Published online by Cambridge University Press: 23 April 2008
This study focuses on the ancient past of coastal Guinea's Rio Nunez region, a coastal rice-growing region virtually unexplored by studies of West African rice and rice farmers. It argues that coastal cultivators have adapted mangrove rice-farming systems in situ for approximately the past 1,000 years, a historical period pre-dating the European travelers' accounts on which the current literature heavily relies. Rather than diffusing from the interior, these specialized farming systems grew organically out of land-use systems. Using the comparative method of historical linguistics and cultural vocabulary, the study establishes different stages in coastal farmers' experimentation, adaptation and specialization in the coastal environment and approximates historical dates when the different stages occurred. And with botanical and biological studies of mangrove vegetation, the study distinguishes between the softer, spongy roots of white mangroves and hard, twisted roots of red mangroves. The interdisciplinary evidence reveals that knowledge of white mangroves was an early formative stage in cultivators' fabrication of coastal land-use systems. Lastly, from an examination of loanwords, the study discusses the important contributions made by Mande groups, who speak the Susu language in the Rio Nunez region, in intensifying mangrove rice-farming systems indigenous to the coast and extending them from zones of white to those of red mangroves. The interdisciplinary methods and sources enable the study to capture localized experimentation and innovation as continuous processes, thereby breaking with the current literature's emphasis on diffusion from the interior to the coast.
2 Roland Portères, ‘African cereals: Eleusine, Fonio, Black Fonio, Teff, Brachiaria, paspalum, Pennisetum, and African rice’, in Jack R. Harlan and Jan M. J. De Wet (eds.), Origins of African Plant Domestication (The Hague, 1976), 441–5; Portères, , ‘Berceaux agricoles primaires sur le continent Africain’, Journal of African History, 3 (1962), 197–9CrossRefGoogle Scholar.
5 Dorian Fuller, ‘Crop-cultivation – the evidence’, in Kevin Shillington (ed.), Encyclopedia of African History, vol. I: A–G (New York, 2005), 326–8.
6 Portères, ‘African cereals’, 441–5; Portères, ‘Berceaux agricoles primaires’, 197–9.
7 Recent literature on the Rio Nunez region and its inhabitants includes the following: Berliner, David, ‘An “impossible” transmission: youth religious memories in Guinea-Conakry’, American Ethnologist, 34 (2005), 576–93CrossRefGoogle Scholar; Berliner, ‘Nous sommes les derniers Bulongic: sur une impossible transmission dans une société d'Afrique de l'ouest' (Ph.D. thesis, Université libre de Bruxelles, 2002); Marie Yvonne Curtis, ‘L'art nalu, l' art baga de Guinée: approches comparatives' (thèse doctorat, Université de Paris-Sorbonne, 1992); Edda Fields, ‘Rice farmers in the Rio Nunez region: a social history of agricultural technology and identity in coastal Guinea, ca. 2000 bce to 1880 ce (Ph.D. diss., University of Pennsylvania, Philadelphia, 2001); Fields, ‘Before “Baga”: settlement chronologies of the coastal Rio Nunez region, earliest times to 1500 ce’, International Journal of African Historical Studies, 37 (2004), 229–53; Gerald Gaillard (ed.), Migrations anciennes et peuplement actuel des côtes Guinéennes (Paris, 2000), 385–402; Bruce L. Mouser, ‘Trade and politics in the Nunez and Pongo Rivers, 1790–1865’ (Ph.D. diss., University of London, 1971); Ramon Sarró-Maluquer, ‘Baga identity: religious movements and political transformation in the Republic of Guinea’ (Ph.D. thesis, University of London, 1999).
8 According to Olga Linares, the acidic nature of coastal soils favors decomposition of many fossilized materials. In addition, agricultural practices in which coastal farmers cyclically turn over the soil have disrupted fossils and artefacts interred in the earth. Sapir, Olga Linares de, ‘Shell middens of Lower Casamance and problems of Diola protohistory’, West African Journal of Archaeology, 1 (1971), 23–54Google Scholar.
9 Roland Portères, ‘Un problème d'ethno-botanique', Journal d'Agriculture et de Botanique Appliquée, 11 (Oct.–Nov. 1955), 538–42.
