For anyone who, despite being human, is by nature not his own but someone else’s is a natural slave. And he is someone else’s when, despite being human, he is a piece of property . . .
Aristotle, Politics 1.1254a 14–16, trans. Reeve 1998
Slavery is the fact that one man is the property or possession of another.
Herman J. Nieboer, Slavery as an Industrial System: Ethnological Researches (2010, originally 1900), 7
Slavery is the permanent, violent domination of natally alienated and generally dishonored persons.
Orlando Patterson, Slavery and Social Death: A Comparative Study (1982), 13
Greek slavery was plainly different from Roman slavery; both were very different from New World slavery on the one hand, and from the slavery attested in anthropologists’ reports on less complex societies on the other hand. But what was the same? Why do we apply the same term, slavery, in such varied societies across such a large span of history and geography? What exactly is slavery?
Non-historians sometimes assume that any extreme system of oppression constitutes slavery. For example, somebody today might condemn a clothing manufacturer for using “slave labor” in its Third-World factories. If asked to explain the use of “slave labor,” that person might point out that the workers are paid poorly, work long hours in appalling conditions, and may be fired if they get sick and cannot work for a couple of days. In this case, the use of “slave labor” expresses severe disapproval; we all deplore slavery, and so we ought to deplore the practice of this hypothetical clothing company.
Historians generally do not approve of oppression, and virtually all condemn slavery severely, but to understand how different societies and institutions work, more precise categories and criteria are needed: “nasty” and “nastier” are not the most helpful terms. Herman Nieboer, an early anthropologist of slavery, provides what has become the traditional historical definition in terms of property in the quotation above – a criterion that goes back to Aristotle in fourth-century Greece. Such a definition does not necessarily imply bad treatment; people can take good care of things they own, and they can take good care of their slaves. Indeed, in most slave societies we can find slaves who, in material terms at least, are treated well. For example, some slaves in Athens managed banks for their masters, sometimes gained their freedom, and, in at least a couple of cases, ended up among the wealthiest men in the city. In the process they almost certainly also made their masters rich. These slaves were probably far too valuable for their owners to risk angering them, much less not feeding them enough or whipping them.
While some slaves were treated well, the majority were not. It is important, however, to look at the situation of slaves in the context of other members of society. For instance, the plight of the free poor in the ancient world was often extreme. They rarely starved to death, but malnutrition was common and the threat of starvation was a constant concern; they suffered through winters with inadequate clothing. In contrast to slaves, nobody else suffered a direct financial loss if a poor free person died. Dependent peasants and serf-like classes of various sorts, such as the Helots subject to Sparta, often lived lives as harsh in material terms as slaves. They were not fully free either: some owed much of their produce to their lords; others also owed days of personal service; and many were not allowed to leave their farms. Like slaves, such classes rarely had any outside power to turn to if they wanted to mitigate their oppression or curb egregious abuses.
While it is thus important to realize how harsh was the lot of most people throughout most of history, this grim context does not make slavery any better. A few key points will neutralize any temptation to underestimate the horrors of slavery. First of all, to treat a person like property is in itself to humiliate and dehumanize him or her. Plenty of evidence for this feeling in the ancient world exists: already a character in Homer’s Odyssey bemoans that “Far-seeing Zeus takes away half of a man’s worth when the day of enslavement comes upon him” (17.323). Second, most people do not buy expensive property, like slaves, just for the sake of treating it well, but rather to use their property and often to profit from its use. Calculating masters wanted their slaves to stay alive, but they also wanted them to work as hard as possible. In many cases they used punishment and the threat of punishment to make their slaves work hard. Third, not all masters were calculating or effective; many appear to have beaten or abused their slaves out of anger or cruelty despite the fact that they were damaging their own financial interests. Slaves had virtually no recourse against this sort of treatment.
