“The Cloth from Bondoukou: textiles of the Ghana/ Ivory Coast border”
full text of article by Duncan Clarke from “Hali” magazine #157 Autumn 2008
The strip woven “kente” cloths of the Asante and Ewe are among the most well-known of African textile traditions1 . Uniquely in sub-Saharan Africa worldwide interest in these dramatic fabrics has stimulated the emergence in Accra of specialist textile dealers serving the export market for both new and vintage local textiles. In turn the presence of these dealers and the supply networks they have established has begun to reveal a number of types of fabric that had previously escaped the notice of scholars and collectors. This article takes a preliminary look at a group of exceptionally interesting cloths collected on the Ivoirian side of Ghana’s western border that have only recently begun to be available. The unsettled security situation in Côte D’Ivoire in the aftermath of the recent civil war has as yet precluded my undertaking any field research into these cloths so this article will confine itself to a consideration of the pieces themselves and the light that can be shed on them by the existing literature. At first glance many of them could be mistaken for unusual variants within the Ewe traditions that flourished in the Volta region of eastern Ghana but closer scrutiny reveals instead a group of cloths that combine affinities with the main Ghanaian “kente” styles with other aspects closer to certain forms of Baule and Gouro textiles from central Côte D’Ivoire2. The cloths were mostly collected and brought to Accra by traders living in Bondoukou 23, a town close to the Ghana border that historically was a key entrepôt on trade routes linking the Akan kingdoms of Ghana with Sudanic peoples to the northwest. It is now clear that at least by the early part of the twentieth century, and probably considerably earlier, weavers in the Bondoukou region were producing textiles of a complexity and sophistication that rivalled those of the Asante and Ewe.
Asante “kente” cloth is perhaps today the best-known product of the narrow strip double-heddle loom used by male weavers across most of West Africa4. Woven primarily in Bonwire and a few other specialist weaving villages close to the Asante capital of Kumase, elaborate silk (and subsequently rayon) cloths were at the apex of a complex hierarchy of court and prestige dress throughout the Asante empire and beyond. The main feature of Asante kente is the framing of geometric motifs woven using a supplementary floating weft between weft faced bands called “babadua” on a warp faced, warp striped, background. Weft faced areas are often aligned against warp faced areas of adjacent strips creating a regular “chequerboard” layout for the completed cloth. Distinct borders at each end of the cloth, created by a more dense and generally more regular alignment of weft faced areas are a feature of most but not all Asante kente. Ewe weavers use the same techniques but their output can be distinguished by a preference for cotton over silk/rayon, a different range of colours, the use of figurative as well as geometric motifs and other minor differences. The Ewe tradition also encompassed more varied styles including a greater variety of strip widths, entirely weft faced cloths, and a number of as yet mainly undocumented local styles. Until the middle years of the twentieth century both the Asante and Ewe also wove simpler striped and checked fabrics from local white and indigo dyed cotton5. Overall one could say that there is a distinct sensibility apparent to each tradition, a different grammar that shapes the overlapping technical vocabulary and allows a viewer with a familiarity with both groups of textiles to distinguish fairly easily between them in almost all cases.
Bondoukou has been cited in the literature on Ghanaian textile history as the source of simple blue and white warp striped fabrics called kyekye that were highly prized by the Asante and regarded as an input into the early stages of the development of the Asante kente tradition6. Bondoukou was the principal trading town of a kingdom called Gyaman established in 1690 by Tan Date an Akan warlord who lead a body of migrants from Akwamu north of present day Accra and consolidated various Akan warriors and their followers into a group that became known as Abron (or across the current border in Ghana as Brong.) For reasons that will become clear later I will refer to the fabrics to be considered here as “Bondoukou cloths” rather than Abron. They are representative of a third, previously unpublished, tradition which shares much of the vocabulary of techniques found in Asante and Ewe textiles but combines them in different and unfamiliar ways – a different grammar to extend the metaphor used above. Like Asante and Ewe cloths these fabrics have weft-faced bands on a warp faced background. They have geometric and in some cases figurative supplementary weft float motifs. Some have distinct border areas at each end of the cloth. Some have the familiar “chequerboard” type layout in which patterned areas and plain warp faced areas are juxtaposed on adjacent strips across the cloth. As with the Asante and Ewe there are smaller women’s wrapper cloths worn singly or in pairs as well as the large cloths shown here that are worn toga fashion by men.
