1 - Atlantic Slave Trading and World History

Summary

Review of the inhumane practices of people in both New and Old Worlds prior to Columbian contact. Slave trading and cruelty were widespread, and slave trading was extensive. Most slaves were female, employed in domestic or agricultural environments (with little evidence of gang-labor), and came from a wide range of geographic areas and cultures. Most were born into slavery or were enslaved as a result of raids and wars in which many men on the losing side were killed. Slave markets existed across Eurasia, though in the pre-contact New World such markets were less common. After 1500, transatlantic trafficking came to draw exclusively on Africa or at least on Black people, probably because of the long isolation of the Americas from the rest of the world, and the inability of its Indigenous population to resist harmful pathogens from the Old World. Before 1820 migration to the New World was dominated by Africans rather than Europeans and by males (in contrast to the female-dominated slave populations of the Old World). White slaves were scarcely ever present in the New World.

We begin with the largest of canvases to ask where the transatlantic slave trade lies in the rich and extensive annals of inhumane behavior around the globe. Europeans and their descendants in the Americas and the Indian Ocean carried off and enslaved close to 13 million Africans from sub-Saharan Africa in the 450 years between about 1450 and 1867.¹ Even after the dramatic decline of the Indigenous populations of the Americas in the two centuries following European contact in 1492, Europeans themselves proved reluctant to participate in transatlantic migration. Indentured servants, convicts, and all those owing obligations to work off labor debts to others always comprised a small share of total transatlantic migrants in the early modern era. Prior to 1820 the share of completely free migrants in this broad picture was much smaller again. Thus, for the greater part of this period merchants seeking to exploit the riches of what for Europeans was a newly discovered continent, resorted to slave labor, a form of coercion that had largely disappeared in their home countries by 1500.

Extraordinarily, faced by a labor shortage, peoples of European descent resuscitated this most violent and inhumane manifestation of all terms of labor, but applied them almost exclusively to the Indigenous populations of the Americas and sub-Saharan Africa. But why did sub-Saharan Africans seemingly accommodate this European prejudice? As Nathan Huggins, the first director of the W.E.B. Du Bois Institute pointed out and as I developed further in an earlier book, it happened because the rich and diverse cultures of the Indigenous Americas and of sub-Saharan Africa could not accommodate a pan-American or pan-African identity before the twentieth century. Just as Indigenous peoples in the Americas did not know they were American, in Huggins’ striking words, Africans sold other Africans because they did not know they were African.² In 1859, the Africans on the Clotilda, among the last brought to the Americas on a slave ship, still “did not call their homeland Africa.”³ Even terms such as “Igbo” or “Yoruba” emerged in the diaspora, not Africa. Sigismund W. Koelle, the missionary who compiled Polyglotta Africana based on interviews with Liberated Africans in Sierra Leone found “that they had never heard it [the term Igbo] till they came to Sierra Leone … and know only the names of their respective districts of countries.”⁴

Such cultural dissonance, and, of course, profit maximization on both sides of the transaction that consigned people to a slave deck, were responsible for more than forty thousand voyages, 1519–1867, dispatched first to Africa and then to a plantation colony. The inhumanity, violence, and disastrous long-term consequences of this traffic are difficult to grasp by those living today in free societies with historically low violent crime rates. First there was the terror and killing inflicted in the capture of enslaved peoples and their transport to the coast. Then, almost 2 million Africans died at sea – a staggering mortality unmatched on transoceanic voyages carrying only Europeans. Most perished from the dehydrating effects of dysentery. Others succumbed to infectious Guinea worm or dracunculiasis, “their entrails filled and gnawed with worms; some small and flat, others … monstrous in length and size; some vomited in a bunch and others in a quantity of twelve to fifteen at a time.”⁵ Human actions were often more deadly than Guinea worms. The tenth slaving voyage of the Dutch West India Company ship, Leusden, in 1738 foundered in the estuary of the Maroni River in Suriname with 664 Africans captive below deck. Fearing the chaos that might result as the ship went down within sight of land, the captain ordered the hatches nailed shut before abandoning ship. But first he made sure that the enslaved people that he himself owned, together with the company’s gold (taken on board at Elmina, now in Ghana), were secure on the boat that carried them to safety.⁶ Such horrors continued to the very end of transatlantic slave trafficking. In 1861, the St. Helena colonial engineer John Charles Melliss was the first British official to board the Ardennes, a newly captured slave vessel. He wrote:

[T]he whole deck, as I picked my way from end to end, in order to avoid treading on them, was thickly strewn with the dead, dying and starved bodies of what seemed to me a species of ape that I had never seen before … (yet) the miserable, helpless objects being picked up from the deck and handed over the ship’s side, one by one, living, dying and dead alike, were really human beings. Their arms were worn down to about the size of a walking stick. Many died as they were passed from the ship to the boat, but there was no time to separate the living from the dead.⁷

In this nearly four-century-long gallery of inhumanity, the murder of the Leusden captives stands out as the worst recorded atrocity of all the 40,000-plus transatlantic slave voyages that occurred before 1867. Unsurprisingly, violence bred violence. In the sixty or so recorded slave revolts in which the captives gained at least temporary control of the vessel, all but one or two of the crew were slaughtered as an incentive to the surviving crew to navigate a return to Africa. In a couple of such instances, a now-captive European kept a journal before expiring, at which point the vessel drifted in the Atlantic as the Africans gradually succumbed to hunger and thirst. In one case, fifteen African survivors plus the sailor’s journal were eventually picked up mid-Atlantic and taken to Bristol in the United Kingdom. From Bristol, the Africans likely continued their journey to a lifetime of plantation labor. The most successful slave-ship rebellion in history occurred on a French ship, the Regina Coeli, in 1858. The vessel’s 391 captives took control of the ship, killed all the crew, and escaped. Many found their way to Monrovia, the capital of Liberia. The almost six hundred shipboard rebellions that we do know about were unsuccessful. Failure was inevitably followed by even more horrific violence, as captain and crew exacted a retribution constrained only by the need to preserve the human merchandise for sale in the colonies.⁸

Survivors of these transoceanic horrors might well have thought that they were not particularly fortunate. They could expect long hours of gang labor on plantations enforced by whipping and other cruelties. If family life was possible, it was insecure. North of Brazil it meant there was virtually no possibility of parents or children escaping servitude, except by running away. It is nearly one hundred years since Ulrich Bonnell Phillips – then the leading White authority on slavery in the US – argued that the slave trade improved the lives of Africans by bringing them to the Americas.⁹ Since then both research into and recognition of the enormity of the horrors of the business have expanded dramatically. There is probably no field of history that can match the gains in knowledge and the shifts in scholarly attitudes that have taken place in the study of slavery in the Atlantic world. We now know the structure, dimensions, and impact of the abominable business. Most importantly, we can grasp something of the experiences of those forcibly removed from sub-Saharan Africa as well as their descendants to a degree previously unimaginable. As I will demonstrate, the study of slavery and the slave trades in the four centuries after 1450 has moved to center stage for both scholars and the general reader.

But these advances have not occurred without some costs. One is the steady erosion of the comparative component in more recent research. Slavery outside this four-century interval and beyond the confines of the Atlantic has not received the same attention. For example, scholars of the Indian Ocean world, where slavery and the slave trade had a much longer history, have referred to the “tyranny of the Atlantic” in the discipline, by which they mean the disproportionate research funding it receives.¹⁰ A second issue is raising the profile of the slave trade in the catalogue of catastrophic cruelties that our species has inflicted on itself. The scale of the Atlantic slave trades ensured that four out of every five people arriving in the Americas from across the ocean before 1830 were African, not European. Cumulatively, migration from Europe exceeded arrivals from Africa only in the following decade. For more than three centuries, the repopulation of the Americas after the demographic collapse induced by Old World disease was by Africans, not Europeans.¹¹ However, only in the temperate Americas – southern Brazil and what became the southern US – did the two flows of people overlap significantly. Overwhelmingly, Africans both departed from and arrived in tropical regions, but the disease environment of Africa was not the same as in the tropical Americas. Moreover, the grueling labor Africans were required to perform in the low-lying wetlands where sugar grew best ensured higher Black mortality. Disease, malnutrition, and sheer exhaustion killed untold numbers. Sugar ensured a much larger inflow of Africans to the Americas than Europeans, but by 1800 only the Black populations of the US, Barbados, and probably Antigua had achieved an intrinsic positive natural rate of growth.¹² In the Americas as a whole, descendants of the smaller European migration came to heavily outnumber their African counterparts. This chapter attempts to set the horrors of the slave trade against the backdrop of the preceding millennium.

A broad comparative approach to what might be called the repeopling of the Americas helps explain why the trafficking of people first from Africa to the Americas (and soon thereafter between ports within the Americas) has had such a great and enduring influence on the modern world. This despite all the famines, conquests, wars, attacks on minorities, and viciously suppressed rebellions that in the previous millennium created such historical shatter zones of human misery. Yet the grim and ongoing legacy of the violent slave trade should not prevent us from locating what was effectively the forced repopulation of the Americas against the backdrop of inhuman behavior in recorded history. The very term “inhuman” seems inappropriate given the prevalence of the practice of slavery in our past. Scholars can cite the occasional thinker who had reservations, but up to 250 years ago, and in most parts of the world long after that, slavery was regarded as an unfortunate condition, one to be avoided, but certainly not one to be abolished. If not quite a universal practice, it was seen as a quintessentially human practice. A first step toward understanding what happened in the Atlantic world – especially to Africa and Africans – is to recognize what this massive, forced migration had in common with what had gone before, as well as how it differed. Even a brief examination of global slavery and slave trades over the last two millennia – instead of the more usual focus on the Atlantic world alone since 1450 – reveals perspectives and contexts usually missing from specialist works.

