Jump to ContentJump to Main Navigation
Slavery and African Ethnicities in the AmericasRestoring the Links$

Gwendolyn Midlo Hall

Print publication date: 2005

Print ISBN-13: 9780807829738

Published to North Carolina Scholarship Online: July 2014

DOI: 10.5149/9780807876862_hall

Show Summary Details
Page of

PRINTED FROM UNIVERSITY OF NORTH CAROLINA PRESS SCHOLARSHIP ONLINE (www.northcarolina.universitypressscholarship.com). (c) Copyright University of North Carolina Press, 2018. All Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a monograph in CSO for personal use (for details see http://www.northcarolina.universitypressscholarship.com/page/privacy-policy). Subscriber: null; date: 21 June 2018

Making Invisible Africans Visible: Coasts, Ports, Regions, and Ethnicities

Making Invisible Africans Visible: Coasts, Ports, Regions, and Ethnicities

(p.22) Chapter Two Making Invisible Africans Visible: Coasts, Ports, Regions, and Ethnicities
Slavery and African Ethnicities in the Americas

Gwendolyn Midlo Hall

University of North Carolina Press

Abstract and Keywords

This chapter attempts to make Africans, who played a crucial role in the formation of new cultures throughout the Americas, more visible. It discusses several studies of African diaspora in the Americas. Such studies can provide a better understanding of when particular African ethnicities started to become victims of the Atlantic slave trade, and of their final destinations in the Americas. The chapter also highlights the Trans-Atlantic Slave Trade Database, the best recent synthesis of the work of the quantification school of Atlantic slave trade studies.

Keywords:   Africans, cultures, Americas, ethnicities, Atlantic slave trade, Slave Trade Databases

We are called, we are named just like a child is given a name and who, like a child, do not have a say in the choice of our own name.

—Olabiyi Yai, in African Ethnonyms and Typonyms, 1978

Qui se disent leur nation Bambara. [They say their nation is Bambara.]

—From a slave sale document in Louisiana, 1799

Studies of the African diaspora in the Americas began mainly during the early twentieth century among anthropologists: most notably Nina Rodriguez in Brazil and Fernando Ortiz in Cuba, and then a generation later by Frances and Melville Herskovits in the United States. Fieldwork was a primary methodology. They often studied communities of African descent in the Americas, linking them with particular regions or ethnicities in Africa by seeking out shared cultural traits. Their work is very useful, informative, and fascinating, and their methodologies are more sophisticated than some recent critics have been willing to acknowledge. Nevertheless, their approach poses problems for the study of the African diaspora in the Americas. Religion, worldview, and esthetic principles—including the styles and social role of the plastic arts, music, musical instruments, and dance—are among the most enduring and resilient cultural heritages. But they are also the most generalized. There are many common cultural features in Africa. It is not always easy to disaggregate which features are characteristic of any particular ethnicity or region. Very few scholars are familiar with a substantial number of African languages. Some seize on a word or a name they know and extrapolate it broadly to prove the presence and influence of a particular African ethnicity in the Americas. But the same or similar names and words exist in several African languages and can have the same, a similar, or a different meaning. The result is sometimes romanticized and inaccurate views of the influence of particular African ethnicities and languages. Swahili becomes the “African language,” but few speakers of Swahili were brought to the (p.23) Americas. “Yoruba” becomes the African ethnicity, although “Yoruba” presence in the Americas before the late eighteenth century (recorded as “Nago” or “Lucumi” in American documents) was not very substantial. Except for Louisiana, where Nago were 4 percent of identified ethnicities, the presence of Yoruba in the United States was insignificant. They were important in St. Domingue/Haiti (Nago) after 1780, in Cuba (Lucumi) during the nineteenth century, and, most significantly, in nineteenth-century Bahia, Brazil (Nago).

During the last few decades, anthropologists have taken the lead in denying the significance of African ethnicities in the Americas. Their arguments include questioning both the accuracy and the significance of these designations in American documents as well as the very existence of African ethnicities in Africa. Based on his studies of documents in Cuba, Moreno Fraginals has written:

The main trend at present (perhaps due to the influence of Roger Bastide) is to attach little importance to ethnic references; the assumption is that they were imposed arbitrarily by the slave-trader so as to trick his clients, or that this must have been affected by the traders' geographical ignorance of the zone in which they were operating or by pseudo- scientific prejudices of the time. We maintain the complete opposite. Our basic assumption is that the slave trade was the business that involved the greatest amount of capital investment in the world during the eighteenth and early nineteenth centuries. And a business of this size would never have kept up a classificatory scheme had it not been meaningful (in overall general terms, in keeping with reality) in designating in a very precise way the merchandise that was being traded.1

We will see that there is strong evidence that Africans often identified their own ethnicities recorded in American documents. The knowledge and perceptions of the slave traders had much less to do with this process than scholars have assumed.

The assumption of timelessness is a basic methodological flaw of anthropology. But it is a problem shared by many historians of the African diaspora. They sometimes project patterns in times and places with which they are familiar to all times and places. In their eagerness to make sense of vast, complex data, scholars sometimes embrace abstract, generalized concepts, which obscure rather than reveal the past. Studies of the African diaspora in the Americas need to be concrete and contextualized. Peoples and their cultures evolved and changed on both sides of the Atlantic. There was no single pattern of creolization either in Africa or in the Americas. Slavery as well as other forms of exploitation of labor varied greatly over time and place. (p.24)

Making Invisible Africans Visible: Coasts, Ports, Regions, and Ethnicities

Map 2.1. African Ethnicities Prominent in South America, 1500–1900

Making Invisible Africans Visible: Coasts, Ports, Regions, and Ethnicities

Map 2.2. African Ethnicities Prominent in North America and the Caribbean, 1500–1900

The study of transatlantic slave trade voyages is not enough. These documents do not list and rarely mention African ethnicities. Documents generated by these voyages can at best give us African coasts or ports of origin and/or ports in the Americas where slaves were first sold. More rarely, the numbers, genders, or age categories of the slave “cargoes” were listed. The proportions of African ethnicities exported from African coasts and ports changed over time. After enslaved Africans arrived in American ports, they were normally sold and then often transshipped to other places, sometimes outside the colony or country where they first landed. Studies based entirely (p.26) on transatlantic slave trade voyages need to be supplemented by studies of the transshipment slave trade as well as other types of documents generated over time in various places in the Americas. These studies can help us discover which Africans from which regions, ethnicities, and genders found themselves where, when, and in what proportions. Such studies far transcend local history. They have important implications for the study of the African diaspora throughout the Americas as well as for African history. Patterns of introduction of African ethnicities into the Americas are mirror images of their export from Africa. They can help us understand when particular African ethnicities started to become deeply victimized by the Atlantic slave trade as well as where they were finally located.

Ethnicities exported from the various African coasts defined and named by Atlantic slave traders changed. The same ethnicities were exported from two or more of these coasts over time. The coastal origins of transatlantic slave trade voyages can give contrasting results with ethnic descriptions of Africans derived from documents generated in the Americas. This book tries to avoid making artificial separations among Africans exported from more than one of these African coasts. It is often best not to treat African coasts defined and named in various ways by European slave traders as entirely separate regions. Firm boundaries among some of these coasts are an illusion. Some of Curtin's African coastal definitions have been cogently challenged. Boubacar Barry has convincingly argued that it is often awkward and questionable to separate “Senegambia” from “Sierra Leone.” Curtin's “Windward Coast” designation presents problems discussed below.

I argue that our best evidence for the distribution of Africans at their final destinations in the Americas is in documents containing “nation” descriptions of enslaved Africans, despite the fact that these ethnic designations are sometimes unclear and equivocal. The Mexican scholar Gonzalo Aguirre Beltran published a pioneering work about African ethnicities in North America. He made systematic studies of American documents over time and place and linked them to developments in Africa as well as with patterns in the transatlantic slave trade. His work focused on Mexico and the Caribbean through the seventeenth century. James Lockhart studied African ethnic designations in Peru during the mid-sixteenth century, but his data are thin. Gabriel Debien and his colleagues studied African ethnicities listed in documents in the French West Indies dating almost entirely from the eighteenth century. Philip D. Curtin relied heavily on the work of these three scholars.2 Colin Palmer's book about Mexico discusses African ethnicities there.3 Subsequent studies have usually been based on larger numbers of African ethnic designations listed in American documents. During the past three decades, (p.27) some scholars have made major contributions to our knowledge of documents listing African ethnic designations. Their work is studied and cited throughout this book.

