Kongomania and the Haitian Revolution

Abstract

The article critiques the scholarly emphasis on the centrality of West-Central Africans in the Haitian Revolution. It argues that two highly influential articles published 30 years ago by John Thornton greatly exaggerated the presence of such “Congos” in the colony, and overstated that of Africans in general. Amplified in subsequent works by Thornton and others, this exaggeration has become the prevailing orthodoxy and the issue has gone entirely unnoticed down to today. To make its point, the article draws on a data set of more than 31,000 enslaved workers of known origin and it attempts to calculate population change on the eve of the revolution. It lays out the way the ethnic composition of the black population varied by crop type and region, and produces for the first time estimates for the whole of Saint Domingue. It additionally makes two excursions into African studies. The first is to investigate the ethnic/geographic origins of the “Congos.” The second relates to the nature of slavery in West-Central Africa and certain items of Kikongo vocabulary. This forms part of a critique of an ambitious article by James Sweet concerning the influence of Kongolese in Saint-Domingue that constitutes the article's final section.

Anotable trend in the last half-century of scholarship on early African American history has been an increasing emphasis on the presence and contribution of West-Central Africans, whom scholars had formerly neglected relative to the peoples of West Africa. The primary impetus for this shift in perception came from quantitative analysis of the records of the slave trade and of colonial slave populations, and it received powerful reinforcement from the work of art historian Robert Farris Thompson.¹ The trend has affected the histories of different parts of the Caribbean, Brazil, and more recently North America.² Its impact has been felt nowhere more strongly, however, than in Haiti. There the prominence of Aja-Fon and Yoruba influence on Haitian culture had long promoted the assumption that these peoples must have predominated in the slave trade to Saint-Domingue during the colonial period, an optic that tended to obscure or marginalize remnants of Kongo culture in the country.³

Although this broad reassessment has been extremely valuable, it has become grossly exaggerated during the last 30 years in the case of Saint-Domingue, the early modern Caribbean's most populous and productive colony and the second most important site of West-Central African migration in the Americas. The purpose of this article is to highlight the misunderstandings that produced this distortion, which has gone entirely unnoticed, and to suggest that it has reached a new extreme in James Sweet's research note “New Perspectives on Kongo in Revolutionary Haiti,” published in this journal a few years ago.⁴

Kongomania

The reassessment of the Kongo contribution began to veer off course with the publication of two articles in the early 1990s by John Thornton. They are justly celebrated for their pioneer revisioning of transported captives in terms of the experience and ideas they may have contributed to the Haitian Revolution, but the articles also greatly inflate the West-Central African presence in the colony. Thornton claims that “in the 1780s ‘Congos’ made up 60 percent of the slaves in the North Province, where the revolution began, and about the same percentage in the south.”⁵ I will argue below that this statement more than doubles the true proportion of Congos in northern Saint-Domingue and misrepresents by a factor of almost three their presence in the south. I also propose that Thornton's accompanying claim that a majority of Saint-Domingue's “inhabitants” in 1791 had been born in Africa is at the very least debatable, and that the assertion they had arrived in the ten years before the revolution is demonstrably wrong.⁶

Thornton cites in support of these statements several articles that offer no support at all and, if anything, demonstrate their implausibility.⁷ In fact, nothing in the growing body of research on Saint-Domingue slave society produced in the preceding decades justified this radical new assessment. Since the 1940s, French historian Gabriel Debien had spearheaded the collection and analysis of primary sources concerning the colony that did much to establish the importance of West-Central Africans in the slave trade, which Philip Curtin signaled in his Atlantic Slave Trade: A Census.⁸ Debien summed up his findings in 1974 in Les esclaves aux Antilles françaises.⁹ Building on this foundational work, subsequent articles suggested that Debien actually understated the Kongolese presence among Saint-Domingue's Africans, because his sample drew too little on the colony's North Province and coffee-growing regions.¹⁰ Yet these findings in no way warrant the picture Thornton presented.

Why he would think Kongo constituted 60 percent of the enslaved in northern and southern Saint-Domingue is hard to comprehend. The slave trade data then available to him indicated that West-Central Africa supplied (only) 51 percent of new arrivals during the decade before the slave uprising, and for 20 years “the majority.”¹¹ The second estimate was based on a very approximate tally by Jean Fouchard; neither one indicated a proportion of 60 percent; and, of course, one cannot deduce the composition of the slave population from that of its recent migrants. This was particularly true of a society with many generations of locally born slaves, known as Creoles. In any event, current estimates of the slave trade to Saint-Domingue put the West-Central African share somewhat lower: 46.9 percent of arrivals between 1782 and 1791, and 49.4 percent between 1772 and 1791.¹²

Determining whether Africans in general formed a majority of the colonial population is not at all straightforward. Saint-Domingue censuses give no information on the ethnic makeup of the slave population and only an imperfect idea of the size of the white, free colored, and slave communities. Interpretation of the censuses involves a host of problems.¹³ Thornton hazards that, on the eve of the 1791 revolt, “perhaps as many as two-thirds of the slaves” were African.¹⁴ On this he cites the Saint-Domingue magistrate Moreau de Saint-Méry, as do most of the colony's historians, including myself in times past. The problem is that no collection of data from this period supports such an assessment—none of the articles cited above or below, nor the new data set I will examine here.¹⁵ It implies that Africans constituted 53 percent of the colony's slaves during the period 1770-91. It seems likely, therefore, that Moreau de Saint-Méry's much-cited estimate concerned adults but not children.¹⁶

The last true colonial census, the one for 1789, counted 434,429 slaves, 30,831 whites, and 24,848 free people of color.¹⁷ If we accept these figures and assume that Africans constituted a maximum of 55 percent of the slaves (and at most 1,000 of the free non-whites), they could not have formed the majority of the colony's inhabitants, as Thornton claims.¹⁸ Yet all the official figures understated reality, as contemporary observers knew.¹⁹ Using a mixture of their estimates, and a generous projection of the slave population's growth, one might suggest the following as a fair approximation of the mid 1791 population: 35,000 whites, 32,000 free coloreds, and 510,000, or even 520,000 slaves. However, this would not increase the African proportion of the total population; it would remain under 49 percent.