10 For examples of the historical and comparative linguistic method applied to the Bantu language group, see Koen Bostoen, ‘Linguistics for the use of African history and the comparative study of Bantu pottery vocabulary’, Antwerp Papers in Linguistics, 106 (2004), 131–54; Ehret, Christopher, ‘Cattle-keeping and milking in Eastern and Southern African history: the linguistic evidence’, Journal of African History, 8 (1967), 1–18CrossRefGoogle Scholar; Ehret, , ‘On the antiquity of agriculture in Ethiopia’, Journal of African History, 20 (1979), 161–77CrossRefGoogle Scholar; Ehret, , ‘Sheep and Central Sudanic peoples in southern Africa’, Journal of African History, 9 (1968), 213–22CrossRefGoogle Scholar; Nurse, Derek, ‘The contribution of linguistics to the study of history in Africa’, Journal of African History, 38 (1997), 359–91CrossRefGoogle Scholar; David Lee Schoenbrun, A Green Place, a Good Place: Agrarian Change, Gender, and Social Identity in the Great Lakes Region to the 15th Century (Portsmouth NH, 1998); Vansina, Jan, ‘New linguistic evidence and the “Bantu expansion”’, Journal of African History, 36 (1995), 173–95CrossRefGoogle Scholar; Vansina, Paths in the Rainforest: Toward a History of Political Tradition in Equatorial Africa (Madison, 1990).
11 A few previous studies use historical linguistics to reconstruct portions of West and West-Central Africa's interior, but not the coast. See Kairn Klieman, ‘The Pygmies were our Compass’: Bantu and Batwa in the History of West-Central Africa, Early Times to c. 1900 ce (Portsmouth NH, 2003); M. E. Kropp-Dakubu, ‘On the linguistic geography of the area of ancient Begho’, in H. Trutenau (ed.), Languages of the Akan Area: Papers in Western Kwa Linguistics and on the Linguistic Geography of the Area of Ancient Begho (Basel, 1976), 63–91; Tal Tamari, Les castes de l'Afrique occidentale: artisans et musiciens endogames (Nanterre, 1997); Kay Williamson, ‘Linguistic evidence for the use of some tree and tuber food plants in southern Nigeria’, in Thurstan Shaw, Paul Sinclair, Bassey Andah and Alex Okpoko (eds.), The Archaeology of Africa: Foods, Metals and Towns (New York, 1995), 139–53; Williamson, ‘Linguistic evidence for the prehistory of the Niger Delta’, in E. J. Alagoa, F. N. Anozie and Nwanna Nzewunwa (eds.), The Early History of the Niger Delta (Hamburg, 1988), 65–119.
12 Among Indo-European languages, the comparative method of historical linguistics is used on languages with ancient written traditions dating back several hundreds of years. In the African context, not even Bantu languages possess written records dating back more than a few centuries. This also applies to proto-languages. See Bostoen, ‘Linguistics for the use of African history’, 133.
13 In general, the following studies have confirmed the integrity of the Atlantic language group: Doneux, Jean L., ‘Hypothèses pour la comparative des langues Atlantiques’, Africana Linguistica, 6 (1975), 41–129CrossRefGoogle Scholar; Fields, ‘Rice farmers in the Rio Nunez region’, 42–6; Fields, ‘Before “Baga”’, 229–53; Konstantin Pozdniakov, Sravnitel'naia grammatika Atlantichskikh iazykov (Moscow, 1993) (translated by Lioudmila Selemeneva, Carnegie Mellon University, Department of English), 3; J. David Sapir, ‘West Atlantic: an inventory of languages, their noun class system and consonant alternation’, in Current Trends in Linguistics in Sub-Saharan Africa (The Hague, 1971), 45.
There are, however, dissenting voices which argue that one linguistic subgroup, Mel languages, are not part of the Atlantic language group. See Dalby, David, ‘Mel languages in polyglotta Africana, part I: Baga, Landuma and Temne’, Sierra Leone Language Review, 4 (1965), 130Google Scholar; Dalby, , ‘The Mel languages: a reclassification of southern West Atlantic’, African Language Studies, 6 (1965), 1–1Google Scholar; W. A. A. Wilson, ‘Temne and the West Atlantic group’, Sierra Leone Language Group, 2 (1963), 26.