The introduction of Africans into the Americas was a complex process. It involved vast regions of the world over nearly four centuries. It must be placed within the context of ever changing patterns over time and place on both sides of the Atlantic as well as at sea. Primary sources exist, and very important books and articles have been published in major European languages: English, French, Spanish, Portuguese, Dutch, German, Danish, and Russian. There are important sources published in Arabic as well as in African languages written in Arabic script.4 There are invaluable oral histories as well. Many studies have been published about various regions of Africa and the Americas at specific times and places. There are other fine scholarly works about the slave trade of specific European and American nations.

Two important syntheses of the Atlantic slave trade and slavery in the Americas were recently published in English. Both books—one by Robin Blackburn, the other by Hugh Thomas—discuss African slavery and the Atlantic slave trade over the centuries, the longue durée. They rely on the literature in major European languages, avoiding the narrow English focus of some historians. Organizing and synthesizing this vast, difficult, and unwieldy body of knowledge is a major achievement. But neither of these books discusses the African ethnicities involved in the Atlantic slave trade, limiting their usefulness for this study.5

Many of the most prominent historians of the Atlantic slave trade, especially those writing in English during the past thirty-five years, have been fascinated by quantitative studies. The first major work using quantitative methods was Philip D. Curtin's pioneering book The Atlantic Slave Trade: a Census, published in 1969. This bold, imaginative, impressively researched work has inspired more than one generation of historians to search for greater precision in estimating the numbers of Africans loaded aboard slave trade ships in Africa and landed in the Americas. Some of these historians were Curtin's distinguished students. But until recent years, few of them followed Curtin's interest in African ethnicities.

There are conflicting interpretations of the number of Africans removed from Africa and landed in the Americas. Joseph E. Inikori has pointed out that many voyages involving smuggling and piracy were obviously undocumented and that many slave trade voyages went directly from the Americas to Africa and back, bypassing Europe. Many of them have been overlooked or undercounted by historians working mainly in large, centralized archives in Europe. Inikori raised Curtin's estimate of 9.55 million Africans put on land in (p.28) the Americas to about 15.4 million.6 Hugh Thomas gives a higher estimate of the numbers of transatlantic slave trade voyages, raising David Eltis's estimate from approximately 40,000 voyages to 50,000. Thomas estimates that about 2 million enslaved Africans left for the Americas from Senegambia and Sierra Leone alone, a much higher figure than most historians allow. This book confirms and explains the substantial undercount of voyages from Greater Senegambia.7

African, Spanish, Portuguese, and Latin American scholars have appreciated Curtin's work. But some of the most prestigious among them have remained reasonably skeptical about the limits of quantification, regardless of how sophisticated, when it is based on uneven and inadequate data. They, as well as Curtin, have emphasized the importance of understanding the historical context in which these voyages took place. Some have pointed out that for various reasons these Atlantic slave trade documents reported falsified information. Many documents are still to be found or are missing. Other voyages were never documented. Thus, quantification alone is a limited tool in the absence of a deepening of our knowledge, understanding, and appreciation of studies relying on unquantifiable sources.8 Many fine historians have assiduously collected, translated, annotated, and prudently used traditional sources despite changing fads and often without the support and recognition they deserve. Their work, some of it cited in this book, is essential in putting flesh on the bare bones of quantitative studies and ensuring that they are used judiciously.

The TransAtlantic Slave Trade Database is the best recent synthesis of the work of the quantification school of Atlantic slave trade studies.9 It is an extremely useful compilation of research carried out mainly during the past thirty years and, especially important, a computerization of much of the known information about transatlantic slave trade voyages. It was published as a relational database on a compact disk with a search engine, which allows for rapid answers to questions about documented and studied transatlantic slave trade voyages as well as comparison of broad patterns over time and place. Some of these results are of transcendent importance. For example, they indicate that individual transatlantic slave trade voyages collected their “cargoes” overwhelmingly from the same African coast, often from only one or two ports. A very important, valid conclusion is that African ethnicities were not nearly as fragmented by the transatlantic slave trade as scholars as well as the wider public have long believed.10

Obviously, many important questions about the transatlantic slave trade cannot be answered by studying The TransAtlantic Slave Trade Database alone. Missing and uneven data about voyages as well as inaccurate (p.29) Interpretations of geographic terms have led at times to inaccurate and distorted conclusions. Contrasting uses of the geographic term “Angola” by British and other European slave traders and “Angola” as an ethnic designation by slave owners add to the confusion. The documented British slave trade has been extensively studied, overemphasizing this trade compared to that of other countries. Manolo Florentino has pointed out that more than twice as many enslaved Africans were brought to Brazil than to the British colonies. Africans arriving on voyages to the British Caribbean were much more likely to be transshipped to colonies of other powers, especially to Spanish and French colonies, than Africans arriving in Brazil, although a relatively small number were transshipped south from Brazil to Spanish Rio de la Plata (now Argentina and Uruguay) and west to Upper Peru (now Paraguay and Bolivia), Lower Peru, and Chile.11 The deficit in Portuguese and Brazilian voyages in The TransAtlantic Slave Trade Database is widely recognized, including by its creators. Post-publication revisions have already added about 7,000 voyages, most of them Portuguese and Brazilian.

The truth is, we do not know and probably never will know how many enslaved Africans were loaded onto Atlantic slave trade ships in Africa and how many disembarked in the Americas. How representative are the voyages included in The TransAtlantic Slave Trade Database? My research into unusually rich documents housed in Louisiana archives reinforces the conclusion that direct voyages between Africa and the Americas have been under- counted. More studies of other American ports should be revealing. This database is a very large sample showing important trends. It should neither be dismissed nor turned into a fetish used uncritically. Its potential for answering pivotal questions has only begun to be explored.

African coasts defined in Philip D. Curtin's influential work have remained largely in vogue and were adopted in full by the creators of the TransAtlantic Slave Trade Database. Curtin's demarcation of territories along the West African coast clearly proceeded from the outside perspective of European and American slave traders rather than from the indigenous view of the peoples inhabiting various African regions. Distinguished historians have challenged two of these coastal definitions. Boubacar Barry defines the region between the Senegal and the Sierra Leone rivers as Greater Senegambia. Cultural inter- penetration among ethnic groups throughout this region was intense over many centuries. Languages of two major language groups, Mande and West Atlantic, are spoken throughout this vast region.12 Mandingo was the lingua franca. Separating Senegambia and Sierra Leone creates awkward problems when ethnic designations in American documents are used to define coastal origins. Timing is an important factor. Some ethnicities found frequently in (p.30) American documents—for example, the Fulbe (listed in documents as Fula, Fulani, Poulard, Peul)—migrated south and east over vast distances. During the 1720s, because of the growing desiccation of their grazing lands, many Fulbe migrated with their herds of cattle from the banks of the Senegal River and established the Fula Alamate in Futa Jallon. By 1780, they were heavily engaged in warfare with the Mandingo in the region defined by Curtin as Sierra Leone. Many Fulbe and Mandingo captives were exported from ports in Sierra Leone.13

The same ethnicities were exported from more than one region or coast. The American data can give us ethnicities present at a particular time and place. They may not correspond to African ethnicities sent from any particular port or coast. The study of African ethnicities in the Americas requires mining of data from both sides of the Atlantic over time. For example, after 1750 African ethnicities normally associated with Senegal narrowly defined arrived in the Americas on voyages that had left from Sierra Leone. Atlantic slave trade voyages from Sierra Leone increased more rapidly than voyages from Senegal narrowly defined. More Temne, Kisi, and Kanga appear in American documents during the last half of the eighteenth century. But the percentage of ethnicities associated with Senegambia narrowly defined does not in fact decline in documents recorded in the Americas. Fulbe and Hausa were exported from African coasts located further south and east over time: from the Slave Coast during the eighteenth century; from the Bight of Biafra as well during the nineteenth century.

