1. A different type of proper name rarely utilized for historical reconstruction is clan/lineage-names. While their etymological interpretation is difficult, their spatial distribution can be detected with little effort (in the field), and yield results independent of other source types. Thus, the distribution of present-day Sonko lineages and lineage branches, and especially that of former royal lineages furnished by Sonkos (map 12), give us clues to the extent of a political entity once ruled by the faran of Sankola (see below).

2. Houis, M., “Quelques données de toponymie ouest-africaine,” BIFAN 20B (1958), 562.Google Scholar I know of only two short treatises dealing with African place names in southern Senegambia: Thomas, L.-V., “Onomatologie et toponymie en pays Diola,” Notes Africaines 71 (1956): 76–80Google Scholar; and Carreira, António, “Duas cartas topogràficas de Graça Falcão (1894–1897)Google Scholar e a expansão do islamismo no rio Farim, ,” Garcia de Orta 11(1963): 189–212.Google Scholar I am not concerned here with Senegambian place names of European origin.

3. The interpretation of the past through the movement of groups of persons. Migrations are to be found at the core of the origines of nearly all social groups around the world. They are, in some historical treatises, still taken at face value and are not, as required by critical historiography, seen as images created to explain political, social, religious, and cultural links between social groups. This imagery does not, of course, preclude the actual historicity of certain migrations. Generally, the link is structured hierarchically, i.e., it is intended to connect one group to another, which is seen as superior. Thus, the ‘inferior’ group, through adoption of a genetic relation with the group/place seen as superior, borrows the latter's high reputation. This practice is based on the constitutive belief in common descent of ethnic and other social groups. This “descent ideology” is instrumentalized by borrowing prestige from a group endowed with historical/mythical prestige and ensuing high status by claiming descent from the latter. Rome borrowed Troy's reputation, by creating an origo of having been founded by descendants of the Trojan Aeneas. And the early medieval Saxons of north Germany created an origo of being descended from members of Alexander the Great's army. Similarly, most groups in western West Africa claim an origin from the highly-reputed empire of Mali, extended in more recent times by a ‘first origin’ from the Middle East (Mecca, Egypt, etc.) reflecting the influence of Islam (seen as culturally superior by Islamic and non-Islamic peoples alike).

4. lam/nam: “king” is Bainunk u-nam, Balanta a-lama, Beafada ñama. Fula lam-do; “chief with land tenure rights” is Serer/Wolof lam(an), “territorial ruler/organizer” is Maninka lamollambo; “to rule” is Songhai lama and Mossi nam (the latter in Delafosse, Maurice, La langue mandingue (2 vols.: Paris 1955) 2:455Google Scholar). jon(k)/etc.: “slave” is yom in languages of Burkina Faso and northern Ghana, Togo, and Benin; donko in Asante; jong in Mandinka, Bambara, Susu/Jalonke, etc.; kome in Sarankole/Soninke; ja:m in Wolof; usamp in Pajadinka; da:m in Balanta; -jo:k in Papel; ujoenk in Kasanga; usong/uso:g in Bainunk. mano/malo/etc.: east of this area (in Niger and Nigeria) sinkafa is the general term for “rice;” in Gabon/Congo/Zaire/Angola it is lose. Words in Koelle, Sigismund, Polyglotta Africana (London, 1854).Google Scholar

5. Bertrand-Bocandé, E., “Notes sur la Guinée Portugaise ou Sénégambie méridionale,” Bulletin de la Société de Géographie de Paris, 11(1849): 265–350Google Scholar; 12(1849): 57-93.

6. This may happen either through gradual adoption of a new language by the indigenous population or by the settlement of speakers of a different language from a different area. In the former case language replacement is effected through direct contact with speakers of the new language, who settle among the indigenous population. A hierarchical relationship of the two cultures (part of which are the languages) is created once one culture is accepted by another (recipient) population as being superior. The bearers of the culture of lower reputation begin to adopt the culture (and language) of a reputation conceived to be higher.

In West Africa, in the latter case, (peaceful or conquering) settlers usually come from nearby areas. Only in very specific, rare circumstances did long-distance relocations of a population take place. One such relocation was the settlement of Muslim long-distance traders (and marabouts in their wake) along the trade routes. Others may have been the armies of major empires. Whether blacksmiths once migrated over long distances, protected by their special status (Mandinka ñamalo), remains to be investigated.

In many cases of language replacement a combination of factors was at work. The expansion of Diola and Balanta into Bainunk and Kasanga eth-nic/culture/language areas was the result of both military conquest and peaceful change through the presence of Diola- and Balanta-speakers among the Bainunk and Kasanga respectively. A hierarchy of cultures, in the eyes of the people concerned, develops under specific conditions. The observance of acculturation processes may give hints to those historical conditions. These subjective aspects of social and cultural processes have rarely been tapped by historians.