Finally, we need to focus on what is typical and what is exceptional, and well-treated, elite slaves were exceptional. Athens was large enough to provide work for perhaps a couple of dozen slaves in bank management; in contrast, the Athenian silver mines may have employed over twenty thousand slaves. A Roman noble may have kept a couple of highly educated, well-treated Greek slaves as status symbols and tutors to his children. But over a thousand slaves may have spent their lives doing hard manual labor on his farms scattered across the Italian countryside. So, even though historians define slavery in terms of property rather than nastiness, classical slavery was still a deplorable, and usually a harsh, system of exploitation.
The definition of slavery in terms of the property relation between the master and the slave is useful because ownership is a familiar concept. Yet Orlando Patterson points out in Slavery and Social Death (1982) that ownership is more complex than it seems. For one thing, people in different societies possess all sorts of property rights in other people. For example, the marriage arrangement in which prospective grooms give “bride-price” to their bride’s fathers implies that the bride was in some sense the property first of the father and then of her husband. This did not constitute anything like slavery: a wife was proud of her bride-price and would be ashamed if her family had not received one. In light of these counter-examples and other issues, Patterson proposed an influential alternative definition of slavery.
Patterson defines slavery as the “the permanent, violent domination of natally alienated and generally dishonored persons” (1982, 13). That slavery is permanent and heritable – the children of slaves typically become slaves – distinguishes it from debt bondage and indenture, both of which can be as harsh as slavery and can involve sale of a person, but are not necessarily permanent and often not heritable. Many systems of oppression ultimately depend on violence, but the violence of slavery is particularly prominent and open. For example, slaveholders at a wide variety of places and times used the whip to punish and intimidate slaves. And most slave systems required a supply of new slaves. Reducing free people to slavery generally requires violence – often war – or the open threat of violence. It is not surprising that the free members of a society think that slaves lack honor – although the slaves’ own views may differ! By “generally dishonored” Patterson means that they are not dishonored on one occasion or in some particular way, but are fundamentally without honor, without respect, within a society.
The most enlightening part of Patterson’s definition is the concept of natal alienation, which is related to the “social death” of his book’s title. Slaves are natally alienated in that they lack all the rights that other people acquire with birth: their claims on their parents, their relations with siblings, and their links with and prestige deriving from their ancestors. In fact, slaves have parents and sometimes have siblings, children, and wives; they remain socially dead because they cannot acquire any such claims or establish any formally recognized relationship other than the one subordinating relationship with their masters. In the words of Brent Shaw, “Chattel slaves are normally denied a viable family life and tradition, and are thereby deliberately cut off from the rich and complex ways in which the identity of the person is attached through the family to community traditions and empowerments” (1998, 14). In the modern world, we are used to the idea that people are born with human rights, but in ancient societies a person’s rights came from being somebody’s son or somebody’s wife or a member of a clan or a village or larger state. Slaves are shut out from any of these rights. Patterson believes that slaves are socially dead in this sense. That may be why they can be treated as things, as possessions, and why defining slavery in terms of ownership rarely leads you astray. It is, nevertheless, important to understand Patterson’s viewpoint for the additional insights it yields.
Despite these satisfactory and illuminating definitions, it would be an oversimplification to imagine that ancient slaves constituted an homogenous category surrounded by a precise border. Some ancient slaves enjoyed perquisites that we would associate with freedom. For example, one group of slaves in classical Athens were called the choris oikountes, “those living apart” (Kamen 2011). They seem to have lived separately from their masters, sometimes in families. They paid their masters a set amount each month but were not regularly subject to their masters’ commands. And, both in Greece and Rome, some slaves had de facto control of personal property – considerable wealth in some cases. Since masters could, at their discretion, reverse these arrangements and sell their slaves away to who knows what fate, both the choris oikountes and wealthy slaves still fit our definition of slaves. On the other side of the line, some groups do not fit the definitions of slavery, even though we would not regard them as fully free. For example, the Helots were a group subject to the Spartans – treated in detail in Chapter 10 on account of their rebelliousness. They were not allowed to leave their farms; they owed a proportion of their produce and, sometimes, personal service to the Spartans. As we’ll see in Chapter 13, the coloni adscripti of the late Empire and eventually medieval serfs endured similar obligations and restrictions – and historians have found other cases throughout the classical world. Such persons could not be sold individually; they lived in their own families and villages. So we do not categorize them as slaves. Still it would be ridiculous to insist that they were completely free. The same is arguably true of many ex-slaves and certainly of those in Hellenistic Greece subject to paramonē, a sort of conditional or deferred freedom that we’ll meet again in Chapter 7 (Zelnick-Abramovitz 2005; contra Sosin 2015).
