Ethnicities of Enslaved Africans in the Diaspora: On the Meanings of "Mina" (Again)

I

The term "Mina," when encountered as an ethnic designation of enslaved Africans in the Americas in the seventeenth to nineteenth centuries, has commonly been interpreted as referring to persons brought from the area of the "Gold Coast" ("Costa da Mina" in Portuguese usage), corresponding roughly to modern Ghana, who are further commonly presumed to have been mainly speakers of the Akan languages (Fante, Twi, etc.) dominant on that section of the coast and its immediate hinterland. In a recently published paper, however, Gwendolyn Hall has questioned this conventional interpretation, and suggested instead that most of those called "Mina" in the Americas were actually from the "Slave Coast" to the east (modern southeastern Ghana, Togo, and Bénin), and hence speakers of the languages nowadays generally termed "Gbe" (though formerly more commonly "Ewe"), including Ewe, Adja, and Fon. Given the numerical strength of the "Mina" presence in the Americas, as Hall rightly notes, this revision would substantially alter our understanding of ethnic formation in the Americas.¹

In further discussion of these issues, this paper considers in greater detail than was possible in Hall's treatment: first, the application of the name "Mina" in European usage on the West African coast itself, and second, the range of meanings attached to it in the Americas. This separation of African and American data, it should be stressed, is adopted only for convenience of exposition, since it is very likely that ethnic terminology on the two sides of the Atlantic in fact evolved in a process of mutual interaction. In particular, the settlement of large numbers of returned ex-slaves from Brazil on the Slave Coast from the 1830s onwards very probably fed Brazilian usage back into west Africa, as I have argued earlier with respect to the use of the name "Nago" as a generic term for the Yoruba-speaking peoples.²

The argument advanced here is that, in its original meaning in West Africa, the name "Mina" did indeed relate specifically to the Gold Coast, or at least to persons who originated from the Gold Coast even if settled elsewhere, though these included speakers of the Ga-Adangme languages of the eastern Gold Coast, as well as Akan;³ and that in the Americas, although the term was sometimes (especially in Brazil) used with an extended reference that included speakers of Gbe languages, it is questionable whether it ever denoted Gbe-speakers as distinct from speakers of Akan or Ga-Adangme. The implication is that the size of the Gbe-speaking element in the "Mina" presence in the Americas has to be substantially scaled down from that hypothesized by Hall.

II

The origin of the name "Mina" is not in dispute. It is Portuguese for "mine," and in this context alludes to the mining of gold. After the Portuguese first reached the coast of modern Ghana in 1471, the name "A Mina"/"The Mine" was soon applied to the area where they purchased gold, initially specifically to the coastal town of Shama. In 1482, however, the Portuguese constructed the fort of São Jorge da Mina, "St George of the Mine," 30 kilometers further east, at a village whose indigenous name was Edina, and subsequently the name "Mina" came to refer primarily to this site. The corruption of this name into its modern form "Elmina" occurred in the period of Dutch occupation of the fort after 1637.⁴

Although "Mina" thus became the name of a specific town, however, the term "Costa da Mina"/"Coast of the Mine" was applied more generally, at least roughly in the sense of the later "Gold Coast". But the application of the term was not precisely fixed, and sometimes included a broader area. In administrative terms, in fact, the Portuguese "captaincy of the Mina," over which the Elmina fort presided, covered the area from Cape Palmas to the river Volta, and thus included the "Ivory Coast" (Cape Palmas to Cape Three Points or thereabouts), as well as the Gold Coast proper. The coast to the east of the Volta, including the later "Slave Coast," by contrast fell under the jurisdiction of the "captaincy of São Tomé," administered from the island so named.⁵ In Portuguese/Brazilian usage from the late seventeenth century onwards, as Pierre Verger has shown, the connotation of the term "Costa da Mina" shifted again, to mean "the Coast leeward [east] of Mina,"⁶ including the section of the coast immediately east of the Volta, otherwise known as the "Slave Coast."