My present conception of culture change in southern Senegambia received its initial impulse by Wright's, Donald “Beyond Migration and Conquest: Oral Traditions and Mandinka Ethnicity in Senegambia,” HA 12(1985): 335–48.Google Scholar

7. R. Hachmann, G. Kossack, and H. Kuhn have detected this “Nordwestblock” by analyzing written sources, findings of archeology, toponyms (hydronyms and place names), and personal names: Völker zwischen Germanen und Kelten (Neumünster, 1962).Google Scholar

8. Classification of West Atlantic: Sapir, J. David, “West Atlantic: An Inventory of the Languages, Their Noun Class Systems and Consonant Alternation,” Current Trends in Linguistics 7(1972), 45–112.Google Scholar Mandinka is a tonal language; for the classification of Mande languages see W. E. Welmers, “Niger-Congo, Mande,” ibid., 113-40.

9. A very common prefix in Niger-Congo, shared by Bantu.

10. In Gambian English “compound,” in colonial use “yard,” denotes the (generally fenced) group of buildings inhabited by the basic social unit. This usually is the extended family. Equivalents are French concession, Portuguese moronça.

11. Following Luther's bible translation into High German, it was gradually replaced by New High German. Today it is spoken only in rural areas.

12. West Atlantic words for the previous examples are in Koelle, , Polyglotta, 32–33.Google Scholar

13. A typical example of Franconian place names from the period of migration into northeastern France: founder's name plus the etyma -court or -ville of Latin origin.

14. Backerenduck: Map of the Kurlanders published in Mattiesen, Otto Heinz, Die Kolonial- und Überseepolitik der kurländischen Herzöge im 17. und 18. Jahrhundert (Stuttgart, 1940) 272Google Scholar; Buguba: Maps from ca. 1471 onwards in: da Mota, Avelino Teixeira, Mar, além mar, Estudos e ensaios de História e Geografia (Lisbon, 1972) 122Google Scholar; Bisanao: Maps of Benincasa atlas; cf. Teixeira da Mota, ibid., 122; Bajabo: Coelho, Francisco de Lemos, Duas descrições seiscentistas da Guiné, ed. Peres, Damião (Lisbon, 1953) 155Google Scholar; Baiab: Bertrand-Bocandé, , “Notes,” 345.Google Scholar

15. First recorded, to my knowledge, by Coelho, Lemos, Duas descrições, 30, 116.Google ScholarBequin on a 1688 map by Barbot (Public Record Office, London: M.P., 1, 493).

16. The earliest recording is: “um braço que se chama dos Hereges [a branch called dos Hereges]”; de Almada, André Alvares, Tratado breve dos Rios de Guiné do Cabo Verde, ed. Bràsio, António (Lisbon, 1964) 57.Google Scholar Alternative etymologies proposed: J. M. Gray took it for a “corruption of the Portuguese word ‘igereja’ meaning ‘church’, A History of The Gambia (London, 1940) 11.Google Scholar Paul Hair linked it with Portuguese herege “heretic” (in an unpublished translation of Alvares de Almada, 1985), guided by Lemos Coelho's wording ‘Aldea dos Herejes,’ possibly itself a misinterpretation (Coelho, Lemos, Duas Descrições, 85Google Scholar). My interpretation as “crocodile place” seems to be corroborated by the name of a country Bãbaro, which, according to its location and its inhabitants (Bainunk) (de Almada, Alvares, Tratado, 57, 64Google Scholar), must be the land Geregia. This may have been the Mandinka version of jire:ji. Mandinka bamba means “crocodile,” bamba-la would be “crocodile home.” The latter resembles Bãbaro (l and r being interchangeable). It was mentioned again in 1665: “The Alcaide of Babarra” (Public Record Office, London, T70/828, 73), but the name bambo also appeared in “Farrenbambo king of ffonia” (PRO: T70/829, 147, 29 January? 1666), i.e., ‘faran bambo king of Fogni/Fonyi.” Geregia/Herejes and the name of its one-time superior, Sonkodu/Sangedegu, came to be erroneously reconciled in the mixed form Sanguirigu or S. Guirigu (which itself probably was the Bainunk form sangerugu/sungurungu of sangedugu/sonkodugu), before Geregia/Jereges/etc. emerged. For want of space, I cannot elaborate here.

17. Adjectives, like nouns, are classified with prefixes. While in Mandinka, “old man” would be constructed by contracting ke: “man,” plus koto “old,” to form ke:koto, in Bainunk both the adjective gi-def “old,” and the noun u-dige:n, “man,” would continue to carry their prefixes and form gidef udige:n, “old man.” Each element, noun and adjective, would continue to be composed of at least two syllables (prefix and stem), rendering them too unwieldy to form fluid compounds.