Two considerations should keep these complexities from dismaying us too much. First, there were many unambiguous slaves and unambiguous free persons. Our definitions of slavery tell us something important about a large proportion of the people in the classical world – not everything about everybody, of course! Second, this same imprecision and complexity affects many other key concepts that historians rightly consider crucial to their craft: capitalism, industrialism, the market economy, democracy, imperialism, law, and others. These terms are all subject to debate, partial, and hard to delimit precisely. They are still valuable, virtually irreplaceable, for historians. We are also better off with a definition of slavery. Although my focus will be those people who fit this definition, unexpected privileges are an important part of their story. And those groups, like the Helots, “between slavery and freedom” will also figure in this book.
Historians of slavery use the expression “slave society,” to indicate societies in which slavery was a central institution (Turley 2000, 62–100). Throughout history, many societies have allowed slavery, but slave societies are far fewer: ancient Greece and Rome and the three New World slave systems – in Brazil, the Caribbean, and United States South – are members of that short list. Historians use different criteria to determine whether slavery is central to a society or not (Bradley 1994, 12–30). These range from the numerical proportion of slaves in the population, to slavery’s economic, cultural, or institutional impact on a society. Do slaves constitute more than, say, 20 percent of the population? Do slaves dominate key areas of the economy? Does the society’s elite derive their wealth from slave labor? Did slavery play a dominant part in the culture and institutions of a society? In Chapter 4, I will use the economic definition of slave society, but the different criteria often go together: slaves become a large proportion of a population when their economic role is important, something that often translates into high visibility in a state’s culture and institutions.
Historians of ancient slavery have far less evidence to work with than do their counterparts who study modern slavery. Historians of slavery in the American South, for example, can draw on a variety of primary sources, some types of which exist in large quantities. They possess full law codes, records of legislative debates, political pamphlets, plantation records, newspapers with news and editorials, posters describing fugitive slaves, memoirs and letters by slave owners, both male and female, and books written by slave owners to justify the practice of slavery to an increasingly hostile world outside the South. All of this evidence comes from slave owners, and so has a particular bias, while accounts by writers visiting the South provide different perspectives.
But historians of modern slavery still have a difficult time reconstructing the slaves’ side of the story. Contemporary narratives by escaped or freed slaves often had an openly abolitionist agenda; they aimed to present slavery in the worst possible light. Such accounts are not likely to dwell on any aspects of slavery that might make their condemnation of the institution at all ambiguous. Historians also worry that some slave autobiographies were in part ghost-written or heavily edited by white, highly educated abolitionists. Their perspective probably aligns closely with the ex-slaves’ abhorrence of slavery, but it could distort the slaves’ feelings, thoughts, and recollections in other ways. In the 1930s, the New Deal’s Federal Writers Project conducted systematic and extensive interviews with living ex-slaves. These had the advantage of recording ex-slaves’ exact words but had the disadvantage of asking people to recall events in the distant past; and most of the ex-slaves interviewed had been young at the time the Civil War brought an end to slavery. So even with a wealth of textual evidence and eye-witness accounts, historians of modern slavery still encounter problems interpreting their evidence, especially when it comes to the thoughts, feelings, and experiences of the slaves themselves.