Set against these similarities there are significant differences. The thread used for the warp striped background on Bondoukou cloths is almost always white and indigo dyed hand spun local cotton, whereas virtually all Asante and Ewe collected in the C20th have been woven from thinner machine spun thread in a much wider range of colours. The cloth strips used on most Bondoukou cloths have a narrow white warp stripe at each selvedge where the strips are sewn together – a feature that is shared with many Baule and Gouro blue and white cloths from central Côte D’Ivoire but that is not typically found in Ghana. Many of the geometric motifs in supplementary weft float differ in scale and pattern from those used by the Asante and Ewe and also have some affinity with central and northern Ivoirian designs. Figurative motifs are quite infrequently used and a majority of those seen to date represent lizards or crocodiles. In many cases the layout of weft-faced bands on the cloth differs from that used in Ghana. On some cloths,1,4, pattern blocks are aligned in rows across the fabric strips in a linear arrangement that visually cuts the cloth into sections with bilateral symmetry around a central focus. The patterns can range from simple weft stripes on a balanced plain weave ground, through weft faced blocks, to complex pattern blocks made up of both weft faced blocks and supplementary weft float motifs, as in the centre of the fabrics shown here. In contrast to this rigid linear layout, other cloths 3 have weft faced bands and in some cases a few supplementary weft float motifs, scattered across the fabric in a seemingly random “off-beat” layout which may be combined with a more regular border. In 3 the border has elaborate braided tassels. A few Asante cloths have similar tassels, a feature that Venice Lamb attributes to contacts with Côte D’Ivoire. In 7 the layout retains aspects of this more random patterning but moves close to the kind of regular alternation of pattern blocks between adjacent strips across the cloth that is typical of Ewe and Asante textiles in Ghana. We can note some figurative designs, apparently representing crocodiles or lizards, executed using the supplementary weft float technique in 1,7 and 8. The next two cloths 8and11 are very close in layout to the typical chequerboard design of many Ghanaian textiles, differing only in the float motifs and the use of hand-spun cotton. With the final cloth 12 we return to a linear alignment across the cloth and bilateral symmetry but combined with new features. Only two examples in this style are known to date, both said to have been collected in a village near Bondoukou called Sanpa. On a plain machine spun cotton ground the cloth has complex borders, a large centrally placed supplementary weft float motif and a distinctively decorated border strip7.
In the absence of detailed field research our comments on these cloths can only be preliminary. With this proviso in mind there are three external sources we can look to in an attempt to shed some light on the cultural background and origins of the textiles collected: references in accounts by early European traders on the West African coast; the late nineteenth and early twentieth century accounts of colonial administrators and ethnographers; and finally visual imagery in the form of photographs and postcards from the same period. References in early European sources demonstrate that the Bondoukou region was an important producer of textiles as early as the sixteenth century. European traders settled in small numbers on the West African coast from the late fifteenth century, building forts and trading posts known as “factories” which served as bulking centres for local produce and distribution points for European goods in support of the maritime trade. The earliest traders were Portuguese but by the seventeenth century the Dutch, English, French and others rivalled them. Although gold, ivory, and later slaves were the key products, both local made and imported textiles were also traded in substantial quantities. African made cloths could be bought cheaply at some points on the coast such as at Benin, and traded for other products, such as gold and slaves, at the Gold Coast forts or further south in Luanda. Much of this locally made cloth was woven quite far away from the coastal points where it was sold to Europeans – the traders and their exotic imported goods competed within and extended an already existing pattern of trade routes rather than initiating a radically new practice. Portuguese and Dutch cloth imports, for example, at times struggled to be competitive against African cloths brought from many miles in the interior. In 1510 a Portuguese captain at Elmina complained in a letter to the King “among the Negroes there is so much cotton cloth of Mandinga