The term “Windward Coast” as used by Atlantic slave traders designated an imprecise location in Africa. It could refer to anywhere from the Gold Coast to Greater Senegambia as well as to what it is widely understood to mean: Liberia and the Ivory Coast. In 1980, Armah, Jones, and Johnson published the first challenge to Curtin's definition of the “Windward Coast,” the coast of present-day Liberia and the Ivory Coast. They concluded that the term was vague, that it did not always conform to Curtin's geographical definition, and that the numbers of enslaved Africans exported from Curtin's “Windward Coast” were substantially exaggerated. Michael A. Gomez included voyages listed as coming from the Windward Coast among voyages coming from Sierra Leone. The region defined by Curtin as the “Windward Coast” posed many difficulties for maritime slave traders because of severe surf and by effective, ongoing resistance to the slave trade by the Kru people, who were highly skilled mariners living there. From the earliest years of the Atlantic slave trade, they refused to supply enslaved Africans to the maritime slave traders. They were highly skilled navigators, boatmen, and swimmers and provoked and assisted revolts among Africans imprisoned aboard (p.31) Atlantic slave trade ships anchored along their shores. The Kru described as mala gente (bad people) were considered very dangerous by the slave trade captains. In order to obtain slaves, the captains had to send crews in small boats to surprise, assault, and kidnap their victims. These raids were dangerous to the raiders and unproductive in collecting slaves.14 When European slave traders had to raid directly for slaves it usually meant there was no existing export market for slaves. They surely would have preferred to purchase enslaved Africans rather than run the risks involved in raiding directly for them. Over time, a few enslaved Africans were exported from the Windward Coast. But the Atlantic slave trade from this coast has been significantly exaggerated. After the Gold Coast began to export large numbers of slaves, starting in the 1650s, the Ivory Coast continued to export ivory rather than slaves.

In The TransAtlantic Slave Trade Database, 252 voyages (41.6 percent of 606 voyages) were recorded as coming simply from the “Windward Coast.” This could mean anywhere from Greater Senegambia/Upper Guinea to the Bight of Benin. The number of slaves bought at the first port of purchase along this “Windward Coast” is missing in the case of 98.5 percent of the recorded voyages. Only 62 of the voyages (10.2 percent) are known to have resulted in the sale of slaves in the Americas. Many of them could have been stopoffs for wood, water, and food and for the purchase of other products. Our most knowledgeable experts advise that we exclude the Kanga from the Windward Coast and include them in Sierra Leone. If we exclude the Kanga, African ethnic designations from Curtin's “Windward Coast” are extremely rare in American documents. Armah, Jones, and Johnson pointed out that there were very few in the lists of African ethnicities of slaves studied by Gabriel Debien and his colleagues. The only possible exception is Cape Lahou—actually a port, not an ethnic designation—which Debien included in the Gold Coast. Debien found 26 Kanga and 25 Cape Lahou among 6,188 slaves in British-occupied St. Domingue in 1796–97.15 David Geggus counted 253 ethnicities from the Windward Coast in his more recent study of 13,344 slaves in St. Domingue. Among them 124 were listed as Mesurade/Kanga. No other ethnicities were listed for this coast. Only one Caplao was found in Louisiana documents: a 25-year-old cook named Joseph who was first sold in 1815. Some of the smaller Caribbean islands listed more “Cape Lahou.” Guadeloupe probate inventories listed 49 “Caplaous” (7 percent of Africans of identified ethnicities) between 1770 and 1789.16 Barry Higman found 231 Cape Lahou among 2,638 Africans of identified ethnicities and no Kanga in the St. Lucia registration lists of 1815; no Cape Lahou and 140 Kanga were among 2,986 Africans of identified ethnicities in the St. Kitts lists of 1817; 473 KwaKwa (Kwa language group speakers whom we include with Lower Guinea), 160 (p.32) Cape Lahou, and 270 Kanga in the 13,398 Trinidad entries of 1813; 62 Kanga and no Cape Lahou in the 1,136 slaves listed for Berbice in 1819. No Africans from the Windward Coast by any possible definition were listed among slaves in Anguilla in 1827.17 The data published after 1980 confirm that Curtin's “Windward Coast” was hardly involved in the Atlantic slave trade unless we include Cape Lahou as part of the Windward Coast; in any case, Cape Lahou became involved late and in small numbers. Well into the nineteenth century, the British relied mainly on Kru mariners along the inhospitable shores of the Ivory Coast for help in suppressing the illegal slave trade. But by then some Krumen had become involved in the illegal maritime slave trade as well, and they might have been instrumental in sending a few Africans to the Americas from Liberia and the Ivory Coast.18

In order to grasp the changing patterns of export of Africans of various ethnicities, we need to transcend the concept of African coasts as defined by European slave traders and look at the changing internal patterns in Africa. Ethnicities living near the Atlantic Coast were very likely to have been shipped from ports on the coasts where they lived. But there is disagreement among scholars about how quickly and profoundly the Atlantic slave trade involved peoples living considerable distances inland from these coasts. Peoples in Africa were often mobile to escape from slave raiding, expanding desiccation, famine, warfare, and state formation. By the eighteenth century, some of the same African ethnicities from the West African Middle Belt located to the north of the coasts of Lower Guinea began to be shipped to the Americas from all three of the Lower Guinea coasts named by Atlantic slave traders. For example, Chamba were shipped from the Gold Coast and the Slave Coast. By the nineteenth century, Hausa were being shipped from the Slave Coast and from the Bight of Biafra.

Documents listing and describing enslaved Africans throughout the Americas are uniquely valuable. But like all forms of historic evidence, they have their strengths and their weaknesses. One of their greatest strengths is that they are voluminous. Slaves were legally defined as property, and as a result there is often more information listed in documents about them than about free people. Many documents contain detailed descriptions of slaves: their names and the names of their masters, their genders, ages, skills, illnesses, family members, personalities as perceived by their masters, origins (including sometimes African ethnic designations), and their prices. When slaves were interrogated, they often identified their own ethnicities (described as nations or castas in documents) or the ethnicities of other Africans. There is a whole world of known, unknown, and still to be studied documents describing slaves and sometimes recording their testimony when they were (p.33)

Table 2.1. Origin Information for Slaves in Louisiana Documents

Origin Identified



“Nation” given



Guinea or Coast of Guinea



Other coastal or port origin



“African,” no further details



“Brut” only, new Africans



Born before slave trade began







Louisiana Creole



British America



NonBritish Caribbean



Native American






Grand total


Source: Calculated from Hall, Louisiana Slave Database, 1719–1820.

Note: Atlantic slave trade voyages excluded. Category “other” (n = 52) not included.

interrogated for various reasons: mainly for running away or for involvement in conspiracies and revolts against slavery. The information about slaves in documents throughout the Americas is so massive that it is only now possible to begin to make sense of it thanks to advances in information technology. This new frontier of historical scholarship can begin to restore the severed links among Africans in Africa and their descendants in the Americas.

Documents generated in the Western Hemisphere are clearly the richest source of information about the origins and ethnicities of enslaved Africans. But they vary greatly in quality and quantity of information, mainly in accordance with the languages in which they were written. In Latin America, all transactions involving slaves were public records maintained in chronological order by notaries working in specific places and can be studied systematically over time and place. Documents written in French are by far the most informative. French documents normally list hundreds of distinct African ethnicities. These listings are rarely port or other geographic designations.

French-language documents are abundant, well organized, and generally well preserved, often in bound volumes. For St. Domingue/Haiti, they are housed in the French Colonial Archives in AixenProvence, France. They have not yet been systematically studied. In Louisiana, the notarial documents are housed in parish courthouses. Almost all of these documents dating (p.34) between 1723 and 1820 have been studied and databased. The richness of information about African ethnicities in French-language documents in the Americas explains why historians who have used them—for example, Gabriel Debien and his team of researchers, David Geggus, and I—are convinced of the significance of ethnic identifications in American documents.