Walter Scheidel has estimated the traffic in enslaved people into the Roman Empire at three or four hundred thousand per year at its peak. That was three times the level of traffic from Africa to the Americas in 1829 – the year of the highest annual total recorded for the transatlantic slave trade.¹³ And when pagan peoples invaded and destroyed the now Christianized Roman Empire, beginning in the late fourth century, it is quite possible that in some years the reverse flow of captive Roman citizens into “barbarian” territories was of similar magnitude.

Historians of the Atlantic and the Americas have tracked the movement and evolution of what Philip Curtin called the plantation complex – the nexus of tropical produce, forced labor, and large estates.¹⁴ Several centuries after Europeans learned of sugar production during the Crusades, that complex moved westward through the Mediterranean Sea, then out into the Mediterranean Atlantic bounded by the Atlantic Islands, before landing in São Tomé in the Gulf of Guinea, then making the leap to New Spain, Brazil, and the Caribbean. After 1450, when sugar production reached Madeira and the Canaries, it began its four-century-long dependence on enslaved labor. Despite these links between the medieval and the early modern world, most historians still see the large-scale forced movements of people – the “many middle passages” – as a post-1600 phenomenon unique to the last few centuries.¹⁵

Northern Europe provides an even bigger discontinuity between medieval and post-1600 worlds for most historians of the Atlantic. Slavery supposedly died out in this region before the English, the Dutch, the Scandinavians, and the French revived and reconstituted it in a particularly violent and exploitative form in their overseas colonies. The larger literature on slavery still sees the medieval period as something of a hiatus between the Roman Empire and the plantations of the European Americas – a hiatus that is firmly rooted in the Western-languages scholarship of the last two centuries. Few recent scholars have explicitly made this argument, but a cursory survey of Slavery and Abolition’s annual bibliography of slavery that began in 1981, and in the 1990s attempted to incorporate all titles published since 1900, confirms that during the twentieth century no less than 95 percent of the listed items in this bibliography dealt with slavery before or after the medieval millennium.¹⁶ Coverage of Asia, the Indian Ocean world, Amerindian societies and Oceania has increased significantly in recent years, but so, too, have publications on coerced labor in all periods and geographic regions. Thus, in a comparative sense, the temporal focus of scholarship on slavery over the past two millennia has changed little.

Humanity’s oldest surviving work of literature, the Epic of Gilgamesh, fragments of which appear on 4,200-year-old clay tablets from northern Iraq, contains references to enslaved status.¹⁷ Slavery is thus in its fifth millennium of recorded history. It certainly did not disappear between 450 and 1450 CE; it merely took on different forms and appeared in different types of socioeconomic structures. Large landholdings and mines worked by forced labor were indeed largely absent in medieval Eurasia. The notable exception was the sketchily documented Zanj people, who were forced to clear salt from marshes in ninth-century Mesopotamia. However, the incidence of enslaved labor in households for both domestic and sexual labor appears to have increased across the Old World. The phenomenon of enslaved peoples in temples appeared in Buddhist societies with few antecedents in the ancient world, and none whatever in the later Americas, unless we shoehorn the vast Jesuit slaveholdings in the Iberian Americas into this category. In addition, the medieval millennium saw the high point of enslaved people used in military contexts. Slave soldiers in the Delhi Sultanate, Mamluk Egypt, and the Janissaries of the Ottomans are well known. But given that every elite soldier in the Middle East and Asia required a support corps of tens and in some cases hundreds of enslaved people to minister to his needs, it is probable that the global share of persons in bondage over the whole course of history peaked at some point in the medieval millennium, rather than in the early modern Atlantic world. Certainly, in the millennium before 1420 it is possible that nowhere in the world was there a settled society (i.e. one based on agriculture) that lacked enslaved women and men.¹⁸

Both the medieval and the early modern eras experienced prolonged periods of economic growth, the first curtailed by global plague, and the second, encouraged by industrialization, continuing into the modern era. Increased urbanization, expanding trade networks, and the growth of centers of learning were common to both. Scholars have argued for a twelfth-century renaissance in Europe and the Arab world, leading to some historians positing the evolution of a first “world system” in the thirteenth and early fourteenth centuries. Still, no scholar foregrounds slavery and the slave trade either as a cause or a consequence of this phenomenon.¹⁹ By contrast, one cannot imagine the later and more widely known era of European overseas expansion without attempts to link slavery with economic expansion – either as cause or effect.

But to what extent was slavery in the two eras sufficiently alike to warrant comparison? Property-in-persons and coerced labor have appeared in such a range of forms in history that no consensus definition of slavery is possible. Joseph Miller’s specification of slavery as a “practice” or a “strategy” rather than an “institution” has proved influential. But if practices are “socially constructed and accepted ways of doing things,” the distinction is not helpful for the purpose of comparing the two periods.²⁰ How is such a practice so defined different from an institution? More useful is anthropologist Edmund Leach’s view of another fundamental human practice – marriage – which also took a wide variety of forms. Marriage, he argued, was “a bundle of rights” with such variations that a universal definition is quite impossible. To take just one aspect, think of the almost infinite variations of polygamy in societies in the Americas alone. Analysis can take place only in a specific cultural context, an assessment that assuredly applies to slavery as well, but one that does not inhibit comparisons.²¹ The type that is most readily recognized today – the chattel bondage characteristic of the European-dominated Americas and classical Greece and Rome – was certainly not the predominant form of slavery in global history. As already noted, in the Indigenous Americas, much of Africa, and the Islamic world, an incorporative form of the institution prevailed. While plantation slavery aimed primarily at extracting labor, the incorporative variety evolved in part as a device to extend the size and prestige of a lineage, a fictive kin-group, or a religion, in a social environment where a fulfilling life without such associations was inconceivable. And if, in the first, it was difficult for a captive or her descendants to exit enslaved status, in the second, full integration of an enslaved person or her descendants into the host society was normally a definite possibility. In this work we avoid a definition of slavery, but we know it exists when a person is traded or captured and subsequently owes lifelong obligations to another.

Much of Asia fits neither model. In China, Korea, the long-lasting Khmer Empire, and seventeenth-century Japan, enslaved peoples had few prospects of changing status. They were created by foreign conquest, the judicial system (convicts), or were simply drawn from the lowest social stratum of “base” people. Nevertheless, the concept of property in persons that characterized slavery in the European Americas and the Islamic worlds never evolved in China. Slavery in the Indian sub-continent was different again, as we might expect given the heavy Islamic influence beginning in the mid thirteenth century. The institution there occupied a large conceptual middle ground between the incorporative slavery of Africa and Islam on the one hand, and East Asian systems of coerced labor on the other. Collective ownership of enslaved peoples by Hindu and Buddhist monastic lineages survived alongside ownership by individuals. Medieval Islamic incursions extended both military and court slavery. But Islamic, Hindu, and Buddhist communities legally sanctioned enslaved holdings and transfers of property-in-persons. They also imposed responsibilities on owners. The law might circumscribe both ownership and treatment. Slaves in some jurisdictions might have the right to inherit, to own, and to dispose of their own property.²²

Here I argue that most enslaved peoples were not obtained either by purchase or conquest. Perhaps the common denominator across cultures and time was the existence of a marketplace in which individuals were traded. In the medieval millennium, the term “slave” or its equivalent was not reserved solely for instances in which an owner assumed title to all the individual rights of the enslaved. The term could also describe other forms of social dependency in which only some degrees of unfreedom were subject to transfer. Many markets must have witnessed buyers and sellers where each had a slightly different idea of what was being traded. Nevertheless, such markets could not have existed without both parties to a transaction sharing a common understanding that their exchange involved a commodified person – as, for example, on the African coast in the early modern period where both the buyers of the enslaved (Europeans) and sellers (Africans) viewed the people they traded as outsiders to their own societies.

Debates over the definition of slavery have not proved very productive and for our comparative purposes here, we do not need a definition of slavery applicable across cultures. It is sufficient to note the wide distribution and vibrancy of markets for the enslaved across the major continents, from ancient history to the early twentieth century. These were clearance centers for future units of labor, sexual exploitation, social prestige, expansion of the kin group, and other purposes, but also, from the enslaved’s perspective, humiliation, deracination, and physical abuse.

Even without a definition of slavery we can discern some contrasting geographic patterns in the movement of enslaved people over the globe in the last millennium and a half. In the early modern Atlantic world, an ocean separated the major markets and the enslaved flowed overwhelmingly from east to west. In medieval Eurasia we do see cases of Chinese captives sold in Black Sea markets, presumably having arrived there in a series of stages. Slaves travelled mainly by land throughout this era, albeit frequently following riverine routes, but seas such as the Irish and North Seas, the Mediterranean, the Caspian and Black Seas and the Indian Ocean all had very active slave trades in different periods. Predominantly, however, the enslaved moved on a north–south axis rather than east–west. A series of empires and powerful states stretched from Spain through the Mediterranean, the Tigris–Euphrates region, the Turkic states, and through the Khmer Empire of southeast Asia to China.²³ This belt comprised several dynasties and was occasionally subject to conquest from the north. It drew captives from the less densely populated and poorer regions such as northern Europe, the Caucasus, and central Asia, as well as from the south (Africa, India during the Muslim incursions, and Korea). Victims of Viking raids ln Iceland might sell in Baghdad.