that it spoils much of the trade”8. Large numbers of simple and cheap blue and white cloths brought from inland provided the daily dress of the inhabitants of the coastal regions of the eastern part of what is now Côte D’Ivoire alongside locally made bark cloth and the cheaper coarser imports. Known as Quaqua (or Kwakwa) cloths they gave their name to what traders referred to as the Quaqua coast. Kea makes an interesting distinction between the dress of common people and the wider options of imported cloth from elsewhere in West Africa, Europe and even India available to the wealthy local traders, chiefs and others9. Among these were “Mandinga cloths and blankets.” The most detailed contemporary comment comes from an account by a Frenchman, Godfrey Loyer, who noted that at “Bighu,” a town many days march in inland, African traders bought Turkish carpets (“tapis de Turquie”) and fine cotton cloths with red and blue silk stripes (“fines étoffes de cotton rayées de soye rouge et bleue”)10. From other similar references to “carpets” it seems likely that these were precursors of the heavy wool blankets from the inland Niger Delta whose presence in the Akan courts of Ghana has been noted on many occasions, and that “Turks” can be read to mean Muslims. On the Quaqua coast, and as far to the east as Elmina and Cape Coast, the main African source of locally produced textiles at this period was Muslim traders of Mande origin (“Mandinga” or “Wangara”) who purchased the textiles mainly in the towns of “Bighu” and “Old Wankyi” some 200 miles in the interior. The significance of this for the present topic is that “Bighu” (also referred to as “Begho” or “Insoko”) was the precursor of the current town of Bondoukou. Garrard , in his study of the Akan gold trade, notes that “A Dutch map drawn in the coastal fort at Mouree in 1629 shows Begho (Insoko) in the far north and states that ‘they have very fine goods; cloths woven like carpets, which are worn among the Akanni”11. Groups of Mande people had migrated south from Jenne and elsewhere in Mali, probably beginning in the fourteenth century, establishing a chain of market towns on the northern edge of the forest belt. Access to southern gold fields was the primary motivation but trade in other products such as kola nuts, textiles, trans-Saharan goods, horses, and slaves also became significant. Begho became the main entrepôt for trade between the Muslim Mande and the Akan peoples of the forest and the Gold Coast, with an Akan quarter known as Nsoko developing. By 1700 following internal unrest its residents and their trading role had largely shifted to a new town of Bondoukou some 30 kilometres to the west.
Bondoukou became the trade centre for an emerging kingdom called Gyaman established around 1690 by the Gyamanhene Tan Date who lead a band of Akan that had originated in Akwamu and been expelled from the Kumase region in unrest that accompanied the emergence of the Asante empire in the late 17th century12. These warrior migrants known as Abron were able to establish control over the indigenous peoples of the region and other Akan immigrants and reach a mutually beneficial accommodation with the Muslim traders of Bondoukou. The royal court alternated with the kingship between two branches of the royal matrilineage based in small villages rather than in Bondoukou. The bulk of the population of the kingdom was made up of farming peoples, notably the Koulango and Nafana. The Asante attacked Gyaman in 1740 and again in 1818, after which it was obliged to pay a heavy annual tribute until the British in turn defeated the Asante in 1873. Samori Toure, a warrior chief engaged in a long struggle against the French, captured Bondoukou in 1895, and the Gyaman kingdom was divided into areas of French and British rule two years later. While textile production in the Bondoukou region has clearly been of some importance for several centuries the brief comments in early sources provide only hints of the appearance of the fabrics produced and offer little evidence of their resemblance, if any, to the early twentieth century pieces we illustrate here. There is however, one fascinating comment from the closing years of the nineteenth century that explicitly contrasts the appearance of Asante cloths with those woven further north. In an account of his journey through the Asante and Gyaman kingdoms published in 1898 the English official R.A. Freeman notes “The scheme of colour is as a rule extremely severe and restrained. While the gorgeous cloths of the Ashanti glow with the hues of the peacock’s plumage, the finest fabrics from the northern cities present a plain ground of creamy white or some quiet neutral shade, upon which a few spots of brilliant colour are introduced with great judgement and severity of taste, and with very fine effect.”