14 Cross-cultural evidence has shown that core vocabulary words are some of the oldest words and the most resistant to change in a language. Some linguists resist the notion that one can identify a set of vocabulary words that are culturally neutral in all of the world's languages. See C. H. Borland, ‘Computing African linguistic prehistory’, in Derek F. Gowlett (ed.), African Linguistic Contributions: Papers in Honour of Ernst Westphal (Pretoria, 1992), 6–11; Borland, ‘How basic is “basic” vocabulary?’ Current Anthropology, 32 (June 1982), 315–16.
15 One way of beginning to classify genetically related languages is by employing ‘lexicostatistics’ to estimate the degree of relationship among daughter languages that descended from a common linguistic ancestor. Applying lexicostatistics begins with analyzing 100-word lists of core vocabulary: words for basic nouns, verbs, body parts and elements in nature. From 100-word core vocabulary lists, comparative linguists identify and count cognates – words with similar meanings and sound sequences – shared by pairs of languages. The presence of two cognates in a pair of genetically related languages implies the existence of an ancestral form of the word in a common linguistic ancestor. Words spoken in present-day daughter languages are derived from the ancestral language. Comparative linguists confirm the proposed cognate vocabulary by comparing sounds and establishing sound correspondences in a broader selection of vocabulary than the 100-word core lists.
16 Linguistic subgroups possess a common set of sounds in addition to sets of inherited and innovated vocabulary, remnants of a common ancestral language once shared by the daughter languages. They are evidence that speech communities speaking constituent languages once shared social, political, economic and cultural institutions and came into contact with other speech communities.
17 What I have called ‘Highlands’, Sapir classifies as ‘Temne, Baga Koba, (Banta), Landuma, Tyapi, other Baga languages (excluding ‘Foré and Mboteni)’, as the first linguistic subgroup to diverge from the Mel branch of Atlantic languages. Sapir, ‘West Atlantic’, 49.
18 Fields, ‘Before “Baga”’; Fields, ‘Rice farmers in the Rio Nunez region’, 57–70.
19 For examinations of loanwords in other regions of Africa, see Christopher Ehret, ‘Agricultural history in central and southern Africa ca. 1000 B.C. to A.D. 500’, Transafrican Journal of History, 4 (1974), 1–25; Ehret, , ‘Patterns of Bantu and Central Sudanic settlement in Central and southern Africa (ca. 1000 B.C.–500 A.D)’, Transafrican Journal of History, 3 (1973), 1–71Google Scholar; Klieman, ‘The Pygmies Were Our Compass’, 101–3, 177–83; Schoenbrun, David, ‘We are what we eat: ancient agriculture between the Great Lakes’, Journal of African History, 34 (1993), 1–31CrossRefGoogle Scholar.
21 Bruce L. Mouser, ‘Qui étaient les Baga? Perceptions européennes, 1793–1821’, in Gérald Gaillard (ed.), Migrations anciennes et peuplement actuel des Côtes Guinéennes (Paris, 2000), 436.
22 Voeltz, Erhard, ‘Les langues de la Guinée’, Cahiers d'étude des langues guinéennes, 1 (1996), 29–30Google Scholar.
23 Susu and Jalonke share approximately 90 per cent of their cognates. Based on the high cognate percentage, Susu and Jalonke are still dialects of the same language. Friederike Lupke, ‘A grammar of Jalonke argument structure’ (Ph.D. diss., Radboud Universiteit Nijmegen, 2005), 14.
24 Identifying loanwords and their source languages is not an easy task, because languages borrow words at all stages of their development. Thus, loanwords can be borrowed into an ancestral language and inherited by its daughter speech communities. Early loanwords borrowed before the language in question underwent regular sound changes are more difficult to distinguish from inherited vocabulary words, because they can also exhibit regular sound changes. But in some ways early loanwords are easier to date, because they may exhibit the regular morphological or phonological changes used by comparative linguists to assign words to a particular linguistic subgroup and to date the divergence of the subgroup using glottochronology. More recent loanwords do not exhibit these regular correspondences, making it more difficult to date their entrance into the language. See Vansina, Paths in the Rainforest, 14–16; Bostoen, ‘Linguistics for the use of African history’, 145.