Further east, the Turkic polities of the Samanids and Ghaznavids conducted extensive raids – the Saminids in central Asia, while the Ghaznavids drew heavily on northern India. Key markets, often bordering on bodies of water, constituted transit or staging points for the enslaved. Examples in the period included Dublin, Tana on the Sea of Azov, Caffa in the Crimea peninsula, Baku in Azerbaijan, and Alexandria and Venice in the Mediterranean. In the Indian Ocean, Aden and Zabid in Yemen, and Zayla in Somaliland hosted large slave markets. The city of Guangzhou in China had a centuries-long presence of Africans, perhaps brought by Arabs, beginning no later than the tenth century – long before Admiral Zheng He’s voyages to the sub-continent. However, the extreme south of China, Korea, Inner Mongolia and Central Asia supplied most of the externally generated enslaved peoples to mid-millennium China, in yet another north–south axis. Within North Africa, the pre-Atlantic traffic moved on both axes, east toward Egypt, the Red Sea and Mesopotamia, and on the better known six major south-to-north routes that Paul Lane describes.²⁴

Both the early modern Atlantic and medieval Eurasia experienced cataclysmic periods of enslavement. European incursions into the Atlantic and eventually the Americas peaked in the 1701 to 1850 period, when 80 percent of the transatlantic traffic occurred. This expansion of maritime bondage resulted in the startling fact that far more enslaved people than free left the Old World for the Americas in this century and a half. The medieval equivalent of this upheaval involved the conquests of the nomads of Central Asia – first the Mongols, then the Khanates, and later the Turks. These events had huge consequences for enslavement as well as, of course, unleashing violent death and social disruption. In neither case were the enslaved an important component at the beginning of imperial expansion. But both generated an apotheosis of violence, the first transcontinental, the second, transoceanic. The military adventures of the Mongols under Chinggis Khan (d. 1227) and Tamerlane (d. 1405) all involved high mortality. Chinggis Khan killed countless citizens of captured cities in the event of their resistance. He enslaved many more besides. Tamerlane’s subsequent conquests in the late fourteenth century led to possibly the largest contiguous territorial empire in history. The later European disruption of Atlantic peoples destroyed three-quarters of the population of the Americas and triggered violence and enslavement in Africa. Just as European empires brought an Atlantic market for enslaved peoples into existence, so the Mongol Empire strengthened and supplied an integrated market for both commodities and enslaved Eurasian people.²⁵

Without claiming cause and effect, it is striking that arguably two of the most catastrophic sociopolitical events in global history are, in some senses, antecedents of the two superpowers of the mid twenty-first century – China in the East and the United States in the West. Finally, both power centers came to abandon outside sources of forced labor. The China that the Mongols helped shape, and of which they had become an integral part by the fifteenth century, had only a tiny fraction of their population enslaved. China no longer drew on external sources for their “base” people. In the West, likewise, Americans, even before independence, quickly came to rely on internal sources for new captives – in the form of natural population growth. Yet the parallels extend only so far. Amid all the claims that slavery was an essential prerequisite of economic growth in America (and an earlier literature that makes similar links for the preceding British case), no one considers chattel slavery when discussing the economic rise of China.

As these comments suggest, comparisons between eras inevitably raise questions of scale.²⁶ We do know the approximate numbers of captives moving around the Atlantic world after 1500, as well as the broad demographic patterns in the major European colonies of the Americas. In Africa, however, there is no such data for the slave trade era, apart from tallies of the enslaved who were forced to leave for the Americas. Instead, demographers depend on backward projections of twentieth-century data, applying various assumptions in their models. The most reliable of these, by Frankema and Jerven, comprise a modification of Patrick Manning’s foundational work.²⁷ For the end of the slave trade era in 1850, they estimate a continental population of 114 million broken down by modern country, which, after allowing for the large swathes of the north, northeastern, central, and southern parts of the continent that could not have supplied captives to the Americas, means approximately 59 million people living in potential catchment zones. In the peak years of 1701 to 1850, an annual average of 60,600 captives left Africa for the Americas. Deaths prior to boarding the slave vessel through violence, or passage through regions with unfamiliar disease environments, amounted, perhaps, to an additional third of this total, but there is no hard data on this phase of the traffic.²⁸ If Africa lost 81,000 people a year to the transatlantic slave trade, then annual population loss was between .01 and .02 of a percentage point, or 1.5 per thousand per year, assuming that the population remained at about 59 million during these 150 years of maximum impact. Including the trans-Saharan and Indian Ocean world (IOW) slave trades in this computation increases these rates slightly, though the IOW traffic probably became significant only in the late eighteenth century.²⁹

In the seventeenth century these catchment regions within Africa would certainly have contained fewer than 59 million people, an assessment based on what we know about the arrival of new crops from other parts of the globe in the preceding centuries. Images of an unchanging African agriculture before the twentieth century are clearly erroneous. As John Caldwell and Thomas Schindlmayr have argued, “the majority of food eaten in sub-Saharan Africa is non-indigenous and was eaten by no one there 500 years ago.” By the early eighteenth century, European slave-ship captains seeking provisions familiar to their captives prior to the voyage were able to buy Asian rice, manioc, and maize on the Western African coast, none of them native to Africa.³⁰ The superior nutritional content of these foodstuffs gradually gaining acceptance across much of the continent would have enabled population growth even during the years when the transatlantic slave trade was at its height. In addition, we now know that the female component of the forced migration was only one third, and this is a key metric in determining the impact of emigration on a population. While some slave trade scholars argue that the slave trade reduced the African population, few modern historical demographers subscribe to this position – in contrast to their mid twentieth-century predecessors.³¹

The dearth of demographic material anywhere in the world before 1600 is such that the early modern Atlantic world seems data-rich by comparison. It is generally accepted that the medieval global population was much lower than its eighteenth-century successor – perhaps 260 million in the year 1000 CE, of whom at least 200 million lived in Eurasia. About one third of this total lived in China under the early Song dynasty (CE 960–1279).³² China, the largest polity in Eurasia, enslaved both Chinese and non-Chinese people at different times in the medieval era, with both state and private individuals as owners. But as noted above, chattel slavery as practiced in the later Atlantic world fits poorly with categories of dependency as they evolved in China. “Base people” comprised 5 percent or less of the imperial Song population. Setting aside eunuchs and concubines – groups that were not considered base – and given that not all base people were enslaved, the incidence of slavery in China must have fallen below the enslaved ratios of contemporary Islamic and Christian regions, as well as eastern societies such as Korea.³³ The incursions of Jurchens, Jins, and Mongols into China that destroyed the Song dynasty would have seen temporary increases in the incidence of slavery.

Meanwhile, Orlando Patterson has estimated that mid tenth-century Western Europe alone was home to 3.4 million enslaved peoples (15 percent of a total population of 22.5 million). The majority were native Western Europeans.³⁴ Islamic populations filled much of the vast space between the receding Byzantine territory and the western boundaries of the Song, and although Islam viewed freedom as the original condition of humanity, enslaved peoples were ubiquitous across the Muslim Middle East, northern India, and much of central Asia. They labored in a wide variety of roles. Scholars have identified servile labor existing in religiously diverse southern India, where slavery was often associated with temples. Reliable estimates are impossible, but if even 10 percent of the Old World population was enslaved in the year 1000 CE, then at 20 million (one tenth of 200 million) there would have been far more enslaved peoples in Eurasia than at any time in the Americas in the four centuries after 1450. But the rate/ratio of enslavement was higher in the Americas. Given that the Americas accounted for just 2.5 percent of the global population in 1804 (a benchmark year chosen by Barry Higman) a different assessment, one built on ratios rather than aggregate totals yields a different picture. In 1804 the relatively lightly populated Americas had approximately 13 percent of its people enslaved – a larger proportion, no doubt, than any continent in the history of the world. To a greater extent than elsewhere, slavery built the Americas, but would the development path have been different if all migrants had been free?³⁵

Turning from populations (stocks) to forced migrations (flows) does not improve the quality of the medieval data that is available. For the movement of enslaved peoples within Eurasia in the medieval millennium, we rely mainly on the records of chroniclers who were often dependents of the great men about whom they wrote. In such records “the numbers of slaves reported as captured … should be divided by ten, or even a hundred.”³⁶ A few Chinese people on sale in Black Sea ports, the aforementioned African captives in Guangzhou, and the Greek woman in the Sultan of Java’s harem noted by Ibn Battuta, testify to a vast trading network matching, if not exceeding, the range of slaving routes in the later Atlantic.³⁷ But they also point to the interconnectedness of markets for the enslaved. Unlike in the transatlantic traffic, no single slave trader carried his human chattel from one end of Eurasia to sell at the other. Enslaved people acquired and sold via war and conquest dominate the sources. Meillassoux’s model of slavery as being associated with predatory states could be generalized across the ages. If one polity had the ability to overcome another, then war, violent death, or some form of social debasement and dependency for the survivors would surely follow – a situation that probably holds for slavery everywhere.³⁸ James C. Scott has argued that “all Southeast Asian states in history have been slaving states” and “wars (were) about the seizure of as many captives as possible,”³⁹ especially women. In the later Atlantic world such conflicts occurred as a prelude to the enslaved embarking on the African coast. While many similar wars occurred in the Americas, especially the Caribbean, the victory of one party over another – say the British over the French in the eighteenth century – may have resulted in existing enslaved peoples changing hands. Nevertheless, the losers in the conflict did not themselves become enslaved. In Eurasia, by contrast, at least until the emergence of increasingly impermeable “no-slaving zones” late in the medieval period, slave owners as well as enslaved peoples on the losing side could normally expect to face servitude (or extended servitude in the case of the already enslaved).