This is a tentative calculation, and I will revisit it below. We are on much firmer ground asking whether a majority of Saint-Domingue's inhabitants in 1791 had arrived from Africa in the preceding decade. At 263,696, the currently accepted estimate of African arrivals during the years 1782 to 1791 is higher than the one available at the time of Thornton's writing, but even so it does not come close to justifying his assertion.²⁰ These migrants could not have made up half of the slave population in 1791, let alone half the colonial population. This is because a large proportion would have died before 1791. The point is repeatedly overlooked by historians who cite these statistics. In American slave populations, mortality was at its highest among newborns and newly arrived Africans.²¹ French commentators frequently claimed that a third of African migrants died during their first year in the Caribbean; some reckoned it was a quarter in six months, or a third, or a half in three years; and planters considered Kongolese particularly vulnerable.²² Modern historians have given mortality estimates of 25 percent in 18 months, and 50 percent in three to eight years.²³ In 1784, 15 percent of African arrivals in Cap Français and Port-au-Prince died even before they could be sold.²⁴ The issue merits more research, but it is perfectly clear why the substantial increase in the slave trade to Saint-Domingue during the 1780s did not reshape the colony's population as much as some have believed.

Although the implausibility of some of Thornton's claims should have been apparent at the time of their publication, they have powerfully shaped the subsequent historiography of the Haitian Revolution. Christina Mobley's dissertation, which appears to have inspired James Sweet's article, informs us that “the population of Saint-Domingue was not only overwhelmingly African but also largely Kongolese.” She asserts that, of 500,000 slaves in 1791, almost 240,000 had arrived in the previous decade and an incredible 40,000 in the eight months prior to the slave uprising—a fantastical figure for which no evidence exists.²⁵ In the leading general history of the Haitian Revolution, Laurent Dubois states that Kongolese “accounted for 40 percent of slaves” on northern sugar estates and “at least half the slave population” on coffee plantations in the North and West Provinces.²⁶ While thus avoiding one of Thornton's extreme exaggerations, Dubois repeats Thornton's misreading of my 1993 article and confuses African slaves with all slaves. Others have made the same mistake.²⁷ At the same time, however, there continued to appear new and larger studies on the composition of the slave population that should have demonstrated how wrong-headed all these assertions were.²⁸ Like their predecessors, these studies were misread or ignored.

In the years following pubication of his two early articles, John Thornton further inflated his estimate of the African presence in Saint-Domingue: to 60 percent of the colonial population (in 1790), and then to “about three-quarters” (in 1791).²⁹ This would mean that the white, free colored, and Creole slave populations combined accounted for barely one-quarter of the colony's residents. “We must assume,” he concluded, that “a significant percentage” of “those not born in Africa” would have become proficient in an African language, often Kikongo.³⁰ This assumption prefigures much more extreme arguments by Mobley and Sweet that I will address below. Here I will note that Thornton offers no rationale for this imagined spread of bilingualism and that it was apparently not common in the Americas. Where locally born slaves looked down on African migrants as “Africa-sheep,” “horses,” “Guinea-birds,” or “saltwater Negroes,” it is hard to see why Creoles might want to learn their languages, especially when acculturated Africans pretended to be Creoles, or to have forgotten their native tongues.³¹

Cuban Esteban Montejo was raised among Africans and in his dictated “autobiography” displays great interest in their cultures, but he did not learn an African language.³² Saint-Domingue already had, in French Creole, a lingua franca and had no need of a second one, since all slaves had to learn Creole to communicate both among themselves and with free people. Apart from the famous case of Toussaint Louverture, who learned the language of his Aja-Fon parents, the historical record seems to be silent on the question of native Domingan slaves acquiring African languages. As the largest African minority in Saint-Domingue, Kikongo speakers did have a better chance than other Africans of forming families among themselves and potentially passing on their language to their children. No doubt this happened sometimes on plantations where Kongolese were concentrated; the insurgent leader Noël Prieur may be an example. The Bakongos’ advantage in this regard, however, was substantially reduced by their extremely unbalanced male:female ratio (much higher than that of the Fongbe speakers), and the birth rates of all enslaved Africans were very low.³³

Oddly, after stating in his Cultural History of the Atlantic World that Africans made up close to 75 percent of Saint-Domingue's inhabitants, Thornton then proposes that the expanded community of Kikongo speakers constituted but one-third of the population. This is impossible to square with his earlier claims about their numbers. More surprising, he avers later in the same book that only half of the slaves were African-born.³⁴ The incongruity of these figures underlines the fact that none was based on any research. To put this question on a firmer footing we need to reconstruct the ethnic makeup of the slave population.

The Ethnic Composition of the Slave Population

Historians of Saint-Domingue are lucky that the colony's property records are quite detailed as to the identities of its enslaved workers; however, they have to be pieced together one slaveholding at a time to obtain an overview. With about 7,000 plantations, and surviving records scattered across several dozen archives, this is time-consuming work. Over the last few decades, studies have appeared of particular regions, periods, and types of enterprise.³⁵ These have revealed fairly distinctive “ethnic profiles” for the colony's three provinces and for different types of plantations. Sugar, coffee, and indigo planters developed contrasting preferences for Africans from different cultures. This was due to a mixture of factors, notably average height, traditional divisions of labor, and epidemiological experience, as well as the labor demands of the different crops.

Congos were viewed positively by mountain coffee planters but negatively by lowland sugar planters.³⁶ Slave traders took these preferences into account, so that the supply of slaves to each province reflected the local distribution of crops.³⁷ The balance between African and Creole slaves varied according to how long different districts had been settled. Table 1 shows the distribution of the Congos and West Africans in relation to these regions and crops.

Table 1 Distribution of “Congo” and West-Central African Slaves in Saint-Domingue (1770–1791) by Region¹ and Plantation Type

1. North = Fort Dauphin to Môle-Saint-Nicolas. West = Port-à-Piment to Petit-Trou. South = Jérémie to Cayes-Jacmel.

2. Congo, Mondongue, Damba(u).

Sources: Too numerous to list here, the sources used in earlier iterations of the database are identified in four articles by Geggus: “Saint Domingue on the Eve of Revolution,” in Haitian History: New Perspectives, Alyssa Sepinwall, ed. (New York: Routledge, 2013), n. 41; “Indigo and Slavery in Saint Domingue,” in Slavery Without Sugar, Verene Shepherd, ed. (Gainesville: University of Florida Press, 2002), n. 55; “The Sugar Plantation Zones of Saint Domingue and the Revolution of 1791–1793,” Slavery & Abolition 20 (1999): n. 8; and “Sugar and Coffee Cultivation in Saint Domingue and the Shaping of the Slave Labor Force,” in Cultivation and Culture: Labor and the Shaping of Slave Life in the Americas, Ira Berlin, Philip Morgan, eds. (Charlottesville: University Press of Virginia, 1993), 95–98. To these I have added more ethnicity data from the plantations listed below [“SDOM” and “Not reg” refer to Archives d'Outre-mer, Aix-en-Provence [hereafter: ANOM], Notariat de Saint-Domingue. Regrettably, renumbering of the notarial archives has rendered some of the “Not reg” citations no longer current.

