two compounds per village in order to consider architectural variability. Different constraints were driving the choice of these two compounds. They had to include the different types of architectural structures (dwellings, kitchens, granaries), and they had to correspond to villagers’ view of what composes a ‘traditional compound’, without the addition of modern material like cement or an iron roof. We conducted a quantitative survey to ensure that the dwellings of the two selected compounds were representative of the dwellings of the whole village.

OPEN IN VIEWER

Table 1.

Table summarizing the architectural data under study.

OPEN IN VIEWER

Figure 2.

Photograph of a Malinke compound at Daloto (V11). © T. Pelmoine.

In each compound, the survey consisted of a plan-view drawing of the building layout, complemented by a semi-structured interview with the household chief. This interview allowed for the documentation of construction techniques and materials used, as well as the function of different spaces, the status of the inhabitants, and architectural changes, if any. Additionally, we were able to observe the ongoing construction of new dwellings in several villages, allowing us to clarify actions and the material used.

The collection of data in the field was then followed by a detailed description of the chaîne opératoire of the construction techniques for the walls and the roofs. This allowed us to analyse the architectural technical system in detail, in order to individualize its different elements (Lemonnier, 1976: 108). Various ‘operations’ (such as actions and gestures) and materials lead to different technical sequences, here referred to as ‘methods’ (Balfet, 1991; Karlin et al., 1991; Lemonnier, 1992). A cartographic and statistical analysis of this architectural variability has completed this study, with the aim of suggesting patterns and their explanation through reference to environmental, socio-economical and cultural constraints.

This article focuses on the variability of materials and techniques, and leaves to other publications the spatial organization of the compounds, a theme that has already been considered preliminarily elsewhere (Huysecom et al., 2016, 2017, Mayor et al., 2018, 2019).

Description of techniques

Cob construction

In West Africa, people refer to the technique of earth construction by the generic term of ‘banco’ (Prussin, 1970), which is somewhat ambiguous as it may apply to constructions in both cob and brick. Cob or ‘piled earth’ constructions may be undertaken simply by building up clumps of earth in successive levels without any formwork (Aurenche et al., 2011).

In eastern Senegal, the preparation of the material takes place at one end of the compound. This material is made of earth, sometimes mixed with other elements such as temper. The raw material can be clayey or lateritic sediments, or come from termite mounds. To this material are often added vegetal elements such as chopped grasses, or more occasionally a concoction of the sap of cissus populnea. According to Herta Haselberger (1963), the straw acts as an adhesive, while the micro-organisms present in the sap catalyse bio-chemical reactions that harden the earth.

Once the ground has been swept, it may be prepared in one or two ways, according to the kind of soil upon which the structure is to be built. If the ground is rocky and impossible to dig, it is often decided to lay a foundation of small paving slabs, 10 cm high and 20 cm thick, so as to isolate the wall. If the ground is soft, a foundation trench of 20 to 50 cm is dug with a shovel and pickaxe. The bottom is filled with 10 to 20 cm of rubble. The upper part is filled by the first course of cob (Figure 3).

OPEN IN VIEWER

Figure 3.

‘Chaîne opératoire’ of the cob wall construction technique. © T. Pelmoine.

We have identified two main methods for raising the walls. The more complex (method 1) consists of simultaneously laying two courses of cob. Two clumps are laid in the first course, followed by a third, placed staggered in a second course. This process is repeated until there is a section of 30 to 40 cm in height. The course is then manually straightened by tearing off protruding bits of earth from the clumps and using this to fill any visible joints. The wall is then scraped by hand in order to form a flat vertical plane, and finally smoothed with a machete blade or split branch. This wall section must then dry for one day before another can be added. The method with the fewest steps (method 2), consists of laying only one course 10–15 cm in height. The mason’s hands always repeat the same series of movements. Each new clump is placed against the last, then drawn up with the palm first diagonally, then vertically in order to cover the joints. The top of the clump is pinched with the thumbs to flatten it and prepare it for the next course. The entire circumference of the structure is built up in a similar fashion. It is possible to lay two courses per day. For both methods, this series of gestures is repeated until a wall of 1.6 to 2.4 m has been raised. The walls of the dwelling are raised in one to two weeks.