Turning to twentieth century ethnography, more light is shed on the multi-ethnic character of the region and of Bondoukou itself. Tauxier, in his epic monograph “Le noir de Bondoukou” (1921) noted that neither of the two major peoples of the Gyaman kingdom, the Koulango and the Abron, were resident in any numbers in Bondoukou town. Instead its population was predominantly Dioula, with small areas inhabited by Huela, Hausa, Senoufo, Nafana, Gbin, Gouro and Noumou. Dioula (or Dyula) is the modern term for the Muslim trading peoples of Malian origin referred to as Mandinga or Wangara in early sources13. The preferred occupations of Dioula families in pre-colonial times were as warriors, Islamic scholars, or long-distance traders but the Dioula were also the pre-eminent weavers of the northern part of Côte D’Ivoire and neighbouring areas, producing large quantities of cloth both for their own use and more significantly as a key commodity for local and long distance trade. In the process they disseminated knowledge of weaving to the Senoufo, Koulango and other farming peoples. Until its destruction by Samori in 1895, the Dioula town of Kong in north central Côte D’Ivoire was particularly known for its high quality weaving. In the twentieth century Dioula weaving for their own use as women’s ceremonial wrappers is characterised by features including warp ikat, supplementary weft float motifs, and the juxtaposing of cloth strips with alternating even blocks of weft colours in a balanced plain weave14. As Muslims, Dioula men’s dress took the form of tailored and embroidered gowns rather that the wrapped cloths derived from the Akan chiefly traditions we see here. However the Dioula wove cloth mainly to sell and were ready to adapt their output to meet different local demands, such as for funerary blankets among the Senoufo. So was it the Dioula who wove these cloths? The historian Terray noted “As for weaving it was done by Abron and especially by Dioula specialists. The former remained in their villages. The latter moved about with their looms or settled in Bondoukou”15. Tauxier on the other hand, mentions weaving primarily in relation to the rural Koulango: “Le tissage ne constitue pas un metier à part: ce sont les cultivateurs koulangos (pas tout, mais un certain nombre) que l’exercent”16. For what it may be worth some of the traders who bring these cloths to Accra have told me that they are woven by Koulango, which would at least seem to indicate that it is the Koulango who are the major weavers in the area today. However it would probably be a mistake to assume that a single ethnic group was responsible for the production of these cloths. Koulango farmers would have grown the cotton, their wives and daughters have spun the thread, while most indigo dyers in Bondoukou were said by Tauxier to be Hausa 17, many weavers were Koulango but others were Dioula or Abron. It was certainly Dioula traders who dominated the cloth trade in the region, but they would have bulked and distributed textiles from other weaving peoples and transported them south down long established trade routes to the chiefs and big men of the Akan courts of eastern Côte D’Ivoire who I would suggest were the main customers for these highly priced prestige textiles.
A limited body of evidence in support of this hypothesis is provided by a consideration of the quite small number of photographs and postcards from the area published in the nineteenth and early part of the twentieth century. Photographs taken in Bondoukou show the men wearing the tailored, and in some cases embroidered, robes that were typical Muslim male dress at the time 518 and can be contrasted with female dress of wrapped cloth. In images such as 6 there are numerous examples of the type of simple blue and white striped and checked cloths the Bondoukou traders have brought to Accra, and a few glimpses of more complex patterning. Note, for example the cloth worn by the woman seated at the upper left, where weft faced bands are apparent. As yet I have not located any early photographs of the Abron court that are sufficiently close up to make any comments on court dress. There are however some important images depicting rulers of other Akan courts in the border region to the south of Bondoukou, in particular those of the immediate southern neighbours of the Abron, known as the Agni. A king at the town of Agnibilekrou in the 1920 or 30s is shown 9 wearing a cloth with white stripes at the selvedge of the strips, weft faced bands and weft float motifs. King Boua Kouassi of the Agni kingdom of Indenie 10, whose court was at the town of Abengourou, is shown (Figure 12) wearing a classic Ewe cloth imported from Ghana, as is the man seated at the right and the standing figure at his shoulder. However the chief seated second from left wears an elaborately decorated Bondoukou cloth with weft faced bands, supplementary weft float motifs, and a decorated border. The cloth worn by the seated man at the extreme left is probably also from the Bondoukou region. What are the implications of the discovery of these cloths for the textile history of the wider region? In design terms they bridge a gap and reveal previously unsuspected visual continuities between the Asante and Ewe traditions of Ghana and the cloths woven by the Akan Baule and their Mande-related neighbours such as the Gouro in central Côte D’Ivoire. Moreover they come from a region that may well have been key to the early history of Ghanaian textile production because it marked the first point of contact between the forest dwelling Akan and Muslim traders from the ancient empire of Mali who were establishing trade routes far to the south in search primarily of gold. Ross has suggested that it was this contact around the sixteenth century that introduced narrow strip weaving into present day Ghana19. A frequently cited Asante tradition notes that a man called Ota Kraben went to Gyaman in the seventeenth century and bought back the first loom which he set up in Bonwire20. There has recently been a discussion in the literature on Ghanaian weaving as to whether the use of two pairs of heddles to alternate warp faced and weft faced areas on the cloth strip was developed by the Ewe or by the Asante21. The existence of a third body of cloths with the same alternation, particularly in a region of such historical significance, clearly raises important questions for this debate. Do the Bondoukou weavers also use two pairs of heddles?22 In the second half of the twentieth century, and perhaps earlier as well, some Dioula weavers from the Bondoukou region spent long periods weaving within Ghana for the local market 23. Are the design similarities between Bondoukou cloths and Asante and Ewe textiles attributable to a shared origin or a later effect of the weavers and their customers’ familiarity with textiles imported from Ghana, or as seems likely, a combination of both factors? Field research will be required before we can suggest answers to these questions.