English-language documents contain the least information about African ethnicities. British colonies did not have notaries to keep transactions involving slaves as public records. Therefore many sales of slaves, inventories of slaves after the death of masters, wills, marriage contracts, and other types of documents were private papers of individuals. Many of these documents have not been preserved. Others are scattered and difficult to obtain and study over time and place. The richest, most reliable information about African ethnicities in English-language documents have been found in the registration lists of slaves in preparation for general emancipation in the British West Indies. African ethnic information about them was recorded in formerly French islands as well as in Trinidad, which was settled largely by French Creole-speaking masters and slaves from Martinique. Barry W. Higman used these registration lists for his sophisticated studies of slavery in the British West Indies. Aside from these lists, newspaper advertisements for runaway slaves are the major source of information about African ethnicities in English-language documents. Daniel Littlefield pioneered their study for South Carolina. His detailed, sophisticated analysis remains unsurpassed. Michael A. Gomez effectively used advertisements for runaway slaves in his study of African ethnicities in the United States. Douglas B. Chambers has collected ethnicity information from newspaper advertisements for runaway slaves in Jamaica.19

In documents recorded in Brazil and in Spanish America, broad African coastal or regional designations predominated, and we find fewer specific African ethnicities recorded than in French documents. While Spanish and Portuguese colonists and officials grouped Africans under large regional or coastal denominations, Africans in these Iberian colonies in the Americas made finer distinctions among themselves, which sometimes emerge in the documents, especially records of court testimony in which Africans identified their specific ethnicities. Although some Africans exposed to others of closely related ethnicity in the Americas tended to overcome their more narrow, localized identities and identified themselves more broadly, they often resisted the extremely broad designations projected on them by Iberian slave traders and colonial officials.20 Except in Santiago de Cuba, where French refugees from St. Domingue/Haiti were a significant portion of the population, slaves recorded in sales documents in Cuba were listed under broad (p.35)

Table 2.2. Africans with “Nation” Designations Sold in Cuba, 1790–1880


Number Sold


Karabalí (Calabar)






Ganga (Kanga)



Lucumí (Nago/Yoruba)












Source: Adapted from Bergad, Iglesias Garcia, and Barcia, The Cuban Slave Market, 1790–1880, 72, table 4.7.

regional categories regardless of their ethnicities: For example, “Karabali” for slaves from the Bight of Biafra; “Mandinga” for slaves from Senegambia; “Ganga” (probably derived from Kanga) for slaves from Sierra Leone; “Ashanti” for slaves from the Gold Coast; “Lucumi” for slaves from the Slave Coast; “Congo” for slaves from West Central Africa.

Moreno Fraginals has found some specific ethnic designations listed among slaves inventoried after the death of masters. We do not know where in Cuba these inventories came from, but if they follow the patterns of sales documents, a very high percentage of specific ethnicity information came from French-language documents in Santiago de Cuba.

In Cuba as well as in Brazil, official mutual aid societies were organized and named in accordance with the broad African coastal or regional designations most familiar to masters and colonial authorities. But they were often reorganized internally in accordance with ethnicities recognized by their African members. In Cuba, for example, the name of an African coast or the largest and best-known ethnicity arriving from that coast was used as the designation for a particular Cabildo de Nacion, even though it likely encompassed quite distinct peoples. The Cabildo de Karabali included the Igbo, speakers of a Kwa language, and the Ibibio, speakers of a Northwest Bantu language. Some of these Cabildos de Naciones disintegrated along ethnic lines.21

In Latin American countries where the conversion of Africans to the Catholic faith was a priority and the Catholic Church maintained records of vital statistics as sacramental documents (births, marriages, and deaths), some information about African ethnicities was put on paper. For the reasons discussed below, even the best of those records need to be used with caution.22 Although the documents concerning adult baptisms reflected the (p.36)

Table 2.3. Africans with Ethnic Designations Recorded on Cuban Sugar and Coffee Estates

Ethnic Designation















































































































Source: Adapted from Moreno Fraginals, “Africa in Cuba,” 190–91, tables 2–4.

(*) Yoruba

(**) Arada/Aja/Fon

(***) Also written “Canca” or “Canga”

(****) Ibibio

African population better than other sacramental records, in Brazil they did not list specific ethnicities, and the broad coastal designations that were recorded did not reflect their proportions among African slaves. In baptismal documents in eighteenth-century Rio de Janeiro, denominations for African origins were few and very broad: Mina, Guinea, West Coast, Cape Verde, and “Contracosta” (meaning East Africa). Slaves from West Central Africa—Kongo and Angolans—were most likely to have been baptized before they left Africa. They would be underrepresented in baptismal records and over- represented in church membership lists because of their exposure to Christianity in Africa. Islamized Africans from Upper Guinea were probably most likely to resist baptism and would therefore be underrepresented in sacramental records. Africans from the Slave Coast, listed in Brazilian documents under the broad denomination “Mina,” were rarely baptized in Africa and were therefore baptized in Brazil in higher proportions. In Brazil as well as in Louisiana, and possibly elsewhere, adult women accepted baptism more readily than adult men. Therefore, Africans from the Slave Coast, listed as (p.37) Mina in Brazil, would have been overrepresented partially because of a higher proportion of women among them.23

Some designations that appear to be ethnicities actually refer to regions. Early Spanish-language documents lumped all Africans from the same region under the name of the most widely known ethnicity from that region. During the fifteenth century, for example, slaves in Valencia, Spain, were designated by the names of the major ethnicities living in large regions of Greater Senegambia. “Jalof” meant all of northern Upper Guinea, “Mandega” meant central Upper Guinea (from the Gambia to the Rio Geba), and “Sape” meant southern Upper Guinea. According to George E. Brooks, “Sape” meant the Bullom or Temne ethnicity.24

Marisa Soares has discovered and studied remarkable documents about the Mahi (Maki) groups organized within the Irmandade da Mina (Brotherhood of Mina) in Rio de Janeiro, Brazil, during the eighteenth century. The Mahi lived north of the kingdom of Dahomey. They were heavily raided after Dahomey captured Whydah and established itself on the Atlantic Coast during the late 1720s, when the Mahi began to arrive in Brazil in significant numbers. The Mina Irmandade divided itself into two major groups: the Dahomeans (Fon) and other lesser-known groups who had been raided by Dahomeans for slaves. These ethnicities identified themselves as Mahi (Maki), Agolin, Savaru, and Sanno. They all spoke the general Mina language of Brazil, “lingua geral da Mina,” demonstrating a seemingly contradictory process of integration and disintegration along ethnic lines. Each ethnicity within this brotherhood elected its own kings, queens, and regents. The first Mahi king, Capitão Ignacio Gonçales Monte, was elected in 1764. He claimed descent from the kings of Mahi. These Mahi rejected the Dahomeans as pagans.25

Portuguese and Brazilian documents commonly identify Africans by the port from which they left26 or, in the case of Mozambique, from the region from which they were exported.27 Thus West Central African ports of origin were often used for names of Africans in Brazil: for example, Fortunato Cabinda, José Benguela. Broad regional designations were often included in slaves' names: for example, Domingos Mina, Vitorino Moçambique.

The meanings of these “nations” recorded in American documents are not obvious. There is no detailed, existing body of knowledge about historical African ethnicities either in Africa or in the Americas. Ethnic designations and identities changed on both sides of the Atlantic during the 400 years of the Atlantic slave trade. It is easy to get lost in the maze of hundreds of nominal designations of a great variety of peoples over four centuries. Scholars often focus on how Europeans in Africa identified African ethnicities. (p.38)

Making Invisible Africans Visible: Coasts, Ports, Regions, and Ethnicities

Mozambique Africans in Brazil. (Johann Moritz Rugendas, Voyage pittoresque dans le Brésil, 1835.)

But the relevance of observations of Europeans in Africa about Africans brought to the Americas is often questionable. There is strong evidence that specific ethnic designations recorded in American documents were often selfidentifications. For example, masters could not possibly have been familiar with the hundreds of ethnic designations listed in French documents.

While many scholars look to Africa to understand how Europeans designated (p.39)

Making Invisible Africans Visible: Coasts, Ports, Regions, and Ethnicities

West Central Africans in Brazil. (Johann Moritz Rugendas, Voyage pittoresque dans le Brésil, 1835.)

Africans of various ethnicities over time and place, our most reliable information is how Africans identified themselves in the Americas. Except in the case of the recaptives from illegal slave trade voyages brought to Sierra Leone during the first half of the nineteenth century, self-identified African ethnicities are rare in historical documents in Africa.28 But they are frequent in the Americas. Some of our best, most detailed information comes from (p.40)
Making Invisible Africans Visible: Coasts, Ports, Regions, and Ethnicities

Each of these four images represents a different African “nation” in Brazil. (Johann Moritz Rugendas, Voyagepittoresque dans le Brésil, 1835. From the website “The Atlantic Slave Trade and Slave Life in the Americas: A Visual Record,” 〈http://hitchcock.itc.virginia.edu/Slavery〉.)

Louisiana. The total slave population of early Louisiana was small compared to that of many other places in the Americas. But the richness of the information, especially about African ethnicities, is certainly unique for documents about slaves who became part of the population of the United States and quite possibly for documents generated in any other place in the Americas.

