1
2	‘Austronesian’ staple food crops in Africa Until the end of the 1990s, unravelling the history of vegetative staple crops in Africa had to be pursued by indirect methods, notably historical linguistics and the study of somatic variation, as no archaeobotanical material was available. Evidence from phytoliths is now available from Cameroun (Mbida et al. 2000, 2001) There has been controversy over the Cameroun date (see e.g. Vansina 2004; Mbida et al. 2005) but they are very much in line with conclusions drawn by researchers, especially linguists, since the 1930s using other lines of evidence.
3	"There are at least two..." There are at least two studies of the vernacular names of bananas and plantains in Africa, Blakney (1963) and Rossel (1989, 1991, 1996, 1998), but their conclusions cannot easily be aligned with the phytolith evidence. This paper attempts to draw together and reformulate the linguistic and culture geographical evidence, as well as considering the possibility that plantains –and perhaps bananas as well- were accompanied by other vegetative crops.
4	Murdock’s hypothesis: the ‘Tropical Food Kit’ Murdock (1959) was the first author who pointed to the historical enigma presented by SE Asian food crops in Africa. At the period when Austronesian navigators were presumably reaching the East African coast (before 2000 BP), its only inhabitants would have been Cushite pastoralists, and Khoesan-related groups with a hunting-gathering economy. Neither of these are likely candidates for the transmission of vegetatively reproducing crops requiring elaborate agricultural skills. Murdock’s answer to this was to postulate a ‘Yam Belt’, a corridor with its Easternmost tip in Southern Somalia, passing North of the Equatorial forest, as far as the Kru and other coastal tuber-growers in the West of West Africa. Murdock’s candidates for the adoption and transmission of these cultigens were a people he calls ‘Megalithic Cushites’, then said to be inhabiting the Highlands of Southern Ethiopia.
5	Murdock’s hypothesis: the ‘Tropical Food Kit’ The cultigens are the cultivated Musaceae, taro or ‘old’ cocoyam (Colocasia esculenta) and the water-yam (Dioscorea alata). These three crops seem to have been well established in West Africa by the time of the first European contacts with the coast. It was proposed that they diffused across the centre of the continent via the Central African rain-forest. Simmonds (1962:137; 1976:213) confidently shows a thick black arrow sweeping across the centre of the continent from East to West schematically representing the diffusion of plantains and bananas. Nevertheless, exactly how and when elements of the ‘tropical food kit’ were introduced to West Africa remains problematic
6	Linguistic and cultural evidence A study by Blakney (1963) listed and grouped the vernacular terms across the continent. Blakney found that the two principal word-stems #-ko and #-to were widespread. Unfortunately, the data that Blakney used failed to consistently distinguish between plantain and banana, and as he seems to have been unaware of their very different distributions, he failed to match any of his widespread roots with either type. Blakney concluded from the broad dispersal of the root #-ko must indicate that it formed part of the core vocabulary of the Niger-Congo language phylum. This is now an extremely problematic assumption. Other authors (e.g. Vansina 1990) argued for an early date for the banana in the equatorial rainforest on the basis on linguistics, although without setting out the evidence in detail. Rossel (1998) studied the vernacular terminology of plantain and banana in the entire continent. Her studies accumulate much fresh data, but reach the rather idiosyncratic conclusion that ‘A westward spread of musa (from Asia) began only in Islamic times and reached Africa not long afterwards’ (Rossel 1998: 52).
7	Enset names and the #-kom root The only wild Musaceae species indigenous to West Africa is Ensete gillettii, an enset with an inedible fruit found in rocky areas across West Africa used mainly for magical purposes or as a famine food. Names for this plant in West-Central Nigeria incorporate the root #-kom and it is likely that this term that can be reconstructed back to proto-Benue-Congo. It has also been borrowed into the unrelated but intertwined Chadic languages. At some point, this name has been transferred either to plantain or to the cultivated Musaceae in general
8	Plantain names and the #-kondo root The vernacular terms that have stimulated the expenditure of the most scholarly ink are those related to proto-Bantu #ko[n]do. These are embedded in languages in NW Bantu, but also appear in Mande and Atlantic languages in the Guinea-Liberia region. It is assumed that the occurrences in Mande, Atlantic. Kwa and Gur languages are all borrowings from Bantu and that this must have occurred as a result of late Portuguese transfers of crops along the coast.
9	Banana names and the #banana root The table compiles the West African names that appear to be cognate with English ‘banana’. In the Mande languages, many of these are contracted and compressed, which speaks of some antiquity. It has been suggested that the source of this word is Indian vannan < purported Sanskrit varana (Blakney 1963:77). However, this is not confirmed by the relevant dictionaries; the nearest form is Sanskrit vanakadalii (वनकदल). Garcia da Orta (1563) mentions palana on the Malabar Coast, and this does look like a convincing source for the Mande names. The Portuguese may have picked up this name in India and carried the small diploids to West Africa, along with the Asian name. This became fixed in English as ‘banana’ and then was borrowed back into Camerounian languages in the colonial era. From Cameroun it spread into the interior, surfacing in Nilo-Saharan languages as an indirect loanword.
