Navigation – Plan du site
Archéologie et sources

On Tobacco Pipes in Asante

Brian Vivian
p. 87-95


Les pipes en terre que l’on trouve, au Ghana, associées aux sites d’habitat du XVIIe au XXe siècle sont particulièrement intéressantes du point de vue esthétique et chronologique. Il y a une quarantaine d’année, Paul Ozanne démontrait la sensibilité chronologique de ces pipes de fabrication locale en développant un modèle de sériation qui est toujours utilisé comme méthode de datation des dépôts archéologiques sur tout le territoire ghanéen. Peu de chercheurs ont tenté d’affiner cette sériation en définissant des types régionaux. L’analyse de pipes provenant de la région achanti indique l’existence de types qui sont des marqueurs à la fois chronologiques et régionaux, et peut-être identitaires.

Haut de page

Texte intégral

Fig. 1 : Miscellaneous representational pipes

1That the pipes discussed here were used for smoking tobacco seems to be beyond doubt, for no such pipes have been found in a pre-European context. Rare are accounts of smoking herbs in pre-contact times (see Effah-Gyamfi 1985). Yet historic documents are full of descriptions that tell of village residents trading for and smoking tobacco and illustrations show chiefs smoking pipes are common place. An example is seen in Isert’s tale of meeting the ruling chiefs at Ada, where two individuals are depicted smoking long stemmed pipes while seated in the middle of a palaver (Isert in Winsness1992: 46). Chiefs and other dignitaries are also portrayed holding tobacco pipes in Bowdich’s famous artistic rendition of greeting the Asantehene in Kumasi (Bowdich 1966[1819]). These pipes are stylistically the same as those found in archaeological excavations.

2By the end of the nineteenth century tobacco pipes became an attractive cultural curio for travellers to collect (see Freeman 1967 [1898]), and numerous tobacco pipes were brought back for display in museum collections across Europe (see Coombes 1994: 153 – note that these are wrongly identified to be from East Africa in the text; rather they are clearly Asante pipes, Rohrer 1946, Valentin 1976). In the 1950s and early 60s archaeological investigations into the prehistory of Ghana were soon to describe and detail the many pipes found in excavation (Nunoo 1957, Shaw 1961, Shinnie and Ozanne 1962, Davies 1977). Researchers readily distin­guished between locally produced and imported pipes, and the stylistic variability in these finds lead to speculation on the origins and development of pipes in West Africa (Shaw 1960, Phillips 1983).

3An analysis of the assemblage from the well stratified site of Dawu lead Thurston Shaw to make the first attempt at defining a pipe typology and establishing a seriation for the site, although he noted “no such classification or development was immediately apparent” (Shaw 1961: 11). Not to be deterred by these results, Ozanne examined the large collections the Department of Archaeology and the National Museum of Ghana had acquired, numbering between fifteen hundred and two thousand pipes. Most of these were surface finds collected by resident farmers and school children from across Ghana. Focusing specifically on the finds from the Shai Hills and the surrounding Accra Plains, Ozanne (1966) defined five basic pipe forms on the basis of obvious morphological differences in the base, bowl and stem. Ozanne in turn relied on the association of these defined types with historically documented sites to establish a regional sequence. In this manner the first regional pipe seriation in Ghana was developed.

4Although it was developed specifically for the greater Accra region, Ozanne did allow that in time the results may be applied to less-known areas of Ghana as more data came to light. He confidently postulated that the other regions of the country were strongly influenced by the development of smoking pipes in the Accra area, particularly in the period circa 1660-1690, and that the practice of smoking was transferred inland from southeastern Ghana (Ozanne 1966: 48).

5Subsequently archaeological studies further inland in Ghana accepted the challenge, and applied Ozanne’s basic typology with varying success (Afeku 1976, Bellis 1972, Boachie-Ansah 1986, Effah-Gyamfi 1985, Shinnie and Kense 1989, York 1973).  In that many of these studies were situated on the northern margins of the forest in Brong-Ahafo Region meant the pipe typology could be examined on a regional basis. Most directly adopted Ozanne’s typology and the site specific assemblages under scrutiny were analyzed accordingly (Boachie-Ansah 1986a, Crossland 1989, Effah-Gymfi 1985, Shinnie and Kense 1989). Some found that the “diagnostic pipe fragments represent the five major smoking types already recognized in Ghana”, and did not refine the accepted typology any further (Shinnie and Kense 1989: 160). Others expressed reservations about the applicability of the Accra typology to assemblages from regions further inland, but found that the results more or less confirmed Ozzane’s findings (Afeku 1976, Crossland 1989, Boachie-Ansah 1986b, Effah-Gyamfi 1985).