The Louisiana Slave Database, 1719–1820, systematically sheds light on the circumstances under which African ethnicities were identified. All extant (p.41)

Making Invisible Africans Visible: Coasts, Ports, Regions, and Ethnicities

Men and women from Benguela and Kongo living in Brazil. (Johann Moritz Rugendas, Voyagepittoresque dans le Brésil, 1835. From the website “The Atlantic Slave Trade and Slave Life in the Americas: A Visual Record,” 〈http://hitchcock.itc.virginia.edu/Slavery〉.)

documentation indicates self-identification or, on occasion, identification by other Africans. The clearest cases of self-identification involved slaves testifying in court proceedings. These interrogations almost always involved recaptured runaways and slaves accused of involvement in conspiracies or revolts against slavery. When they were sworn in, they were normally asked, among other standard questions, “What is your nation?”

(p.42) When asked to identify other slaves, they often identified their African “nation” as part of the description. Since there were substantial numbers of Africans of the same ethnicity and region throughout Louisiana, deliberate misrepresentation of African ethnicities would pose problems. Although Africans were often multilingual, language use among new Africans might make misrepresentation difficult. Some Africans no doubt identified with larger, better-known ethnicities found in the Americas that were closely related to them. But the great variety of ethnic designations found in Louisiana documents would minimize, but not exclude, this possibility.

The 8,994 records in the Louisiana Slave Database containing specific African ethnicity information involve 217 different ethnicities, many of them spelled in a variety of ways. Among them 96 have been identified, although identifications remain to be further refined by input from other scholars. There are 121 ethnic designations (consisting of only 152 individuals) whose “nation” was recorded but cannot as yet be identified. Most of the identified ethnicities and all of the unidentified ones were represented by very few individuals, often only one. Among the 8,842 Africans of identified ethnicities (excluding “Guinea” and ethnicities listed but unidentified), 96.2 percent (n = 8,508) were clustered among 18 ethnicities ranging between a low of 66 records for the Edo of the Bight of Benin to a high of 3,035 for the Kongo of West Central Africa.29 Although the nominal designations of the most frequent ethnicities varied over time and place, all of them can be found in substantial numbers in documents throughout the Americas. African ethnicities described in very few documents are of particular interest to specialists in African history, but they are too few in number for use in studies of the distribution of African ethnicities in the Americas.

Table 2.4 shows the changing patterns of introduction of the most frequent African ethnicities in Louisiana over time and by gender. By selecting smaller time periods and particular locations, we can study the pattern of the appearance of African ethnicities in Louisiana documents in great detail. These studies, including age studies, can throw light on the patterns of export from Africa.

The Louisiana Slave Database, our most detailed, sophisticated tool about African ethnic designations in the Americas, indicates that when new Africans were first sold, their ethnicities were rarely identified. The longer Africans remained in Louisiana, the more likely were their ethnicities identified. If these patterns in Louisiana can be generalized, it appears that Africans in the Americas, not slave traders or masters, identified African ethnicities recorded in American documents. Sales documents, especially of newly arrived Africans (described as “new Africans” in English, brut in French, bozal in Spanish), (p.43)

Table 2.4. Eighteen Most Frequent Ethnicities by Gender in Louisiana, 1719–1820





Bam ana









% of Total













% of Total













% of Total













% of Total













% of Total













% of Total













% of Total













% of Total













% of Total













% of Total













% of Total













% of Total













% of Total













% of Total













% of Total













% of Total













% of Total













% of Total









Percentage within Gender




Source: Calculated from Hall, Louisiana Slave Database, 1719–1820.

(p.44) tend to lump Africans into only the few best-known ethnicities or coastal origins when they give such descriptions at all. Enslaved Africans who had been in the Americas for several years and were then resold had their specific ethnicities recorded more often than newly arrived Africans. The inventories of Africans on estates after their masters died offer much more numerous and detailed ethnic designations than any other type of document. In probate inventory documents generated after the deaths of masters as well as in sales documents listing and describing slaves in Louisiana, several notaries explained that they could not list the African “nation” of particular slaves because these slaves did not know their nation: for example, Quebra, “does not know his nation;” Marie, “does not know her nation;” Francine, age twentythree, a woman from Guinea, “does not know the name of her nation.”30

Aside from this impressive, but still largely anecdotal evidence, overall patterns revealed by calculations drawn from the Louisiana Slave Database make an even more convincing case that enslaved Africans themselves—rather than masters or appraisers—normally identified their ethnicities. The longer Africans remained in Louisiana, the more likely it was that their specific African ethnicities would be identified. Of new Africans sold in Louisiana, 81.7 (p.45) percent (n = 2,860; t = 3.499) were listed simply as brut or bozal, and 3.0 percent (n = 106) gave their African coastal origins only. Only 15.3 percent (n = 553) listed specific ethnic designations. The mean age of new Africans listing numeric ages was 19.2 (n = 2,867). When Africans with recorded ethnicities and not listed as brut or bozal were sold, their mean age was 26.4 (n = 3,946). They were mainly resold slaves who had been in Louisiana for an average of seven years. When Africans with recorded ethnicities were inventoried and appraised in probate documents after the deaths of masters, their mean age was 34.5 (n = 4,489). They had been in Louisiana for an average of 15 years. The mean age of all Africans with numeric age and specific African ethnicity information recorded was 31.2 (n = 8,226). Slaves with recorded African ethnicities were 68.3 percent (n = 4,750; t = 6,955) of all African slaves listed in probate documents; 46.4 percent (n = 3,448; t = 7,435) listed in sale documents; and 46.6 percent (n = 796; t = 1,709) listed in all other types of documents. These figures exclude towns, villages, coasts, ports, or other geographic designations. The largest number of coastal designations by far were listed as Guinea or the Coast of Guinea: a total of 1,052 records, which have been excluded from the calculations given above.

To recapitulate, new Africans were the least likely to have their ethnicities recorded in documents. Africans who had been in Louisiana for several years and were resold were much likely to have their ethnicities recorded. Africans who had been in Louisiana for many years and were inventoried and sold after the master's death were most likely to have their ethnicities recorded. A credible explanation is that new Africans had the greatest difficulty communicating and therefore could not identify their ethnicities. But after they had been in Louisiana for a number of years, they could communicate better. The longer they had been in Louisiana, the more likely they could communicate and identify their ethnicities. It is reasonable to conclude that the African ethnic designations listed in Louisiana documents overwhelmingly involved self-identification by Africans. The role of maritime slave traders in this process was apparently quite limited. Neither masters nor appraisers could possibly have been familiar with the hundreds of ethnic designations listed in these documents. We can safely minimize their role in identifying the African ethnicities of slaves brought into Louisiana and probably into other places in the Americas as well. This surprising pattern of when, where, and how African ethnicities were identified calls into question the assumption of very rapid loss of particular African ethnic selfidentifications during the process of creolization. The extent to which these patterns can be confidently generalized throughout the Americas will have to await databased studies from other colonies, nations, and regions.

(p.46) It is obvious that names for places, regions, and peoples in Africa need to be examined carefully. Few historians who study African ethnicities in the Americas view them as “tribes.”31 “Tribe” is a static term with heavy overtones implying primitiveness. It assumes that all Africans identified themselves according to kinship when in fact the basis for group identification among Africans varied greatly and changed over time and place. In the process of state building, strong matrilineal traditions were a seriously destabilizing force, especially in polygamous societies. For example, in Angola, in order to destroy the power of matrilineal descendants, the Imbangala (Jaga) prevented their wives from having children, sometimes killing their wives and their biological children, or excluding them from their communities. They adopted captured children who were not related to them by blood. Elite men sometimes married slave women to thwart the ambitions, demands, and conflicts among their cowives and children. In the Upper Niger region during the eighteenth century, the expanding Segu “Bambara” state was constructed and consolidated by substituting age-group affiliation and personal loyalty for descent.32 Thus traditional hierarchies based on kinship and descent were destroyed as new polities were created demanding new loyalties.

Many Africans had a broad, politically based identification with ancient and more recent empires, kingdoms, and smaller polities. Some groups of Africans were assigned nominal designations by other African groups but of course maintained their own names and identifications among themselves. Some Atlantic slave traders referred to peoples by using port, regional, coastal, and other geographic designations. For example, the Portuguese named the Bissago Islands after the Bissago people who lived there.33 Common, mutually intelligible, and closely related languages were important but not necessarily decisive factors in identity. For example, the Bamana (“Bambara”) and the Mandingo spoke mutually intelligible Mande dialects but they had historic and religious conflicts resulting in the maintenance of separate identities in the Americas as well as in Africa. Africans from the Slave Coast in Brazil spoke the common lingua geral da Mina created in Brazil, but those raided and enslaved in Africa by the kingdom of Dahomey distinguished themselves from and remained hostile to the Dahomeans.