Once dry, the wall must be protected by a render. The women make the choice between termite earth, clayey or organic soil. This is then mixed with cissus populnea sap for a waterproofing effect. The occasional addition of liquified cow dung can reinforce this effect. Once the wall has been rendered, the exterior base is further buttressed by stones or packed earth. Currently, however, cement is often preferred.

As far as the interior ground surface is concerned, the natural soil is dug, humidified, packed down with a wooden plank, then coated with the same concoction used to render the walls. If the soil is rocky, it is coated with this render without the preceding preparation.

Moulded brick construction

Also known as ‘banco’, the moulded brick technique is widespread in the area. This method consists of filling a bottomless parallelepiped mould with earth, then turning it out and allowing it to dry in the sun (Aurenche et al., 2011). In eastern Senegal, multiple kinds of soil can be extracted with a shovel and pickaxe to make the bricks. Using lateritic earth does not require any additional materials. In contrast, the use of a clayey earth, generally sourced from river banks or dried swamps, requires the addition of finely-chopped grasses. More rarely, this clayey earth is mixed with termite earth and cissus populnea sap. Termite dirt can also be used on its own, or mixed with chopped grasses. These various materials are mixed with water to form the construction earth.

It is possible to make the bricks at the mining site, once the earth has been mixed and left to settle. In this case, a wooden mould is placed on the ground and filled manually with the construction earth. The earth is packed by hand, then levelled off and immediately turned out of the mould by drawing it upwards. These bricks are then allowed to dry on their bed, then on their stretcher faces for approximately 15 days, before being taken back to the compound.

On the day of construction or the day before, a pit is dug within or outside the compound in which earth and water are mixed to make an earthen mortar (Figure 4). The ground surface intended for construction is then swept and the wall foundations are outlined in the soil with a rope attached to a branch.

OPEN IN VIEWER

Figure 4.

‘Chaîne opératoire’ of the brick wall construction technique. © T. Pelmoine.

Brick structures are for the most part built on foundations. These are simple and allow the wall to be well anchored in the soil. They are dug with shovel and pickaxe, and laid as one to five courses of bricks sealed with an earthen mortar.

The raising of the wall is quick and can be completed in several hours. Community members lay the bricks on a bed, leaving very loose joints (2 to 5 cm) between each brick. The space between the bricks is filled with mortar, then a thick bed of mortar is spread on the surfaces of the bricks in order to prepare them for a second course. This operation is repeated until the wall is built. Any mortar squeezed out from the bed is used to fill in badly sealed joints. The builders then leave the wall to dry for several days.

A wall primer may be made either of clayey earth, termite earth, or organic earth, but these are never mixed. Cissus populnea sap and/or cow dung is almost always added. Following this primary coating, a second decorative and protective coat may also be applied. This is typically made of natural white earth or earth whitened with ash. The exterior wall base is generally buttressed with bricks or stones, or by packing in a bevelled earthen reinforcement.

The interior ground surface, whether it be natural soil or coated with a render of termite earth, is packed down with a plank of wood. This packed surface is then given a protective coating of white, black, or clayey earth, in all cases mixed with cow dung and also occasionally with cissus populnea sap. Today, the ground surface is more and more frequently coated with cement.

Wattle-and-daub construction

This method consists of protecting a wooden scaffold with a mixture of clayey earth and various other materials, primarily plant fibres. In contrast to the cob and brick methods, here the earth is not load bearing, it is used only for its insulating qualities (Dewulf, 2015).

In eastern Senegal, wattled walls are constructed solely of ‘humid’ bamboo, which has been freshly cut not far from the village (Figure 5).

OPEN IN VIEWER

Figure 5.

‘Chaîne opératoire’ of the wattle-and-daub wall construction technique. © T. Pelmoine.