For the Aztecs, the Viking raiders, the Umayyad Caliphate, the Latin kingdoms in the aftermath of Christendom’s resurgence, the Ghaznavids in India, and the string of Mongol victories to the east, enslavement followed from raids and conquest. The same would be true later with Europeans in the Americas. But prisoners of war and civilians treated as booty – the spoils of war – fed into the Eurasian network of markets for the enslaved. Nevertheless, war captives cannot have been the most important source of slave labor. What other sources were there? Conquered groups frequently supplied enslaved people to the victors as tribute. Self-enslavement brought on by extreme deprivation appears to have been universal though not numerically significant, as were judicial enslavement and kidnapping. In fact, all five of the enslavement mechanisms that appear in Sigismund Koelle’s report of his interviews with Liberated Africans in nineteenth-century Sierra Leone are to be found in medieval sources, even though the documentary record suggests that armed struggle and its consequences was the principal route to enslavement.⁴⁰

Yet this catalogue of pathways to oppression omits the probable major source of slaves in global history: birth. Almost all discussions of slavery in specific contexts explore the ways in which individuals became enslaved. The current conventional meaning of the verb “to enslave” implies that people had at some point lost their free status. In a reality that modern scholars today almost never recognize, the implications of the hereditary nature of slavery throughout history, at least until the twentieth century, guaranteed that birth was the main source of slaves in almost all societies. Even the word “captive” is problematic. In eras when slave raids and war-generated captives were the dominant sources, many of those caught up in such tumult already had the status of slaves. Medieval and classical sources provide countless instances of strong states raiding their weaker neighbors, but we hear of the status of the resulting enslaved persons only if they were high-born or later became well-known. People already enslaved in the target society were those most likely to be captured, or more precisely, recaptured.⁴¹ In the transatlantic context given the high incidence of slavery within Africa, many captives dispatched to the Americas would have been already enslaved.

A quick survey establishes that most captives were born into servitude. In the European-controlled Americas three different demographic scenarios played out for people of African descent. First, in the relatively well-documented US case, fewer than half a million captive Africans arrived via the maritime slave trade over 250 years. But by 1865 at least 10 million people had experienced slavery on the North American mainland. More than 90 percent of them were born into bondage, rather than captured and then sold.⁴² Second and by contrast, the slave labor force of the Caribbean sugar colonies experienced a much harsher demographic regime. Except for Barbados after 1760, constant inflows of Africans were necessary to prevent the labor force from declining in the face of the exigencies of sugar cultivation. Yet, just prior to emancipation in 1833, and twenty-five years after the slave trade ended, only one in five of the half-million enslaved persons in the British Caribbean could have been African-born.⁴³ Third, in Brazil, too, few of the enslaved freed in 1888 could have been born in Africa.⁴⁴ Most scholars have still not engaged with the fact that almost two thirds of all those taken from Africa to the Americas spent the balance of their lives in Latin America.⁴⁵ Here, unlike the US and most of the Caribbean, it was more possible for some slaves to change status over their lifespan via the practice of self-purchase, although manumission rates were always low outside the major cities.⁴⁶ The fertility of the Brazilian slave population was lower than that of the Black population of the US South, but high compared to most other regions in the Atlantic world for which data is available. Robert Slenes has used a range of reliable data from the 1870s to calculate an annual fertility rate of 36/1,000. With an enslaved population in 1850 of around 2 million, about 72,000 people per year were born into slavery in Brazil.⁴⁷ Annual arrivals from Africa to Brazil between 1801 and 1850 – the peak half-century of the Brazilian slave trade – averaged only 41,000. In the Americas, of course, the slave trade was vital to the expansion of the various plantation economies, but creoles, not Africans, must have dominated the slave labor forces at most points in time.⁴⁸

Slavery in the Americas, and the traffic in people that made it possible, receive far more scholarly attention than does slavery in the rest of the world. Such scholarly focus is perhaps warranted given the high ratio of enslaved people in these places compared to elsewhere, and the high land-to-labor ratios on the continent. Estimates of the pre-contact populations of the Americas have varied widely but even before the “great dying” that European settlement and its accompanying diseases triggered among Indigenous peoples, population densities were likely much lower than on other continents.⁴⁹ Imperial systems in the Americas before 1492 certainly maintained enslaved populations that drew in part on tribute slaves supplied by fringe client states.⁵⁰ In the aftermath of European conquest, population densities, already low by global standards, fell dramatically, with repopulation occurring via immigration and positive rates of natural growth. Even the Indigenous populations of the Americas experienced natural growth from the late seventeenth century. The inflow of people to the Americas occurred under the auspices of five major transoceanic imperial systems (English, French, Spanish, Portuguese, and Dutch), all of which depended on enslaved labor. As already noted, before 1820 four out of five of the arrivals from the Old World were bonded Africans rather than European free migrants prior to 1820. Except for parts of Oceania, no other part of the known world experienced demographic disaster on the scale of the early Americas. Plagues and famine decimated populations, but nowhere did slavery play such a large role in the subsequent demographic recovery as in the Americas.

In the medieval millennium, plantation slavery was rare, but the same pattern of births constituting the main source of slaves held. Most empires in history have relied on outside populations for their servile labor in their early expansionary phases. But over the course of their imperial existence, they drew on natural increases of their slave populations. Unfortunately, the sources for the medieval millennium are dominated by descriptions of the enslavement process, particularly when this was a by-product of war and imperial expansion. Hard data on natural population growth in medieval empires does not exist. We do, however, have sources that point to the predominance of females in most slave populations. In medieval India, Leslie Orr has found that “[m]ore than half of all [temple] inscriptional references to slaves, where gender is noted, concern the gift or sale of women or of female-headed families.” The late medieval Deccan state, Orr continues, “possessed tens of thousands of slaves, mostly women, who had been captured and brought to court from the neighboring non-Muslim kingdoms.” Among the Mongols, Michal Biran argues that “female slaves were more valuable, as they could be used not only for herding sheep and cattle but also for … domestic chores that transformed the produce of the nomadic economy into useful goods … and as wives or concubines.”⁵¹ Genetic analysis of modern near-eastern Arab populations shows the African component arrived via females beginning in the early sixth century. In Mamluk Egypt “female slaves outnumbered male, probably by a significant margin,” a profile that emerges even more strongly among Jewish owners in pre-Mamluk Egypt’s Geniza documents. Across the late medieval Mediterranean world, in response to plague pandemics, the slaves in demand were “especially female … who could be employed as domestic servants.” In northern Europe, David Wyatt argues that while scholars continue to focus on the male agricultural laborer, the medieval traffic in slave women barely declined. The enslaved were principally attached to households where they cohabited with their masters.⁵²

Shaun Marmon makes clear the implications for natural population growth of a majority-female slave presence in a predominantly domestic context. “Since the sexual exploitation of female slaves is inherent to the institution of slavery, in all historical contexts, all female slaves were potential sexual slaves” and “slaves only gave birth to slaves.”⁵³ As graphic illustrations of this point, consider just three examples of very different cultures spread across 850 years. The arrival of the Rus (of Swedish origin) in Bulgar, a major slave market in Eastern Europe in 922, was famously described by Ibn Fadlan as being “accompanied by beautiful slave girls for trading. One man will have intercourse with his slave-girl while his companion looks on … Sometimes indeed the merchant will come in to buy a slave-girl from one of them and he will chance upon him having intercourse with her but will not leave her alone until he has satisfied his urge.” Also, Geniza documents from the relatively sedate urban environments of twelfth- and thirteenth-century Egypt show Jewish households stressed by the failure of Jewish law to recognize concubinage to the point that marriage and engagement contracts came to include a “slave woman clause,” specifying that “the groom will not retain a slave woman whom she (the bride) dislikes.” Across the ages, it does seem that household rather than agricultural slavery is what most enslaved people experienced. Finally, Thomas Thistlewood’s plantation record in Jamaica contains a depressing record of sexual assault and rape that was all too common in a plantation society. From his arrival on the island in 1750, through to his death thirty-seven years later, he “engaged in 3,852 acts of sexual intercourse with 108 different slave women,” all meticulously recorded in his diary.⁵⁴ The absence of consent is the common thread in these three examples. Finally, to return to the Miers and Kopytoff assimilative model useful for understanding post-conquest Islamic – and indeed Amerindian – slavery, it might be assumed that host societies found it easier to integrate and assimilate women and children rather than men.

If most of us were asked to identify the central difference between the predominantly Eurasian slavery of the medieval millennium and the large slaveholdings that characterized the succeeding era in the Atlantic world, we would likely answer, first, that slavery became more extensive in the later period, and second, that the African component of the global slave population increased to the point that the enslaved became overwhelmingly, perhaps exclusively, Black after 1500. But both points are questionable. The enslaved in the Americas always accounted for a small share of the global enslaved population. Barry Higman estimates that in 1800, the global population of chattel slaves had reached 45 million, or 5 percent of the Earth’s total. However, only 6 million lived in the Americas.⁵⁵ Given that slavery was common in the Indigenous Americas both before and after the destructive European incursion, and that the large Islamic band of polities stretching into southeast Asia continued to thrive after 1500, the transatlantic slave trade by itself, massive though it was, may not have drastically changed the global ratio of enslaved to free.

