1The “Crossing Borders Project” was established in 2006 by the Section of Archaeology and the Service of Linguistics at the Royal Museum for Central Africa (Tervuren, Belgium). The general aim of this joint project is to work towards the integration of archaeology and historical linguistics (Livingstone Smith et al. 2006). In Africa, as in other parts of the world, the two disciplines have a long and troubled history. Notwithstanding the positive aspects of cross disciplinary thinking, this is particularly true as regards the Bantu languages expansion phenomenon. Beyond the problems of shaky interpretations and circular reasoning, much of the trouble comes from the lack of a proper methodology (for an overview of this topic see: Blench 2006; Blench & Spriggs 1999; Eggert 2005; Nurse 1997). In fact, the sheer geographical and temporal scale of the expansion of Bantu languages makes it difficult to control the different parameters of the debate.
2The linguists have at their disposal a relatively large set of data covering the southern half of the African continent. However, the same area is only partially known from an archaeological point of view. As put by Eggert, “with rare exceptions (i.e. parts of East Africa), we are unable to even sketch the broadest outlines of Bantu territory archaeology” (Eggert 2005:321). As regards the temporal scale, the radiometric dating systems used by archaeologists provide an absolute chronology, while the chronological system used by historical linguists is essentially relative. Correlating these chronological frameworks is obviously difficult. Finally, at such broad temporal and geographical scale, it is difficult to understand the social mechanisms underlying the convergence of material and non-material culture traits. In a way, models of interpretation are as broad as the territories and times they are covering.
3To address these problems, we decided to limit the methodological, chronological and geographical scope of the project “Crossing Borders”. The idea is to anchor our cross-disciplinary work in the present (study of modern day pottery traditions) and to focus the chronology on the last millennium and the geography on Southern Central Africa.
- 1 The roughing-out process is the first step of pottery building.
- 2 This approach is labelled ethno-archaeology. It is essentially a simple anthropology of material cu (...)
4Firstly, the methodological integration is achieved thanks to a third party: anthropology. In the Bantu area, as in other areas of the world, studies of modern ceramic production were devoted to the relationship between social identities and material culture (see for example: Bostoen 2005; Gosselain 2000, 2002). The strength of these studies was to consider the social mechanisms underlying specific spatial configurations of cultural traits. Fieldwork enquiries on pottery manufacture encompassed the identity and biography of the artisans (social and linguistic affiliation, pre- and post marital movements, learning, etc.), the socio-economic context in which the craft is performed (organisation of production and distribution, etc.), as well as observations concerning the complete manufacturing process. This allowed for a much more comprehensive view of the relationship between people and their material or non-material cultural productions. Gosselain’s (1998; 2000) examination of pottery building methods in sub-Saharan Africa, for example, showed the need for detailed descriptions of the roughing-out process1 and their relationship with social and linguistic identity (for examples outside of the Bantu area, see for instance: Dietler & Herbich 1989; Gelbert 2003; Herbich 1987; Livingstone Smith 2000; Mayor et al. 2005; Sall 2005; Sterner & David 2003). Studies devoted to the relationship between identities and pottery chaînes opératoires showed the value of an integrated and comparative approach and developed new tools to explore the past2.
5Secondly, as regards the spatial constraint, the project is centred on Katanga (DRC). The area is known as the home of the Luba Kingdom, one of the great savannah kingdoms. It is also thought to be a border area between the eastern and western stream of Bantu languages –Bantu languages are divided in several large units related to distinct waves of languages expansion (Bastin et al. 1999; Bostoen & Grégoire 2007).
6Finally, considering our chronological framework, the project is focused on the last millennium. Indeed, the area of study is also renowned for its wealth of archaeological sites excavated between the 1950s and the 1970s (de Maret 1999). The excavations have resulted in a culture historical sequence spanning from the 9th to the 19th century ad. This sequence is characterised by the continuity of some elements of ceramic styles and the emergence of new ones throughout the sequence. Although archaeologists from Nenquin (1963:272-273) to de Maret (2004) have already considered the relationship between archaeological cultures and present populations, these stylistic variations remain to be explained in regard to population dynamics.