These difficulties, however, are small indeed compared to those that face a historian attempting to reconstruct systems of slavery that existed between 3000 and 1500 years ago. Some of these challenges are the same as those that face any historian of the ancient world. The primary challenge is that writing, and thus textual evidence, was far less fundamental to classical Greece and Rome than it is in the modern world. Especially once we look beyond the elite, literacy was low and the culture essentially oral. One reason for this is that, in the absence of printing presses, writing something down was often expensive and time-consuming. So less was originally committed to writing, and spoken words, of course, leave no trace. Texts inscribed on stone, epigraphic evidence, often survive to the present, but most ancient writers employed perishable materials such as papyrus. For these to survive after the fall of Rome, through the Middle Ages, to the early modern age and the printing press, most texts needed to be recopied every few centuries. Papyrus texts preserved in the Egyptian desert constitute one exception, but overall only a tiny fraction, probably far less than 1 percent, of what was written down in classical times has survived to the present day. For example, Antisthenes was a disciple of Socrates and, reportedly, the son of a slave woman. We know that he wrote a treatise On Freedom and Slavery, but all that survives is one short quotation from this in a later source. If we had the whole of On Freedom and Slavery, we might know more about what Athenian intellectuals, and perhaps even slaves, thought about the institution of slavery, opinions that are otherwise poorly attested.
While less was written down and little of that survives, that is not all: what has survived is often not what historians of slavery would find most valuable. For example, we only occasionally have ancient statistics or enough evidence to allow us to derive our own statistics. As a rule, classical states did not even collect the sorts and amount of statistics that a modern nation does. Ancient Greek and Roman governments sometimes counted their people, usually when they wanted to know how many soldiers they could muster. Only very rarely do they seem to have (perhaps) counted the slaves. So historians need to make indirect arguments about the numbers of slaves and the proportion of the population that they constituted. For instance, starting in the sixth century BCE, Greek states sometimes taxed slave sales and manumission, the formal liberation of a slave (Zelnick-Abramovitz 2013, 21–27). The Romans too had exacted a 5 percent tax on manumissions starting in the fourth century BCE (Livy 7.16.7). Somebody must have known, and probably wrote down, slave prices in the process of collecting some of these taxes; such records did not need to last and, written on perishable materials such as papyrus, have disappeared long ago. We do possess several large collections of Greek inscriptions on stone commemorating manumissions, one set of which (from Hellenistic Delphi) even includes prices, invaluable evidence. Nevertheless, the big picture is that only rarely are ancient historians well informed even about basic issues such as slave prices and numbers.
Certainly, nobody sat down and wrote what they thought social historians from a different culture a couple of thousand years later might want to know. Rather, people usually wrote texts for their own immediate purposes. So ancient authors did not normally describe the way their world worked but simply assumed their readers knew. Just as a modern novelist would not bother explaining that most adults in the United States have jobs located someplace other than their homes and thus they “go to work,” rich Romans rarely felt the need to specify where in the house their slaves slept; everybody knew. As in this case, the things that must have been most obvious at the time are often those that historians today have the most difficulty figuring out.
Because ancient Greek and Roman authors rarely felt the need to describe the institution of slavery, we usually have to learn about slavery when it comes up in passing in texts about something else entirely. On the one hand, we do possess a couple of books of advice for rich Greeks and Romans about how to manage their farms as more or less absentee owners. Such books needed to treat slavery directly since the management of slave workers was a large part of running a farm. These works are crucial for the historian. On the other hand, we need to use evidence from all sorts of sources with no particular focus on slavery: tragedies, epic poems, epitaphs (inscriptions on gravestones), the text of laws and ancient commentaries on them, law court speeches, novels, bureaucratic records on papyrus, and ancient historical narratives. Inscriptions on stone, for example, often record official information, such as lists of slaves among the property confiscated from some condemned men in Athens, while epitaphs may bring us closest to the actual words or sentiments of slaves themselves. Even here we cannot always be sure whether it was the slave, his or her surviving family, or the slaveholder who put up the monument and decided what words were to be inscribed. These types of sources can help us understand ancient slavery, but they all are difficult to interpret for one or more reasons – as we shall see when we investigate particular cases throughout this book.