As for the African-descended share of the enslaved, just more than a quarter of the global total of enslaved persons were of African descent in 1800. At any point earlier than 1800, the African American component of the global total was inevitably smaller. But what was the pattern in the Old World? Some scholars argue that the exodus of Africans via the trans-Saharan and Indian Ocean came close to matching Atlantic departures, but such claims are dubious, given that evidence of the eastern and northern branches of the African slave trades cannot compare to what is available for the Atlantic traffic.⁵⁶ Islamic regions constituted by far the largest market for captives in medieval Eurasia.⁵⁷

What did Africa contribute to the enslaved labor pool of this vast Eurasian region? With the single exception of the area occupied by the Zanj, a variously described Bantu-speaking people in what is now southern Iraq in the ninth and tenth centuries, it seems that people of African descent never constituted a majority population of any part of medieval Islam. We might point to the huge diversity of enslaved people that the region contained, a diversity that only gets wider when we recognize that the term “African” itself encapsulates a vast range of ethnolinguistic groups.⁵⁸ Islamic states drew on Circassians, Tartars, Rus, Slavs, Mongols, Turkic peoples, Latin Europeans, and many others. The major group traded in Cairo, the largest slave market in the Mediterranean through most of the medieval millennium, were brought in from Black Sea regions, though about half the slave women in medieval Egypt’s Geniza documents were Black or Nubian.⁵⁹ For the early period, Michael McCormick’s magnum opus, Origins of the European Economy: Communications and Commerce, AD 300–900, focuses on European slaves entering the Islamic world, but Jeff Fynn-Paul points to several non-European groups that McCormick overlooks who passed along well-established trade routes only one of which connected with sub-Saharan Africa.⁶⁰ Early Islam’s (the Abbasid Caliphate) use of African soldiers has attracted considerable scholarly attention, but not all of these were slaves. By the later Mamluk era in Egypt and in the Delhi Sultanate in North India the African component was much smaller, replaced by purchases from the Black Sea region, Turkic areas, and Central Asia.⁶¹ Less is known about the origin of the numerous non-combatant slaves that supported the military elite. Africa may have quickly become the sole external source of slaves when the Europeans took over the Americas, but Africans appear to have had no significant presence in the United Mongol Empire (1206–1260) or the subsequent Khanates of the Mongol Commonwealth.

Manuals on buying slaves offer further insights into the origins of Islamic slaves. A treatise by an eleventh-century Baghdad physician identifies “women of India, Sind, the Maghrib, several regions of sub-Saharan and coastal Africa, three towns of Arabia (Mecca, Medina and Taif), Yemen, Qandahar (Afghanistan), Allan (Caucasus) and the Daylam region (northern Iran) … as well as Turkish, Greek (presumably Byzantine) and Armenians.” Other guides to buying for the Mamluk era offer stereotypical descriptions of a similarly wide range of peoples.⁶² The etymological evidence also points away from Africans, but this time not toward a multiethnic mélange of peoples, but to a specific alternative group. In the later Middle Ages, both Latin languages and Arabic began to roll together the terms for a slave on the one hand and, on the other, the name of the ethnic group on which both Latins and Arabs had come to largely depend. That term was not “African,” “Black,” “negro,” “nigra” or any derivatives of these words, but rather Slavic peoples, whose name forms the stem of the word slave in all western languages – the Latin term for “Slav” being esclavus. Arab terminology, too, quickly came to integrate the status of slave with the ethnicity of most people who occupied that status. Thus, “saquiliba” (Arabic صقالبة) denoted both Slavic and enslaved.

None of these sources allow us to pin down either a number or a specific ratio, much less a trend over time, but they do make it unlikely that Africa was a dominant source of slaves for the medieval Islamic world. Nor does this contradict evidence of the stigma of black skin that periodically surfaces in the medieval sources, and arguably intensified in the later Middle Ages. Adherents of all three Abrahamic religions do show increasing evidence of prejudice against Black people over the millennium. The bizarre story of the Curse of Ham gained currency, wherein Abraham condemns the offspring of Canaan, the son of Ham, to be slaves in perpetuity for a trivial offence on the part of Ham. A discourse on color symbolism begins in Christendom and Islam in which Black is equated with evil and White with good. But the Black descendants of Cush, brother of Canaan, are deemed innocent in early rabbinic writings, and writers continue to see black skin as a function of climatic factors as did their classical-era predecessors.⁶³

Turkic and Sudanic peoples formed the bulk of the elite Mamluks, but other states that used elite enslaved military units in India and the Middle East drew on those of African descent. More important, none of the societies adhering to the Religions of the Book had legal systems based on somatic norms. Antarah ibn Shaddad, born a Black slave, became a famous knight as well the best-known poet in the pre-Islamic near east. The freedman, Bilal ibn Rabah, likely an Ethiopian, was the Prophet’s close companion and is regarded as the first caller to prayer. In Christendom, one of the three kings attending the birth of Christ had become portrayed as Black by the end of the medieval period. And the well-known cult of St. Maurice, the black knight, could scarcely have flourished in the face of such discrimination. Thirteenth-century literary romances Parzival by Wolfram von Eschenbach and the anonymous Dutch Moriaen illustrate the increasing European engagement with and ambivalence toward “Ethiopians,” and perhaps a growing epidermal racism. But nothing in the medieval Eurasian written record in any way suggests that only Blacks should be enslaved, and thus nothing foreshadowed the situation that emerged in the Atlantic world in the century after 1420.⁶⁴

After 1500, with the acceptance of Christianity or Islam by the peoples of eastern Europe, the steppes and much of northern Africa now complete, the Latin West and the Arab world did draw more heavily on Africa for enslaved people. Even though slaves usually accounted for only a small share of most Asian populations (Korea was an exception), the typical slave prior to the nineteenth century was Eurasian and female – not African (or Slavic) and male – given the enduring importance of Asia in global demography. The modern world may visualize peoples of African descent when they think of slaves in a historical context, but in fact this group has never come close to forming most of the world’s population of enslaved peoples. The extraordinary slave ratios in the eighteenth-century Caribbean populations were highly exceptional in the context of global history. What also follows is that despite the transoceanic slave trade, the enslaved population of the Old World always vastly exceeded that of the New.

The exclusive association of black skin with slavery, discussed more fully below, was not established until the early seventeenth century, and never took firm root in the Islamic world and points east. The first sale of a sub-Saharan African as a chattel slave certainly occurred BCE. Among the last were those accompanying their Nigerian owners to Saudi Arabia on a hajj in the 1930s (and perhaps later), where they were sold – having been forced into the role of an animated traveler’s check for their owners.⁶⁵ For most of this span of more than two millennia, Africans were no more likely than people on other continents to be enslaved and sold into a long-distance slave trade. When Jacopo Tintoretto (the elder Tintoretto) painted his stunning and widely acclaimed “San Marco libera uno schiavo,” known in English as “The Miracle of San Marco,” in mid sixteenth-century Venice, the slave in the foreground was White, and two of the onlookers in the background were not only Black, but well-dressed merchants to boot.⁶⁶ The transatlantic slave trade was already four decades old when the painting was finished in 1548. In European eyes a slave could be white and a Madonna, black.⁶⁷

From the early seventeenth to the late nineteenth century in a large part of the Western world, a slave had to be someone of African descent. Even the Portuguese, whom much of the literature has viewed as having the most malleable of attitudes toward skin color, came to associate slavery with African descent. In 1670 captured Portuguese soldiers in the hands of the Soyo army in West Central Africa were offered the option of becoming slaves or being killed, and “they replied proudly that ‘Whites will never serve Blacks’; they were then put to death.”⁶⁸ While the corollary never held, being Black usually ensured a status somewhere below full citizenship, especially throughout the Americas. In modern-day slavery this association of skin color with enslavement, at least in the popular perception, seems to have at last disappeared. While de facto slavery still exists in parts of Africa, the victims of human trafficking – and most are still predominantly female – can be of any origin. In terms of the multiethnicity of modern-day slaves – except perhaps for concentrated numbers of temporary debt slaves in South Asia – we seem to have returned to the enlightenment of the early Middle Ages.⁶⁹

Why, four centuries ago, would the diversity of people enslaved by the Latin West shrink to comprise Africans alone? Slavery itself was entrenched in both the medieval millennium and in the four centuries of the Atlantic world that seamlessly emerged from it. In one sense, as we have seen, it did not become less diversified, given Africa’s more than 1,500 distinct spoken languages, hundreds of distinct systems of beliefs, and huge genetic and phenotypic variation shaped by a wide range of climates, diets, and exposures to infectious disease. But in geographic terms the point cannot be doubted.

Three factors help account for the narrowing of Western eligibility requirements to sub-Saharan Africa. Historians have discussed the first two at some length, while the third requires more attention. First, more than three-quarters of the people carried off across the Atlantic and Indian Oceans disembarked in colonies producing sugar and its derivatives. As already noted, the cultivation of sugar cane began BCE in the Pacific and gradually spread after the Crusades via India and the Arab world into the Mediterranean. Over the course of a further two centuries, Europeans spread the complex across the Atlantic and the real price of refined sugar declined substantially as the use of slave labor rapidly expanded. Where long-distance trade had mostly centered on luxury products, after 1500 the produce of tropical plantations, chiefly sugar, gradually came within the range of ordinary consumers. Consumer choice drove the expansion of the plantation complex. Stimulants such as coffee, tea, sugar, tobacco, and a range of alcoholic beverages had become mass consumption items by the end of the era. Working classes came to see such items as necessities, even though their overall impact on health was likely negative. Prior to the advent of sugar beets in the nineteenth century, refined cane sugar was a commodity that Europeans could access only through long-distance trade. The human craving for a cheap sweetener thus lies at the foundation of the sugar plantation complex, as well as the slave trade that sustained it.⁷⁰ Accounting for the demand for slaves, however, explains nothing about where those enslaved peoples were to be captured.