Anne: SDOM 857, 13.1.1779; Ardisson: Not reg 1387, 31.10.1783; Auguié de Lascary: Bibliothèque nationale, Paris, Manuscrits, NAF 22,367, fol. 151; Auriol: National Archives, London (TNA), HCA 30/304; Beloeil: Jérémie Papers, University of Florida, Gainesville [hereafter: JPUF], 5-115; Boccalin: SDOM 1281, 30.3.1786; Bouché (1779): JPUF, 5-26; Bourgeois: SDOM 1518, 29.2.1780; Bourgogne: SDOM 117, 26-28.3.1791; Brieux: JPUF, 1-125, UF; Brisson: JPUF, 20-3; Buyter: TNA, HCA 30/273; Canonge/Delaud: JPUF, 6A-131; Carnave: SDOM 258, 7.3.1792; Cavayé: Bibliothèque Mazarine, Paris, Ant. Ms. 21/2; Chapui: SDOM 1155, 23.4.1777; Chatules: Archives départementales (AD) de la Vienne, Poitiers, F3, Papiers Gilbert; Chollière: Not reg 1246, 17.7.1783; Clément/Baudu: SDOM 869, 22.3.1788; Cognac: SDOM 1348, 21.1.1779; Collas de Mogent: SDOM 758, 24.1.1785, 2.5.1785; Cosson de la Sudrie: AD de la Dordogne, Périgueux, 8J25, 14.5.1782; Couderq: in Bernard Foubert, “Un Agenais à Saint-Domingue,” Revue de la Société Haïtienne d'Histoire et de Géographie 183 (1995): 1–23; Coueson: SDOM 755, 19.5.1777; Courtois: SDOM 789, 15.6.1787; Daguzan/Dupetit: SDOM 525, 20.4.1779; Dangluze: JPUF, 5-100; David de La Chapelle: SDOM 616, 25.2.1777; De Saux: Not reg 1396, 14.1.1780; De Sevré: SDOM 868, 18.9.1787; Désir: G. Debien & M.-A. Ménier, “Toussaint Louverture avant 1789,” Conjonction 134 (1977): 67–80; Desmortiers: SDOM 525, 12.3.1779; Dessources/Jacquemin/Labatut: SDOM 858, 12.12.1779; Diobonne: SDOM 1397, 11.5.1781; Dolle & Raby: Archives nationales, Paris, MC/Momet/XVI/853; Douret: JPUF, 1-55; Drouillard (Antoine): SDOM 1518, 21.6.1781; Dubuc Saint-Olympe: SDOM 117, 2 and Jan. 14-20, 1791; Ducasse: ANOM, SDOM 1397, 12.6.1781; Dumaine: in Roseline Siguret, “Esclaves d'indigoteries et de caféières au quartier de Jacmel (1757–1791),” Revue Française d'Histoire d'Outre-mer 55 (1968): 190–230); Dumoulin children: JPUF, 19-94. 18.8.1782; Dupetit (2 properties): Not reg 1386, 20.3.1783; Dupon: SDOM 758, 29.11.1788; Dupont de Boisguy: AD d'Ille-et-Vilaine, Rennes, 2E d 93; Falret: SDOM 1551, 1784; Faurestier: Not reg 1386, 17.3.1783; Favereau: SDOM 1551, 14.2.1784; Fignoux: JPUF, 6a-44; A.J. Fouché: SDOM 191, 8.8.1785; Gauffreau: SDOM 788, 9.2.1786; George & Mallet: JPUF, 19-60; Gilles (2 properties): Not reg 1384, 18.7.1781; Glier: SDOM 1348, 8.2.1779; Goux: ANOM, Greffe 21, fols. 362-363; Grandmaison: Not reg 1156, 13.10.1787; Gueye: SDOM 1012, 20.2.1786; Hamelin/Ducasse: JPUF, 6a-3; Jouon/Bussière: Bibliothèque Municipale, Versailles, Ms Lebaudy 141; La Bachelerie: SDOM 17, 11.7.1791; Labole: TNA, HCA 30/280, 3.10.1777; Labut: SDOM 616, 4.11.77; Lamard: ANOM, Greffe 175, 3.1.1791; La Mare aux Oies: SDOM 117, 28.2.1791; Lamaud (2 properties): SDOM 1713, doss. 5, 3.1.1791; Lasplace Daulhem: Notsdom 1713, dossier 6; La Tremblaye: Archives nationales, Paris, MC/ET/XXX/512; Lecoq de la Terronière: JPUF, 11-28; Lefebvre Deshayes: JPUF, 20-9; Lefèvre (Etienne): LDS microfilm, Charleston Consulate, vol. 5, 28.12.1808; Legendre: SDOM 1348, 28.2. and 27.4.1779; Legurié: TNA, HCA 30/304; Lehoux: Boston Public Library, Ms Haiti 66-161; Lelevain: JPUF, 6a-66; Lemaire (Limonade): Not reg. 1390, 31.5.1786; Le Métais Milon: Eric Pouillevet, “Une caféière du Nord de Saint-Domingue,” Généalogie et Histoire de la Caraïbe (Sept. 2012); Le Paroy (5 properties): ANOM, 164 APOM, 1786; Les Faurier: SDOM 290, 5.9.1786; Ligny: JPUF, 6a-64; Lo: SDOM 1551, 5.11.1784; Maffrand: JPUF, 9-7; Maigné de Salanauve (2 properties): Not reg 1155, 25.2.1777, 12.3.1777; Maret Dumeny: SDOM 616, 4.5.1777; Martelly: AD de la Gironde, Bordeaux, 73 J 60, doss. 319; Martin: SDOM 117, 9.2.1791; Menbrède de Laglaire: SDOM 1281, 21.5.1786; Merceron: SDOM 117, 23-27.8.1791; Mercier (Bayardel): JPUF, 6a/10; Merot dit Clermont: Not reg 367, 28.10.1784; Mondion [ex-Chavannes]: SDOM 291, 2.1.1787; Monnereau/Pinot: Not reg 1245, 20.12.1782; Mosneron: J. L. Donnadieu, Un grand seigneur et ses esclaves (2009), 115, 313–314; Moulin de Récy: SDOM 865, 11.7.1778; Oudart: SDOM 788, 26.6.1786; Pauvert (2 properties): JPUF, 5-135B&C; Payas: JPUF, 1-122; Père: SDOM 117, 21.11.1791; Pigal: JPUF, 5-127; Rifflet: SDOM 616, 1.12.1781; Robert de la Bressaudière: AD de la Sarthe, Le Mans, 1Mi4(R1); Robion (3 properties): TNA, HCA 30/304; Tangui: SDOM 1306, 24.12.1777; Testas: JPUF, 20-11, 19.3.86; Toiry: SDOM 1124, 4.12.1786; Tripier (2 properties): JPUF, 6a-60; Vallée de la Frenaye: JPUF, 5-145; Vignault: JPUF, 6a-26; and 14 additional lists from Gabriel Debien, “De l'Afrique à Saint-Domingue,” Revue de la Société Haïtienne d'Histoire et de Géographie 135 (1982): 7–73.