6Although these studies brought questions to bear on the dating, accuracy and order of the seriation, less often did they attempt to provide a revised classification. Rather attention was in large focused on defining the geographic direction from which tobacco smoking was introduced from. Only Afeku (1976) and Effah-Gyamfi (1985) defined their own typologies which recognized specific regional variants different than those found in Accra. However these typological variants were loosely defined on the basis of surface collections and were never firmly fixed in time or space. More recent studies have shifted their attention toward Africans in the Atlantic world, and specific cultural historic questions regarding the applicability of the Accra pipe typology were not considered (Decorse 1989, 2001, Stahl 2001).

Tobacco Pipes in Asante

7Questions concerning the applicability of the Accra typology to those pipes found in Ashanti Region have largely been avoided, essentially due to the paucity of archaeological research in the region. For prior to the 1980s only the village sites of Ahinsan and Mampongtin had been excavated, and final reports on these were never properly completed (Davies 1977a, 1997b). This is not to suggest that there was nothing known of Ashanti pipes, for they are common in museum collections (Rohrer, E. 1946, Valentin 1976), and also attracted the attention of various colonial administrators (Rattray 1927, Bullwinkle 1949). A glance at these various finds shows them to be very different than those pipes in the Ozanne typology. This fact was not lost on Ozanne (1966: 2) who stated “the varieties [of pipes] found in Ashanti are still mostly incomprehensible”.

8Only recently have pipes from Ashanti region gained further attention (Vivian 1998). This is due to the Asante Archaeological Research Project, which under the direction of Peter Shinnie, has excavated four village sites in Amansie and Adanse Districts, south of Kumasi (Shinnie 1988, Shinnie and Vivian 1991; Vivian 1990, 1996). These excavations have resulted in the recovery of some seventy pipes and pipe fragments from stratified and well dated contexts. This archaeological assemblage has been augmented by another forty (+) pipes that were picked up by local inhabitants in Amansie District, and ten pipes collected from various locations around central Kumasi.

9A second assemblage of Ashanti pipes was also located at the Pitt-Rivers Museum. Bullwinkle collected these pipes from across Ashanti Region while in the employment of the colonial service in the 1930s. In later years he gave this collection and a detailed report on the pipes written to gain promotion in the colonial service to the Pitt-Rivers Museum (Bullwinkle 1949). Collectively, the pipes recovered by the Asante Archaeological Research Project in combination with the Bullwinkle collection constitute a single regional assemblage of well over two hundred pipes that facilitates the recognition of regional Ashanti pipe forms. Analysis of this assemblage allows for distinct pipe types to be defined and temporal affiliations to be assigned.

Representational Pipes ca. 1850 - 1930

10Representational Pipes (fig. 1) are perhaps the most distinctive Ashanti pipe there is, and are common place in Museum Collections throughout Europe and Ghana (Coombes 1994: 153; Rohrer 1946, Valentin 1976). It is Malcom McLeod (1981: 85) who actually coined the term ‘representational’ pipe, in reference to those found in the British Museum. Historical references to this type of pipe are widespread, and mention of representational forms is found in Rattray (1927: 301) and Freeman (1967 [1898]: 124).

11Representational Pipes are easily identified by the sculpted bowl, which may take on any one of various anthropomorphic, zoomorphic or geometric forms (hence the name of the pipe type). As with most of the Asante visual arts, these forms are directly associated with proverbs (Rohrer, E. 1946, Bullwinkle 1949). The stem is typically attached to the bowl at an acute angle (35-550) and is generally round, sometimes square, and is always highly decorated with incised geometric patterns. The bowl is always presented sitting on a footed base or plateform. On occasion this plateform may be flat (as seen in Rattray 1927: Fig. 253), but more often than not it is a triangular shape (as seen in profile) that emulates the base of an Akonfromfi chair (see McLeod 1981: 120). References can be found to this type of pipe being caste in metal, but most are sculpted from clay.

12Although rarely found in an archaeological context, representational pipe forms appear to have existed for much of the 19th century, as Thomas Bowdich (1821) was presented with one during his visit to Kumasi in 1817. Although probably not common to begin with, by the later decades of the nineteenth century they appeared to become increasingly popular and widespread, as is exemplified by their common presence in European Museum collections from this period (Rohrer 1946, Valentin 1976). By the late 1930’s Bullwinkle tells of meeting individuals who knew how to make such pipes, but says that this style had generally fallen into disuse by the 1920’s (a period when pipe smoking in general declined in popularity).

13The geographic distribution of these pipes appears to be largely restricted to Ashanti Region. Despite the many excavations that have taken place throughout Brong-Ahafo Region none have been reported from here, or from further north. Nor have any been reported from excavations from around Volta Lake. The only reported find from the coastal regions is at Elmina, where a snail shaped pipe found was found in contact with a floor dating to 1873 (Decorse 1989: 174).