However, it is questionable whether, as is commonly assumed, ships from Brazil scheduled for the "Costa da Mina" took in all their slaves from the Slave Coast. The Dutch, who claimed the right to regulate Portuguese trade by issuing of passports, did indeed seek to restrict their activities to specified ports east of the River Volta–Popo (nowadays Grand-Popo), Ouidah, Jakin (modern Godomey), and Ekpè, all situated within the modern Republic of Bénin.⁷ It is doubtful, however, whether this restriction could ever have been fully effective, and Portuguese-Brazilian ships did purchase some slaves west of the Volta. Between 1680 and 1683, for example, the Portuguese usurped possession of the Danish fort of Christiansborg at Accra, and it is clear that in this period the Portuguese, in common with other Europeans, bought slaves, as well as gold, there.⁸ It was still reported that the Portuguese were buying slaves (albeit only "a few") at Accra in 1688.⁹ In 1697 the chief of the English fort at Accra complained that "[w]e are hardly ever clear of Portuguese vessels," which was hindering the purchase of slaves by the English.¹⁰

Portuguese ships also bought slaves on the sections of the coast immediately neighboring Accra, in the kingdom of Agona to the west, and in "Allampo," i.e., the Adangme country between Accra and the river Volta to the east. The factor on an English ship seeking slaves at Winneba in Agona in early 1681, for example, noted that on his last visit there he had bought slaves in company with a Portuguese ship.¹¹ In "Allampo" to the east, various reports in the 1680s-1700s alluded in general terms to the Portuguese, as well as other European nations, purchasing slaves there.¹² In detail, English records refer in early 1681 to the presence of two Portuguese ships at Allampo, which by implication were buying slaves, and later in the same year an English ship trading for slaves there complained of competition from a Portuguese ship, which was paying higher prices. In the following year another reported two Portuguese ships there, again presumably trading for slaves, and in 1683 an English slaver at Allampo faced competition from a Portuguese ship, which bought between 70 and 80 slaves there before proceeding east of the Volta to complete its cargo.¹³ It is clear, therefore, that some (albeit probably small) proportion of the slaves brought to Brazil from the "Mina Coast" came from the eastern Gold Coast, rather than from the Slave Coast.

On the Gold Coast itself, when "Mina" was used in an ethnic sense, its basic meaning was people from Elmina specifically, as opposed to other communities in the region. But outside the Gold Coast, the term came to have a more inclusive meaning, referring to persons from the Gold Coast in general. In 1731, for example, when it was reported on the Slave Coast that "a great army of Minas" was marching to fight against Dahomey, the dominant power in the region, the reference was to forces from the kingdom of Akyem, which were currently operating east of the Volta.¹⁴

The question is complicated, however, by the fact that there was a "Mina" diaspora along the coast to the east, into the Slave Coast. European ships trading on the Slave Coast regularly brought with them canoes purchased on the Gold Coast, and also hired crews of canoemen there, in order to communicate with the shore.¹⁵ The practice is first documented in Dutch trade in the 1650s, but whether this was a Dutch innovation or copied from earlier Portuguese practice is unclear.¹⁶ Around the same time, Gold Coast merchants began using ocean-going canoes to trade independently with the Slave Coast, buying locally-made beads and cloth for resale on the Gold Coast.¹⁷

Although such canoemen and traders normally returned to the Gold Coast on completion of their business, some of them settled locally on the Slave Coast. European factories on the Slave Coast often employed persons from the Gold Coast, some of whom were recruited as freemen, but others slaves or pawns. In 1694, for example, an English ship trading at Ouidah noted that most of the slaves employed in the English factory there were "Gold Coast negroes," who were considered superior soldiers to the local people, and also that, although there was no Dutch factory there, there was a settlement of persons "which call themselves Mine-men," who assisted Dutch ships.¹⁸ In this context, the "Minas," from their association with the Dutch, were evidently from Elmina specifically, as opposed to those in the English factory, who would have been recruited from Cape Coast, where the English headquarters on the Gold Coast was situated. In later references, however, the category "Mina" has been expanded to include both: a French source of 1716 thus describes the slaves in the English factory at Ouidah, as well as of a Dutch factory that had now been established in the Hueda capital Savi, as being "almost all inhabitants of the Gold Coast, or Minas."¹⁹ To the present day there are families in Ouidah who acknowledge their descent from canoemen from the Gold Coast who served the European factories there,²⁰ and these are still sometimes referred to as "Minas" in local usage.

In other cases, on the western Slave Coast, canoemen or traders from the Gold Coast established their own independent communities. The best-known instance is "Little Popo" (modern Aného, in Togo), so called in distinction from the original "Popo" (Grand-Popo, in Bénin). This was a settlement of canoemen from Elmina, first attested in the contemporary record in the 1650s, among the places on the Slave Coast visited by merchants from the Gold Coast, who came there by sea in canoes.²¹ Another Mina settlement on the Slave Coast was presumably (from its name) "Elmina Chica [Little Elmina]," alternatively called Adina (i.e., Edina, the indigenous name of Elmina), on the coast just west of the eastern boundary of modern Ghana, although this is not attested until the nineteenth century.²²