As a first step in testing Thornton's claims about the presence of Kongolese and Africans, I have assembled a new database drawn from inventories of 126 sugar estates, 176 coffee plantations, and 50 indigo and cotton plantations dating from the period 1770 to 1791. Far larger than any previous study, it contains details on 31,382 slaves of known origins, and it represents most of the colony's 52 parishes.³⁸ Table 1 presents the results and distinguishes slaves labeled “Congo” from a slightly larger group that corresponds to maximum possible arrivals from West-Central Africa. It includes, along with Congos, “Mondongues” and “Damba(u).” Mondongos were a small ethnic group well known in American slave societies for their filed teeth, facial markings, and carnivorous reputation. They originally spoke a non-Bantu language and lived northeast of the Bakongo among the Tio.³⁹ Dambau might refer to Ndembu, the southernmost province of the Kongo kingdom, which was inhabited by a mixture of Kimbundu, Kikongo, and Lunda speakers.⁴⁰ In Saint-Domingue, they appeared almost exclusively on northern sugar estates.

The term “Congo” apparently applied to speakers of dialects of the Kikongo language cluster, but because considerable controversy has surrounded the distribution of slave trading within West-Central Africa, there has been much uncertainty as to captives’ origins and whether the term also covered slaves from other language groups. Although French trading was concentrated on the Loango coast, north of the mouth of the River Congo, some have hypothesized a significant French penetration of the Portuguese monopoly to the south, and the scholarship has sometimes evinced an almost romantic attraction toward the idea of long trade routes extending far into the interior. Marcos Almeida and Daniel Domingues have shown the shortcomings of the latter approach in regard to the nineteenth-century slave trade, but they focus on Angola.⁴¹ Since a tiny number of French ships certainly did trade at Luanda and Benguela, it is likely that at least some Kimbundu and Umbundu speakers reached Saint-Domingue and were regarded as Congo.⁴²

Christina Mobley formerly asserted that most Congo in Saint-Domingue came from regions to the south and east of the Kongo kingdom, but she has switched to claiming they in fact originated in the northern Kongo culture area, from the Loango coast and its Mayombe hinterland.⁴³ She thus continues to reject John Thornton's emphasis on the kingdom's civil wars as the main source of captives. He argues that “at least half” of those sold on the Loango coast came from the kingdom (primarily south of the River Congo), but he also notes that the kingdom probably replenished its population to an unknown extent by importing outsiders.⁴⁴

The evidence from Saint-Domingue offers little support for an expansive reading of the term “Congos” but it is hardly conclusive. If the general association, widely attested in the Americas, between Congo slaves and Catholicism points toward the kingdom's heartland south of the River Congo, all the colonial descriptions of magico-religious practices are more suggestive of the north Kongo region.⁴⁵ The Bantu vocabulary found in colonial texts and modern Haitian Creole appears to be principally Kikongo, as some scholars have long assumed, and Mobley's recent work now details also with an emphasis on the north.⁴⁶ A more direct view of the issue comes from colonists’ property inventories, which on very rare occasions identify different Kongo subgroups.

The present data set has yielded 93 such cases, to which I added another 55 taken from fugitive slave notices in the colonial newspaper Affiches Américaines (1788 and 1790). The largest subgroup identified, constituting 27 percent of the cases, was Nsundi (Moussoundi, etc.), who lived south of the River Congo in the north-central province of the old kingdom. Those from the Kongo heartland (Congo Franc, Congo Bord de Mer, Solongo, Moussicongo) added another 20 percent. By including the Mazinga and Sanga, who lived north of the River Congo or alongside it, we can say that about 52 percent of the sample originated in the old kingdom. Mayombe was the second largest subgroup, with 21 percent, and together with the Mazinga, Sanga, Kamba, and Loango, Gabingue (Cabinda), and Malembe, all those from north of the river made up 37 percent of the total. Another 13.5 percent were Yaka and Mbala (Mayaque, Monbala, Diaga, Mossoco), who came from the eastern frontier of the Kongo kingdom or beyond, and spoke languages from outside the Kikongo cluster, though related to it.⁴⁷ The sample is small and invites caution, as we cannot tell why a colonist chose to assign a particular individual to a subgroup while applying the generic “Congo” label to others.⁴⁸ Whatever its limitations, the evidence leans somewhat more toward Thornton's reading of the term than toward Mobley's.

As can be seen in Table 1, the demand for Kongolese workers was strongest in the coffee sector and especially in the colony's north, where about half of the colony's coffee plantations were located.⁴⁹ Although most sugar planters tried to avoid buying them, this was hard to do in the north where coffee planters and Congos dominated the regional slave trade. Despite their prominence among the African-born, however, Kongolese were considerably less prominent in the regional slave population as a whole, because the north was also the most creolized of the three provinces, the one with the most American-born slaves. The expanded category of West-Central Africans added only about two percent to the number of Congos on sugar estates, and on all plantations in the west even less. The negative reputation of Mondongos and sugar planters’ selectivity was doubtless responsible. Coffee and cotton planters, hurrying to bring new lands into cultivation, were less selective. In the north, Mondongos made up more than four percent of their African slaves and on the south coast, neglected by French slave traders, they constituted between four and eight percent in the non-sugar sector.

Even substituting the broader category of West-Central Africans for that of Kikongo speakers, it should be apparent from Table 1 that the claims of Thornton, Dubois, and Mobley are very wide of the mark. To ascertain by how much requires estimating the demographic weight of each regional/crop sector relative to the slave population as a whole. This can be only an educated guess, because information about non-plantation slaves is very sparse, and although the census provides plantation numbers, establishing average workforce sizes is problematic.⁵⁰ I derived averages initially from an expanded data set of 275 sugar plantations, 459 coffee plantations, and 127 indigo/cotton plantations from the period 1780 to 1791, but the result was a population of implausible size, because the records of large plantations have survived more often than those of smaller ones. This was especially true of the indigo sector, which remains easily the least documented and understood part of Saint-Domingue society.

Fortunately, in the 1820s, when the French government was discussing financial compensation for colonists’ losses during the revolution, there were two independent investigations of the structure of property ownership in the colony. The Indemnity Commission and the veteran colonial bureaucrat Charles Wante did not agree on all the details, nor whether the true slave population in 1789 should be estimated at 470,000 or 480,000, and they did not look into regional variations, but they both reckoned that plantation slaves constituted 85.5 percent of the total.⁵¹ The slaves who worked in market gardening, jobbing gangs, haulage, ranching, and industrial and maritime trades, about whom we know least, were thus relatively few in number.