25 Though a complete analysis of the Northern branch of Mande is beyond the scope of this study, some analysis can be made from the core vocabulary lists that I collected during my fieldwork. Present-day Jalonke words possess the definitive marker -na, -nna in their nominal inflections. However, Susu words do not end in -na, -nna, because the Susu language appears to have dropped the inflection, possibly since its divergence from proto-Susu-Jalonke. Lupke, ‘A grammar of Jalonke’, 109–14.
26 Christopher Ehret, ‘Testing the expectations of glottochronology against the correlations of language and archaeology in Africa’, in Colin Renfrew, April McMahon and Larry Trask (eds.), Time Depth in Historical Linguistics (Cambridge, 2000), 373.
27 Colin Renfrew, ‘Introduction: the problem of time depth’, in Renfrew et al. (eds.), Time Depth in Historical Linguistics, ix–xiv.
28 Sheila Embleton, ‘Lexicostatistics/glottochronology: from Swadesh to Sankoff to Starostin to future horizons’, in Renfrew et al. (eds.), Time Depth in Historical Linguistics, 143–66.
29 Jan Vansina, How Societies Are Born: Governance in West-Central Africa Before 1600 (Charlottesville, 2004), 4–5, 8.
30 Ehret, ‘Testing the expectations of glottochronology’, 373–99.
31 Vansina, How Societies Are Born, 4–5, 8.
33 In order to reconstruct words to the proto-Coastal language, they must meet two criteria. First, cognates of the words must be present in the Nalu and Mbulungish languages whose non-contiguous speech communities are located the farthest apart. Second, the words in question must exhibit regular sound correspondences. See Schoenbrun, A Green Place, a Good Place, 49.
For a listing of regular sound correspondences for the Coastal linguistic subgroup, which confirm reconstructed vocabulary, see Fields, ‘Before “Baga”’, 229–53.
34 White mangrove (Avicennia africana) – Nalu: -yof; Mbulungish: -yɔp.
35 According to Teixeira da Mota, Donelha uses a term tarrafe in Portuguese Creole, which is derived from the Arabic tarf, to describe A. africana and Laguncularia racemosa. See André Donelha, Descrição da Serra Leoa e dos rios de Guiné do Cabo Verde (1625) (notas por Avelino Teixiera da Mota, Description de la Serre Leoa et des rios de Guinée du Cabo Verde (1625)) (translated by P. E. H. Hair) (Lisbon, 1977). Santo also includes Rhizophora racemosa on this list. See Santo, J. do Espirito, ‘Nomes vernáculos de algumas plantas da Guiné Portuguesa’, Boletim cultural da Guine Portuguesa, 18 (1963), 458Google Scholar.
Burkill et al. also define the Portuguese Creole word tarrafe as A. africana and describe island people in Western Senegal using the ‘germinated seeds of Avicennia as a famine food, but, these when uncooked or improperly prepared are actually poisonous’. See H. M. Burkill, J. M. Dalziel and J. Hutchinson, The Useful Plants of West Tropical Africa Being an Appendix to the Flora of West Tropical Africa (London, 1937), 453–4, 85–7.
36 Donelha, Descrição da Serra Leoa, f.123, 99.
37 Andrè Alvares de Álmada, Brief Treatise on the Rivers of Guinea (c. 1594): Part II, 15/11, Appendix III, 9/3.
38 M. Sow, A. Diallo, N. Diallo, C. A. Dixon and A. Guisse, ‘Formations végétales et sols dans les mangroves des rivières du Sud’, in Marie-Christine Cormier-Salem (ed.), Dynamique et usages de la mangrove dans les pays des rivières du Sud (du Sénégal à la Sierra Leone): actes de l'atelier de travail de Dakar du 8 au 15 mai 1994 (Paris, 1994), 51–6; Peter J. Hogarth, The Biology of Mangroves (Oxford, 1999), 36–45; P. B. Tomlinson, The Botany of Mangroves (Cambridge, 1986), 12–20, 96–8; Jordan, H. D., ‘The relation of vegetation and soil to development of mangrove swamps for rice growing in Sierra Leone’, Journal of Applied Ecology, 1 (May 1964), 209–12CrossRefGoogle Scholar; Linares de Sapir, ‘Shell middens of lower Casamance’, 26.