Plateform Pipes ca. 1800 - 1870

14In overall form, Plateform Pipes (fig. 2) are near identical to and can be seen as a precursor of the more elaborate Representional Pipes mentioned above. Plateform Pipes are distinguished by the lack of representational or figurative forms. Rather the bowls are sculpted into simpler saucer, oval or barrel shapes. Nonetheless the base of the pipes always rests on a triangular shaped ‘Akonfromfi chair’ base, the same as that described above. The bowl is usually elaborately decorated with cross hatch impressions and often features small appliqué knobs or bosses that are visually depictions of cowrie shells.

Fig. 2 : Miscellaneous plateform pipes

15Although these pipes are less often found in the ethnographic collections made at the end of the nineteenth or early twentieth century, they are common in surface collections such as that Bullwinkle made, and frequently found in archaeological excavations of 19th century settlements (Shinnie 1988, Shinnie and Vivian 1991; Vivian 1990, 1996). In fact these pipes are so common they can be considered a stratigraphic marker of this time period.

16As with Representational Pipes, rarely have Plateform Pipes been found beyond the limits of Ashanti Region. Effah-Gyamfi (1985: 182) reports on five triangular base pipes found in the area of Bono Manso which appear to fit into this category (based on the single one illustrated). He notes that these were either surface finds or recovered from the later occupation levels of Amuowi, and none were found in the earlier occupations identified at the site of Bone Manso. No other examples of Plateform Pipes have been reported of from excavations north of this, nor do I know of any examples of like pipes documented from the coastal regions.

Footed Pipes ca. 1720- 1830

17Footed Pipes share many attributes in common with the previous pipe types described above. The bowl and stem are shaped in the same manner as Plateform Pipes and decorated in much the same way. However, these are clearly distinguished by the base, which is characterized by a lower flat base with two small triangular protrusions at the front of the pipe. These protrusions or ‘feet’ on the pipe are suggestive of the footrests that are commonly used by Asante chiefs. While the bowl and stem is decorated in much the same fashion as Plateform Pipes, more often than not the bowl is round (as opposed to the saucer and barrel shapes that become common with Plateform Pipes).

18Although yet to be reported on in full, excavations at Asantemanso, Anyinam and Esiease found examples of Footed Pipes to be common place in eighteenth and early nineteenth century deposits (Shinnie and Vivian 1991). Davies (1977) documented several other examples at Mampongtin, all found in the upper levels of the midden. In his analysis of pipes from Begho, Afeku reports on one pipe that fits within this type category, which he refers to as Type 5b (see Afeku 1976: 51). This pipe he correctly relates to those Davies found at Mampongtin and another Ameyaw found on the coast, near Acheampong. Ozanne (1966: 50) provided a more complete description of the pipe Ameyaw found and recognized it to be Asante in origin. He went on to conclude that the introduction of this pipe form to the Accra area dated to sometime between 1724 and 1730.


19To conclude, as Ozanne had surmised tobacco pipes from Ashanti can be placed in a chronological order. The three types identified above are all associated with the later centuries, and in excavations of village middens there is much overlap in their occurrence in the mixed cultural deposits associated with this time period. However, these types are readily distinguished by the basal morphology and other distinguishing features. Where as all three types where readily recognized as Asante by Ozanne and others in more recent years, the development from a footed base, to a triangular plateform base and the of transformation of the bowl from simple round forms to elaborate sculpted forms shows a continuum in the development of Asante pipes over the last two centuries. A lenticular graph would illustrate the development sequence of these pipe forms to closely follow the time periods provided. That these pipes are Asante seems certain, given the lack of like pipes found in other excavations throughout Ghana. Considering these to be a distinctive marker of Asante culture, these pipe types provide powerful index fossils that are indicative of time and may inform on the level of cultural influence the Asante Kingdom exerted on neighbouring regions.

Haut de page


AFEKU, I., 1976, A Study of Smoking Pipes from Begho, Legon, B.A. Honours Thesis, Departement of Archaeology.

BELLIS, J., 1972, Archaeology and the Culture History of the Akan of Ghana, Ph.D., Departement of Anthropology, Indiana University.

BOACHIE-ANSAH , J., 1986a, An Archaeological Contribution to the History of Wenchi, African Occasional Papers No.3, Calgary, University of Calgary Press.

—, 1986b, “Smoking Pipes and the Dating of Post 16th Century in Ghana: the evidence from Ahwene Koko”, West African Journal of Archaeology, 16, pp. 53-70

BOWDICH, T., 1966 [1819], Mission from Cape Coast to Ashantee, London, Frank Cass and Co.