18. Bertrand-Bocandé, , “Notes,” 314, 336.Google Scholar

19. Possibly the NE-Bainunk and Kasanga locative prefix a- is derived from ka-. Loss of an initial plosive also occurred with the prefix for persons (singular): gu- is only (occasionally) retained in SW-Bainunk, while it is u- in other dialects and in Kasanga.

20. Coelho, Lemos, Duas Descrições, 172.Google Scholar The stem somini may be identified as Sameni/Samin/Simini, which recurs as a place name (Samine in Boudié, twice in Kasa, Masomine near Mansabà). Possibly it derives from the name of an ancestral shrine known among the Konyagi as S ameni, and among the Kasanga and Ijaher-Bainunk as Simini.

21. Another one is the term bolong, “small waterway, marigot.” It does not occur in the core area of Maninka and is an adoption of bulom for “creek, river, swamp, etc.” found in many West Atlantic languages (it even became an ethnonym for a group populating the mangrove swamps of Sierra Leone). Certainly both the widespread distribution and the fact, that the geographical features it describes are absent or rare in continental Mande, have led to its adoption by Mandinka. The general loan process went in the opposite direction: West Atlantic languages adopted loanwords from Mandinka, due to the reputation of Mandinka being the language of the Mali empire (cf. the multitude of Latin loanwords in Germanic languages).

22. Possibly ga-jaga. The name of the Jahanka-traders' town of Jaga/Jaha(-ba) on the western fringe of gold-producing Bambuk was first mentioned by Fernandes, though (erroneously) as the capital of Mali, and by Enciso, Valentim Fernandes, Description de la Côte Occidentale d'Afrique (Bissau, 1951) 37Google Scholar; Hair, Paul, “Some Minor Sources for Guinea, 1519-1559: Enciso and Alfonce/Fonteneau,” HA, 3(1976) 27.Google Scholar With a locative prefix ka/ga-, possibly used in neighboring Silla (where West Atlantic Fula was spoken), we would have gajaga. Jahaba and Gadiaga, both centers of long-distance traders, may thus share the stem jaha/jaga.

23. Kasan was the major emporium for the rich Gambia salt trade. Boats of the king of Niumi and of his Niuminka fishermen subjects brought salt from the Felane salines (south of the lower Saloum) to Kasan, where it was sold to Julas, who, in turn, traded it into the salt-hungry interior of the continent (:cf. Gomes, , De la première découverte de la Guinée, ed. Monod, Théodore, Mauny, R., Duval, G. (Bissau, 1959) 43Google Scholar; Coelho, Lemos, Duas Descrições, 110Google Scholar; Durand, J.P.L., A Voyage to Senegal (London, 1806) 40Google Scholar; Park, Mungo, Travels in the Interior Districts of Africa (London, 1799) 4.Google Scholar

24. Fernandes, Valentim (Description, 74, 84, 88)Google Scholar mentioned piçaa, pisoo, and bissa, an “idol” with these features and an oath formula in the Sierra Leone area around 1500. The Diola south of river Casamance and the neighboring Bainunk call the sacral royal and community “talking drum” (a slit gong generically termed bombolong from the river Gambia to Sierra Leone) kabisa and ebisa respectively.

25. Roche, Christian, “Ziguinchor et son passé (1645-1920),” Boletim Cultural da Guiné Portuguesa, 28(1973): 37.Google Scholar

26. de Almada, Alvares, Tratado, 63.Google Scholar

27. J. Doneux, Personal communication, Brussels, 24 October 1990.

28. For meanings of Temne funco, fungo (historical) and a-funk, a-punkan (modern) see Paul Hair's unpublished notes (1984/1986, chapter 14) on Alvares de Almada. The grain store played a social and sacral role in Sierra Leone (de Almada, Alvares, Tratado, 120Google Scholar; funco) and among the Bainunk of certain areas, who held meetings of lineage elders under the store and buried their elders there.

29. Diola mentioned east of Fogni (in Sangedegu) by Coelho, Lemos, Duas Descrições, 144.Google Scholar

30. Cadamosto, , Le navigazioni atlantiche del Veneziano Alvise da Mosto, ed. Leporace, T.G. (Venice, 1966), 79, 85, 97.Google Scholar

31. “e este de Guambea que tambem na lingoa dos Mandinguas hà nome Guabuu,” Pereira, Duarte Pacheco, Esmeraldo de situ obris, ed. Mauny, Raymond, (Bissau, 1956), 64.Google Scholar

32. Published in The Voyages of Cadamosto and Other Documents on Western Africa in the Second Half of the Fifteenth Century, trans. and ed. by Crone, G.R. (London, 1937), 137.Google Scholar

33. Gambú; also appears as an—obviously ethnonymic—surname in a Santo Domingo slave inventory of 1547, together with Bañol (Bainunk) and Mandinga: Blanco, Carlos Larrazabal, Los negros y la esclavitud en Santo Domingo (Santo Domingo, 1967), 82.Google Scholar

34. The u behind G may be neglected. This spelling followed a contemporary fashion; Diogo Gomes, another early voyager, wrote his first name as “Dioguo.”