Other emigrations from the Gold Coast to the Slave Coast involved refugees displaced by the expansion of the kingdom of Akwamu in the 1680s. Numbers of refugees from the "Allampo" or Adangme country east of Accra settled east of the Volta, many of whom were absorbed into the Anlo state, in what is nowadays southeastern Ghana.²³ Further east again, in modern Togo, refugees from the Akwamu conquest of Accra in 1680 settled at Glidji, on the north bank of the lagoon across from the already established Mina settlement of "Little Popo," which acknowledged the authority of the displaced Accras at Glidji.²⁴ Europeans commonly referred to the Glidji kingdom, as well as its coastal port, as "Little Popo," but this is not in fact its indigenous name, which is "Gen" (or in French spelling "Guin"), which in turn is a variant of "Ga," the indigenous name of Accra.²⁵

The earliest detailed description of Little Popo, by an English visitor in 1687, explicitly distinguished between the ethnicity of the two communities, the royal capital inland occupied by the displaced "King of Accra," and the coastal village whose inhabitants were "all Mine people, or at least the rulers."²⁶ Down at least to the 1740s, the people of "Little Popo" were commonly referred to in European sources as "Accras," reflecting the dominance of the Ga community of Glidji.²⁷ From the 1720s, however, they were alternatively termed "Minas."²⁸ This latter usage may have reflected awareness of the composite character of the state, including both Accra and "Mina" (in the narrow sense of Elmina) elements, but more probably it reflects the extension of the term "Mina" to include people from the whole of the Gold Coast noted above.

In the first half of the nineteenth century, emigrations from Aného established two further "Mina" communities, Porto-Seguro (modern Agbodrafo) to the west and Agoué to the east (in modern Bénin).²⁹ The French Roman Catholic mission that operated in this area from the 1860s regularly employed the term "Minas." In this the missionaries claimed to be following local usage, but it seems likely that in fact they were reproducing that of the Brazilian settlers on the coast, with which the French mission was closely associated.³⁰ The first head of the mission, Fr. Francesco Borghéro, applied the name "Minas" to the entire western Slave Coast, including Keta (the main coastal port of Anlo) and Elmina Chica to the west and Grand-Popo to the east, as well as Porto-Seguro, Little Popo (including Glidji, as well as Aného) and Agoué.³¹ This extension of the term to include Anlo/Keta and Grand-Popo, however, was probably merely a confusion; at any rate, it did not persist.³²

The name "Minas" remained in use under French colonial rule in the twentieth century, and indeed down to the present, to refer to the inhabitants of Agbodrafo, Aného and Glidji, and Agoué; although there remains an awareness locally that it should properly be applied only to the Elmina element in Aného (and its offshoots), as distinct from the Gas at Glidji.³³ It is not quite accurate, however, to state as does Hall, that these people "call themselves" Mina.³⁴ Rather, this is an external, European coinage, which is nowadays used locally only when speaking in French. At a conference held at Aného in 2000, I recall a member of the local community insisting that "the Minas exist only for scholars," the self-appellation of the people in their own language being "Gen."

The Adangme settlers in Anlo and the Ga at Glidji spoke languages which were closely related to each other, but distinct from that of the Mina settlers at Aného, which was Fante, a member of the Akan group–though doubtless many of them were bilingual in both Ga/Adangme and Fante.³⁵ Over time, however, these immigrants became linguistically assimilated into the populations among whom they had settled, which were Gbe-speaking. The Adangme in Anlo thus adopted the local language, Ewe.³⁶ The Ga and Fante of Little Popo/Gen also adopted the local language, which in this area was probably Adja to the north of the lagoon but Hula, the dialect of Grand-Popo, to the south, but in this case giving rise to what is nowadays recognized as a distinct new dialect, Gen.

This process of linguistic assimilation, however, evidently took some considerable time. In the early nineteenth century, for example, it was still noted that at Aflao, just east of Anlo, "a mixture of Adampe [Adangme] and Kerrapay [Krepi, i.e., Ewe] is spoken, accounted for by the emigration of a large body of the former people."³⁷ At Agoué as late as 1859 the Methodist missionary Peter Bernasko, who was from Sekondi on the Gold Coast, was able to converse with a local chief in Fante, while in Glidji Ga was reportedly still being spoken as late as the late nineteenth century.³⁸ This persistence of bilinguality needs to be borne in mind in considering the formation of ethnic identities among transported enslaved Africans in the diaspora. To the extent that African-American "nations" were constituted primarily on the basis of common language, clearly many slaves taken from the western Slave Coast in the eighteenth century would have had in effect a choice of ethnicities, being able to communicate with both Akan-speakers (and/or Ga-Adangme-speakers) and Gbe-speakers.