39 Sow et al., ‘Formations végétales’, 51–7.
40 Hogarth, The Biology of Mangroves, 4–11.
42 Schoenbrun, A Green Place, a Good Place, 49.
43 Red mangrove – Nalu: m-mak/a-mak; Mboteni: ε-ma, e-ma/a-ma.
44 Seasonal stream – Mbulungish: i-pal/a-ppalleŋ; Mboteni: pɔl/sam-pɔl; large seasonal stream – Mboteni: pɔlmeni.
45 Mosquito – Mbulungish: ɔ-bo/ɔ-bolleŋ; Mboteni: a-bɔ.
46 Crab – Mbulungish: i-nep, e-nep/εnippel, ε-nippel; Mboteni: a-nep/a-neppel. Type of crab – Mbulungish: i-laŋ/ayel-laŋ; Mboteni: a-laŋ/alaŋŋel.
47 Linares de Sapir, ‘Shell middens of Lower Casamance’, 41.
48 Fields, ‘Before “Baga”’; Fields, ‘Rice farmers in the Rio Nunez region’, 100–10.
49 Mound – Mboteni: e-nεk/a-nεk'; Ridge – Nalu: ma-nek/a-nek; Mbulungish: ε-nεk/ki-nεk, ta-nεk tafεt/ma-nεk mafεt; Mboteni: e-nεk/a-nεk; Sitem: a-nek/-nek; Landuma: ta-nεk/ma-nεk.
50 To transplant – Nalu: ma-cεεp, m-cεεpan; Sitem: ki-cεp. To transplant rice – Nalu: -cεp; Sitem: pa-cεεp. To sow rice with finger – Nalu: ma-cεεp; Sitem: kɪ-cep tecir. Seedling – Nalu: m-kicεεpa/aŋ-kicεεpa.
51 Linares, Power, Prayer and Production, 19; Linares, ‘Diminished rains and divided tasks: rice-growing in three Jola communities of Casamance, Senegal’, in A. Endre Nyerges (ed.), The Ecology of Practice: Studies of Food Crop Production in Sub-Saharan West Africa (Amsterdam, 1997), 49.
52 Shortest shovel – Nalu: ma-kumbal/a-kumbal; Mboteni: faa aŋkumbεl; Sitem: aŋ-kumbεl. Short shovel used to turn soil for the second time in ridges – Mboteni: porbal aŋkumbεl.
53 In comparison to Nalu's 3 noun classes, Sitem has 15. Toŋ-kumbεl/aŋ-kumbε appears to fit into class five, t-, tV-, tVŋ-, tɪŋ-/m-, mV-, mVŋ-, mɪŋ-, m-, which includes inanimate words like ‘drum’, ‘hut’, ‘skirt’ and ‘seedling’. The noun classifiers suggest that toŋ-kumbεl may alternatively be a Sitem word.
In spite of the paucity of source materials, linguists agree that Nalu possesses only three noun classifiers, m-, ma-/a- for singular/plural inanimate objects and m-, ma-/b-, bε- for singular/plural animates. Based on the morphological data, ma-kumbal exhibits the noun classifiers we would expect for an inanimate object in the present-day Nalu language. In addition, linguists agree that Mboteni is one of a few Atlantic languages in which the noun class system does not operate. Thus, based on the morphological evidence, the Mboteni language likely borrowed faa-aŋ-kumbel from Nalu in that it exhibits a fossilized noun class.
54 Medium-size shovel used to make mounds – Mbulungish: -ki-taŋɡbanyi/ci-taŋɡbanyi; Susu: kitangbanyi.
55 Interviews in the Nalu village of Kukuba with Ibrahima Camara, 17 Dec. 1998; Souleyman Camara, 19 Sept. 1998; Saliou Bangoura, 10 Dec. 1999; Mohammed Ndjongo Bangoura, 2 Dec. 1998, 4 Dec. 1998.
56 Linares, Power, Prayer and Production, 19; Linares, ‘Diminished rains and divided tasks’, 49; Paul Pélissier, Les paysans du Sénégal. Les civilisations agraires du Vayor à la Casamance (Saint-Yrieix, 1966), 738–41.
57 Walter Hawthorne, Planting Rice and Harvesting Slaves: Transformations along the Guinea-Bissau Coast, 1400–1900 (Portsmouth NH, 2003), 153.