One last problem is crucial. Ideally, scholars of ancient slavery would have copious evidence by masters and by slaves, both male and female, as well as by people who were neither masters nor slaves. The most obvious bias in our surviving evidence, however, is that wealthy men, slaveholders almost without exception, produced virtually all of it. In their writing, they occasionally reveal their thoughts about slavery in general and about their own slaves in particular. They sometimes even represent the lives and words of slaves: for example, slaves play important roles in Greek tragedies. But it was wealthy free men who wrote these plays.1 Rarely do we hear directly from women and almost never from slaves themselves. So historians are faced with the problem of figuring out how slaves actually spoke, thought, and acted based on how their masters depicted them – and slave masters were hardly disinterested and objective observers of their slaves. This bias in our sources makes it hard to avoid a top-down view of ancient slavery and limits our insight into slaves’ perspectives or the active role slaves sometimes played in shaping their lives.2 In this book, I have tried whenever possible to understand or at least imagine the slave’s perspective, but these attempts must often remain speculative or superficial due to our one-sided evidence.
That wealthy men wrote the vast majority of classical literature sometimes introduces another distortion: slaves may be overrepresented, especially domestic slaves in comparison to the free poor. The lifestyle of the wealthy in both Greece and Rome depended on domestic slaves; their houses were teeming with them. Domestic slaves were a part of the everyday life of the rich and thus figure often in the literature they produced. In contrast, the free poor often played less of a role in the lives of the affluent. In some times and places we know less about them than about slaves: the free poor can disappear almost entirely from the world of our literary texts. For example, the theory of natural slavery in Aristotle’s Politics seems to assume that people are divided in two classes: natural masters and natural slaves. But what about all the people who were neither, who were free but not masters? The dichotomy of slave versus master was one way that ancient writers understood and simplified their world, which was, in fact, more complicated.
The example of ancient comic plays illustrates all these points. Both Greek and Roman comedies usually took place within the context of prosperous households and therefore included slave characters. Free Athenians who were not rich enough to own slaves hardly figure at all. The purpose of comedy is to amuse and entertain, so we cannot assume that the slave characters and the way they acted were typical at all. Many modern comedies are far from realistic about life today – I’m sure you can think of examples. Ancient playwrights were just as imaginative and willing to be unrealistic in their quest for laughs. The plays of Aristophanes (ca. 446–386 BCE), for example, included fantastic elements: in The Birds, a city of birds controls the world by intercepting the smoke from religious sacrifices, which the gods need to live on; another is set in Hades, the underworld. Can we trust Aristophanes to present realistic slaves? Even later comic writers whose plays were set within more or less typical households probably never worried, “Am I being unrealistic here?” Plays were judged on how well they entertained the audience, not on their veracity.
Writers of comedy, who were free men, addressed themselves to an audience composed also of free men, many of whom were slave owners. In early Athenian comedies, we find simple jokes at the expense of slaves. These make it obvious that slaves are not the intended audience, since it’s doubtful that slaves enjoyed jokes about terrified slaves getting beaten. But much of the humor and attraction of the presentation of slaves in comedy is more complex: for example, many comedies include a stereotypical character, the “clever slave,” who is smarter than his master and often outwits him. The audience may even have identified with these slave characters, who were, nevertheless, base or trivial in their goals and mentality, a characterization that reassured audiences of slaves’ inferiority.