—, 1821, An Essay on the superstitions, customs an arts, common to the ancient Egyptians, Abyssinians and Ashantees, Paris.

BULLWINKLE, L., 1949, An Essay on Ashanti Pipe Bowls and Tobacco, manuscript, Pitt-Rivers Museum, Oxford.

COOMBES, A., 1994, Reinventing Africa: Museums, Material Culture and Popular Imagination, New Haven, Yale University Press.

CROSSLAND, L., 1989, Pottery from the Begho-B2 Site, Ghana, African Occasional Papers No.4, Calgary, University of Calgary Press.

DAVIES, O., 1977, Excavations at Mampongtim, 1955: a corpus of early eighteenth century Ashanti pottery, Pietermaritzburg.

DECORSE, C., 1989, An Archaeological Study of Elmina, Ghana: trade and Culture Change on the Gold Coast between the Fifteenth and Nineteenth Centuries, Ph.D., University of California, Los Angeles.

—, 2001, An Archaeology of Elmina, Washington, Smithsonian Institution Press.

DUPUIS, J., 1966 [1824], Journal of a Residence in Ashantee, London, Frank Cass and Co.

EFFAH-GYAMFI, K., 1985, Bono Manso: An Archaeological Investigation into Early Akan Urbanism, African Occasional Papers No.2, Calgary, University of Calgary.

FREEMAN, R; A., 1967 [1898], Travels and Life in Ashanti and Jaman, Londres, Frank Cass and Co.

NUNOO, R.B., 1957, “Excavations at Asebu in the Gold Coast”, The Journal of West Africa Science Association 3, pp. 12-47.

OZANNE, P., 1962, Tobacco-Pipes of Accra and Shai Hills, unpublished manuscript, Legon, Department of Archaeology, University of Ghana.

PHILIPS, J., 1983, African Smoking and Pipes, Journal of African History 24, pp. 303-319.

RATTRAY, R., 1927, Religion and Art in Ashanti, Oxford, Clarendon Press.

ROHRER, Von E., 1946, “Tabakpfeifenköpje und Sprichwörter der Asante”, Jahrbuch des Bernischen Historischen Museums 26, pp. 104-121.

SHAW, T., 1960, “Early Smoking Pipes: in Africa, Europe and America”, Journal of the Royal Anthropological Institute XC, ii, pp. 272-305.

—, 1961, Excavation at Dawu, Legon, University College of Ghana.

SHINNIE, P., 1988, “Excavations at Asantemanso, 1987”, in Nyame Akuma 30, pp. 11-12.

SHINNIE, P. et F. KENSE, 1989, Archaeology of Gonja, Ghana: Excavations at Daboya, Calgary, University of Calgary Press.

SHINNIE, P. et P. OZANNE, 1961, “Excavations at Yendi Dabari”, Transactions of the Historical Society of Ghana 6, pp. 87-118.

SHINNIE, P. et B. VIVIAN, 1991, “Asante Research Project”, Nyame Akuma 36, pp. 2-6.

STAHL, A., 2001, Making History in Banda, Cambridge, Cambridge University Press.

WINSNESS, S. (éd.), Letters on West Africa and the Slave Trade. Paul Erdmann Isert’s Journey to Guinea and the Caribbean Islands in Columbia (1788), Oxford, Oxford University Press.

VALENTIN, P., 1976, “Les pipes en terre des Ashanti”, Arts d'Afrique Noire 19, pp. 8-13.

VIVIAN, B., 1998, Putting Pipes in Their Place, contribution non publiée, 14th Biennial Meeting of the Society of Africanist Archaeologists, Syracuse, NY.

—, 1996, “Recent Excavations of Adansemanso”, Nyame Akuma 46, pp. 37-42.

—, 1990, “Origins of Asante Research Project: 1989-90”, Nyame Akuma, 34, pp. 19-22.

YORK, R., 1973, “Excavations at New Buipe”, West African Journal of Archaeology 3, pp. 1-189.

Haut de page

Table des illustrations

Légende Fig. 1 : Miscellaneous representational pipes
Fichier image/jpeg, 28k
Légende Fig. 2 : Miscellaneous plateform pipes
Fichier image/jpeg, 29k
Haut de page

Pour citer cet article

Référence papier

Brian Vivian, « On Tobacco Pipes in Asante », Journal des africanistes, 75-2 | 2005, 87-95.

Référence électronique

Brian Vivian, « On Tobacco Pipes in Asante », Journal des africanistes [En ligne], 75-2 | 2005, mis en ligne le 15 décembre 2008, consulté le 30 mars 2017. URL :

Haut de page


Brian Vivian

Department of Archaeology, University of Calgary, Canada.

Haut de page

Droits d’auteur

Société des africanistes

Haut de page
  • Logo Société des africanistes