35. The name of the Ijaher/Ijagar-Bainunk is derived from the name Jagara, mentioned around 1500 for the king of the Bainunk of São Domingos and his land (Fernandes, , Description, 68, 70Google Scholar). The same applies to the land Jarra, called Jagra in all sources before the mid-nineteenth century. In southern Senegambia, jagara designated the members of the royal lineage (de Almada, Alvares, Tratado, 100Google Scholar). Variants were titles such as “Jagrofa,” “duke,” and “Gagarafe (ministro do Estado)” (Dias, António, “Crenças e costumes dos indígenas de ilha de Bissau no século XVIII,” Portugal em África 2(1945): 224Google Scholar; “Relatório do Feitor da Fazenda Real e capitão-mor de Bissau, José António Pinto, (…), de 1793 a 1797” in Carreira, António, Documentos para a história das ilhas de Cabo Verde e ‘Rios de Guiné’ (séculos XVII e XVIII) (Lisbon, 1983), 166.Google Scholar In northern Senegambia jagara(f) were members of the council electing the king (and “Capitães-Generais:” de Almada, Alvares, Tratado, 32Google Scholar); in northeastern Senegambia a slave in charge of royal land management; and in Kabu (Fula jarga) the village chief. Other countries named after their rulers: the Sarankole land Galam (Gadiaga) seems to be (Fula) ga-lam “land/place of the king.” Wuli may have been named after its royal Wali lineage. Reinhard Wenskus has suggested that the name of the north Germanic “Teuton” is derived from a word for “king;” “Pytheas und der Bernsteinhandel” in Untersuchungen zu Handel und Verkehr der vor- und frühgeschichtlichen Zeit in Mittel- und Nordeuropa, I: Methodische Grundlagen und Darstellungen zum Handel in vorgeschichtlicher Zeit und in der Antike (Göttingen, 1985), 101.Google Scholar

36. Pereira, Pacheco, Esmeraldo, 65.Google Scholar

37. Gomes, , Découverte, 39.Google Scholar

38. Attempts to identify Bati with Badibu (Boulègue, Jean, “La Sénégambie du milieu du XV siècle au début du XVII siècle,” unpublished thesis, 33–35Google Scholar, according to da Mota, Teixeira, Mar, além mar, 248Google Scholar) are not convincing. I believe that Pating was meant: all early Portuguese sources name rulers by adding their title to the name of their country. The nasalization of Pating's final i may have merged with the nasal m of mansa to form Battimansa. The latter may thus have been formed of Bating and mansa. Gomes met the Battimansa on the “right” bank, travelling upstream (Gomes, , Découverte, 41Google Scholar), thus on the south bank. The Portuguese classified the river banks as seen in the direction of travel. Batti is to be sought further upstream: Cadamosto, (Navigazioni, 98)Google Scholar located it “about sixty or more miles from the river mouth.” Cadamosto was inclined to underestimate distances on the river. He reckoned James Island to be 10 (italian) miles (2.5 port, leagues or 12.5 to 15 km) from the river mouth. In fact, the distance is about 40 km. Cadamosto's “sixty miles” would be equivalent to six times the distance from the river mouth to James Island (“ten miles”), which would give us 240 km, the distance from the mouth to Kudang or Kasan/Kuntaur on the middle Gambia, closer to Kabu than to Badibu. Further evidence for Pating is the relative location of the Battimansa upstream of the area ruled by the Farosangoli (ibid.) who was subject to Mali (Gomes, , Découverte, 41Google Scholar). I identify the latter with the faran of Sankola, whose influence reached the Gambia west of Niani. This faran ruled over one Frangazick (faran of ga-zick), who styled himself the former's “nephew” (uncle/nephew a euphemism in the guise of kinship terms still employed for senior/junior power relationships), from whose dominions Gomes continued his upstream voyage on the Gambia to Niani, Wuli, and Kantora (ibid., 34-36). The Frangazick may thus have resided in Niamina (a Sonko state). The same identification with the faran of Cissoko, Sankola by Sékéné-Mody, “La royauté (mansaya) chez les Mandingues occidentaux, d'après leurs traditions orales,” BIFAN, 31B (1969), 328.Google Scholar Different identifications: da Mota, A. Teixeira, Mar, além mar, 191, 246Google Scholar; and Brooks, George, Kola Trade and State Building: Upper Guinea Coast and Senegambia, 15th-17th Centuries, (Working Paper no. 38, African Studies Center, Boston University (Boston, 1980), 16).Google Scholar