We can reduce the size of this blind spot further by adding the only existent analysis of an urban population, that of Cap Français, the largest town, and applying it to all the towns in the north.⁵² Table 2 employs this material together with the sectoral totals advanced by Wante and the Indemnity Commission, adjusted by me, and produces a sample that should accurately reflect the ethnic composition of some 88 percent of Saint-Domingue's slave population and 96 percent of that in the north. It shows that Congos overall made up 24 percent of the slave population.

Table 2 Distribution of Congos Among Saint Domingue Slaves, 1789

Sources for Table 2 include:

1. Plantation numbers and the assumed total population (480,000) are from the 1789 census. Average plantation populations are estimates.

2. Regional population refers only to the sectors identified. The proportion of the regional populations omitted from the analysis was probably between 8 and 12%.

3. Congo plantation sector percentages, and regional boundaries, are from Table 1.

4. Urban data are from Geggus, “Slaves and Free Coloreds of Cap Français” and “Urban Development in Eighteenth Century Saint Domingue.”

5. Using the 1789 census and the regional boundaries defined in Table 1, the regional populations were weighted: north 0.379; west 0.394; south 0.227.

Table 3 shows the distribution of enslaved Africans and locally born Creoles. Foreign Creoles from other colonies (French, Dutch, Spanish, and Anglophone) formed another half percent of the slave population. Of the 31,704 slaves of known identity in the rural and urban samples, only 7 were described as Indian. Africans, it seems, formed just over half of the slave population. They were at their most prominent in the rapidly developing south, most densely concentrated in the mountains of the west, and least prominent in the north, where the Haitian Revolution began.

Table 3 Distribution of Saint Domingue Creole and African Slaves, 1789

Source: All data and the definitions of regions and regional populations derive from Tables 1 and 2. In Table 3, Creole = born in Saint Domingue. The gap between Creole and African totals and 100 percent is due to the number of Creoles born in other American colonies.

Creoles predominated on the large sugar estates of the northern and western plains. In the long-settled north, they narrowly outnumbered Africans in the countryside, despite the region's focus on coffee, where Africans always dominated. About 60 percent of Creoles in the coffee zone and in the south's sugar sector were children (aged under 15), but elsewhere a clear majority were adults. The only place in the colony where Creole adults outnumbered African adults was in the western half of the North Plain—precisely where the Haitian Revolution began.⁵³

It remains to be seen how closely this depiction of the 1789 population corresponds to that of mid 1791, at the outset of the slave uprising. The question is important, because the number of African arrivals in Saint-Domingue reached an all-time high during the intervening period. Unfortunately, no census was taken in these years; the head counts commonly cited as being for 1790 or 1791 were really arbitrary adjustments of the 1788 census, marred by various misprints.⁵⁴ From July 1789, as the final census count was closing, through August 1791, slave ships disembarked in the colony some 84,300 captives.⁵⁵ In view of the mortality estimates discussed above, it may be reasonable, even conservative, to assume these migrants suffered losses of 25 percent during their first year after arrival, and another 10 percent during their second year, especially as this period saw a worsening subsistence crisis exacerbated by the severest drought in living memory.⁵⁶ Their numbers may thus have been reduced to about 67,000 by August 1791.⁵⁷ In the meantime, the mid-1789 slave population of 480,000, experiencing normal vital rates of 5 percent mortality and 1 percent natality, would have shrunk to 442,370. Hence the Haitian Revolution began in a population that probably contained fewer than 510,000 slaves.

That population would have been more African than in 1789 but not by much. Creole slaves always enjoyed positive rates of natural growth—otherwise they would not have existed, let alone steadily multiplied decade by decade. If they had merely maintained their numbers after 1789, the African proportion of the slave community would have risen from 53.5 to 56.2 percent. Even then, with the resident white population numbering at least 35,000 and free people of color perhaps 32,000, Africans would still have constituted just under half of the colonial population.⁵⁸ Plantation data from the last three years of the data set indicate a change of similar or lesser magnitude. Although the number of indigo/cotton plantations in the sample was too small to permit a comprehensive analysis, we see a small general increase in the African presence that was most pronounced in the southern region. The prominence of Congos among the Africans grew in the south but in the north and the west it actually declined somewhat, along with West-Central Africa's dominance of the local slave trade in these years.⁵⁹

Extreme Kongomania

This question of numbers needs close attention because the strength of Thornton's arguments about the Kongolese contribution to the Haitian Revolution depends a good deal on demography, and this is even more true of James Sweet's “New Perspectives.” This ambitious article claims that Kongolese ways of thinking caused “ideological transformations” among Saint-Domingue colonists, who readily assimilated “the world of Kongolese ideas.” Citing Dubois's (erroneous readings of) slavery statistics, Sweet states that, “the presence of Kongolese ideas should come as no surprise.” He repeats, seemingly as accepted fact requiring no documentation, the opinion that nearly two-thirds of the slaves were African, that half of those had arrived in the five years before the revolution, and that the majority were Kongolese, which he later renders as “the predominance of west central Africans in Saint-Domingue on the eve of the Revolution.”⁶⁰

Even if the demographic underpinnings of Sweet's argument conformed more closely to reality, the weakness of several aspects of its articulation would raise questions as to its credibility. Kongolese ideas, he writes, “penetrated into the very fabric of eighteenth-century St. Domingue society, reaching even French planters like Baudry [des Lozières], Moreau de Saint-Méry, and others.”⁶¹ Who the “others” were, however, Sweet does not say, nor does he show how the prolific chronicler Moreau de Saint-Méry (who was not a planter) participated in this syndrome. Sweet's thesis actually rests on the single, and singular, case of Baudry des Lozières. He was a lawyer who fled Saint-Domingue during the revolution and later published a list of Kikongo words and phrases that he had learned from his own slaves.⁶² The fact he could recall some 800 words and phrases after ten years of exile, and claimed to have used them to promote good relations with his slaves, leads Sweet to depict Baudry as having extensive conversational ability in Kikongo. His reading of Baudry's translations causes him to praise his “depth of cultural understanding.” He was “inexorably drawn into the social and intellectual worlds of the enslaved Africans on his plantation.” And because of the apparent intensity of the Frenchman's engagement with the language, Sweet concludes: “it appears that it might have been a lingua franca.”⁶³

This seems highly improbable, since colonists’ inability to communicate with their newly purchased slaves was Baudry's stated reason for publishing his dictionary, and in his surviving manuscripts, which Sweet did not consult, Baudry makes clear that very few whites in Saint-Domingue could speak an African language.⁶⁴ As is well known, Moreau de Saint-Méry, one of the colony's leading intellectuals, and the Creole planter Louis Drouin de Bercy each published accounts of colonial Vodou that included chants in Kikongo, but neither man could identify the language. Moreau referred to his example as “an African song,” and Drouin provided a translation that is entirely spurious.⁶⁵