The second factor is geographic. The sugar complex made its way into the Atlantic, as the Portuguese learned how to navigate the oceans of the world. In the age of sail the great gyres and the associated wind patterns determined all transoceanic transportation. The Earth’s rotation ensures that a single vast spiral dominates each of the northern and southern sectors of the great oceans. North of the equator in the Atlantic the gyre turns clockwise, and to the south, and separated from it by the doldrums, the gyre moves counterclockwise. As the Portuguese discovered, to sail west (or in returning, east) across the Atlantic one must first sail south (or north). To reach the Americas from Africa south of the equator, one must first sail north (or, in returning from the New World, south).⁷¹ Despite the circuitous route, Western sub-Saharan Africa was closer to the Americas than any other well-populated landmass, and during the sixteenth century, ocean-going technology had made the transatlantic crossing routine. It became suddenly quicker and less costly to transport three or four hundred people from the well-populated forest zones of West Africa to northeast Brazil, than to form a caravan to cross the desert to the Mediterranean coast.

If consumer demand for sugar and the Atlantic’s oceanic patterns – two factors scholars do emphasize – were pre-requirements of a transatlantic slave trade, they were not sufficient in themselves to make it happen. The third and most important factor in narrowing European conceptions of who could be enslaved was a sense of the otherness of Black Africans that appears to have held among all Europeans who aspired to trade with Africans on their Atlantic shores. As I argued in a previous book, Europeans had become reluctant to impose slavery on other Europeans, even if those others were convicted murderers, rebels, heretics, or prisoners of war. In Jeff Fynn-Paul’s terms, by the end of the Middle Ages both Christendom and the Islamic powers of North Africa and the Middle East had become “perfect no-slaving zones,” by which he means co-religionists could not be enslaved.⁷² We know that galley slaves based in Spanish America could be Moriscos, North Africans, or orthodox Christians, but were their offspring also slaves? A documented instance of a white-skinned chattel slave, complete with subsequent progeny who were also enslaved, has yet to emerge from the archives of the Atlantic world. By contrast, light skinned mulattos could be slaves, as could their children.

A fuller understanding of the shift toward exclusively African enslavement requires us to step outside both the medieval millennia, and the 1500–1800 era, and look briefly at the evolution of Homo sapiens. As is well known, a range of hominids, including Homo erectus, Neanderthals, and eventually Homo sapiens, evolved first in Africa before populating the Euro-Asian landmass. Homo sapiens emerged over a period from about 500,000 to 300,000 years ago, but at separate locations. Advances in DNA analysis have, rather counterintuitively, destroyed the consensus on the timing of the exit from Africa that used to exist.⁷³ Considering the above discussion, however, the first point to note is that if all hominids, including Homo sapiens, originated in Africa then they must have been black-skinned. The obsession over the causes of black skin that has preoccupied Western intellectuals over the last millennia shows a Eurocentric preoccupation with the wrong question. The issue should rather have been what caused white, red, or yellow skin.⁷⁴

Black migration out of Africa meant a series of journeys that were overwhelmingly one-way and comprised a series of constant goodbyes. The final stages by sea through to the settlement of Hawaii and New Zealand occurred about one thousand years ago. We now know that Homo sapiens interacted with earlier versions of humans who had migrated well before the main exodus of Homo sapiens, Neanderthals and Denisovans, especially. “Interaction” in this context no doubt meant a great deal of violence over millennia in addition to extensive interbreeding.⁷⁵ As the last ice age ended about 11,500 years ago, rising water levels impenetrably sealed off huge sections of global populations from each other. The cultural implications of these tectonic events were enormous and continue to shape social tensions in our modern world. Migration meant that the number of languages and cultures in the world peaked – probably just prior to 1500. Ocean navigation enabling return voyages, as evidenced by the exploits of Admiral Zheng He and Columbus, initiated the reintegration of global populations, a process still in its very early stages in the twenty-first century. Cultural diversity, as evidenced by language loss, has been in decline ever since return voyages became not only possible, but routine, and will no doubt continue.⁷⁶

The isolation of the Americas and the late arrival there of Homo sapiens also ensured that for four centuries after 1492 America continued to have by far the largest land-to-labor ratio in the temperate world. Historically these are not the conditions under which individuals usually work voluntarily for others for any extended period.⁷⁷ Slaves certainly existed in probably every Indigenous society in the Americas, but even in the Indigenous empires of Central and South America, they were very much in the minority. To believe that America’s resources would have entered the global economy without empire, slavery, racism, extreme violence, and unequal income distribution is to ignore global history. As explored further in Chapter 2, a transoceanic slave trade could have happened without Western Europe and without Africa, but not without America.

The key difference between the initial dispersion and the later reintegration of peoples was the time span involved. Whereas migration by land had occurred over millennia, and new languages and cultures emerged gradually, transoceanic migration brought peoples who were very different from each other, culturally and physically, into sudden contact. Oceans had ensured that many societies had developed without awareness of continents other than their own. The reactions in northwestern Europe to peoples and cultures previously beyond reach differed from the reactions of their Iberian neighbors to the south, who were still accustomed to living with multiethnic enslaved people. Consider the contrast with England, which became the first country in Europe to permanently expel its Jewish population (in 1290), then became the first to expel Black people (in 1599 and 1601). Despite repeated incursions by various mainland European peoples stretching from the Bronze Age to the Norman Conquest, Geraldine Heng identifies England as the “first racial state in the West.” It appears almost inevitable that a few decades later its colonies in the Americas presided over one of the most oppressive and, in racial terms, rigidly defined slave regimes in recorded history. In this they were joined by the Dutch and the French.⁷⁸

By contrast, the Reconquista in Iberia, and the slave-trading prowess of both Genoa and Venice throughout the Mediterranean and Black Seas, ensured that the multiethnic conception of slave eligibility, characteristic of the medieval millennium and earlier, was abandoned much more slowly in southern Europe than in the north. People of African descent comprised 10 percent of the Lisbon population in 1480, not all of whom were enslaved. Official arrivals of enslaved Africans to the port averaged 2,000 a year between 1490 and 1516.⁷⁹ But at the same time the degredados, sent to a lifetime of labor in overseas Portuguese possessions, and the galley slaves at work in the sixteenth-century Spanish Caribbean, were certainly Europeans, many of whom were orthodox Christians or Jews.⁸⁰

Among the maritime powers of Europe, the Portuguese, Spanish, and Italians had the strongest connections with sub-Saharan Africa in the late medieval period. Not surprisingly, they also developed slave systems and laws in the Americas based on the Castilian code, Siete Partidas, of 1265. This contained an escape hatch or two to freedom missing from similar codes in the slave colonies of their northern neighbors. David Wheat has argued that the pre-plantation Spanish Caribbean (lasting in Cuba’s case until well into the eighteenth century) was settled by Blacks and Whites together. Blacks were certainly early victims of the slave trade. But by the early seventeenth century, large free populations of African descent in major cities, and the rural areas that provisioned them, have led Wheat to describe Blacks as “surrogate settlers.” New Spain did have a modest sugar complex in the first century after the Spanish reached the mainland. But most Blacks, enslaved and free, mostly “Latinized,” and living in both the familiar towns (pueblos) and surrounding rural areas (partidos), had quickly come to form most of the non-Indigenous population of the early Spanish Americas. The practice of coartación (or self-purchase) remained in place through to the abolition of slavery in the late nineteenth century, even as the Spanish sugar and Brazilian coffee plantation complexes reached their respective apogees.⁸¹ Self-purchase notwithstanding, slaves held in the Americas by Western European powers quickly came to be exclusively African and exclusively employed in the export economy and the activities that the export economy enabled. Africans in the Spanish Americas came to be clustered around export activities, urban centers enriched by such activities, or in areas that were strategically important in protecting routes to Europe – as in early Cuba and the Rio de la Plata. Differences between Iberian and non-Iberian systems in the Americas were slight compared to differences between slavery in the Americas and slavery in the rest of the world.

The horrors of the slave trade were thus in part the result of people separated by vast and unnavigable bodies of water developing different perceptions of how they saw themselves in relation to others. Slavery and the traffic in slaves evolved without much discussion or self-reflection within the societies that adopted these practices. If an opportunity to acquire captives occurred, it would be taken. Questions of morality arose only over eligibility for enslavement, not the practice itself. For major groups such as those living in Christendom and under Islam, and in various east Asian imperial polities, the identity point is obvious, particularly when counterposed with the inhabitants of meso-Americas and sub-Saharan Africa, where no such overarching identity existed. As already noted, no Indigenous American identified themselves as American. No one living in sub-Saharan Africa identified as African. Until the late eighteenth century throughout the world eligibility for enslavement was always the central question, not the practice of slavery itself.

To summarize, three factors shaped the cataclysm that was the transatlantic slave trade: first, consumer desire for the products of plantations; second, rapid diffusion of new maritime technologies facilitating reintegration of global populations; and third, the fact that Africans and Indigenous Americans had a much narrower conception of who was ineligible for enslavement than had Europeans. But almost as important in explaining what happened in the Atlantic world after 1500 were the differences in perceptions of self within Christendom. Britons and Iberians may have seen themselves as Europeans, and thus ineligible for enslavement, but they had different codes, both formal and informal, governing their relations with Africans. It was inconceivable that the English would have occupied Jamaica or the Dutch, Surinam, with the same settlement pattern that prevailed in Spanish Cuba, much less that they would have institutionalized self-purchase for their slaves and overseen the emergence of colonial populations comprising an intermixture of indigenes, people of African descent, and Europeans. Miscegenation has negative connotations in English and Dutch completely missing from its Spanish-language counterpart, mestizaje. Strikingly, several Latin American countries have embraced the concept of the mestizo as central to the formation of new national identities in the era of independence. English-speaking countries in the Americas would have found this impossible.