How commonly persons from the "Mina" communities on the Slave Coast were sold into trans-Atlantic slavery is uncertain, though they are unlikely, as Hall notes, to have represented a significant proportion of transported slaves.³⁹ The ports of the western Slave Coast–Keta in Anlo to the west and Grand-Popo to the east, as well as Little Popo, themselves accounted for only around 6% of slave exports from the region.⁴⁰ However, their inhabitants were more likely to be enslaved through capture in warfare or kidnapping, and to be sold elsewhere on the coast. In fact, the only reference of which I am aware to "Minas" as an ethnicity of slaves purchased by Europeans on the Slave Coast is the account of the French captain Des Marchais describing the slave trade at Ouidah, the principal port in the region, in the 1720s. The "Minois" are here distinguished from two Gbe-speaking groups, the "Aradas," i.e., Allada, the dominant state in the region prior to the rise of Dahomey, and "Foin [Fon]," the people of Dahomey itself, immediately inland from Allada, but otherwise no indication is offered of their identity. The published version of this text, edited by Père Labat, explains that the "Minois" were "from the Kingdom [sic] of Saint George of the Mine," i.e., Elmina, but this is evidently no more than a misinformed editorial gloss.⁴¹ More likely, in this context these are the "Minas" of Little Popo, as Hall also seems to agree.⁴² Some (probably small) proportion of those identified as "Minas" in the Americas therefore probably did come from the "Mina" communities on the Slave Coast.

III

With respect to the nomenclature of African ethnicities in the Americas, it should be stressed at the outset that it is incontestable that in some contexts the term "Mina" did mean specifically people from the area of the Gold Coast. This is clear, for example, from the analysis of African ethnicities by Christian Oldendorp, a missionary active in the Danish West Indies in the 1760s.⁴³ Oldendorp explicitly and unambiguously linked the "Amina nation" to the Gold Coast. He clearly meant by this a larger grouping than Elmina town, describing it as "the most powerful nation on this [Gold] coast," whose territory was "very large and has many villages."

The only communities specifically named by Oldendorp as sub-groups of the "Amina" are the Kwawu ("Quahu"), located in the eastern interior of the Gold Coast, who are described as "one tribe of the Amina;" and the Accra ("Akkran"), "a people who belong to the Amina, whose language they also understand." This latter phrasing suggests awareness that the language of Accra, Ga, is distinct from Akan, but implies bilinguality on the part of the Accras. On the other hand, the Kyerepong ("Akripon"), another eastern Akan-speaking group, which formed part of the state of Akwapim, although said to "also speak the language of the Amina," seem to be regarded as a separate people. Other Akan-speaking groups mentioned, including the Asante ("Sante"), Akani ("Okkan"), Etsi ("Atti"), Assin ("Assein"), Adanse ("Adansi"), and Akyem ("Akkim"), seem also to be regarded as neighboring but distinct from the "Amina."

By distinction, in Oldendorp's conception the Gbe-speaking peoples of the Slave Coast are classified separately, as the "Papaa [i.e., Popo] nation." This was primarily the name by which Europeans referred to two communities on the western Slave Coast, mentioned earlier, "Great Popo" and "Little Popo" (nowadays Grand-Popo and Aného), although this was not their indigenous name (that of Grand-Popo being Pla or Hula, and that of the oldest quarter of Aného being Plaviho, "[House of] Little Pla"), and its origin is unclear: it may be a name by which the Yoruba to the east referred to the Gbe-speakers (or some of them), picked up by the Europeans in the earliest days of the trade.⁴⁴ Its use in a wider sense, to include persons from other communities in the region, is also attested in the British West Indies, where it was applied generically to slaves exported through Ouidah.⁴⁵ Oldendorp lists as peoples which "belong to the Papaa kingdom": Allada ("Arrada"), which had been the most important eastern Gbe-speaking kingdom down to the early eighteenth century, and the Fon ("Affong") of Dahomey, which conquered and replaced it in the 1720s; Tori ("Attolli"), a small state adjoining Allada on the south, also conquered by Dahomey; and Kpessi ("Apassu or Apeschi"), a Gbe-speaking community in the west (in modern Togo); and also the "Nagoo," i.e. the Nago or Yoruba, neighbors of the Gbe-speaking peoples to the north and east. The inclusion of the Yoruba, whose language is in fact distinct from Gbe, here again probably reflects bilinguality. The Nago were also regarded as a sub-group of the "Popos" in the British West Indies.⁴⁶ The people of Hueda ("Fida"), the coastal kingdom which controlled the port of Ouidah, also conquered by Dahomey in the 1720s, by implication also belonged to the "Popo nation," since they are described by Oldendorp as closely related to the Allada.⁴⁷ On the other hand, two of the major western Gbe-speaking groups who are named, the Adja ("Atje") and Ouatchi ("Watje"),⁴⁸ seem to be regarded as separate from the "Popos."