Designations for African regions as well as for ethnicities varied among European slave traders. British traders in Africa referred to all of West Central Africa as Angola and English-language documents in the Americas tended to list all West Central Africans as Angolans. English slave trade documents indicating port of origin in West Central Africa are vague, further obscuring the origins of English voyages to West Central Africa. Among the 933 English voyages to West Central Africa recorded in The TransAtlantic Slave (p.47) Trade Database, 68.7 percent (t = 641) list the principal port of slave purchase simply as Angola.34 French- and Spanish-language documents tend to list all West Central Africans as Kongo: for example, 93 percent (n = 4,561) in the lists from eighteenth-century St. Domingue studied by David Geggus.35 Michael Gomez's calculations from Gabriel Debien's collection of ethnic designations in St. Domingue during 1796 and 1797 indicate that all West Central Africans were listed as Kongo (n = i,65i).36 In the Louisiana Slave Database, 97 percent (n = 3,152) of Africans from West Central Africa were listed as Kongo. Only 25 were listed as Angolans, 18 of whom had been brought in from South Carolina in 1783 by Dr. Benjamin Farar. Although all West Central Africans were normally recorded as Angolans in British colonies, runaway slave ads in Jamaican newspapers after 1775 (1776–95 and 1810–17) described 499 of them as Kongo while only 27 were described as Angolans. Five were described as Angolans and none as Kongo before 1776, indicating perhaps increasing precision as more West Central Africans entered Jamaica over time.37 Several ethnicities from West Central Africa recorded in significant numbers were not listed simply as Kongo or Angolan; they were, instead, listed as Mungola in Jamaica, Monjolo in Brazil, Mandongo in Louisiana and Cuba, and Mondon- gue in St. Domingue. But relatively few documents created in the Americas list a significant number or variety of distinctive West Central African ethnicities. The partial explanation for this is perhaps the specific pattern in which fishing communities along the Kongo River were populated, which resulted in close linguistic and kinship relationships among peoples living over wide geographic areas.38 Mary Karasch has found an array of specific ethnicities from West Central Africa described in travelers' accounts in Rio de Janeiro during the nineteenth century.39

Some designations that scholars have taken to mean a port or a coastal designation had other meanings. In Louisiana and no doubt elsewhere in the Americas as well, “Senegal,” a coastal designation, meant Wolof. During the first years of the African presence in Louisiana, Le Page du Pratz, director of the Company of the Indies, noted that the Wolof were called Senegal by the French colonists but they continued to be called Wolof (“Djolaufs”) among themselves.40 “Mina” normally did not mean slaves coming through the fortress/port of São Jorge da Mina (Elmina) on the Gold Coast. It was a designation referring to different ethnicities over time and place, but it certainly sometimes meant people from Little Popo, originally Akan speakers who had migrated from west of the Volta River. They were often designated Mina- Popo in Brazil and Cuba. Popo was usually written “Pau Pau” in English. The designations “MinaNago” and even “MinaCongo” were sometimes found in Brazil.41

(p.48) Some Africans in the Americas, when asked to identify their “nation,” replied with the name of their village or district. But this did not necessarily mean that they lacked a broader self-identity or that they were so isolated as to have no concept of other peoples. Isolation was far from universal among Africans involved in the Atlantic slave trade. Ancient, extensive trade networks involving the sale of products made throughout Africa linked coastal with interior regions and across the Sahara and the Indian Ocean. This active trade predated the Atlantic slave trade by many centuries. Except for bringing in slaves for sale in the ports, the Atlantic trade weakened commercial links with the interior. Warfare, conquest, kidnapping, and slave raiding escalated at the expense of production and internal trade. As the price of slaves rose over the centuries and cheap East Indian and then British textiles conquered the African market, African production tended to lose its economic incentives.42 The maritime slave trade created enclave economies along the coasts geared toward supplying the ships, especially with food for the voyages. Although the Atlantic slave trade introduced new products from all over the world to exchange for a variety of African products—including gold, copper, and other currencies, ivory, gum, pepper and other spices, textiles, kola nuts, rice, millet, sorghum, yams and other foods as well as slaves43—a blanket rejection of Walter Rodney's argument that Europe underdeveloped Africa is not warranted.44

Patterns of self-identity among Africans differed from region to region in Africa and the Americas. For example, in Greater Senegambia, the Wolof and the Mandingo came from stratified societies with a long tradition of state formation and self-identity. Fulbe herdsmen and warriors, heavily Islamized and quite mobile, relocated great distances south and east of their original home along the middle reaches of the Senegal River to protect their herds from drought and desiccation. In their migrations through West Africa, the Fulbe were also active in warfare and the capture and sale of slaves. These ethnicities identified themselves clearly as distinct peoples living in extended geographic areas. Africans from small, local communities in parts of Upper Guinea, along the Bight of Biafra, or in parts of West Central Africa where bureaucratic, stratified state systems were less common generally identified themselves in terms of their immediate vicinity. To cite an example from a document from Opelousas, Louisiana, dating from 1802: Celeste, the thirteen-year-old Creole daughter of an African slave couple, was accused of assaulting her master with an axe and nearly killing him. When her parents were called on to testify, they both explained that they did not know their ages and professed no religion. There is little doubt about the father's ethnicity; he was a self-identified Mandingo, a heavily Islamized ethnicity. Celeste's mother testified that her (p.49) “country” was called “Yarrow” (Jarrow?), which was the name she used for herself.45 A slave in Maryland, Yarrow Mahmout, lived to be very old while openly practicing the Muslim faith. The personal or place name Yarrow might have had a religious significance.46

It is obvious that over four centuries the meanings of these nominal designations changed. Some peoples were assigned names by their neighbors, or by other Africans. But when one group of people was named by another, they nevertheless had various bases for self-identification and their own names for themselves. Africans had group identities in Africa as well as in the Americas. To deny that ethnicities existed in Africa and assume that the many and varied African ethnic designations recorded in documents in the Americas did not originate in Africa but were created in the Americas is worse than being named by others. It denies the roots of peoples in Africa, including their names, homogenizes them, and renders them invisible. Africans are the only peoples who have been subjected by scholars to this level of denial.47

It is understandable that some scholars throw up their hands in the face of the bewildering number of African ethnic designations with changing spellings, pronunciations, and meanings recorded in several languages over many centuries in vast regions on both sides of the Atlantic. Nevertheless, we can go far by concentrating on the relatively few African ethnicities most frequently listed in documents throughout the Americas, despite the fact that they were often recorded under various names. We have seen that the best evidence we now have indicates that these ethnic designations were normally selfidentifications by Africans in the Americas rather than identifications by slave traders on either side of the Atlantic or at sea.

The study of African ethnicities in the Americas has been widely neglected during the past three decades. The influential Mintz-Price thesis was first published in 1976.48 It claimed extreme diversity among and random distribution of Africans brought to the Western Hemisphere. This work has several virtues, including its partial emphasis on changes over time. It emphasizes the importance of the early, formative period of creolization in the Americas, although no pattern of creolization can be confidently generalized. Unfortunately, the most influential conclusion of the Mintz-Price thesis is flawed by its static approach to the Atlantic slave trade as well as its projection of patterns supposedly found in one, small place in the Americas to all of the Western Hemisphere. It collapses time and then calculates the percentage of Atlantic slave trade voyages arriving in Suriname from various African coasts. In generalizing this flawed finding, the thesis concludes that the impact of particular African regions and ethnicities on the formation of Afro-Creole cultures in the Americas was nonexistent or insignificant. But Africans from (p.50) the same regions and ethnicities arrived in various places in the Americas in waves and were often clustered over time and place. Dynamic perception and comparative analysis reveal clustered patterns while static perception gives the false impression of random, dispersed patterns. The Mintz-Price thesis had a chilling effect on studies of African ethnicities in the Americas. Such studies became almost a heresy among influential scholars in the United States. The influence of the thesis spread abroad as well, although it was greeted with less enthusiasm in Latin America than elsewhere.49 The work of historians—especially minorities and/or women working in the United States—who even mentioned African ethnicities has yet to get the recognition it deserves. Much of that work is discussed and cited in this book.