The centuries-long evolution of identity and its implications for the exclusive enslavement of Africans in the Atlantic world, as well, eventually, as in much of Islam, helps explain the continuing impact of Atlantic slavery on the modern world. Of all the many past horrors, the trafficking of people first from Africa to the Americas, and soon thereafter between ports within the Americas, has had one of the most enduring influences on the modern world. Contemporaneously, famines, conquests, wars, attacks on minorities, and suppressions of rebellions also created vast zones of human misery.⁸² But the cursory overview of global history presented here does underscore that those historical perspectives on slavery and the slave trade changed over time. According to the Old Testament, all but one of the twelve tribes of Israel were enslaved in Egypt. Enslavement and exodus are core elements of Jewish ritual, but enslavement of ancestors is not one of the many issues that separate Jews from non-Jews today. In China, Korean women were prized as concubines for centuries, as well as the eastern Khanates both before and after the Mongol conquests. These women were forced into sexual slavery across east Asia and beyond. But while the Japanese exploitation of Korean “comfort women” in World War II remains a bitter issue between the two countries, the earlier and vastly larger scale of exploitation perpetrated by China does not. And Western intervention in the Balkans in the 1990s was not complicated by memories of long centuries of Latin slave trading in Slavic regions.⁸³ More recently, the mark of slavery in parts of what is now coastal British Columbia was a flat head brought about by applying pressure to the head of infants.

Figure 1.1 Deformation as a mark of enslavement? Reproduced with the permission of the US National Anthropological Archives, Smithsonian Institution, BAE GN 03084.

“A flattened head was a mark of beauty among some indigenous groups living along the Lower Columbia River in the nineteenth century. Such groups saw round-headed individuals as potential enslaved persons. Indigenous groups in British Columbia by contrast saw flattened heads as marking the potentially enslaveable. The flattening was the result pressure applied to an infant’s forehead and occiput. Epidermalization of slave status – the norm in the Atlantic World between 1600 and 1888 – has been highly unusual over the several millennia in which slavery across the globe has flourished.”

Yet among the Wakashan and Chinookan-speaking Indigenous societies in the Lower Columbia River area to the south, a flat head signified membership of a group that owned slaves. Warfare and enslavement had created this situation, yet resentments, indeed, memories, of these physical distinctions of enslaved and enslaver do not appear to inform modern interactions of these two groups.⁸⁴ By contrast, the enslavement of Africans in the Atlantic continues to feed into deep modern social inequities, most recently with the higher viral infection rates during pandemics among African Americans.

Is it still useful, though painful, to ask when the Holocaust and the transatlantic slave trade will come to be treated with the detachment that historians assume in evaluating, say, the slave trade generated by the Roman, Mongol, Islamic, and Latin empires? The short answer to this question is no time soon. A slightly longer answer, on Atlantic slavery at least, is that perhaps it might happen when Black/White income differentials disappear, or when Black/White ratios in prison populations – as well Black/White victims of police shootings of unarmed people or awaiting execution on death row – come to reflect Black/White ratios in the general population. This view does not come from a belief in progress (though in material terms this cannot be questioned), and it certainly does not mean that racism will eventually disappear. Sadly, anyone looking back over 3,000 years of racist stereotyping in the written record must conclude that, while overall violence has declined, the target populations of discrimination, often based on physical appearance, will not disappear. Instead, they may well continue to shift, just as they have throughout global history.

It will not have much impact on current racial inequalities, but it is worth recognizing that every one of the nearly 8 billion people on the Earth today are descended from Black people. It is equally certain that all of us have both a slave and a slave owner in our ancestral heritage. The Legacy of British Slavery group at University College, London, began life with a mission to identify all British slave owners. In addition, www.slavevoyages.org has now developed user interfaces that do the same for owners of slave ships of every nation. These show the very wide distribution of participation in ventures designed to sustain and enforce coerced labor. Bondspeople were frequently owned collectively, thus spreading slaveholding further across the community. Monasteries and temples in South and Southeastern Asia owned thousands of slaves, many given to monks as a testament of lay piety. Kinship groups were even more important collective owners in Indigenous African and American societies. In Brazil, many enslaved people were themselves owners of enslaved people.⁸⁵

For evidence of the slave status of non-Africans, researchers must dig deeper, but no modern ethnic or national group can possibly be free of such ancestry. The enslaved normally accounted for a small share of a society. But as already discussed, slavery was ubiquitous, ancient, and female. The temporary nature of bondage in many historical societies has also ensured wide descent lines. While, say, 10 percent of a population may have been enslaved in each society, they would not have been the same 10 percent over a lifetime, and not all manifestations of slavery in Asia were hereditary. Orlando Patterson describes the phenomenon of “social leakage,” by which he means the high manumission rates in some slaveholding societies. In the Islamic World, much of non-Islamic Africa, and the Indigenous Americas, gradual erosion of slave status over time can be observed as slaves (or their descendants) slowly integrated into the community to be replaced by others. But the central reason for the widespread descent from slave status of all modern populations was the majority-female component of most servile groups. Records dating to the second millennium BCE indicate that some two-thirds of bond-persons in Babylonia were female.⁸⁶

It is therefore not surprising that presidents George W. Bush and George Herbert Walker Bush had a direct ancestor, one Thomas Walker, who owned and captained slave ships sailing from Bristol between 1784 and 1791 before he emigrated to the new republic.⁸⁷ What has received less attention is that, some centuries earlier, the West Country of England, the ancestral home of the Bush family, very likely participated in other kinds of slave trafficking, as either slave traders or as victims of slave traders, or both. Meanwhile, Icelandic children were sold in the Bristol region as late as the early fifteenth century, and, before 1100, English slaves captured by Irish raiders were sold in other parts of Europe and in the Mediterranean. Irish-Scandinavian Dublin had a major slave market in this era. Bristol, for a few decades the major slaving port in Europe for transatlantic slaving ventures, thus had a centuries-long experience of both importing and exporting captives, well before it began to send slave ships to Africa in the late seventeenth century. According to the Domesday Book survey of 1086, one in five people in England’s West Country were enslaved (as opposed to an overall average for England of about one in ten). And as late as the early twelfth century, monks visiting Bristol were warned that Irish merchants on ships in the harbor were likely to kidnap and carry off the unwary to slavery. The families that organized the later African trade, including the eighteenth-century Walkers before they emigrated, could easily have had ancestors enslaved in England and counted in the Domesday Book. Some perhaps were dispatched to Ireland. Others, according to a widely cited letter of Pope Gregory I, were sent to the slave markets in late sixth-century Rome.⁸⁸

It was not just the Irish Sea and its surrounding territories that saw slavery and slave trading continuing without abatement for centuries after the fall of Rome. As we have seen, slave routes crisscrossed the Old World (and probably parts of the New) during the 450–1450 era. For the global slave trading that preceded European contact with the Americas, large blind spots remain, and will forever lie beyond the light of scholarship. We will never know, for example, the demographic impact of the disruptions stemming from Mongol conquests, or the extent and direction of slave trading in the early Indian Ocean. In some important respects, however, the advantage lies with the earlier period. For the traffic that kept the slave markets of Eurasia supplied for a millennium before 1450, the sources typically tell us much more about the origins of enslaved people than is available for early modern Africa, as well as the conflicts that led to their capture. While we know the region of embarkation for more than half of the vessels that brought captives to the Americas, for fewer than 1 percent of these arrivals do we have homelands in the interior of Africa. Except for seventeenth-century Angola, we have only sketchiest information on the wars and raids that generated the Africans dispatched to the Americas. Prior to 1807, at least, African records that might reveal this information simply do not exist. While scholars now recognize the “many middle passages” in the making of the modern world – the convict trade from Europe, contract laborers from Asia, as well as the enslaved from Africa – only a few medievalists focus on the largely terrestrial movement of forced labor across Eurasia that preceded these migrations. These, too, were “middle passages,” albeit not transoceanic. Medievalists are much more likely to reference events in the post-1500 Atlantic world than are scholars of the later Atlantic to draw parallels with, say, the Mongol Empire. Yet a comparative review of earlier slave-trading activities and their transoceanic African successors does provide both context and continuities between medieval slave trades and the cataclysm that subsequently engulfed the tropical Atlantic world.⁸⁹

The broad pattern running through all written records on slavery is that power imbalances between polities or world religions usually resulted in violence, and at least the partial enslavement of the weaker of the two. The first reference from Mesopotamia describes women and child captives being distributed among members of the court and the army generals of Babylon or sold to those who could afford them after a successful war. Most adult males were presumably killed after capture. If we set to one side the necessarily short-term Nazi enslavements in the Holocaust, then the last such case of a power imbalance triggering enslavement occurred in the more remote reaches of Polynesia. Beginning in 450, Polynesian farmer migrants with identical technologies began sailing east. After centuries-long island-hopping, their descendants reached New Zealand and the tiny Chatham Islands, 500 miles west of New Zealand at about the same time – 1000 CE. The two groups of migrants then evolved in isolation of each other over the next eight centuries. The Maori used intensive farming to establish a technically advanced, warlike, and highly stratified society in New Zealand. Meanwhile, the Chatham Islanders – the Moriori – reverted to hunter-gatherer status in the face of poor soil conditions, but abundant shellfish, nesting seabirds, and seals. In the 1830s the Maori learned of the existence of the Moriori from European sealers. They thereupon invaded the Chatham archipelago and slaughtered or enslaved all the estimated 2,000 inhabitants. This enslavement lasted until 1863. “It was” said a Maori leader, “in accordance with our custom.”⁹⁰