"Mina" was, however, as Hall correctly notes, certainly a term which was applied in Brazil, in some contexts, to speakers of Gbe languages. The "Casa das Minas [House of the Minas]" in São Luis in the province of Maranhão, for example, is a cult-house of specifically Dahomian origin, venerating spirits associated with the royal dynasty of Dahomey.⁴⁹ A vocabulary of "the common language of Mina" compiled in the province of Minas Gerais in 1741 is likewise of a Gbe language, basically Fon, although incorporating some elements from other dialects.⁵⁰ The vocabulary itself gives the vernacular equivalent of "gente mina [Mina people]" as "Guno," i.e,. Gunnu, "Gun people," this being the name of the most southeasterly Gbe-speaking group, straddling the modern Bénin/Nigeria border (including the towns of Porto-Novo in Bénin, and Badagry in Nigeria).⁵¹

In Rio de Janeiro a "Congregation of the Mina Blacks" formed in the 1740s comprised a number of different "nations," though all reportedly speaking a common language: "Dagomés," "Makis," "Sabarus," "Agolins," and "Ianos"—i.e., Dahomey; Mahi, its immediate northeastern neighbour; Savalou, north of Dahomey; Agonlin, to the east (whose principal town is Covè); and Oyo (Ayonu, "people of Oyo," in Fon), in the interior to the northeast. Of these the Oyo speak Yoruba rather than a Gbe dialect; their incorporation with the Gbe-speaking "Minas" presumably here again reflects bilinguality. In 1762 the congregation split, with the Mahi, Savalou, Agonlin, and Oyo breaking away to form a separate "Congregation of the Mina Blacks of the Kingdom of Maki."⁵² It is noteworthy that this split did not occur along linguistic lines (Gbe-speakers against Yoruba), but instead followed political alignments in Africa, where the four secessionist groups were all enemies of Dahomey and/or victims of its territorial expansion.

These instances, however, seem to reflect the use of "Mina" as a very general term, including a number of different peoples, rather than that the term meant Gbe-speakers in particular. The classic study of Africans in Brazil, by Nina Rodrigues, written in 1906, observes that in Maranhão all Africans were referred to generically as "Minas."⁵³ In Rio de Janeiro in the eighteenth century, three principal categories were employed to designate the origins of enslaved Africans: "Guiné [Guinea]," "Mina," and "Angola." These were clearly broad geographical terms, rather than specific ethnicities: "Guiné" and "Mina" evidently representing respectively the western and eastern sections of the West African coast, and "Mina" being by implication the entire West African coast from the Ivory Coast eastwards.⁵⁴ Given known patterns of shipping and shifting African sources of slaves in the trans-Atlantic trade, the Rio "Minas" would probably have been predominantly Gbe-speaking in the mid-eighteenth century, but would have become predominantly Yoruba-speaking by the nineteenth century.⁵⁵

In the province of Bahia, however, where the African population was overwhelmingly West African and ethnic terminology was consequently more differentiated, the "Minas" were only one among several West African "nations" recognized, distinct both from the Gbe-speakers, who were here called "Gêges," or in modern spelling "Jejes" (and who included both the Fon/Dahomians and the Mahi),⁵⁶ and from the Yoruba, here also called "Nagôs."⁵⁷ According to Rodrigues, in Bahia two sub-groups of the Minas were distinguished: the "Minas-Santés," i.e., the Ashanti, and the "Minas-Popos," in the latter of whom he correctly recognized the Akan/Ga speakers who had settled east of the Volta at Little Popo.

In Cuba likewise, in the classic inventory of African ethnicities compiled by Fernando Ortiz, originally published in 1916, the "Minas" also seem basically to represent people from the Gold Coast.⁵⁸ At any rate, the bulk of the Gbe-speakers here were again put into a distinct category, the "Ararás" (a variant of the name "Allada"), whose component groups included the Dahomians ("Dajomé"), Mahi ("Magino"), Savalou ("Sabalú"), and perhaps Covè (Cuévano"), as well as others not immediately recognizable ("Agicon," "Nezeve"); and the Yoruba were also separate, here known as "Lucumís". The existence in Cuba too of a brotherhood (cabildo) of the "Minas Popó Costa de Oro," listed in 1909, evidently refers again to the Gold Coast immigrants settled at Little Popo, as Hall also concludes.⁵⁹ An earlier cabildo, of the "Mina Guagui" nation, documented in 1794-1812, may also refer to Little Popo, under its indigenous name of Genyi, "Gen country."⁶⁰