As with the other construction techniques, the ground is cleaned and the outline of the building is drawn in the soil. Bamboo construction requires a deep base in the form of a series of narrow postholes 30 to 60 cm deep, dug with a pickaxe. In each posthole, two bamboo stalks are positioned vertically before the hole is filled. Once the circumference is completed with vertical poles, the bamboo stalks are woven horizontally, forming the wattling. At this stage, the door is cut with a machete.

In this region, the wattling is often filled in with rubble. In a pit near the construction area, this temper is prepared by mixing clayey earth with termite earth and, almost always, with the addition of straw and occasionally cow dung and water. This mixture is sometimes left to soak for several days. When the material is ready, the builders take the clumps of mud in their hands and throw them against the wattling until it is completely covered. Then they flatten and scrape the surface with the palms of their hands.

Following this, the wall is rendered with back, clayey, or termite earth. Cow dung is sometimes added to the mixture, as well as Cissus populnea sap. This surface may be further protected with a new layer of ashy or white earth. Finally, the exterior wall base is buttressed by bricks or packed earth.

The interior ground surface is prepared similarly to that of the brick dwellings previously described.

Roof construction

The roofs are of common rafter construction, covered with plant fibres (Figures 6 and 7). The roofs are simply placed on the walls, without any special preparation.

OPEN IN VIEWER

Figure 6.

‘Chaîne opératoire’ of the roof construction techniques. © T. Pelmoine.

OPEN IN VIEWER

Figure 7.

Drawing of the three roof-covering construction methods: (a) straw spread on the roof, covered by a net of plant fibre; (b) straw woven in strips covered by a solid wood frame; (c) straw held in a clamp. The numbers in black circles refer to the elements and steps of the ‘chaîne opératoire’ described in Figure 6. © T. Pelmoine and D. Glauser.

The roof frames are commonly made of ‘dry’ bamboo. If this is not available, the builders gather straight, non-ramified wood of various species (e.g. Anogeissus leiocarpus, Bombax costatum, Borassus flabellifer, Erythrophleum guineense, Mitragyna inermis, Pterocarpus erinaceus and Strophantus hispidus). The assembly takes place on the ground, beginning with the laying out of three or four bamboo stalks, which serve as rafters, lashed together at the top with a double-ring of braided plant fibres. Once this base is fortified, additional bamboo stalks are slid into place. Next, a second ring is fixed 50 cm below the previous one. This procedure is repeated until the roof frame is complete. Once finished, it is installed on top of the dry building walls.

With regard to the covering, there are three principal methods. The first (method A: Figures 6A and 7A) consists of throwing bundles of straw on the frame, before spreading them evenly across the entire structure from the bottom to the top. Once the entire frame is covered, the straw is fixed in place by a cord made of plant matter and simple bamboo scaffold. The life expectancy for such a covering is one to five years. This method occasionally replaces straw with palm leaves.

The second method (method B; Figures 6B and 7B) takes longer to build, but is more solid. Straw is laid out on the floor in a long strip two to four cm thick. Plant fibres (or wire) are woven around the most solid part of the grasses. The strip is then rolled up and allowed to dry, before being unrolled on the base of the roof frame. Subsequent strips are laid from the bottom to the top of the frame, with each new one covering two thirds of the strip below. Once the frame is covered, the covering may be left as such, or can be fixed in place by one or two bamboo stalks attached to the top in order to prevent the wind blowing it away. In method A, it is also possible to make a cord net for this same purpose. The covering has a life expectancy of some 3 to 8 years.

The third method (method C; Figure 6C and 7C) is the most solid. A sheaf of durable straw, 10 cm thick, is packed together and placed, grain facing upwards, on the lower part of the roof frame. Next, a stalk of bamboo, cut lengthways, is placed in order to clamp the straw between itself and a ring of the roof frame below. Then, the additional straw is slid under the stalk all around the roof. A second layer may then be placed on the first, covering at least two thirds of the latter. The builders repeat this operation until they have 12 to 15 layers. The life expectancy of this type of covering varies between 7 and 20 years, depending on the quality of the straw and the thickness of the installation.