Between Ancient Mesopotamia and the Chatham Islands of 1863, and no doubt well before the written record began, not much human migration occurred without violence. It is unlikely that Neanderthals negotiated their peaceful takeover by Homo sapiens. It is equally improbable that the hunter-gatherers of the Philippines voluntarily yielded their territory to Austronesian-speaking farmers several centuries prior to the written record.⁹¹ The mix of Neanderthal and Denisovan genetic material that we carry today was probably the result of the familiar pattern of raids and warfare that dictated only the women and children of the Neanderthal and Denisovan target population would survive. The maritime manifestation of both human occupation and reoccupation of the globe remained the same from Homer’s Greece through to Irish, Viking, and Islamic raiders and farmer migrants in parts of Indonesia. With one dramatic exception all used the same violent measures. Jonathan Gottschall’s generic description of the earliest of these five examples applies to them all:

Fast ships with shallow drafts are rowed onto beaches and seaside communities are sacked before neighbors can lend defensive support. The men are usually killed, livestock and other portable wealth are plundered, and women are carried off to live among the victors and perform sexual and menial labors.⁹²

For several millennia Gottschall’s description held, but the big exception was the African slave trade. For the most part, the embarkation points for European and Arab slave merchants on the African coast were places where captives could be purchased, rather than bases for launching raids to capture them. Raiding and other sources of captives, of course, occurred, but off-site and certainly off-camera, as it were. The role of the major African embarkation points was akin to that of Venice or Caffa described earlier.

The Europeans who took over or invaded the early modern Atlantic and maritime southeast Asian worlds had a more complex self-justification for their actions than did the Maori. This involved religious proselytization and an Ulrich Bonnell Phillips-like conception of Africa as a place so savage that slavery elsewhere would save not only souls, but bodies. However, in the end, such justifications are little more than “according to our custom,” particularly with respect to eligibility for bondage. In both the East and the West, Europeans frequently came across pre-existing institutionalized slavery and slave trading. Writing of the traffic that the Dutch found (and quickly joined and expanded) when they arrived in the Indonesian archipelago in the seventeenth century, François Valentijn described the slave trade as the “world’s oldest trade.”⁹³ In the West, early expansion into the Atlantic tapped into an ancient trans-Saharan traffic in people. In terms of enslavement, Europeans became the Mongols of the oceans. Like the Maori, wherever power imbalances permitted, Europeans organized a slave trade as in the early traffic in indigenes from southern to northeastern Brazil. Where the imbalance was not pronounced – as on the Western African coast described in Chapter 5, and in addition to most of the Asian landmass – they bought rather than raided for people.

Ordinary Europeans, rather than rapacious capitalists and rulers, were ultimately responsible in first, responding strongly to the lower prices for products that plantation slavery ensured, and second, refusing to migrate in large numbers to the tropical Americas. Relative freedom for Europeans meant extreme oppression for many Africans. And the cataclysmic decline in the Indigenous population of the Americas linked European consumers directly with the African slave trade. Europeans had the option of not signing up for plantation labor, except, it may be assumed, for wages that would have greatly reduced both plantation profits and output. It is striking that the early modern commercially minded states and their elites refrained from enslaving even marginalized Europeans. Instead, European expansion meant the destruction of two-thirds of the Americas’ Indigenous population by the second half of the seventeenth century. It meant the forcible removal of 12.75 million Africans to another continent. And it meant many more dying in the disruption caused by the traffic.

Nevertheless, the slave trade as it evolved in the Atlantic world was simply a maritime – and much better documented – version of what had been happening on land for millennia. Europeans moved far greater numbers of people over far greater distances than had ever before happened in global history. These numbers and distances were not exceeded until the mass migration of Europeans to the Americas began in the mid nineteenth century. Yet in neither the medieval nor early modern eras was slave trading among the worst examples of man’s inhumanity to man. Historians can easily point to cataclysmic events in the last two millennia of global history that had a much greater demographic impact than the transatlantic slave trade: the 40 million Amerindians who died in the sixteenth century alone, or the 20 million people who died in the 15 years of the Taiping Rebellion, to cite just two relatively short-term catastrophes that bookend the era of Atlantic trafficking.

Four novel features help explain the enduring impact of the Atlantic slave trade on modern life – as opposed to other atrocities in history. It was not the number of enslaved people that separates slavery in the Americas from slavery in the Old World, but rather the exceptionally high proportions of such populations that were enslaved in most American regions. Such societies have been rare in history. So it makes sense that the only successful slave revolt in global history – in St. Domingue, now Haiti – should occur in such an environment. A second novelty was the 340-year exclusive focus on a single geographic source for captives: sub-Saharan Africa. This was a place of dizzying cultural complexity. In 1850, Africa was inhabited by just 5 percent of the global population but was home to 25 percent of all global languages.⁹⁴ As already argued, such cultural diversity ensured that no one living in Africa without contact with Arabs, Europeans, or Asians could possibly have conceived of themselves as African, sub-Saharan African, or having an identification with any region larger say than early Mali or the Sokoto Caliphate. Even among non-Africans, the term “African” only gradually came to replace other generalized nomenclatures such as “Blacks,” “Blackamoors,” or variants of the word “Negro.” The system in which they were held came to be known as “racial” slavery, even though it is absurd to refer to Africans or their descendants as comprising a race. This helps explain why, against all the historical evidence, for most people slavery is still associated primarily with people of African descent.

A third novel feature is that the transatlantic traffic is the only large-scale slave trade in history that, at least until its last three decades, was predominantly comprised of adult males. New World plantation owners preferred to buy males. African sellers preferred to retain females.⁹⁵ Europeans solved the security issue thrown up by attempts to move large numbers of men across long distances by adding defensive features to slave vessels. Finally, the fourth novelty is the persistent question of the connection between historical practices of slavery and slave trading and the emergence of part of the modern economically developed world. Some scholars, and increasingly most of the modern media, see a strong causal connection between slavery and the economic rise of the West, even though three thousand years of recorded history suggests that a reverse of this cause-and-effect relationship would be much more plausible. By this, I mean that a global historical perspective points to prosperity invariably enabling slavery, rather than slavery having had an essential role in enabling prosperity.⁹⁶ All these distinctive features are issues that are taken up more fully in the chapters that follow.

Finally, a note on terminology and methodology. The foregoing survey makes clear the continuity of enslavement practices across the ages and suggests that the peak number of enslaved people on the globe probably occurred before the plantation complex moved into the Mediterranean Atlantic, much less the Americas. In the last two decades, historians have begun to use the terms “first slavery” to refer to slavery in the Americas before circa 1800 and “second slavery” to what happened after (or because of) industrialization.⁹⁷ These phrases reflect the ingrained Eurocentrism of the World-systems school of scholars, first established by Immanuel Wallerstein.⁹⁸ Slavery in the Americas before 1800 shared far more with its post-1800 successor than with its pre-1500 antecedent (and with its very extensive, non-European contemporary systems of bondage). How would the experience of working in the cane fields of eighteenth-century Jamaica have differed from the same environment in nineteenth-century Cuba? The history summarized in this introduction establishes that the global peak of slavery occurred before 1492. How could slavery in the Americas between 1500 and 1800 possibly have been the “first slavery”? Forcing the history of slavery into such a framework and viewing prosperity as the result of slavery when it was more often its cause, distorts historical reality.

Many recent historians have expressed concern over what they see as the dominant role of statistics in the study of the slave trade. As the outlines of both the transatlantic and intra-American slave trades have come into focus in the last few decades, revealing new issues and paradoxes, historians from James Walvin to Stephanie Smallwood have worried that the figures fail to convey the “human” element in the terrible sequence of events that transferred millions from one side of the Atlantic to the other. Marcus Rediker, calls such an approach “a violence of abstraction” – by which is meant writing about slavery without showing its devastating human toll – that has “dehumanized … reality.”⁹⁹ For Toby Green, quantification inhibits scholars from “thinking through … the cultural, political and social consequences of this phenomenon [meaning the slave trade].” Vincent Brown sees quantification as rendering “the deadly migration of Africans somewhat like the chalk outline of a murder victim.” Most critics are less melodramatic and simply resort to tired variations of “these people were not mere commodities,” as they most certainly were not.

Yet trying to imagine scholarship on the slave trade and, indeed, the Atlantic world, in the last half century without knowing the size, direction and mortality/morbidity of this largest forced migration in history is rather like battling disease without the discipline of epidemiology, or climate change without long-run data on temperature. No discipline other than history would tolerate such a position. Would our appreciation of the horrors and sufferings of the slave trade be greater today if we did not know its dimensions? My argument rests firmly on the belief that one must know what a phenomenon is before one can think through “the cultural, political and social consequences” of that phenomenon. This in turn cannot happen without data as Jennifer Morgan’s critique of such data sources makes plain.¹⁰⁰ Scholars have agonized over drawing on archives created by exploiters and racists. But given the ubiquitous acceptance of slavery until relatively recently, everyone was at least a potential exploiter including, in Brazil, some enslaved people and former slaves. For every abolitionist-inspired Solomon Northup, there were many millions of White, Black, Indigenous American, and Asians whose experience of enslavement before 1800 did not erode their subsequent willingness to become slave owners. They, too, helped create the pre-1800 archives.¹⁰¹