7In practical terms the aim of the project was to examine pottery traditions, with a specific focus on pottery building techniques, and languages in the Luba territory and surrounding area. To improve the comparative methods for studying the relationship between material culture and language, pottery traditions were used as an interface between disciplines and between the past and the present. Methodological integration, in practice, starts with joint fieldwork. The data pertaining to pottery making and languages were collected simultaneously, with the same informants. The analytical results (roughing-out techniques, languages, etc.) are mapped so that linguists and archaeologists in the project use the same framework for the data, the spatial patterning of cultural behaviours. Convergence and / or discrepancies between technical and linguistic behaviour can be assessed taking into account the social contexts in which they are performed. To build up a more detailed picture of linguistic variations, dialects were to be studied within the Luba area, while Kiluba was to be compared with neighbouring languages. Finally, the analysis of the technology of pottery excavated in the Upemba depression was planned to provide time depth to this overview (Livingstone Smith & Vysserias, accepted).
8A first fieldwork in 2006 (Livingstone Smith 2007) and additional ethnographic data suggested the existence of an important technical boundary in the distribution of shaping techniques in Southern Katanga (Livingstone Smith, submitted). What is the nature and extent of this boundary? Is it possible to relate this boundary to other aspects of society, such as languages and other cultural practices?
9To answer these questions we decided to proceed in two steps. Firstly, prior to field work, we re-appraised the distribution of primary forming techniques at a sub-continental level and established that this hypothetical technical border was indeed observed at a broader geographical scale. Secondly, new fieldwork was planned in the area to the northeast of Likasi and south of the Upemba depression to explore this question in more detail.
- 3 A variety of factors may cause artisans to move from their place of origin, such as marriage, resid (...)
10Joint linguistic and ethno-archaeological fieldwork was conducted in Katanga (DRC) between September 21st and October 28th 2008 (Fig. 1). Data on pottery production and languages were collected in a series of locations in and around Lubumbashi (Kimpakasa, Nwansha, Malambwe and in the “Zambia” area of the city Lubumbashi itself), in the region of Likasi (Kabanga Banga, Kitana, Kapede 2), near Fungurume and around Kiubo (Kiubo, Kialwe). Enquiries were also led around Mwashia (Ntondo, Lubuko) and on the road between Lubumbashi and Kasenga (Kumanwa Km 41, Malambwe). Artisans belonging to the Luba, Sumbu, Lomotwa, Nwenshi, Sanga, Lamba and Bemba groups were interviewed. Enquiries aimed at collecting information on the social context of the activity. Specific attention was paid to elements influencing the spatial and temporal distribution of pottery traditions, such as time and place of learning and pre- and post marital movements3. We also tried to map the potter’s knowledge of other artisans and traditions. The direct observation of pottery manufacturing processes was an important part of the interviews (Gosselain et al. 1996; Gosselain & Livingstone Smith 1997). From the linguistic point of view, the lexicon pertaining to pottery making was systematically collected in all visited populations. Furthermore, we decided to record other specialised vocabulary such as vocabulary related to metallurgy, salt production, kinship systems, as well as basic vocabulary and grammatical features. Because of our interest in the Luba Kingdom, we also worked with speakers of the different Kiluba variants to complete the data collected by a member of the project in 2007 (Birgit Ricquier, field notes). Finally, for comparison purposes, speakers of Cokwe, Cibemba, Ruund, Wundembo, Kisongye, Kiyashi, Cikanyok, Kitaabwa, Kishila and Kihemba were also interviewed in Lubumbashi.
Figure 1 - Location of the enquiries and the main languages spoken in southern Katanga
Enquiries made prior to 2006 in red, 2006 fieldwork in orange, 2008 fieldwork in brown.
© A. Livingstone-Smith
11The linguistic data collected in the field are still being processed, but several comments can be made of the spatial distribution of pottery building methods (the sequence and technique of roughing-out), technical vocabularies pertaining to pottery making and the social context in which they are performed.