Spatial distribution of the techniques

The description of wall and roof construction techniques has revealed a significant variability in the different chaînes opératoires. To summarize, the walls may be constructed by one of three main techniques: wattling, bricks, or cob, with the latter following one of two distinct methods. There is an equally significant diversity of methods for roof construction, including two methods for the construction of the roof frame, three for the installation of roof coverings, and five for fixing these coverings in place.

Regarding the spatial distribution of wall construction techniques in eastern Senegal (Figure 8), the northern zone is largely characterized by the use of sun-dried moulded bricks. Only the two villages located on the eastern bank of the Faleme build both in brick and in bamboo wattling, and the easternmost village in Mali, builds solely in wattling. Concerning the central zone, in its northern half the most used technique is that of bamboo wattling, although bricks are also very occasionally used. In its southern half, in contrast, bricks are the primary construction material, while wattling is only seldom used. Only in the southernmost village are bricks and cob used together. Finally, the southern zone is characterized by the near total use of cob construction.

OPEN IN VIEWER

Figure 8.

Map showing spatial distribution and frequencies of wall construction techniques by village and by ethnic group. © T. Pelmoine.

Regarding the roof construction techniques, specifically the covering on the roof frame (Figure 9), the northern zone is characterized by the throwing of straw directly onto the roof structure (method A) and the use of a lightweight scaffold to hold it in place. In the central zone, the most common method is the weaving of strips of straw that are then installed on the roof frame (method B). Only two villages do not follow this pattern, with one preferring the northern method, and the other preferring that used in the south. Finally, the southern zone is the most complex, with two methods being well represented: that of the thick beds of straw held in place by a bamboo clamp (method C), and where straw is spread out on the roof frame (method A) and covered here with a solid bamboo scaffold. The two easternmost villages in this area use a hybrid method combining, sometimes on the same roofs, straw clamped in place by bamboo stalks and woven strips of straw (typical of the central zone).

OPEN IN VIEWER

Figure 9.

Map showing spatial distribution and frequencies of roof covering construction techniques by village and by ethnic group. © T. Pelmoine.

There are thus several major differences in wall and roof construction techniques that allow the three zones of the area to be distinguished. While the southern zone is somewhat homogeneous in its wall construction technique, the northern zone is extremely so regarding roof construction technique. The central zone includes wall and roof construction techniques different from the ones present in the other zones and shows some homogeneity in the roof-covering technique.

Analysis of factors behind variability

This demonstration of variability of architectural techniques in eastern Senegal, as well as its spatial mapping, has led us to question the factors behind certain technical choices in an area characterized by diverse physical, cultural and historical environments.

The role of cultural factors

In general, nearly all the groups studied today have a preferred method for constructing their walls and roofs. During the surveys, when we asked the inhabitants to name the architectural elements that best reflected their identity, they most often cited their techniques of roof construction. Our analysis indeed confirms that roof construction techniques are closely correlated with cultural identity but, furthermore, that other construction techniques are also affected by cultural factors. However, the numerous borrowings that have taken place over the course of time have blurred the image, while at the same time testifying to the interwoven links between different groups and their contacts with the exterior (see Table 2).

OPEN IN VIEWER

Table 2.

A synthesis of historic changes of construction techniques by ethnic group.

Wall construction techniques

Our detailed analysis shows that it is difficult to discern unequivocal links between wall construction techniques and a particular culture, given that different cultural groups use the same techniques.

In our region of study, only the second method of the cob construction technique is correlated exclusively with Bedik masons, the earliest established group in the area. These masons work as often for the Bedik as for the neighbouring Fulbe groups of the Bandemba region, who have forged close links with the former after two centuries of interbreeding with Bedik women (Langaney, 1972). The Bedik explain their choice by virtue of their refusal to build in wattling, in order to avoid any association with their former oppressors, the Manding slave traders or Fulbe Muslims from Futa-Jalon (Rançon, 1894). Indeed, their reclusive attitude towards identity partly explains the perseverance of their traditions and their reticence towards adopting new techniques.