39. Balaba, mythical foundress of Kabu's ruling matrilineal Sane lineage is said to have dwelt in a cave (= autochthonous origin) near Mampating (Sidibe, Bakary, “The Story of Kaabu: Its Extent,” paper presented to the Conference on Manding Studies, London 1972, 9)Google Scholar, capital of Pating (other traditions mention her for Pachana or as a former name of Puropana). This (certainly ahistorical) idealized ancestral figure was obviously named after the common southern Senegambian Balaba (or Buloba, Buleba, etc.), a shrine spirit of the ancestors often imagined as a snake. It served divinatory ends and was frequently with royal lineages as those of Kasa (Balaba-kisa, my interviews with Salifu Jingali, Boulomp, 15.12.1987, and Alaji Jingali, Aniak, 18.5.1987, tape-copies at Oral History Division, Banjul, -ki:s is a Wolof conjuration suffix) and Bissau (Baloba: Pinto, , “Relatório,” 166Google Scholar). Among the Diola (e.g., for the procurement of rain): Thomas, L.-V., Les Diola (2 vols.: Dakar, 1959), 2:591Google Scholar; Thomas, L.-V.: “Samba Diatta, ‘roi des Blis’,” Notes Africaines, no. 115 (July 1967), 87Google Scholar; and among the southwestern Bainunk it is a fétiche parlant. It also appears in toponyms: ex-NE-Bainunk villages Kambaleiba (Jasin, Darang, near Jarol), also Buloba (Manjak, near Canchungo), Rios Balaba and Buloba in Beafada lands.

40. de Almada, Alvares, Tratado, 55, 70Google Scholar; Donelha, André, Descrição da Serra Leoa e dos rios de Guiné do Cabo Verde (1625), ed. da Mota, A. Teixeira, English trans. Hair, P.E.H. (Lisbon, 1977), 119.Google Scholar

41. First mentioned in the fourteenth century by al-Bakri for Walata, (Corpus of Early Arabie Sources for West African history, ed. Levtzion, N. and Hopkins, J.F.P. [Cambridge, 1981], 284).Google Scholar A source of ca. 1500 defined theforoes (farans) as “corregedor ou gouernador delrey” (Fernandes, , Description, 42Google Scholar), i.e., “magistrate or governor of the king” of Mali.

42. After Mali's collapse and the political emancipation of its former provinces, which shed their vassal status, the title faran came to mean “emperor, king over kings” (first: de Almada, Alvares, Tratado, 43Google Scholar). But it also retained its older sense of “vassal chief,” thus coming to bear an ambiguous value. In this latter sense it is found in the numerous southern Senegambian Farankunda place names, but also in Jifarong (Kiang) as well as in Coelho's, LemosCafarão (*ka-faran) (Duas Descrições, 30, 156)Google Scholar, which is probably Iarang, a former capital of a northern Balanta state (Bertrand-Bocandé, , “Notes,” 345Google Scholar: Jarang). Gomes' Frangazick (*faran-gasik, see above) was also such a lesser ruler.

43. Birasu was the successor to its eastern neighbor Sankola as a regional power. The faran of Birasu was first mentioned by Fernandes, (Description, 69).Google Scholar Access to trade with the Portuguese (ibid.: horses) on the upper Rio Cacheu may have added momentum to its rise.

44. The shift of the stem's initial consonant according to grammatical context; cf. Kasanga aca:r, “knife,” gasa:r, “knives.”

45. Coelho, Lemos, Duas descriçôes, 36, 154.Google Scholar

46. Cf. elup,“[mud] house.” This etymology is corroborated by the Bainunk parallel case bilit/gulid, “house, compound, wall.” These words are mainly used in areas with a tradition of mud structures (in the southwest, close to the Diola with equal building techniques). The latter stem recurs in Gujaher-Bainunk u-lod, “potter” (who works with clay/mud).

47. Adam Jones used ma- (for liquids) prefixed hydronyms to reconstruct the distribution of Mel-speakers in Sierra Leone before the expansion of Mande languages: Jones, , “Who were the Vai,” JAH 22(1981), 169–70.CrossRefGoogle Scholar

48. In the course of a series of such inventories, the U.S. Ministry of the Interior published: Senegal. Official Standard Names (Washington, D.C., 1965)Google Scholar; Portuguese Guinea. Official Standard Names (Washington, D.C., 1968)Google Scholar; Gambia. Official Standard Names (Washington, D.C., 1968).Google Scholar

49. Geregia was first mentioned by Alvares de Almada, Tratado, 57, as Hereges, and Conjongolong was mentioned only by Bertrand-Bocandé, , “Notes,” 314, 336.Google Scholar

50. Ibid., 63.

51. Among other sources: Le Brasseur, J.A., “Détails historiques et politiques, mémoire inédit (1778),” ed. Becker, Charles and Martin, Victor, BIFAN, 39B (1977), 105Google Scholar; «Public Record Office (London): T 70/3, 41; Debat, 29 Jan. 1764; « Map, Bibliothèque Nationale (Paris), portefeuille 111, division 2, pièce 37: “Côte d'Afrique, (…), par Levant, 1776.”