10	Musa names and the #boro root The #boro root has a curiously disjunct distribution. There are scattered occurrences as far apart as Sierra Leone and Kenya occurring in very different language families. This is another name spread by the Portuguese, as many (though not all) of its attestations are coastal. However, its origin is unknown and there are no early textual references to this term. It could well be associated with the introduction of a particular cultivated subtype, although there is no clear evidence for this
11	Related material culture: musical instruments There are some items of material culture related to plantains that seem to be related to their diversification in the Bight of Biafra area. A musical instrument connected with plantains has a distribution suggesting an origin in this zone. The second is the plantain-stem xylophone, the distribution of which maps very approximately against plantain diversity. The wooden bars of the xylophone are laid transversely across fresh Musa stems. No analogous instrument is reported from Indonesia, suggesting that the instrument evolved subsequent to the introduction of the plantain. This xylophone is today found in areas where the banana is the staple, but the map suggests very strongly that West Central Africa is its original nucleus of distribution.
12	The plantain-stem xylophone
13	Distribution of the plantain-stem xylophone
14	Other vegetative crops: water-yam and taro The water-yam, Dioscorea alata, another cultigen of SE Asian origin, is cultivated throughout West Africa and sporadically in East Africa and Ethiopia, as well as on Madagascar. The water-yam has a long dormancy period (Martin 1976), a feature that makes it an ideal plant to transport on long ocean voyages, as it avoids the necessity of keeping a plant alive while en route. This must have been an important factor in its choice as a major staple in Oceania. Chevalier (1936:522 ff.) concluded that the water-yam was long-established in West Africa, although he offers no hypothesis about the route of its introduction. Timitimi (1970) shows that the Kolokuma recognise eighteen cvs. of D. alata while Raponda-Walker & Sillans (1961:150) list three major subgroups and numerous other varieties grown in Gabon. If this is compared with other tubers introduced by the Portuguese, such as the fertile and easily bred sweet potato, such a shallow time-depth seems unlikely.
15	Taro Taro seems to be of an importance similar to the water-yam in the Bight of Bonny area. Knipscheer and Wilson (1980) map the cultivation of cocoyams in SE Nigeria, and shows that in some areas their importance is that of a co-staple. Lyanga (1980) states that the cocoyam is the second most important staple in Southern Cameroun. Karikari (1971) describes cocoyam cultivation in Southern Ghana. However, an account quoted by Mauny (1953) shows that taro was well established in Senegambia by 1500, too early for Portuguese navigators to have been instrumental in its diffusion.
16	SE Asian food-crops and the Bantu expansion The early introduction of these humid-zone cultigens must have had important consequences for African prehistory. The region of greatest somatic diversity of these crops corresponds well to the area of the Bantu, Bantoid and Benue-Congo-speaking peoples. Johnston (1919-1922) originally proposed the idea that the Bantu homeland was to be located in present-day Cameroun. The date generally advanced for this is >3000 BP. Although the route travelled by SE Asian cultigens remains quite obscure, it seems credible that their impact on existing agricultural societies in the Bight of Benin must have been considerable. It is likely that a combination of iron technology and three new high-yielding staples that could be grown successfully in the tropical rain-forest permitted the second, and most dramatic, phase of Bantu expansion. New finds in Southern Cameroun now provide direct evidence for agricultural tools in the rainforest (Eggert et al. 2006). Moving South and East, presumably along the waterways the Bantu seem to have rapidly colonised the equatorial forest. The conjunction of these crops and at least some iron tools to make easier the clearing of the forest may have been the combination of factors that permitted the colonisation of half the continent in a relatively short period of time.
17	Summary of the linguistic evidence for the history of Musaceae in Africa I a) Ensete gillettii is established as an indigenous magical plant in West Africa and as such has an old reconstructible root, *kom, in Benue-Congo languages. b) Plantains are introduced by an unknown route to West Central Africa >3000 BP and the #kom root is transferred to them. It is likely that taro and water-yam are introduced at the same period. c) The plantain becomes a crucial cultigen in the exploitation of the Central African rainforest and thus one of the engines of the Bantu expansion.
18	Summary of the linguistic evidence for the history of Musaceae in Africa II d) Compounding #kom produces a variety of names for plantain, including #kondo and #kombo which diffuse through the Bantu area e) The Portuguese coasting trade diffuses plantains to the West along the coast, along with the Bantu name, which appears as #konto and #kodu as far as Senegambia. Another name, #boro, may also be diffused by the Bantu f) The name kondoŋ, borrowed into Fulfulde, is then rediffused back to agricultural societies in West Africa as an irrigated garden crop
19	Summary of the linguistic evidence for the history of Musaceae in Africa III g) The few banana cultivars are brought by the Portuguese from India and Brazil, along with their trade name, palana, a name of Indian provenance. h) This is borrowed into Mandinka as bàrandá and thence diffused into other Mande languages, where it undergoes phonological transformation and shortening. Forms like Vai ɓàànà are likely to have been borrowed into English as ‘banana’. i) Banana is then re-introduced into languages of Anglophone Cameroun in the colonial era, and borrowed into neighbouring languages, eventually spreading into Chad.
20	Finally Two further observations; Despite the great accumulation of data in Rossel (1998) the linguistic evidence does not support her conclusion of a late spread of plantains associated with Islam. There is, moreover, no purely linguistic evidence for an east-west spread of the plantain across the continent as proposed by Murdoch, Simmonds and De Langhe in various forms. The introduction of the ‘tropical food kit’, despite its enormous impact on the peopling of Africa, remains unresolved, and only further phytolith analysis is likely to shed light on this issue.
21