12The 2008 fieldwork confirmed that there are a series of technical variations in southern Katanga. The pot forming techniques used by the artisans fall into three large categories: 1) cylindrical coiling, 2) drawing of a single ring, 3) drawing of superimposed rings (see Livingstone Smith, submitted, for a description of these techniques).
13The first technique consists in the superimposition of coils to build the wall of the vessel and the use of a disk shaped slab to build the bottom (Fig. 2). The second technique, drawing of a single ring, consists of drawing a doughnut shaped ring of clay upwards to build the walls, shoulder and neck of the vessel (Fig. 3a). When the upper part of the vessel is formed, shaped and decorated, the rough-out is turned upside down and the excess clay is drawn to build the bottom of the vessel (Fig. 3b). The third technique, superimposition and drawing of several rings, consists of the superimposition of several, doughnut or wheel shaped, rings of clay to build the walls of the vessels (Fig. 4). As with the drawing of a ring technique, when the upper part of the vessel is finished, there are variations in the way the bottom is made by drawing excess clay. This variability is related to the shape and number of rings, and the extent to which the rings are drawn upward to shape the body of the vessel. However, all the potters using this method, except one, drew the clay to form the bottom of the rough-out after the upper part of the vessel was finished. Thus these roughing-out methods differ by the shape of assembled pieces, the mode of clay deformation, as well as the sequence of operations used to build the vessels.
Figure 2 - Cylindrical coiling
A series of coils are superimposed to build the vessel walls. A slab is crushed in the centre of the rough-out to form the bottom.
© A. Livingstone-Smith
Figure 3 - Drawing of a single ring
(A) A doughnut shaped ring of clay is drawn upwards to build the walls, shoulder and neck of the vessel. (B) When upper part of the pottery is finished, the rough-out is turned upside down and the excess clay is drawn to build the bottom of the vessel.
© A. Livingstone-Smith
Figure 4 - Drawing of superimposed rings
Several, doughnut or wheel shaped, rings of clay are superimposed to build the walls of the vessel. The bottom of the rough-out is made by drawing the excess clay, when the upper part of the vessel is finished.
© A. Livingstone-Smith
14These techniques are distributed in three technical zones from north to south (Fig. 5).
Figure 5 - Schematic distribution of the three technical zones
© A. Livingstone-Smith
15Cylindrical coiling is only performed in the northern part of the study area (from Kabongo to the Upemba depression) or by artisans who moved out of this area. Most of these artisans speak Kiluba.
16Potters using the superimposition and drawing of several rings live in the central part of the study area (south of the Upemba depression and down to the area of Lubumbashi). They are mainly Sanga, Sumbu, Nwenshi and Lamba, but the technique is also used by three Luba potters who moved into this area.
- 4 A single occurrence of this technique was also observed further north in the area of Fungurume. In (...)
17Finally, the drawing of a single ring technique is used by potters in the southern part of the study area, just north of Lubumbashi. The three Bemba potters in question are settled there, but were born in Zambia4.
18The linguistic data are still being processed, but a linguistic boundary is also evident in the study area.
19Bostoen (2005) has highlighted the existence of two main terms used for pottery clay. The stems °-bʊ́mbà and °-mà, derive respectively from the proto-Bantu stems *-bʊ́mb- and *-mà-, “to make pottery”. As deverbal nouns (or nouns derived from verb stems), they can literally be translated as “material used to make pots”. In spite of this historical link, the verbs and nouns have to be considered as totally independent lexemes (or lexical units), because the distribution of the names for pottery clay diverge considerably from the distribution of the verbs they are derived from (Bostoen 2005:160-161). As regards our recent fieldwork, all the potters we interviewed in the field use a reflex of the stem *-bʊ́mb- “to make pottery”, while reflexes of both °-bʊ́mbà and °-mà are found to designate the “pottery clay”.