58 In Guinea-Bissau, some groups of farmers still use fulcrum shovels without metal blades. The Jola, on the other hand, always cap a kajandu with a steel blade, which, interestingly, is usually made by a Mande blacksmith. Personal communication with Olga Linares at the 2004 African Studies Association conference in New Orleans and via email, dated 11 Feb. 2005.
59 Interview with ‘President’ Mohamed Yongo Bangoura in the village of Binari, 29–30 April 1998.
60 Hawthorne, Planting Rice and Harvesting Slaves, 45–6.
61 Andrè Alvares de Álmada, Brief Treatise on the Rivers of Guinea (c. 1594): Part II (Liverpool, 1984), 132–3, 129.
63 Fonio – Nalu: m-pindi/a-pindi; Mbulungish: pundε/cu-pundεlεŋ; Mboteni: pundu, pundo; Sitem: pundu; Landuma: pende/ya-pande; Susu: fundeyi; Jalonke: fundema; Mende: funde. Sorghum – Mboteni: cu-mank; Susu: mεŋɡi; Jalonke: mεŋɡina.
64 Mound – Nalu: m-tukuŋi/a-tukuŋi; Mbulungish: tukunyi/ci-tukunyi; Susu: tukuŋi; Jalonke: tukuŋma tekina.
65 Medium-sized shovel – Mbulungish: ki-taŋɡbaŋ/ci-taŋɡbaŋŋel; Susu: kitaŋɡbanyi.
66 Shovel (generic)- Nalu: m-kɔp/a-kɔp; Mbulungish: kɔp/ci-kɔppel; Mboteni: kɔp/su-kɔp; Susu: kɔfi, kɔpi.
Long shovel (described by present-day elders to have been up to 2 meters long and to have been used by young men to show off strength and virility and attract the attention of potential mates): Nalu: m-kɔp lanna/a-kɔp lanna; Mbulungish: kɔp kokilannε/ci-kɔppel kokilannε; Landuma: k-ɔpi/c-ɔpi; Susu: kofi kuye. (The current evidence suggests that they combined the borrowed word with adjectives in their own languages to form compound words describing the size of the shovel.)
67 Shovel blade – Nalu: ma-fanc/a-fanc; Mbulungish: ε-fεnc/ε-fεnceel; Sitem: a-fenc; Susu: fεnsi.
The root word of fensi is probably -fac, a word meaning ‘iron’ and ‘iron pot’ spoken in the proto-Highlands language, the linguistic ancestor of present-day Temne, Landuma, Sitem and Kalum languages. Temne: a-fat; Landuma: a-fac; Kalum: a-fac/ε-fac. Proto-Highlands diverged into its daughter languages c. 1000 ce.
European travelers' accounts provide independent evidence that Temne-speakers, particularly of the Scarcies River region in Sierra Leone, possessed and traded in iron ore, but they do not comment on its quality or quantity. Nor do European observers provide evidence that Temne-speakers possessed iron-smelting technology before the arrival of Luso-African observers. See Álmada, Brief Treatise on the Rivers of Guinea (c. 1594): Part I, 15/8; Álmada, Brief Treatise on the Rivers of Guinea (c. 1594): Part II, 15/8; Donelha, Descrição da Serra Leoa, 235 n. 113; Valentim Fernandes, Description de la côte occidentale d'Afrique (Sénégal au Cap de Monte, Archipels), trans. T. Monod, A. Teixeira da Mota and R. Mauny (Publicações do Centro de Estudos da Guiné Portuguesa, 11) (Bissau, 1951), 76, 95, 166 n.149; John Matthews, A Voyage to the River Sierra-Leone on the Coast of Africa (London, 1966), 52.
Based on this evidence, my hypothesis is that coastal dwellers in the Rio Nunez region may have gained access to some iron ore which came from regions inhabited by Temne-speaking people through trade networks with Temne-speakers. The Susu's iron-working technology fashioned the iron ore into blades for the indigenously made fulcrum shovel. Future research is necessary on this important question.
68 Hawthorn, Planting Rice, Harvesting Slaves, 11.
69 Linares, Power, Prayer and Production, 20. Linares's characterization of the complexity and local nature of coastal rice farming among the Jola mirrors my fieldwork observations in coastal Guinea.