As for the Fulbe of Bandemba, they find the Bedik technique of cob construction more durable than their former bamboo wattling technique, and for them there is no inconvenience in hiring Bedik masons to construct their houses. Nevertheless, in their villages, there remain some wattled structures that testify to their former technique. Intermarriages with Bedik women have thus led to modification of wall construction techniques, but not those of the roofs. This may be explained by the fact that women participate only in the construction of the walls, and so it is only here that there is the possibility for the transfer of knowledge.

The Fulbe of Tamgue also build using the cob technique. This group arrived from Futa-Jalon after the wars of Alfa Yaya around 1880. In the Futa-Jalon, they constructed using wattle-and-daub (Richard-Molard, 1953). All the current villages settled at the west of the Gambia River opted for the cob construction technique for its durability, and got Bedik masons to construct it for them, according to method 2. However, the masons of the village of Dian-Héri (see Figure 8, V56), located to the east of the Gambia River and inhabited solely by artisans, build according to the cob technique method 1, like their Jalonke neighbours. Considered in the literature to be former slaves (Dupire, 1963: 266), probably of Manding origin, they may have borrowed the technique from the Jalonke, or kept their former knowledge, as it used to be the technique formerly used also by the Malinke and Jakhanke, both linked with the Manding groups.

The dominance of the cob technique for wall construction in the south can be explained partly by the refusal of the Bedik to change their technique, and partly by the multiple episodes of adoption by the Fulbe of techniques practised by the Bedik (method 2) and the Manding (method 1).

In the north, like the Fulbe, both the Jakhanke and Soninke nowadays build exclusively in mud brick. Only the Malinke, to the east of the Faleme, build partially in bamboo wattling, as do their Malinke neighbours of the central zone, and those further east, in Mali.

It must first be mentioned that the moulded brick construction technique seems to be a recent introduction to eastern Senegal. All our interviewees confirmed that moulded bricks were adopted around 1960, when locals began to imitate the cement blocks of colonial buildings. This technique is considered more modern, less expensive and more rapid. This implies that those villages that construct nowadays in brick used another technique only two or three generations ago. We must nevertheless also point out that the Jakhanke in the northern zone have used bricks for at least 100 years, although they are non-moulded and only for mosque construction.

Therefore, in the studied area, brick construction techniques have supplanted former techniques in all groups. Only a few isolated dwellings that date back several decades are made of cob, confirming oral testimonies mentioning its existence in this area among Manding groups. Some Fulbe oral sources, confirmed by a written source (Rançon, 1894), also note the existence of walls built of straw at the time of their arrival in the zone, as well as in their non-permanent farming villages. During our surveys, the single dwelling built in this way was inhabited by a mentally challenged person. This adult was in fact presented by the inhabitants of Goundafa (V4) as a community member different from others in the sense that his development stopped between infancy and adulthood. This person had to eat with children, but could not sleep with his mother any more, hence he had his own dwelling built in straw. This technique, considered as simple and obsolete, paralleled this individual’s social status within society.

Roof construction techniques

Roof construction techniques in general seem to be more resistant to being replaced by newly borrowed techniques than those of the walls. For the roofs, it is thus easier to articulate the relationships between techniques and culture, even if they are not always unequivocal.

One regularity concerns the frames: if they are constructed on a base with three rafters, the builders are from the Fulbe groups; if the base has four rafters, the builders are Bedik and from the Manding groups. Unfortunately, we do not know the origin of this difference and it would be premature to suggest hypotheses that refer to aspects of their history.

Concerning the roof coverings, the regularities linking materials and identities are as follows: if the covering is constituted of straw spread out on a frame (method A), then the building must belong to a Fulbe or an immigrant from another group living in the Bundu territory, formerly dominated by the Fulbe. If the roof covering is made of strips of woven straw (method B), then it is Manding, and if the covering is made of straw held in place by a bamboo clamp (method C), then it covers a Bedik dwelling.