20Spatially, the reflexes of the stem °-mà are mainly found in the northern part of the study area in the Kiluba, Kisanga, Kinwenshi, Kisumbu and Kilomotwa languages. In contrast, reflexes of the stem °-bʊ́mbà are found only in the southern part of the study area used by potters speaking Cibemba and Kilamba only.
21Considering the correlation between technical behaviours and technical vocabulary, it appears that linguistic areas fit partially with technical areas. People using cylindrical coiling in the north use the reflexes of the stem °- mà for clay, while people in the south east of the study area use the drawing of a single ring technique and reflexes of the stem °- bʊ́mbà. People using the superimposition and drawing of several rings in the middle of the study area are split linguistically in two sub-zones. In the northern one, potters use reflexes of the stem °- mà, while in the southern one potters use the reflexes of the stem °- bʊ́mba. Thus, in other words, this central area (where people use the superimposition and drawing of several rings) is split into two linguistic domains.
22The context of pottery production in the area provides clues to understand the distribution of pottery techniques and vocabularies.
- 5 Mr. Bernard M. K. of Lubumbashi (30/09/2006) is the only one to have described another technique fo (...)
23Most of the potters interviewed were elderly women (between 40 and 80 years old) living in rural areas, although a few live in Lubumbashi. Pottery is generally a secondary activity, to say the least, as most of their income is provided by farming. Only a few vessel forms are still being produced on demand. Artisans that make pottery for the market are extremely rare. The most common products are small pots for preparing medicines. Water storing jars are less commonly made. In general, most of the potters, except in Kabanga Banga, are isolated and rarely know of the work of other potters5. In some cases, however, potters may know about fellow artisans in the area. For instance, artisans in Kabanga Banga (north of Likasi) told us about potters working in Kitana, reputedly a great centre for pottery. When we went there, we met a single potter who was over eighty years old. According to an informant in Lubuko near Mwadingusha, there are also potters near the Catholic Mission of Lukafu (some 25 km from where she lives). When looking for potters south of Lubumbashi we were told about a pottery centre, a village called “Village 41”, on the road to Kasenga some 60 km away. Again, when we visited that village, we met a single potter. Thus, today, except for potters who actually worked in the same village, the artisans’ knowledge on other traditions is sketchy. Finally, a few potters may know of other artisans working in very distant areas where they used to live (up to 400 km away), but from which they recently moved due to insecurity or for economic reasons (this is the case for Luba potters in the area of Lubumbashi).
- 6 She reports that she started making pots after her mother visited her in her dreams (Lubuko, 5/10/2 (...)
24While most artisans claim that they were taught by their mother or a close family member (arguing that pottery is a family craft), a number of artisans claim to be self taught. In most cases they say that they copied someone and/or that they started making pottery after their tutor passed away. For instance, Mme Ilunga W. of Lubumbashi (Lubumbashi, 15/09/2006) says she was taught by her aunt and her elder sister in Malemba Nkulu, but she only started making pots when she was in Lubumbashi. In a similar manner, Mme Elizabeth O. of Lubuko says that she started making pots in 1996, years after her mother had passed away6. This situation may explain the relative degree of variability in pot forming practices.
25A detailed study of marital practices and kinship systems remains to be done in Katanga, but most informants say that there use to be a significant difference between southern populations (i.e. Sanga, Sumbu, Nwenshi, Lamba, Bemba, etc.) and northern populations (Luba). The Sumbu, Sanga, Bemba and Taabwa are matrilineal and matrilocal populations favouring cross-cousin marriage. The Luba are said to be patrilineal and patrilocal, and do not practice cross-cousin marriage. This is true for the central Luba, but not for oriental Luba – they are said to follow matrilineal kinship systems (see Petit 1993:239-241; 2000:20). Most of the informants state that the distinction between matrilineal and patrilineal populations is disappearing as the law of the state is taking precedence over custom. Apparently, marriages are now possible between matrilineal and patrilineal populations, provided that arrangements are made before the wedding to avoid subsequent conflicts. In the past, however, these marriage and residence practices may have played a role in maintaining specific spatial distributions of pottery techniques.