I would like to thank Dr Kerstin Bauer for her comments on an earlier draft of this article.
1. V. Lamb, West African Weaving (London, 1975) D.Ross, Wrapped in Pride: Ghanaian Kente and African American Identity (Los Angeles, 1998)
2. S. Vogel, African Art, Western Eyes(New Haven, 1997), J.Etienne-Nugue & E.Laget, Artisanats traditionels, Côte d’Ivoire (Dakar, 1985), B.Sumberg, ‘Panther Skins and Loaves of Bread – the Tie-Dyed cloths of Oumé’, Hali 124, 2002; pp88-93; B. Sumberg, A History of Cloth Production and use in the Gouro Region of Côte d’Ivoire. PhD thesis, University of Minnesota, 2001.
3. Bondoukou is located at 8.0° N, 2.8°W. Note that this is not the same Bondoukou as is found by Google Earth. It is not clear how widely these traders have ranged in the surrounding areas in order to collect their cloths.
4. See J. Picton and J. Mack, African Textiles. Looms, Weaving and Design, London 1979.
5. These became increasingly rare in the C20th.
6. See Lamb, op.cit.; Ross, op.cit.
7. Possibly Sampa, a village on the Ghanaian side of the international border. The other cloth collected there, in the Metropolitan Museum, New York, has a yellow ground, more elaborate design, and several large figurative motifs. Letter of Manuel de Gois, Captain of Elmina to the King, 18 April 1510. Cited in T.F. Garrard “Akan Weights and the Gold Trade” (1980:26)
8. ‘Letter of Manuel de Gois, Captain of Elmina to the King’, 18 April 1510, cited in T.F.Garrard, Akan Weights and the Gold Trade, London 1980, p.26.
9. R.A. Kea “Settlements, Politics and Trade in the C17th Gold Coast.” Baltimore 1982, p.298.
10. G. Loyer “Relations du voyage du royaume d’Issiny” in “L’Etablisshment d’Issiny, 1687-1702” in P. Roussier ed., L’Etablissement d’Issiny, 1687-1702, Paris 1935.
11. Garrard op.cit., p.44
12. The history of Gyaman has been the focus of the French historian Emmanuel Terray. For a short summary see “Bondoukou in Gyaman” in K. Arnault & E. Dell, Bedu is my Lover, Five Stories about Bondoukou and Masquerading, Brighton 1997.
13. There is an extensive controversy in the literature about the precise origins, current reference etcetera of all these terms.
14. The Dioula are key figures in the history of West African weaving whose role has until recently received comparatively little attention in the literature on West African textiles. See K. Bauer, Kleidung und Kleidungspraktiken im Norden der Côte d’Ivoire, Berlin 2007 for the first extended study of Dioula weaving. Baeur based her research in Kong and carried out limited fieldwork in Bondoukou town. Her study focuses on mainstream Dioula cloth styles and use and does not consider the types of fabric illustrated here.
15, E. Terray, “The Political economy of the Abron Kingdom of Gyaman” in Institute of African Studies Research Review 12/1, Legon 1980 pp1-36.
16. L. Tauxier, Le noir de Bondoukou, Paris 1921 p.147
17. ibid, p.136. However R.A. Bravmann, Open Frontiers; The Mobility of Art in Black Africa, Seattle 1973, p.48 and Bauer, op.cit., note that the Dioula word for ‘dyer’ is the same as that for ‘Hausa’ and suggest that the Bondoukou dyers are Dioula.
18. The main early published source of photographs taken in Bondoukou is M. Monnier, Mission Binger – France Noir (Côte d’Ivoire et Soudan)” Paris 1894. Some images are reproduced in Bauer op. cit.
19. H. Cole and D. Ross, The Arts of Ghana, Los Angeles 1977 p. 38.
20. R. S Rattray, Religion and Art in Ashanti, Oxford 1927, p.220.
21. See M. Kraamer ‘Ghanaian Interweaving in the nineteenth century: A new perspective on Ewe and Asante Textile History’ in African Arts XXXIX (4) Winter 2006
22. Kerstin Bauer, who worked with Diouls weavers in Bondoukou and Abron or Koulango weavers in Kanguélé, a nearby town, did not see the use of a second set of heddles (personal communication, 2008.) However it is likely that weavers who produced these more complex cloth types in the past were familiar with the technique.
23. Kerstin Bauer, personal communication, 2008.