In the French Caribbean colony of Saint-Domingue (modern Haiti) in the second half of the eighteenth century the "Minas" were likewise distinct from the main body of Gbe-speakers, here called generically "Aradas" or "Radas" (i.e., "Allada"), though sometimes particular Gbe-speaking groups were separately identified, notably the Adja ("Adia") and the Hueda ("Foeda"), the latter probably referring here to the community of refugees displaced by the Dahomian conquest of the Hueda kingdom in the 1720s, now settled to the west of their original homeland, in the territory of Grand-Popo, and known as Hueda-Henji.⁶¹

However, interpretation of the Saint-Domingue data is complicated by the fact that they include references to "Coromantis" (and other variants), as well as "Minas," the former being a term regularly used in English sources for slaves from the Gold Coast.⁶² It is unclear whether these are alternative names for the same group, or a distinction between them is intended; and if the latter, what the basis of this distinction might be. Conceivably the use of these two terms might reflect a differentiation of ethnicity between Akan-speakers from the western and central (and interior) Gold Coast, and Ga-Adangme from the eastern Gold Coast, although if so it is not clear which name would have designated which group. Alternatively, and perhaps more probably, however, "Coromantis" might represent slaves from the Gold Coast delivered by English, as opposed to French, ships.

Gwendolyn Hall originally formulated her argument in relation to ethnic designations in the African-American population of Louisiana, of which she has compiled an invaluable database.⁶³ In Louisiana as well, the "Minas" were a distinct category from the "Aradas." Hall implicitly argues that the "Minas" here represented the western Gbe-speakers–Ewe, Ouatchi, Adja, Gen, Hula, etc.–as distinct from the Fon of Dahomey and their immediate neighbors; but if so, their numerical strength in Louisiana (nearly three times the numbers of the "Aradas")⁶⁴ is surprising, given the very limited role which the ports of the western Slave Coast played in the Atlantic slave trade, as noted earlier. Also, since presumably at least some of the Africans in Louisiana were from the Gold Coast, it is not clear where, if not among the "Minas," they were classified. It may further be noted that a leader of the "Mina" community in Louisiana in the 1790s to whom Hall refers, Antonio Cofi Mina, bore a transparently Akan name, Kofi, given to male children born on a Friday.⁶⁵ This is not inconsistent though with his having originated from the Slave Coast (or indeed, with his being a Gbe-speaker), since the use of such Akan day-names spread along the coast with the Mina diaspora as far east as Ouidah.⁶⁶

IV

Hall's argument for classifying the "Minas" as Gbe-speakers leans heavily on a passage in the work of Alonso de Sandoval, a Spanish missionary active in evangelizing African slaves in what is today Colombia in the early seventeenth century, which discusses at considerable length the different ethnicities (he uses the term castas, "breeds") of African-born slaves he encountered in Spanish America.⁶⁷

The critical passage of Sandoval's work on which Hall's argument is based, referring to slaves brought to America from/through the island of São Tomé, reads as follows: "[t]he breeds [castas] which [they] ordinarily bring from those parts are Minas, Popoos, Fulaos, Ardas, or Araraes, which is the same [lit. "which is all one," que todo es uno], Offoons; also Arda breed [casta Arda]; . . . ."⁶⁸ The intended syntax, and therefore the precise meaning, of this passage is not entirely clear. In particular, does the phrase "which is the same" refer only to the immediately preceding terms "Ardas, or Araraes," or, as Hall assumes, to the entire preceding list of "Minas, Popoos, Fulaos, Ardas, or Araraes?" And where do the immediately following "Offoons" fit into this schema? Hall's interpretation implicitly excludes them from the "Ardas," but the list culminates in a repetition of the name "Arda" in an implicitly generic form, "casta Arda," which presumably includes all the preceding groups. It seems likely that in its first occurrence "Ardas, or Araraes" means specifically Allada, but in its second, "Arda breed," represents a generic term, including the other neighboring and/or related groups specified earlier. In either case, however, Sandoval's inclusion of the "Minas" within the wider "Arda" group seems clear.