26The spatial distribution of pottery building techniques in Southern and Central Katanga materialises three technical provinces: cylindrical coiling + slab bottom in the north; drawing of several superimposed rings + drawn bottom in the middle and drawing of a single ring + drawn bottom in the south. A first appraisal of the data concerning the technical vocabulary shows that there is a partial fit between production behaviour and vocabulary. The northern and southern technical provinces are distinct from a linguistic point of view, while the middle technical province is further divided into two zones, each following the vocabulary in the adjacent technical province.
27In terms of technology, this specific spatial distribution seems to be due, as in other cases on the African continent, to a multilayered set of social boundaries limiting the movement of women. The data collected on the social context of production show that part of the variability may be due to individual factors. The potters are rather old and isolated and several artisans claim to have learned “just by looking”. The interaction between “master” and “apprentice” is thus often reduced to simple observation. As most of the potters claim to have started practicing late in their thirties or forties for economic reasons, there can be a rather long period of time between the moment when a potter learns how to make pottery and the time when she actually starts making it. This temporal distance between observation and practice must surely be a cause of technical drifts. What is transmitted is the general idea of a technique, not a strict template. Such a situation may explain the variation observed in the way to perform the drawing of several superimposed rings. Variation may also be facilitated by the absence of communities of practice, as most artisans are isolated and rarely know of other potters. Thus, one could argue that the technical distinctions observed today may only reflect personal variations around a “technical” theme, particularly with the drawing of several superimposed rings technique. However, three points indicate that this situation is related to a larger scale social phenomenon.
28Firstly, the distinction involves the shape of assembled pieces, the modes of deformation and the sequence of operations. In other words the distinction is based on a complex set of operations, a method (sensu Roux 1994), not just on a technique. Secondly, old photographs from the twenties and forties show potters performing the superimposition of rings in the exact manner observed today (Cabu s.d.). Thirdly, some of the building methods are also used outside of the study area, from the Great Lake region to the coast of Angola (see map in Gosselain 2000:202). Thus, technical variations in central and southern Katanga today are not an epiphenomenon related to the choices of isolated individuals.
29From a linguistic point of view further analyses will examine whether the technical boundaries correspond to differences in pottery lexicon and more general linguistic variation (for instance closeness of languages spoken by potters who use the same or different techniques).
30In conclusion, the technical variations observed in the study area are complex, they are rooted in time and they cover a wide, but discreet, geographical area. Clearly, these variations are not due to the whim of old potters, but to traditions! In fact, fieldwork enquiries also indicate that these boundaries involve not only pottery practices, but also marital strategies, kinship systems and even ritual practices (such as circumcision). A key issue in the future of the project will be to understand the mechanism underlying the convergence of such a series of distinct cultural traits, often belonging to distinct spheres of activity. At this stage, we contend that the two great kingdoms in the area (the Luba Kingdom and the Bemba Kingdom), played a fundamental role in generating the clear cut social spaces – pottery practices, language, marital strategies and kinship systems – observed in Katanga.
31The data recently collected in the field are not yet fully processed, but there is no doubt that this case study will provide new insights on the history of south central Africa in particular and on the spatial distribution of material and non-material culture in general.
This research was funded by the Royal Museum for Central Africa (Tervuren, Belgium), thanks to an “Action I” project of the Belgian Federal Science Policy Office. The Musée National de Lubumbashi (MNL) provided administrative support and the director and employees of the MNL did their best to facilitate our work. Thanks to Donatien Muya Wa Bitanko, Henri Bundjoko, Désiré Kapata, Philippe Mikobi, Nicole Matanga Sapato, Jean Maisa. Our thanks also go to Michel Hasson, Vanessa Anastassiou, Père Léon Verbeek, Chief N’Kubitue of Mulungwishi, Jerry Kalonji, Sergio Castro, Pierre Manda Mundje, Kent Fowler and Nic David, as well as Sophie de Ville and Laetitia Dujardin. Finally, last but not least, we would like to thank Els Cornelissen and Jacky Maniacky for their support!