In the northern zone, roofs made with straw spread on a frame are common to all Fulbe dwellings. The local Jakhanke and Soninke populations have also adopted this method due to its rapidity. Only the manner in which the straw is held on the roof changes from area to area, on essentially climatic grounds, as previously mentioned. In the central zone, coverings made of strips of woven straw are all built by Manding groups. It is also this method that was used two generations ago by the Jakhanke, the Malinke and the Soninke in the northern zone, before it was abandoned upon contact with the Fulbe. In the southern zone, only the Bedik use method C. The result obtained by this method is visually similar to that of method B, but provides greater stability, a characteristic that probably makes it attractive to high-altitude villages subject to heavy winds.

Finally, it is worth noting the idiosyncrasy of Jalonke roofs, which represent a hybrid form thanks to the influences of multiple neighbouring groups. In the villages of Malinda (V60) and Wamba (V61), the majority of coverings have two bottom rows made of straw clamped in place with a bamboo rod (method C ‘Bedik’) then, once this structure is stable, woven bands of straw are settled for the top part of the roof (method B ‘Manding’). In Malinda, roofs with this covering may stand adjacent to some others, made of straw, spread over the frame (method A ‘Fulbe’). The latter are held in place by nets, which are particularly common in the north. Historic movements of the Jalonke, a Manding group that has seen much displacement because of multiple military defeats, may explain this variability and hybridity. Indeed, the inhabitants of Malinda come from villages located in Guinea, belonging to the former kingdoms of Futa-Jalon and Sangalan, where they lived in proximity to Fulbe groups. In contrast, the inhabitants of Wamba, part of the Jalonke kingdom of Tembafou, were in close proximity to Bedik groups, before the latter retreated into the Bandemba hills. These hybrid roofs thus reflect the diverse relationships that the Jalonke have built with their Fulbe, Bedik and Malinke neighbours over the course of time.

Discussion

The examination of the different factors at play in the choice between different techniques raises several interesting issues, notably in relation with historical dynamics.

Firstly, it can be clearly deduced that the dynamics related to the adoption of new techniques differ if one considers wall or roof construction techniques. Indeed, wall construction techniques seem to pass more freely from group to group, as we have noted for the cob technique, which transfers from the Bedik and the Manding to the Fulbe of Bandemba and Tamgué, and for the moulded brick technique, which is replacing little by little nearly all pre-existing techniques (see Figure 10). A snapshot of the current situation therefore seems to show large distribution zones for wall construction techniques, poorly correlated with the distribution of cultural groups. Only the interviews and data from the literature are available to trace back the history of these techniques and their influences. On the other hand, the technique of covering one’s roof with straw for protection seems on the whole more resistant to outside influences and is often named by the communities as one of the principal traits that differentiates them from the inhabitants of other groups. These techniques continue to be closely linked to identity in the central and southern zones, whereas the past one or two generations in the northern zone have seen the transfer of the technique from Fulbe to Manding groups (Figure 11).

OPEN IN VIEWER

Figure 10.

Current distribution of wall construction techniques and their historic variability. © T. Pelmoine and A. Mayor.

OPEN IN VIEWER

Figure 11.

Current distribution of roof covering construction techniques and their historic variability. © T. Pelmoine and A. Mayor.

Secondly, the mechanisms behind such transfers differ according to the historical context. In the north, the theocratic kingdom of Bundu (zone 1) brought several Manding groups under its power, such as the Jakhanke, the Soninke, and the Malinke. These populations were obliged to convert to Islam in order to escape slavery, and group exogamy became permitted. The subsequent social mixing provided fertile ground for the transfer of knowledge, leading to socio-cultural and technical homogenization. The consequences for material traditions were significant, as can be seen from the architecture, which varies very little from village to village, regardless of the cultural identity of the inhabitants. Here, not only are the walls constructed in the same way, but also the roofs, since all groups have simply adopted the most rapid means of construction. This phenomenon does become weaker, however, to the east of the Faleme, towards the Malinke kingdom of Bambouk, where villages have notably preserved certain traditional techniques, such as wattling. The weaker authority of the Fulbe, as well as the relative abundance of bamboo in this remote area may explain this resistance. This situation can be conceptualized as a zone of technical homogeneity for walls and roofs, in a heterogeneous cultural context. The technical borrowings are probably due to social blending and attraction for modernity.