The various groups named as making up Sandoval's "Arda breed" are evidently arranged at least loosely in geographical order from west to east. Setting aside for the moment the problematic "Minas," the first name, "Popoos," as has been seen, was one given by Europeans to a coastal port on the western Slave Coast, nowadays Grand-Popo, although in the Caribbean it was sometimes given a wider application, to include other Gbe-speaking peoples. Sandoval's own account, while clearly referring to this coastal port, represents it as included in a much larger "kingdom of the Popos" extending west to the river Volta, whose ruler resided some distance inland.⁶⁹ This probably alludes to Tado, the center of the Adja people, from which several other Gbe-speaking groups, including Grand-Popo, trace their traditional origins.⁷⁰ The second name "Fulao," which is also attested in other sources as a place of trade in the early seventeenth century, seems likely to represent Pla/Hula, which was, as noted earlier, the indigenous name of Grand-Popo, but from Sandoval's more detailed account of its location (midway between Popo and Allada),⁷¹ it clearly refers here to Glehue (modern Ouidah), later the coastal port of the kingdom of Hueda, which, at least according to some versions of local tradition, was originally a settlement of Hula immigrants from Grand-Popo.⁷² The name which follows Allada, "Offoons," presents more of a problem. An obvious identification would be with Fon, i.e., Dahomey (cf. Oldendorp's "Affong"), north of Allada, but this was not in Sandoval's time a place of any great importance, if indeed it existed at all.⁷³ Alternatively, however, it may be a New World coinage, deriving from a common greeting, in modern Fon a fon [dagbe]?, "Have you woken [well]?"⁷⁴ This would tend to suggest that "Offoons" is an alternative generic name for the wider "Arda breed," although admittedly Sandoval's syntax does not contrive to convey this clearly.

But what of Sandoval's "Minas"? The reference is unlikely to be to the "Mina" settlement at Little Popo, which, as seen above, probably did not yet exist. Hall implicitly identifies them with the westernmost Gbe-speaking groups, the Ewe and Adja–although on the face of it, given the geographical extent which Sandoval attributes to "the kingdom of the Popos," both the Adjas and the Ewes would seem to be subsumed into his "Popos."⁷⁵ However, Sandoval's own detailed geographical description elsewhere shows that he located the Minas on the Gold Coast: "From this Cape [Palmas] to the nation, which we call Mina . . . is a hundred and twenty leagues, peopled with many large villages, fifty leagues from which is the River Volta," which in turn marked the beginning of the "Kingdom of the Popos."⁷⁶ The distances mentioned seem to indicate Elmina itself, but evidently Sandoval"s "Mina nation," with its "many large villages" was a wider area, as it was for Oldendorp later. On the face of it therefore, it would seem difficult to cite Sandoval's text in support of the view that "Mina" designated Gbe-speakers, or indeed any other group living east of the Volta.

V

A key assumption in Hall's argument is that the "Mina" slaves referred to by Sandoval in the 1620s could not be from the Gold Coast, because at the time when he wrote no slaves were being purchased by Europeans from that area.⁷⁷ On this, Hall's view accurately represents the consensus of current scholarly literature. In the earliest period of European trade on the Gold Coast, when it was dominated by the Portuguese, slaves were in fact an import rather than an export, purchased by the Portuguese from elsewhere on the West African coast (especially the Benin area) to be exchanged on the Gold Coast for gold.⁷⁸ This import trade apparently still continued in the early seventeenth century.⁷⁹ In consequence, it is commonly believed that slaves did not become an item of export from this area until much later. The most detailed study of seventeenth-century European trade on the Gold Coast, by Robert Porter, suggests that slave-trading on the Gold Coast was initiated only in the 1640s, by English traders. He in fact explicitly identifies a cargo of 100 slaves purchased by an English ship at the beginning of 1647 at Winneba on the eastern Gold Coast, as "the first occasion on which slaves were obtained from the Gold Coast by either the English or the Dutch."⁸⁰

This view, however, is demonstrably inaccurate. When the Dutch West India Company entered the slave trade after 1635 (to supply Pernambuco in Brazil, recently taken from the Portuguese), some of the slaves it carried were obtained on the Gold Coast. The first two Dutch ships to deliver slaves to Brazil in 1637 purchased 38 slaves from the Gold Coast out of a total of 874, the remainder being taken from Allada and the Bight of Biafra.⁸¹ Over the period from 1637 to 1645, around 17% of the slaves that the Company shipped from West Africa to Brazil came from the Gold Coast, an average of about 200 per year.⁸² Most of these were recorded as obtained from Mina, with smaller numbers from Mouri, the original Dutch headquarters 14 kilometers to the east, and from Accra.⁸³ The reference to Mina specifically may be misleading, however, since it is possible that slaves shipped from Mina had been initially collected elsewhere on the coast.