A growing number of historians of Africa, the transatlantic slave trade, and the African diaspora in the Western Hemisphere are focusing on the patterns of introduction of Africans of various coastal origins and ethnicities through-out the Americas over time and place. These studies have shown that Africans were often clustered in the Americas rather than randomized or deliberately fragmented. This conclusion has been reinforced by the publication of computerized relational databases created from massive collections of documents generated by transatlantic slave trade voyages as well as by the publication of other databases created from more varied types of documents generated in the Americas.50 These innovative tools help us refine studies of the pattern of introduction of Africans over time and place as well as other key questions about the slave trade and slavery in the Western Hemisphere. They allow us to avoid the mistake of distorting this complex history by collapsing time, and they enable us to begin to better evaluate and transcend the limitations of previously available evidence.

Much historical interpretation has, up to now, relied heavily on anecdotal evidence collected from travelers' accounts, which were sometimes plagiarized, falsified, or sensationalized to appeal to a public hungry for information about “exotic” peoples and places; administrative reports, which were often self-serving distortions by more or less well informed and observant bureaucrats; and reports of missionaries, which were often better informed because the authors had closer and more sustained contact with the peoples they were writing about, but such reports need to be carefully studied and used judiciously because they are often marked by prejudice and stereotypes. Thus the secondary literature has been constructed to a great extent from sources of uneven quality. Many questionable conclusions have been accepted as truth and repeated by historians from one generation to the next. This writer believes that the creation of databases constructed from large numbers of generally less self-serving documents organized in time series can help save (p.51) history from nihilistic tendencies within the postmodernist school, which subjectivize history and reduce it to literary criticism. (To the postmodernist, one persons myth is often as good as another's.)

Some scholars advance and defend the concept that ethnicity has nothing to do with Africans but is a mere construct dating from the end of the nineteenth century, when the colonial period began in Africa. According to this argument, European missionaries, anthropologists, and administrators invented African ethnicities in order to create conflicts among the peoples they ruled. Joseph C. Miller, grounded in the history of West Central Africa, where distinctive ethnicities were less developed than in some other West African regions and the terms “Kongo” and “Angola” were widely used by Europeans for a great variety of peoples, properly calls for more contextualized studies by historians who posit ethnic continuities among Africans brought to the Americas.

There is only an element of truth to the argument that European colonizers of Africa constructed African ethnicities. During much of the era of the Atlantic slave trade, and in most locales where it was conducted, maritime slave traders and their hierarchies of administrators and missionaries lacked access to the interior regions of Africa. They were often confined to the coast, and sometimes to ships docked offshore. Within Africa, creolization developed as a normal process of contact among peoples incorporating new cultural and linguistic groups into other groups and polities. During the Atlantic slave trade, increasing warfare and raids among ethnicities for captives to sell into the Atlantic slave trade aggravated ethnic conflicts. The creation and promotion of rigid, mutually exclusive antagonisms among ethnicities in Africa were a product of policies of social control as European colonizers advanced into the interior of the African continent.

Indeed, the political motivations for the denial of the existence of African ethnicities in Africa are laudable. The Cold War was a hot war in Africa as the dominant world powers sought to use African clients in proxy wars against each other. Ethnic conflicts continue to be aggravated and manipulated to exploit the natural resources of Africa, tear and mutilate her social fabric, and destroy her peoples. Ethnic conflicts are often promoted by outsiders, including arms merchants and multinational corporations allied with some of the elite or would-be elite within various African states to facilitate the exploitation of gold, diamonds and other precious gems, copper, uranium, oil, and other natural resources.51

Scholars who deny the historical existence of African ethnicities prior to the late nineteenth century sometimes acknowledge that their conclusions are based on studies of conditions in Southern Africa, often in places where the (p.52) Atlantic slave trade was of minor significance. Nevertheless, some of them confidently project these conclusions backward in time over four centuries to regions from which the vast majority of Africans were brought to the Americas. They then conclude that African ethnicity in all times and places was a European construction imposed on Africans. The Atlantic slave trade obviously took place during precolonial times when, except in West Central Africa, the European administrative presence was often confined to Atlantic and Indian Ocean ports and fortresses and sometimes merely to ships docked offshore. Ethnicities existed in Africa and interacted widely not only before the colonial period but also long before the Atlantic slave trade began. The great Senegalese scholar Cheikh Anta Diop emphasized the fundamental unity among all Africans and their common descent. But he never denied the existence of African ethnicities.52 Boubacar Barry described an ongoing process of creolization in Africa among ethnicities with ancient identities. He wrote that in Greater Senegambia, long before the Atlantic trade began,

people switched ethnic groups and languages. There were Toures, originally Manding, who became Tukulor or Wolof; Jallos, originally Peul [listed as Poulard in Louisiana documents], became Khaasonke; Moors [listed as Nar(d) in Louisiana documents] turned into Naari Kajor; Mane and Sane, originally Joola surnames, were taken by the Manding royalty of Kaabu. There was, in short, a constant mixture of peoples in Senegambia, destined for centuries to share a common space. Senegam- bia, in some respects, functioned like a vast reserve into which populations in the Sudan and the Sahel habitually poured surplus members. In their new home the immigrants created a civilization of constant flux…. Nowhere in this Senegambia … did any Wolof, Manding, Peul, Tukolor, Sereer, Joola, or other ethnic group feel they were strangers.53

The denial of the history and even the existence of many African peoples on both sides of the Atlantic reinforces the concept of the generic African, distancing and dehumanizing Africans in the minds of peoples throughout Europe and the Americas. It severs the ties between Africans who remained in Africa and those who were shipped to the Americas, as well as the ties between their descendants.

Finally, let us look at the great abundance and variety of African names of slaves recorded in the Louisiana Slave Database. These 5,647 distinctive African names remain to be fully studied and explained. The possibilities are open ended. The vast majority of slaves with African names—5,980 (57.7 percent)—fell into the category of slaves of unidentified ethnicities or birth-places. We do not know if they were Africans or Creoles. In order to respect (p.53)

Table 2.5. Distribution of African Names among Louisiana Slaves by Origin

Where Born

Total Number

With African Names

Percentage with African Names





British colonies




NonBritish Caribbean




Louisiana (Creoles)




Louisiana (Native American)












Source: Calculated from Hall, Louisiana Slave Database, 1719–1820.

what was literally contained in the documents, these individuals were not coded as Africans. If they had been, the proportion of identified birthplaces as well as the proportion of Africans in the Louisiana Slave Database would have been much higher. This decision is buttressed by Philip D. Morgan's findings that in the British mainland colonies, which became part of the United States, Africanborn slaves usually had Anglo names but they passed on African names to large numbers of their Americanborn children.54

Unless the names included a clear ethnic designation, such as Louis Congo or Samba Bambara, “nation” designations, rather than personal names, were relied on to identify African ethnicities. Naming patterns were fluid on both sides of the Atlantic, and Africans often changed their names. Some Africans adopted the names of others in order to honor them. Enslaved Africans sometimes took the name of a friend or a shipmate or someone they met shortly after landing as a means of identification with this person or out of respect. African names spread among a variety of African ethnicities and regions. Africans of various ethnicities used the same personal names.55 Names with particular meanings among certain ethnicities can be found among other ethnicities: for example, Samba, Comba, Kofi, and other Akan names representing the day of the week on which the person was born. A few Creole slaves took an African ethnic designation as their name, or part of their name, as a way of identifying with the ethnicity. There is the case of a Creole slave, Joseph Mina, who took the ethnic name of the Mina slaves who reared him.56 A few other names of Creole slaves included African ethnic designations: for example, Edouard dit Kanga, Felipe alias Bambara, Louis Kiamba, Senegal, and Maniga. The most startling case was François dit Congo, a four-year-old quadroon slave who was sold in 1817 with his mulatto mother under the condition that both of them be immediately freed, although it was illegal to free (p.54) anyone under the age of thirty by that date. Here was a second-generation Creole who was three-quarters white with an African ethnic designation as part of his name. But these cases are rare enough to ignore in calculations.