Three Malinke warrior kingdoms administered the central zone: the Dantila, the Sirimana, and the Beledougou (zone 2). The population of this territory, essentially made up of Malinke villages, as well as a few Jakhanke and Fulbe ones, is relatively homogeneous culturally, with intermarriages with other groups being rare. Changes in technique are thus mainly due to the pull of modernity, seen in the progressive adoption of bricks instead of cob in villages where bamboo is scarce or absent in the surrounding environment. However, cultural identity can be clearly recognized by the mode of roof construction, since woven straw coverings are almost exclusive in the Malinke villages, while being completely absent from other areas. This situation is a zone of homogeneity for roofs and heterogeneity for walls, in a homogeneous cultural context. Changes are due to the attractions of modernity and reluctance to borrow from neighbouring groups.

The Bedik Country in the southern zone (zone 4) may be thought of as a landlocked place, located at the periphery of the surrounding Malinke, Jalonke and Fulbe kingdoms. Due to the frequent slave raids and campaigns waged to convert the local population to Islam, security in this area was once very low. This heightened danger resulted in an identity withdrawal of the communities into themselves, and the perpetuation of group endogamy. Here, the barrier to the adoption of new techniques is symbolic: the Bedik have fiercely guarded their cob construction technique as a marker of identity. But some of the Fulbe of Bandemba who settled there married Bedik women and hired Bedik professional masons to build their houses, while keeping their own way of building the roofs. It can thus be conceptualized as a zone of homogeneity for walls and heterogeneity for roofs, in a heterogeneous cultural context. Here, one group is attached to its traditions, while the other is borrowing part of the needed techniques after long-lasting social mixing.

Conclusion

Our study of vernacular architecture in eastern Senegal has provided rich and complex information for understanding the evolution of architectural techniques and the history of relationships between ethnic groups in this understudied region of West Africa. While the architecture of this area may seem, at first, homogeneous and ordinary, in-depth field documentation has shown much variation in both the earthen walls and straw roof coverings. The identification of different chaînes opératoires and drawings of architectural layouts from a significant number of villages located in diverse cultural and environmental contexts have led to the first comprehensive documentation useful for further architectural, historical and heritage studies in Senegal and West African savannas. Mapping and statistical analyses of the observed variability in building techniques have revealed complex correlations between spatial distribution and proxies like environment and culture, due to numerous borrowings between groups and changes in technical choices during the last two to three centuries. The history of the different kingdoms in the region during the second part of the Atlantic era, the contrasting attitudes of the authorities towards matrimonial exchanges between ethnic groups, as well as their influence on threatened marginal groups, explain the presence of favourable or unfavourable contexts for adopting new techniques. Finally, the multiple patterns of regularity highlighted in our database provide archaeologists with references that may help to interpret their material remains. These regular architectural patterns will be tested on the proto-historic settlement sites and endogeneous fortifications excavated in the Faleme River valley in the framework of the project ‘Human Population and Paleoenvironment in Africa’ (Aymeric, 2019; Huysecom et al., 2016, 2017; Loukou, 2018; Mayor et al., 2018, 2019). It will be also possible to compare these results with variability of other categories of material culture, like ceramics for example (Cantin and Mayor, 2018).

More broadly, the precise descriptions and illustrations of the chaînes opératoires of construction are intended as a useful reference grid for all architectural studies dealing with earthen dwellings and vegetal roofs, wherever they are in the world. The global methodology for unravelling the variability of techniques and the underlying dynamic through space and time may also provide leads for researchers outside the field of African architecture. Moreover, our results, in terms of the mechanisms of borrowing and technical choices in architecture, introduce a new case study for developing the anthropology of techniques, based on various other categories of material culture.