52. I define state as the territorial, institutionalized, and effective rule of one or more lineage(s) over members of others. This wide definition seems suited to southern Senegambian conditions. It lacks criteria otherwise agreed on in political sciences, such as a minimum degree of centralization of political power, the existence of a social stratum separated from production to serve for administration, or even script, and monumental architecture.

53. Further place names occurring in the former and present areas of the Bainunk, Kasanga, and Beafada, partly also in Kabu, are: Many Bisari/Bisara/Bisora (possibly Tobor/Niamone-Bainunk and Fula sa:r, “settlement”), Gambisara, Mambisara. Payinku/Paiunco (Pachana capital, in Gubisseco, old name of Diambati in Sonkodu) and possibly Bayanka (Geregia) and Badiongo (Kasa). Bugampor (ex-Ijaher-Bainunk village near São Domingos, and in Guinala). A stem yab/jab/cab in Baiab/Bajabo/Faja (capital of ex-SE-Kasa, now Balanta), Patiabor (capital of Sama, co-capital of Boudié), Baiaba (Sandu province of Wuli), and possibly a marigot Badiapour (off the lower Casamance). Stem -ai (Kai-ai with prefix ka- in Niani, “Guiai” with prefix ji- in Guinala, according to Donelha, , Descrição, 174Google Scholar, a land Pai Ai east of Corubal, a village Paiaicunda in Mana). Simbor (capital of Sama, in Woye/Oio, in Bugafara east of Kasa, in Jimara). Guidel/Guidali/Tyidéli (south of Ziguinchor, Forreà, S-Tumana, Nyampaio). Su(m)bundu (Jarra, Sonkodu, Pakao, Sankola).

54. Alvares de Almada, Tratado, 70; Donelha, Descrição, 167; de Sandoval, Alonzo, Naluraleça, policia sagrada y profana, costumbres, ritos, disciplina y catecismo evangelico de todos los Ethiopes (Seville, 1627), 6.Google Scholar

55. It formed the toponyms of, among numerous others, the island of Bulama off Quinara/Guinala and the ethnonym Bulom (Sierra Leone), and was borrowed by Mandinka (bolong, “marigot”) and by Portuguese Creole (bolanha, “rice paddy”). Cognate of the stem lom is in northern Guinea-Bissau, ra:m (cf. Bainunk gu-raw, gu-rang and Kasanga a-dan “marigot”), forming the ethnonym B(u)rame (formerly for the Manjak/Papel/Mankanya, today restricted to Mankanya) and the name of the Bayot-Diola mangrove village Arame. The shifts from voiced to unvoiced consonants are also found in the Moyenne and Haute Casamance and in northeastern Guinea-Bissau: prefix ba- > pa-, ji- > ci-. See Pating, Pakaoi (Kombo: Bakau), Mampalago, Tyidéli (S-Bainunk Guidel = jidel), Tiarap, Tiéti.

56. de Almada, Alvares, Tratado, 95Google Scholar; and Donelha, , Descrição, 159.Google Scholar

57. My interview of 24 January 1987 (copy of tape at Oral History Division, Banjul).

58. Mattiesen, , Überseepolitik, 272.Google Scholar

59. At least six Manduars are known: in Kombo, Kiang, Geregia, Jame/South-Fogni (2), and Pakao (royal residence). The stem war, “stone,” is Bainunk (also in other West Atlantic languages). Possibly it denotes a sacral stone (at the core of many shrines). Manduar might be named after the shrine itself, or after its guardian. Analogous place names among the Diola: Elinkin (linkin, “stone”). In Fulup-Diola elinkin denotes a royal shrine, as does plak, “stone,” in Manjak. Could Manduar be the “place of a royal stone shrine guardian (=king)?” At the end of the sixteenth century, the king of Fogni was mentioned as “afarão [=faran]; called Jaroale” (de Almada, Alvares, Tratado, 57Google Scholar) and as Faram Jariale, rei de Tunhi or Fonhi (letter of 5 August 1696, in Barcellos, Christiano José de Senna, Subsidios para a história de Cabo Verde e Guiné (Lisbon, 1899–1913), 9: 141Google Scholar; and Arquivo Histórico Ultramarino, Lisbon: Papéis avulsos, 249, 9 November 1697). Because of its recurrence after more than a century “Jaroale” may have been a title rather than a name, possibly derived from *ja-ro-wal. Prefix ja- in Senegambian languages could denote functions/offices (see jagaraf above, jabundanes=players of an instrument in Guinala: de Almada, Alvares, Tratado, 98Google Scholar). Prefix ra/ro- is occasionally used in Bainunk to denote rituals (e.g. ran-kub “male initiation”). *jarowal may have been the “person in charge of the stone ritual,” in the case of a royal stone shrine, the king. The stem war, “stone,” recurs in ro/lo-wal/bal, a component found in place names in the region: Tankroal in Kiang (in historical sources, Tankoroval etc.) and on the upper Rio Cacheu (derived from *tan-ka-ro-wal?), in Corubal/Gulubali (a Beafada-land) with prefix ka-/ko-, in “Durubali” (near Pirada-Kansala), and possibly in the Beafada clan name Malobal.