Baudry's command of the language may also not have been so impressive as Sweet suggests, nor was his focus on it as particular as appears at first sight. His ability to reproduce his Kongo vocabulary after ten years of exile was not solely a feat of memory but owed much to the fact that he had been able to salvage his written notes on the language when he fled the colony in 1792. He also benefited, he discreetly mentions, from the knowledge of a ship's captain who knew the African coast well. The captain evidently was responsible for the extensive geographic background in the “Dictionnaire ou Vocabulaire Congo“ and doubtless for the inclusion of many trade goods in the word list, and possibly for more. The reason Baudry had written notes on Kikongo was that, for ten years, he had been preparing a study of “all the jargons of Africa” as part of a projected colonial encyclopedia. He also claimed knowledge of Arada (Fongbe) and Igbo, and in later life he compiled a “Glossa Polygène” of 18 languages, mainly European.⁶⁶ He was thus something of a philologist; his engagement with Kikongo grew out of his general linguistic interests rather than any special status the language had in Saint-Domingue.

Baudry's linguistic activity is probably better described as phrase-book compilation than philology. His recommended model for language learning was to combine basic vocabulary and auxiliary verbs in the infinitive form without worrying about any aspect of grammar or syntax. He in fact states that African languages do not have grammar. Recalling his experience of exile in North America as one of linguistic isolation, he adds that his own aural comprehension and pronunciation skills were very poor.⁶⁷ It is also relevant that, although Baudry spent about 17 years in Saint-Domingue, he was for most of that time a lawyer in Cap Français. He did not acquire a plantation until 1788, and then had just a couple of years there before he was caught up, as a paramilitary leader, in the revolutionary struggles. His opportunities for sustained and close interaction with Africans were thus not extensive. All these factors combined raise serious doubts as to whether Baudry's knowledge of Kikongo would have given him the conversational ability, still less the deep cultural understanding, that Sweet ascribes to him.

Several of the specific details of Sweet's discussion are similarly far-fetched. Baudry uses the word ndoki (witch, supernatural power) to translate brigand, the term French colonists used to describe insurgent slaves (and metropolitan French applied to insurgents in Europe). For Sweet, this means that Baudry saw the revolution through Kongo eyes as a disorienting social rupture explicable only by supernatural influence. However, he fails to note that Baudry used ndoki to translate “poisoner” and “scoundrel” as well, and there is no evidence he was even aware of the word's supernatural connotations. Far more likely is that his interlocutors had taught it to him as a smear, in its broader connotation of “those having evil designs.”⁶⁸ The planter's rendering of “France” as m'poutou, which apparently derives from a corrupted form of “Portugal,” causes Sweet to remark, astonishingly, that the Kongolese homogenized all white Europeans, and thus that with France indistinguishable from Portugal, the French Revolution's influence on Saint-Domingue's slaves is questionable. Yet, whatever the word's etymology may have been, there is no doubt that Baudry understood pouto to mean “country, territory, land”—including pouto fioté (“land of the blacks”).⁶⁹ The insurgent slaves’ astute and well-known attempts to manipulate Franco-Spanish rivalry during the Haitian Revolution do not indicate any confusion in this area, quite the opposite.

The word Sweet pays most attention to is vika. In slightly variant forms, the term was widespread in West-Central Africa and in meaning it seemingly spanned the field “slave, dependent, servant.”⁷⁰ Sweet's argument is that enslaved Kongolese in Saint-Domingue chose to describe themselves as vika because it denoted a serf-like status of dependent superior to that of “bought slaves.” The distinction between slaves born into a society and outsiders acquired by purchase or capture is generally seen in African Studies as highly significant with important implications as to the likelihood of being (re)sold. On the spectrum of terms for slave-like status, the precise valence of vika, however, is far from obvious. Eminent authority Jan Vansina thought the word “came to refer to the traded slave”; Baudry used it to translate “captive” as well as “slave,” and Thornton guessed that in seventeenth-century Kongo it might have designated the worst type of slavery.⁷¹ French slave trader and author Louis de Grandpré did not claim “that the term for slave was not . . . vika,” as Sweet seems to imply; de Grandpré simply did not mention it at all.⁷²

While acknowledging that the term is “admittedly ambiguous,” Sweet builds on this unstable foundation to hypothesize that Kongolese brought to Saint-Domingue a conception of slavery centered on reciprocity that featured “the ‘right’ to run away from cruel masters, and even master exchange.”⁷³ All the evidence of ritualized master exchange in West-Central Africa dates from the mid and late nineteenth century, and its prevalence is uncertain, although it was apparently important enough for putative traces of it to show up in Brazil and Curaçao.⁷⁴ The two eighteenth-century sources Sweet cites on this issue make no mention of any customary “rights,” and they emphasize, rather than the initiative of fugitives, the fractious competition between landholders to lure away one another's laborers in a region of low population density.⁷⁵ Anthropologist Wyatt MacGaffey argues that the idea of rights was completely alien to Kongo slavery, and he reports that vika could be sold or even killed for disobeying their master.⁷⁶

If Sweet's picture of Kongo slavery invites skepticism—and I can claim no expertise in this area—his claims regarding its influence on Saint-Domingue and the Haitian Revolution are even more strained and implausible. In the first place, two basic questions are not addressed at all. Were the Congos in Saint-Domingue more likely to have been vika dependents in Africa rather than war captives or traded slaves? And why would former vika dependents expect to maintain their “rights” once they had been reduced to the status of “bought slaves” and classic outsiders by the Atlantic slave trade?

In Sweet's view, the expectations of deported Kongo led them to demand from French planters “‘freedoms’ [that] apparently included broad patronage networks, master switching, and even marronage metaphors related to cattle.”⁷⁷ The reference to “patronage networks” alludes to cases of escaped slaves who wished to return to their master without being punished and who were able to use an intermediary (typically one of the master's neighbors or relatives) to intercede on their behalf. This practice was well established in Saint-Domingue and it merits more research, but it has no discernible connection to Congo slaves. Enslaved Creoles were undoubtedly best positioned when seeking pardons, because of their kin connections and longer time in the colony. Occasionally it was the slave owner herself who surreptitiously arranged for an intermediary to request a pardon.⁷⁸

Although Sweet claims that master exchange was the “essence” of vika status, he offers no evidence that it ever existed in Saint-Domingue, and nor, to my knowledge, has anyone else.⁷⁹ Instead, he alludes to two advertisements for missing Congo fugitives who, after being sold, supposedly “each fled back to the plantations of their previous masters in the hopes that these more benevolent patrons would reclaim them.” On further examination, however, it transpires that this (ascribed) motive was merely suspected in one case, whereas in the other the slave was simply “believed to be located in the district of his ex-master.”⁸⁰ As it was extremely common for runaway slaves to linger in the vicinity of the plantations they had fled to be near friends and on familiar ground, this constitutes a rather cavalier use of evidence.