We shall return to the complex and controversial clever slave character in Chapter 11. While it might be tempting to say that we learn nothing from such a problematic source as comedy, we can often make plausible guesses about what is a comic distortion and what is realistic. For example, that every family in Athenian comedy owns at least a couple of slaves does not mean that even poor Athenians owned slaves. Nevertheless, if we consider that even families depicted in comedy as poor or middle-class possess slaves, we can reasonably conclude that slave ownership was not confined to a small class of rich Athenians. Not just in the case of ancient comedy, but whenever we are trying to learn from an ancient text, we need to consider its original author, the audience it addressed, and why it was written.
So far we have been looking at textual evidence, but archaeologists study the ancient world by looking at material remains including sculpture, pottery, houses, farms, tombs and graves, temples and other public buildings, mines, and workshops. What material evidence can contribute to the study of ancient slavery and how easy it is to interpret varies greatly. Some material remains vividly confirm what we already know: many texts refer to the chaining and binding of slaves and, sure enough, archaeologists have found a variety of manacles and fetters from around the classical world (Thompson 1993). Sometimes archaeological evidence can shed light on issues that our texts do not illuminate. I’ve already mentioned the mystery of where slaves sleep. Excavations of large Roman houses sometimes suggest a separate, crudely finished area with tiny, low rooms in which slaves lived. We cannot, however, be sure whether all slaves lived there or whether personal attendants slept on the floors nearer to their masters. And most houses show no sign of designated slave quarters (George 2011, 387–390). Another case involves archaeological surveys of the remains of ancient walls, ditches, farms, and surface pottery fragments in the countryside. From this evidence scholars can try to estimate the average size of farms in different periods and places, crucial information for evaluating the role of slaves in agriculture. But material remains were not produced for the benefit of future archaeologists, and their interpretation can be as tricky as textual analysis.
Knowing what people thought about themselves, their world, and their place in it is just as important as knowing what people did, their actual practices, but it requires a different approach. Take, for example, a passage from the late Roman Empire in which Vegetius warns slaveholders against allowing slaves to ride their horses, because they may injure them.3 We can infer from this that slaves sometimes rode horses and that sometimes the horses ended up injured. Why else would Vegetius bother to warn about it? But we would also like to know whether slaves were given permission to ride as a perk or had to ride the horses as part of their job, to exercise the horse when the master didn’t want to. And how much of a problem was the laming of horses? These are examples of the issue of determining practice, what people did.
But Vegetius also claims that slaves ride horses recklessly out of competitiveness when they race each other. And he adds that the slaves don’t care about their masters’ loss when a horse is injured; they are delighted when this happens. These statements describe the thinking and (hostile) feelings of slaves. They are probably more interesting and important for our understanding of ancient slavery than the likelihood that some slaves exercised their masters’ horses. In this case – as often – we do not get the slave’s perspective directly, but rather what a member of the slaveholding class imagined slaves felt. One reason to be suspicious is a significant oversimplification of the whole situation: a master might suffer the loss of a horse; the slave responsible was unlikely to escape harsh punishment and might not be delighted for long. Indeed, Vegetius urges masters to curb reckless riding with the utmost severity – and this in a society where slaves could be whipped for much less.
Last, but not least, Vegetius is likely to have harbored all sorts of prejudices against slaves. His generalization about their careless riding may reflect stereotypes rather than an accurate judgment. To take a modern example, comments about “women drivers” being worse than men persisted long after insurance statistics proved the opposite. One can easily imagine a skilled slave groom indignantly insisting that he is actually much better with the horses than his young master, who, for example, rides infrequently and, when he does, is always trying to show off to his friends.
When it comes to slaves, both kinds of questions, about practices and about ways of thinking, are hard to answer. In the absence of direct evidence from slaves, arguments about their thinking are necessarily speculative; they are worth attempting nonetheless. For instance, in Chapter 6, we’ll see that slaves at Athens often retained a strong and proud sense of their original ethnic identity – many of them were from Thrace or from one of the western provinces of the Persian Empire. But the views of slave owners about slaves and slavery, the topic of Chapters 11 and 12, are better known thanks to the comparatively large volume of textual evidence they have left us.