The widespread survival of African names among Louisiana slaves and those in British colonies is unusual in the Americas. In Brazil and in Spanish America we find fewer African names in lists of slaves because enslaved Africans had normally been baptized and given Christian names either in Africa or shortly after they arrived in the Americas. But in Rio de Janeiro during the nineteenth century, an African port designation was often added to the slave's Christian name. African names were more likely to be found in British colonies where the Christianization and baptism of slaves were less common. In Louisiana 10.9 percent of slaves coming from English-speaking colonies or countries had African names. The African names among slaves in St. Domin- gue were found in highest proportion among Africans from Senegambia.57 In Louisiana as well, Africans shipped from Senegambia retained African names out of proportion to their numbers in the slave population. For example, the Bamana were 5.5 percent of the most frequent African ethnicities but had 10.3 percent of the African names. The Mandingo were 10.9 percent but had 12.7 percent of the African names. The Wolof were 7 percent of the most frequent ethnicities and had 9 percent of the African names. The proportion of African names was higher than could be expected by their numbers among the Nar/Moor and the Fulbe as well. Many of these names were Africanized Islamic names. Africans from the Bight of Benin did not have a higher than proportional retention of African names although they often resisted Christianization. One reason could be that few of them were Islamized. At the other end of the scale, the Kongo were 35.7 percent of the most frequent ethnicities but had only 29.6 percent of the African names. This difference could reflect the fact that many more Kongo than other ethnicities had been baptized and Christianized in Africa and their names were passed down through the generations.

It is a humbling thought that Africans continued to identify with their particular African ethnic and regional origins long after they arrived in the Americas. Many of them retained their African names decades after they arrived, and some of them passed them on to their children born in the Americas. This evidence would indicate that African ethnic and regional identities survived for a longer period of time than most historians and anthropologists believe. In order to understand the process of creolization in various regions in the Americas, we indeed need to ask, “Which Africans?”


(1) . Moreno Fraginals, “Africa in Cuba.”

(2) . Aguirre Beltrán, La población negra de México; Lockhart, Spanish Peru, 173. For a summary for the French West Indies, see Debien, “Les origines des esclaves des Antilles” and “Les origines des esclaves des Antilles (conclusion).”

(3) . Palmer, Slaves of the White God.

(4) . Niane, “Introduction”; Talbi, “The Spread of Civilization in the Maghrib and Its Impact on Western Civilization.”

(5) . Blackburn, The Making of New World Slavery; Thomas, The Slave Trade.

(6) . Inikori, “Unmeasured Hazards of the Atlantic Slave Trade.”

(7) . Curtin, The Atlantic Slave Trade, 268 (table 77); Inikori, “The Known, the Unknown, the Knowable and the Unknowable”; Inikori, “Africa in World History,” 82; Thomas, The Slave Trade, 809, 862.

(8) . Barry, “Senegambia from the Sixteenth to the Eighteenth Century”; Moreno Fraginals, “Africa in Cuba”; Vila Vilar, “The Large-Scale Introduction of Africans into Veracruz and Cartagena.”

(9) . Eltis et al., The TransAtlantic Slave Trade Database.

(10) . Eltis, Rise of African Slavery in America, 244–46; Law and Strickrodt, Ports of the Slave Trade.

(11) . Studer, La trata de negros; Garcia Florentino, Em costas negras, 23.

(12) . Barry, La Sénégambie du XVe au XIXe siècle.

(13) . Brooks, Eurafricans in Western Africa, 293–94.

(14) . Aguirre Beltrán, La población negra de México, 119.

(15) . Armah et al., “Slaves from the Windward Coast”; Debien, Les esclaves aux Antilles françaises, 45, 46, 67.

(16) . Geggus, “Sex Ratio, Age, and Ethnicity”; VanonyFrisch, Les esclaves de la Guadeloupe, 32.

(17) . Higman, Slave Populations of the British Caribbean.

(18) . Brooks, The Kru Mariners.

(19) . Gomez, Exchanging Our Country Marks; Littlefield, Rice and Slaves; Chambers, Jamaican Runaways.

(20) . Reis, “Ethnic Politics among Africans in NineteenthCentury Bahia.”

(21) . Howard, Changing History, 27, 37, 39, 74.

(22) . For some exceptionally useful sacramental records with rich information about African ethnic designations, see Tardieu, “Origins of the Slaves in the Lima Region in Peru.”

(23) . Soares, Devotos da cor, 80, 83–84.

(24) . Buhnen, “Ethnic Origins of Peruvian Slaves”; Brooks, Eurafricans in Western Africa, 167.

(25) . Soares, Devotos da cor, 78, 92–93, 201–30.

(26) . Lovejoy, “Ethnic Designations of the Slave Trade.” For a careful, sophisticated discussion of the various meanings of racial and ethnic designations of slaves in Brazilian documents, see Karasch, Slave Life in Rio de Janeiro, 3–28.

(27) . Medeiros, “Moçambicanizaçao dos escravos saídos pelos portos de Moçambique”; Alpers, “‘Moçambiques’ in Brazil.”

(28) . African Ethnonyms and Toponyms. For a discussion of their recent study of recaptives in Sierra Leone and Havana, see Eltis and Nwokeji, “The Roots of the African Diaspora,” and “Characteristics of Captives Leaving the Cameroons for the Americas, 1822–1837.”

(29) . Hall, Louisiana Slave Database. Recoded in the SPSS.sav file supplied in the CD publication and the website as the recoded field AFREQ.

(30) . Records of the Superior Council of Louisiana, 1738:04:11, 1743:09:09:06, Louisiana Historical Center, New Orleans; Original Acts Pointe Coupée Parish, December 6, 1802, New Roads, La. This information can also be found in the comments fields of the records under these dates and places in Hall, Louisiana Slave Database.

(31) . An inappropriate term used throughout Mullin, Africa in America.

(32) . For conflicts among children of cowives, see Niane, Sundiata. For the Segu “Bambara” state, see Hall, Africans in Colonial Louisiana, 42–45. For West Central Africa, see Vansina, Kingdoms of the Savanna, 139–40; and Miller, Kings and Kinsmen, 128–73.

(33) . Costa e Silva, A manilha e o libambo, 153.

(34) . Calculated from Eltis et al., The TransAtlantic Slave Trade Database.

(35) . Geggus, “Sex Ratio, Age, and Ethnicity.” For the Kongo identity of slaves described in South Carolina documents as “Angola,” see Thornton, “African Dimensions of the Stono Rebellion.” Calculated from Hall, Louisiana Slave Database.

(36) . Gomez, “African Identity and Slavery in America.”

(37) . Calculated by Chambers from his Jamaican Runaways.

(38) . Harms, River of Wealth, River of Sorrow, 111–42.

(39) . Karasch, Slave Life in Rio de Janeiro.

(40) . Le Page du Pratz, Histoire de la Louisiane, 1:342–45.

(41) . Acosta Saignes, Vida de los esclavos negros en Venezuela, 152–53.

(42) . Inikori, “West Africa's Seaborne Trade.”

(43) . Niane, “Introduction”; Niane, “Relationship and Exchanges among the Different Regions.”

(44) . Rodney, How Europe Underdeveloped Africa.

(45) . Procédure criminelle contre la nommée Celeste de Jacob Beam et le nommé Urbin nègre (Criminal procedures against the named Celeste of Jacob Beam and the named black man Urbin), Original Acts Opelousas Post, March through June 1802, Louisiana State Archives, Baton Rouge.

(46) . Diouf, Servants of Allah, 60, 78, 87,180.

(47) . Gomez, “African Identity and Slavery in America.”

(48) . Mintz and Price, An Anthropological Approach to the AfroAmerican Past.

(49) . Palmié, “Ethnogenetic Processes and Cultural Transfer.”

(50) . For relational databases published on CDs and websites and in publications using unpublished databases, see the database section of the bibliography.

(51) . Twelfth Report of the Secretary-General on the United Nations Organization Mission in the Democratic Republic of the Congo, October 18, 2002. The report blames multinational corporations from several countries, including the United States, for robbing billions of dollars of natural resources from this country while provoking genocidal warfare there. Two and a half million lives were lost there over the past few years.

(52) . Diop, “A Methodology for the Study of Migrations.”

(53) . Barry, La Sénégambie du XVième au XIXième siècle, 35.

(54) . These records can be recoded easily and included among Africans of unknown ethnicities, or their ethnicities can be extrapolated from their names if the user of the Louisiana Slave Database so wishes; Philip D. Morgan, Slave Counterpoint, 454.

(55) . Communication from Dr. Ibrahima Seek, November 1999.

(56) . Hall, Africans in Colonial Louisiana, 359–61.

(57) . Communication from David Geggus.