It is likely in fact that most of the slaves that the Dutch obtained from the Gold Coast in this period came from its most easterly section, especially from Accra. The Dutch geographer Olfert Dapper, writing in 1668, but probably reporting earlier (1640s?) conditions, states that Accra supplied not only gold, but also around 300 slaves per year, though he does not state explicitly that these were purchased by the Dutch; and also that the Dutch "sometimes" traded at Beraku, between Winneba and Accra, for slaves that were brought there by African merchants from Accra.⁸⁴

This evidence of Dutch slaving on the Gold Coast in the 1630s and 1640s, however, might represent a recent innovation rather than a continuation of patterns of commerce from the earlier Portuguese period, to which Sandoval was referring. Sandoval himself, in his very brief reference to trade, does not mention slaves as sold on the Gold Coast, but he also fails to mention gold, referring only to "ivory, cloths and other things in which their land abounds." This description can only be understood as referring to the Ivory Coast to the west, which was a source of African-made cloth, as well as ivory, rather than the Gold Coast proper.⁸⁵

However, other Portuguese documents of the early seventeenth century do indicate some trade in slaves on the Gold Coast. A report of 1607 describes the trade of the Costa da Mina (from the context, referring here to the entire coast from Cape Palmas to the Volta, so including the Ivory Coast as well as the Gold Coast proper) as consisting of "much gold, ivory, silver, malagueta [pepper], gum, civet, cats [live civet-cats], wax, slaves." Admittedly, this passage shows some signs of confusion: the reference to silver, if not merely an error, can refer only to trade with Europeans on the coast for specie in the form of silver dollars; while malagueta, gum and wax were normally obtained further west along the coast, though maybe some of these commodities were collected from elsewhere and bulked up at Elmina for final shipment. A later document of 1620, however, proposing countermeasures against Dutch piracy and commercial competition, is more precise. In advocating an expedition to destroy the Dutch fort at Mouri, it alludes to its interference with Portuguese ships which came to buy "black slaves, ivory and oil [presumably palm-oil], gum and civet."⁸⁶ How early this trade in slaves began is uncertain; the Portuguese had made contact with Accra for the first time as early as the 1520s, and maintained a fort there between 1557 and 1572, but the (admittedly fragmentary) sixteenth-century accounts of trade there refer only to gold.⁸⁷

In the light of this evidence, it seems probable that small numbers of slaves were already being purchased on the Gold Coast in the early seventeenth century, if not earlier, specifically from the eastern Gold Coast around Accra. Consequently, the "Mina" slaves whom Sandoval encountered in South America could have been from the Gold Coast, as the term "Mina" normally implied. They would be for the most part have been more likely Ga-Adangme rather than Akan-speakers, however.

This conclusion of course serves to raise the further question of why, if Sandoval's "Minas" spoke Ga-Adangme, they should have been classified with the "Arda breed," who were otherwise all speakers of Gbe languages. In answer, it should first be noted that this is by no means a unique instance. Sandoval himself, in discussing the "Lucumies" to the east, who were basically Yoruba-speakers, refers to the Bariba of Borgu, the northwestern neighbors of the Yoruba in Africa, as a sub-group of them, "Lucumies Barbas," although they in fact speak a different language.⁸⁸ In Cuba later, as reported by Ortiz, the "Lucumí" or Yoruba nation likewise included the Tapa ("Tacua"), which is the Yoruba name for the Nupe, their northeastern neighbors, who likewise speak a different language.⁸⁹ Contrariwise, in the Danish and British West Indies, as noted earlier, the Nago/Yoruba were included within the "Popo" nation, who were otherwise Gbe-speakers, while in Rio de Janeiro the Yoruba-speaking Oyo joined Gbe-speakers in the "kingdom of Mahi."

Such instances of the aggregation of peoples who were linguistically distinct but geographically adjacent (in Africa) are best explicable on the assumption that many people in these groups were bilingual, so that smaller groups could be assimilated into larger ones in the Americas. If the "Mina" slaves of the early seventeenth century were specifically from the eastern Gold Coast, divided from the westernmost Gbe-speakers by the Volta river, which was easily crossed by canoe, the existence of a sufficient degree of bilinguality for this process of incorporation seems a reasonable hypothesis.

In the longer run, however, the focus of the slave trade within the Gold Coast shifted westwards, with Cape Coast and Anomabu becoming the principal ports of embarkation: over the entire history of the trade, Accra and "Alampo" together accounted for only around 9% of shipments from the Gold Coast.⁹⁰ By the eighteenth century, therefore, most of the slaves taken from the immediate hinterland of the Gold Coast would have spoken Akan rather than Ga-Adangme, and more critically would have been from places geographically non-contiguous with the Gbe-speaking peoples east of the Volta. Thus the later "Mina" in Bahia, the West Indies, and Louisiana, in addition to being sufficiently numerous to constitute themselves as a separate "nation," would have been less able to assimilate into the Gbe-speaking groups ("Popo," "Jeje," "Arará," or "Rada").