60. Bertrand-Bocandé, , “Notes,” 62.Google Scholar

61. Schaffer, David M., “Pakao—A Study of Social Process Among a Mandingo People of the Senegambia,” (Ph.D., Oxford, 1976), 177.Google Scholar

62. Houis, , “Quelques données,” 572.Google Scholar

63. Personal communication, 9 April 1988.

64. Coelho, Lemos, Duas descrições, 31, 142Google Scholar, mentions only this settlement for Portuguese trading on the upper Casamance. The identity of Boudiémar with Coelho's Bujé is confirmed by Bertrand-Bocandé, , “Notes,” 312Google Scholar, who wrote of the former as having formerly been the major Portuguese factory on the upper Casamance.

65. Conceivable parallels in other West Atlantic languages: Wolof dak, Bijago ande:gen, Landuma da:di (words in Koelle, Polyglotta).

66. The name of a Diola village in Buluf, Diatok (ja-tok), seems to confirm this interpretation. The original Bainunk inhabitants of Diatok are reputed to have settled at the quarter called Birko. As we shall see below, this place name fits nicely into the general pattern of the vast majority of such place names belonging to seats of kings and royal lineages. In Diatok we possibly find a royal residence (Birko) and a place of meetings presided by the king (Diatok). Admittedly, a different (Diola) etymology also makes sense: jatok, “river.” The name of a kingdom JattuCo on a Kurland map (Mattiesen, , Überseepolitik, 272Google Scholar), seemingly identical with Fogni/Fonyi, might also be interpreted as ja-tuk-o.

67. da Mota's, A. Teixeira introduction to Donelha, , Descrição, 35.Google Scholar

68. de Almada, Alvares, Tratado, 70, 95Google Scholar; Donelha, , Descrição, 167.Google Scholar

69. Bertrand-Bocandé, , “Notes,” 63.Google Scholar

70. Jobson, Richard, The Golden Trade (London, 1625), 237–38.Google Scholar While -way may have been the frequent first name Wali, Mandinka suma is the title of the head of that branch of the royal lineage which furnishes the successor to the throne, and hence it is the title of the successor. Thus Jobson may have been wrong in terming Sumaway a king, although he certainly was the local political authority, and his lineage at Bereck was royal.

71. Bertrand-Bocandé, , “Notes,” 314, 336.Google Scholar

72. Ibid., 62.

73. Oral History Division, Banjul, Tape 130A: Interview with Ibrahima Mane at Brikama Nding (Fuladu), The Gambia, probably 3 March 1972, p. 7 of transcription.

74. Bertrand-Bocandé, , “Notes,” 64.Google Scholar

75. Reputed to have been founded, together with Fanka (a later capital), by the ‘first’ ruler of Sankola, Mansa Bourama Sonko, interview with Sonko, El'haji Bakari, Sankola-Berekolong, , in Cissoko, Sékéné-Mody and Sambou, Kaoussou, Recueil des traditions orales des Mandingues de Gambie et de Casamance (Niamey, 1974), 252.Google Scholar

76. Galloway, Winifred, “A Working Map of Kaabu,” unpublished paper presented at the Colloque International sur les traditions orales du Gabou, Dakar, 19-24 May 1980, p. 1.Google Scholar