The mysterious “marronage metaphors related to cattle” proves to be a confused attempt to blend the etymology of “maroon” (derived from sixteenth-century Spanish ranching) with three highly unusual advertisements for Congo escapees who each “allegedly walked around with bridles in their hands claiming that they were searching for their master's cattle.”⁸¹ Rather than assume that the bridles (actually halters: licous) were a prop to camouflage the slaves’ escapes, Sweet proposes that “these Congos carried the bridles as symbols of their dependent status as runaway vika looking to attach themselves to new masters.” Yet, as cattle were quite rare in the Kongo culture area, it is questionable whether this would have been a resonant metaphor for the slaves any more than for the colonists, especially as the three men in fact said they were looking for chevaux, horses not cattle.

Adding to the confusion is the Kongo term bika, which is glossed in eighteenth-century sources as “to leave, let go of,” and “to wait.” Sweet states that the three fugitives with the halters “exemplify the full field of meanings” of both bika and vika, although he does not elucidate any etymological relationship between the words. This is regrettable, because he further suggests that, “operat[ing] side-by-side,” the two words “open onto a field of meaning that might explain the prevalence of petit marronage” (short-term absenteeism) in Saint-Domingue.⁸² The argument appears to be that, because of their Kongo background, many of the slaves who fled their plantations in Saint-Domingue went in search of a new master. As fugitives numbered well over 2,000 per year by the mid 1780s, the historical record would surely not be so quiet on the matter, if this were true.⁸³ Even less convincing is the claim that the same Kongo influence explains colonial “tolerance of slave gatherings at markets and provision grounds.” Broadly favored by the white population and common in much of the Caribbean, these institutions had originated long before West-Central Africans were numerous in the colony.

In Sweet's brief account of the black revolution in Saint-Domingue there is much to criticize. I will pick two or three points. First, is the idea, popularized by Laurent Dubois, that the August 1791 uprising began as a reformist movement that only later turned revolutionary. Reformist proposals certainly did issue from rebel camps during the uprising's early months, but their most radical demand, that the French abandon the colony, was actually the very first they made. Sweet's contention is that the slave leaders’ failure to demand the abolition of slavery reflected the dominance of Kongo values among the insurgent masses; they remained accepting of a dependent lifestyle until, in the fourth month of the rebellion, the colonial authorities broke off negotiations. However, as many historians have observed, there was another reason those negotiations failed: their vociferous rejection by the insurgent masses when they belatedly realized that their Creole leaders were planning to sell them out. The reformism came from the leaders, who were elite slaves and freemen of color, not their African followers.⁸⁴

In an extraordinary flight of fancy, Sweet asserts that the colonists’ rejection of the December 1791 negotiation “crushed the customary rights of the maroon[;] . . . ‘cows’ could no longer be allowed to roam freely.”⁸⁵ This makes no sense at all. The negotiations had nothing to do with marronage, and while colonists had been forced to accept absenteeism as the cost of doing business there never was a customary right to roam freely, as innumerable recaptured fugitives who were whipped, loaded with chains, branded, or had an ear severed could attest. Eager to believe that French colonists absorbed “Kongo ways of being,” Sweet imagines that slaveowners were “apparently” forced to comply with their slaves’ wishes, at least until shortly before the Haitian Revolution. Then, “evidence suggests,” successful marronage became increasingly difficult, and this helped cause the uprising. That evidence, however, is environmental (the clearing of mountain forests) and diplomatic (an extradition treaty with Santo Domingo).⁸⁶ It has nothing to do with the “curtailing” of traditional “concessions,” or “the violation of the masters’ reciprocal obligations,” or the failure of a “social contract.” That is entirely imaginary.

Sweet notes that the contribution of fugitive slaves to the Haitian Revolution has sparked a “robust” debate. It is remarkable that, although he ignores most aspects of that debate and does not discuss any of them, he still feels able to pronounce that “viewed through the Kongo optic presented here, it is clear that maroons played a crucial role in the revolution.”⁸⁷

Conclusion

The work of Thornton, Mobley, and Sweet on the Haitian Revolution exhibits a chain of errors, each building on their predecessor's mistakes, that moves from exaggeration toward fantasy. To understand how prominent scholars could have produced this questionable scholarship, several factors are worth considering. The root problem, on which I have concentrated here, is one of numbers. Numbers have been out of fashion in the writing of history for a long time. Some frankly disdain them in favor of concepts they deem more sophisticated. Yet, as this article seeks to show, a careless approach to quantification can seriously distort reality.

That Congos were not 60 percent of the slave population in northern Saint-Domingue but less than half that number should matter to anyone seeking to understand Haiti's landmark revolution. Scholarship that asserts that “Central Africans [formed] the majority of the population on the North Plain”—when they were in fact less than a fifth--does not inspire confidence.⁸⁸ Those who wish to depict the Kongolese as central to the revolution ought to consider that in the region where the slave uprising began, the western half of the North Plain, their presence was unusually small compared to most other parts of the colony. On the other hand, the slave population of that region stands out as the most creolized in Saint-Domingue and the only place where Creole adults outnumbered African adults. This surely has some bearing on the fact that, from beginning to end, the main leaders of the black revolution were American-born Creoles. This, however, is often overlooked.

According to the best available estimate, slaves labeled as “Congo” made up some 24 percent of Saint-Domingue's enslaved population when the Haitian Revolution began. They were not, as Laurent Dubois believed, “the largest group among the slaves.”⁸⁹ Creoles were almost twice as numerous and constituted by far the largest ethnic group in the colony. Kongolese were, nonetheless, the second largest. This is reflected in their impact in the areas of magic and religion, which has drawn scholarly attention in the last 40 years. While Aja-Fon culture has been more influential in Haiti, that is usually attributed to its earlier transplantation.⁹⁰

Although locally born slaves far outnumbered Congo migrants, the numbers of Creole and Congo adults must have been quite similar, because close to half of enslaved Creoles were children. As the Congos’ male:female ratio was much higher than the Creoles’, it seems likely that the armies of the Haitian Revolution contained somewhat more Kongolese than Creoles. Yet this is not certain, because after 1791 Africans were more likely than their Creole counterparts to remain in agriculture. What is quite clear is that the 1791 uprising was organized and led by American-born slaves. Beyond the well-known cases of Boukman, Jean-François, Biassou, and Jeannot who early emerged as the dominant leaders, everything discovered about the revolt's planning stage implicates persons who were either of mixed racial descent or who had occupations then typically held by Creoles.⁹¹