77. de Almada, Alvares, Tratado, 104.Google Scholar

78. Bertrand-Bocandé, , “Notes,” 69.Google Scholar

79. Inquérito Etnogràfico, ed. da Mota, A. Teixeira (Bissau, 1947), 113.Google Scholar

80. Possibly two villages by the name of Bakendik/Bakarindik, both belonging to Niumi's ‘first’ (oral tradition) royal lineage (Jame), must be added to the list. One was mentioned as Backerrenduck on a map of the 1650s (Mattiesen, , Überseepolitik, 272Google Scholar). Its original name may, perhaps, be reconstructed as *ba-ka-ran-duk (see above). If this is correct we would be left with another duk stem name. Its affiliation with the Buduk group of place names is also hinted by the alternative name Bourco of its ruling Jame lineage in traditions collected in the 1770s (Le Brasseur, , “Détails historiques,” 113–14Google Scholar), possibly after their residence. The identity of Jame and Bourco follows from Le Brasseur's listing of kings with the surnames Bourco, Sonko, and Mane. Today the Jame, Sonko, and Mane are named as royal lineages, the Jame being considered the original, or ‘first’ one. Burko survives as a rare clan name north of Niumi (in Joal and on the les du Saloum; Becker, Charles and Martin, V., “Les familles paternelles sereer,” BIFAN 44B (1982): 388Google Scholar), and as the clan name of persons figuring in a tradition about Niumi's neighbor Jokadu (Wright, Donald, Oral Traditions From the Gambia (Athens, 1979), 1:101Google Scholar).

81. Between the lower courses of the Casamance and the Rio Cacheu, plus narrow coastal stretches north of the former (Bliss, Karone) and south of the latter (Bote).

82. Presumed to be between the Rio Mansua and the north bank of the Rio Geba estuary.

83. Attested by a village Poudoukou and a. faro (drainage gully, swamp) Borouko (off the Lingourou tributary to the Kogon), both in Guiné-Conakry. Was this area part of Beafada territory, or did its Landuma inhabitants adopt the name from the neighboring Beafada?

84. This conclusion is corroborated by other place names: Diaroumé/Jarume is an alternative capital of the former Bainunk kingdom of Jasin (Mane lineage), a former royal compound in ex-Bainunk Fogni (see above), as Jarumai on Sofanyama Bolong in Wuropana/Niamina (early colonial maps), as Iarom in Woye, and as Jerume in Niani. The Joreng place names also delineate the former Bainunk territory.

85. Maps of the 1460s and 1470s depict a “terra farsangalli” between the mouths of these rivers and north of the Geba (cf. da Mota, Teixeira, Mar, além mar, 122Google Scholar). The regional power of the Farosangoli is attested by his title faran, the extent of his rule as shown on the maps, and by Gomes' and Cadamosto's writings about the “great prince” and “the principal lord” of the lands on the Gambia respectively (Gomes, , Découverte, 34Google Scholar; Cadamosto, , Navigazioni, 97Google Scholar).

86. No proof available for this assumption. Gomes' and Cadamosto's use of the Mandinka titles mansa and faran for the rulers of Niumi, Niani, Wuli, Pating, and Sankola may be due to Mandinka interpreters, to Mandinka as the lingua franca, or to the loanword nature of the titles. See Gomes, , Découverte, 35–37Google Scholar; Cadamosto, , Navigazioni, 97, 98, 104, 109.Google Scholar

87. Possibly as a result of the late fifteenth century Fula military campaign through southern Senegambia (details and dating: da Mota, Avelino Teixeira, “Un document nouveau pour l'histoire des Peuls au Sénégal pendant les XVème et XVIème siècles,” Bolelim Cultural da Guiné Portuguesa 24[1969]: 781–860).Google Scholar

88. Selected evidence for state structures in Kabu before its (Mandinka) culture change: continuity of autochthonous clan names of ruling lineages (Sane, Mane), political structure (matrilineality of royal lineage/clan versus Mali's patrilineality), the term ñanco for members of the royal lineage/clan (in the form naco also among the non-mandingized Manjak, but unknown in Mali).

89. Kabu examples: Mampating (prefixes ma- and ba/pa-), Payinku, Mankorsi.

90. I infer this from dominant matrilineal elements in the kinship system both within Kabu's ruling lineage/clan and among the Pajadinka, and from Kabu's location between the closely-related languages Beafada and Pajadinka.

91. Gomes, , Découverte, 34Google Scholar; Cadamosto, , Navigazioni, 97.Google Scholar

92. Coelho, Lemos, Duas descrições, 14, 37, 114–15.Google Scholar In the seventeenth century this land incorporated Geregia. The Europeans of the seventeenth and early eighteenth centuries, who had no communication with core Sangedegu in the interior (on Soungrougrou river), called Geregia Sangedegu.

93. From certain place names, social features and politico-sacral bonds between the two banks of the upper Rio Cacheu (names of royal lineages, and Bertrand-Bocandé, , “Notes,” 347–48Google Scholar), the ethnonym Txà/Betxà (ca beca:) of the Balanta on both banks of the Rio Cacheu (and its possible derivation from ‘Kasa’), and other evidence I assume a former Kasanga identity of the Balantas of Woye. And in the neighboring area of Mansabà many place names prefixed ma(n)- are encountered (Morès, Mansomine near Mansabà, Mambonco, etc.). Although also present in the northern Bainunk area, it is particularly frequent in the former northern territories of the Beafada.