During the first year of the slave uprising, district commanders in the central zone of insurrection all appear to be Creole, and often free blacks or men of color.⁹² Free blacks, most or all of them Creoles, were also prominent among camp commanders.⁹³ Very few African leaders appear at this time in the archival record: Zagada, who commanded a camp of Hausa and Arada, and the free black Gracia Lafortune who presumably was Yoruba, like his followers. Of the leaders generally regarded as Congos, only Sans-Souci is mentioned at this time, first as the second in command of a mulâtre libre and then as a camp commandant.⁹⁴ Macaya and Pierrot emerged only in 1793, and few others appeared in the north for the rest of the 1790s. Considering the heavy concentration of Kongolese in the northern mountains, this is remarkable. When the slave uprising spread into the northeast region, little studied by historians, it was controlled largely by free men of color until Jean-François succeeded in asserting his influence there.⁹⁵

Elsewhere in the colony, free men of color were much more powerful and insurgent slaves less independent. As in the north, rebellions in the sugar zone were led by Creoles, such as Hyacinthe in the Cul de Sac plain and Armand, Martial, and Gilles Bénech around Les Cayes. In the mountains, local leaders were usually Africans like Gilles Bambara and the Yoruba Alaou, but the ethnic identity of most is unknown. The only Congo I am aware of is Laplume from the mountains behind Port-au-Prince; a former subordinate of Alaou, he became a French general in 1796.⁹⁶ The charismatic shaman Romaine la Prophétesse has occasionally been claimed as a Kongolese maroon, but he was a mixed-race landowner from Santo Domingo, and his Marianism was of Spanish, not Kongolese, origin.⁹⁷

Similarly, despite the much-commented case of Macaya, who proclaimed himself a subject of the kings of France, Spain, and Kongo, there is no evidence to connect with Africa the royalist stance more generally affected by the northern slave insurgents. Those who most prominently and effectively played this role were the main Creole leaders. Macaya's fellow Congo Pierrot exhibited few royalist tendencies and, like Laplume, was more easily won over to the French Republic.

Forming a quarter of the slave population, West-Central Africans undoubtedly bulked large among the foot soldiers of the Haitian Revolution and, as Thornton argued, their putative military experience presumably compensated for the Creoles’ complete lack of it. They do not, however, stand out in any respect during the revolution's first decade. Only during the conflict's final 18 months did a plurality of Congo leaders gain prominence, and it was in a rather ironic manner. In the summer of 1802 about a dozen African leaders in the northern mountains successively went into revolt, refusing to be disarmed by the French occupiers and the Creole generals then collaborating with them. Contemporaries called them “Congos,” which in this period became a generic term for African. Given their location, most doubtless were Kongolese but some, like Yayou and Petit Noël Prieur, were locally born. Although their revolt began what became the war of independence, most continued fighting not only the French but also the Creole generals led by Jean-Jacques Dessalines after they finally took up the cause. A few (Jacques Tellier, Cagnet) switched sides and joined the French, so great was their hatred of the Creole elite.⁹⁸

For Haitian nationalists, these “Congos” were anarchic traitors. Thomas Madiou, the country's first historian, called them previously unknown “parvenus” who dominated in only three parishes—which was an exaggeration but largely true. Evoking a clash between barbarism and enlightenment, Madiou depicted the “Congos” as an obstacle to the achievement of national independence that had to be eliminated by the Creole generals, supported by the mass of Creole plantation workers.⁹⁹ Nowadays, this Congo episode is more likely to be seen as embodying the essence of Haiti and its revolution, and there is truth in this. But it should not obscure the fundamental point that the Haitian Revolution was launched and brought to fruition by Creoles. From 1791 to 1804, they exploited the positions of authority they had occupied in slave society, using their kin networks and greater knowledge of the outside world in a contentious struggle with an African population that at first outnumbered them.¹⁰⁰ Both groups, of course, shared in the ending of slavery and colonial rule, but none of the 37 signatories of the declaration of independence were Africans.¹⁰¹

The tendency to underplay the role of Creole slaves in the revolution, shared by the modern historians discussed here, forms part of a broader trend in slavery studies to assert the resilience and value of all things African, which has coincided with declining acceptance of the creolization thesis of Sidney Mintz and Richard Price. Dubois is explicit in viewing the Haitian Revolution as “an African revolution.”¹⁰² Perhaps this ideological leaning helps explain how published data on the composition of the slave population has so often been misread, with the Creoles screened out and references to “African slaves” interpreted as meaning all slaves.

The numbers problem extends not only to population figures but to more basic aspects of Saint-Domingue's economic structure as well. Thornton's statement that “the majority [of the colony's] slaves worked in sugar estates in the northern half of the island” should startle anyone who has more than a passing acquaintance with its history. Published materials, easily accessible for two centuries, would have suggested a figure of 12 percent; those presented here suggest 10.5 percent. Dubois, author of leading studies of Haiti and its revolution, similarly writes that “most” slaves worked on sugar estates (as opposed to 30 percent).¹⁰³

To some, this may seem an agricultural factoid, a banal matter of commodities, but 60 years of comparative slavery scholarship makes clear there was no more important influence on enslaved workers than whether they worked on sugar estates. It had enormous implications regarding the conditions in which slaves lived, and thus for historical analysis of the revolution. The influence of this sort of misinformation is lamentable. Saint-Domingue was by 1791 the world's major exporter of coffee, and after 50 years of development, coffee was the economy's leading sector in terms of population, land use, and exports. However, a recent study by a younger historian presents it as essentially a sugar colony where the mountains were largely uncultivated until they were opened to coffee by insurgents during the revolution.¹⁰⁴ A more common error has been to associate coffee with the South and West provinces and with free people of color, whereas white-owned plantations in the north were always dominant.¹⁰⁵

Inadequate attention in the historiography to such sizeable aspects of the Haitian Revolution is rarely accompanied by a sound grasp of its complex political narrative. This points to a final factor deserving mention: the history of the Haitian Revolution has generally come from writers possessing a limited acquaintance with its primary sources, and even its secondary literature, as well as little knowledge of the prerevolutionary period. Most general histories of the revolution appear to have been their author's first publications on the subject (something surely unimaginable in the case of other major revolutions). Africanists coming from outside the field, Thornton and Sweet are not unusual in this respect. The expertise they bring from their respective specialties is insightful and stimulating, but it inevitably comes at the cost of “local knowledge.”

Because the numbers of Kongo in Saint-Domingue were much smaller than John Thornton imagined, their military and political influence presumably was correspondingly smaller, and the enthusiastic claims regarding their linguistic and cultural influence made by Thornton, Mobley, and especially Sweet seem highly implausible.