Even when we are able to draw convincing conclusions from a piece or a body of evidence, we encounter the problem of typicality. In the case of some basic facts, ancient historians may know that something holds true for all ancient slaves: for example, that they were legally treated as property. Occasionally, historians may know to which slaves, at what times and places, a certain statement applies. More often, however, our evidence allows us to draw a conclusion of limited scope, and we need to consider carefully how widely we can apply it. This is particularly true when our evidence consists of a story about something that happened on one occasion.
Such anecdotes sometimes provide our only evidence for a practice, but stories often seem interesting and worth recording because they are not usual. To use a well-known example: newspapers do not bother to report every time a dog bites a person, but if a person bites a dog, that gets news coverage. According to this whimsy, historians of the future might come to the completely wrong conclusion by counting newspaper references to person and dog bites. We definitely see this “man-bites-dog effect” in what gets reported by ancient historians. A startling case is that of the wealthy Roman Vedius Pollio, who, according to the story, would feed disobedient or incompetent slaves to the lamprey eels he kept in the fish-pond on his property (Seneca, De Ira 3.40). On the face of it, this seems like outlandish behavior, more like biting a dog than getting bit by one. We also learn about one reaction to his behavior: when Vedius was about to kill a slave whose only offense was to drop and break a glass, the emperor Augustus, who was present, prevented him, had all the glasses in the house broken, and the fish pond filled. On the one hand, all this suggests an egregious practice that evoked disapproval and punishment. On the other hand, even the loss of fine glassware and a fishpond seems like a light punishment for a person we would regard as a deranged serial murderer. The story ends up revealing the low value placed on the lives of slaves. More often, however, we lack the evidence to suggest how people judged an action, and it is far more difficult to make inferences from anecdotes.
Even when we have a substantial body of evidence about a practice at one time and place, it can be hard to decide whether things were similar for other contemporary slaves or for slaves living at other times. For example, classical Athenian references to slave families are rare. But manumission inscriptions from Hellenistic Delphi seem to reveal a world in which many slaves had familial ties of one sort or another, a topic we’ll explore in Chapter 7. Did slaves have better prospects of family life in the Hellenistic period? Was slavery at Delphi somehow special or different? Or, does this mean that slaves throughout Greek history had family ties, but that we only find good evidence of them in this one time and place? In this case, some limited form of the last option is probably correct, but we also need to consider the possibility of selection bias in the evidence from Delphi. Only slaves who gained their freedom left manumission inscriptions, on which family members would sometimes be mentioned. The life experiences of such slaves were probably better in many ways than those slaves who never managed to obtain liberty. Such favored slaves may have been atypical in their ability to form and keep family ties just as they were atypical in their ability to buy their freedom.
To generalize from one time and place to another presents more obvious problems as well. Modern historians often emphasize local variations within the system of slavery: slavery in Virginia was not the same as slavery in the deep South; conditions on neighboring plantations varied greatly depending in part on the personalities involved. The experience of ancient Greek slaves is no more likely to have been the same at all times and places. The case for typicality is often worth making, especially if there is no evidence to the contrary. But one must remain cautious about what we really know and what we merely suspect might be the case: that similar practices took place at other times and places.
Historians often make inferences such as the following: if slaves ran away whenever they could, they must have hated slavery. The starting point for drawing such a conclusion seems to be common sense: people flee from what they hate and not from what they enjoy. Such reasoning is fine for some purposes; slave flight on the whole does indicate dissatisfaction with slavery. But an unstated assumption behind arguments from common sense is that all people think and act alike, and this is not the case. For example, most modern religions have a strong moral element that dictates how people should behave and treat one another: the Bible contains not only narratives of God and humans but also the Ten Commandments. It might therefore seem a matter of “common sense” to ask whether ancient Greek religion seemed to condemn or to justify slavery. But ancient Greek religion was more a matter of ritual and mythology – including stories in which gods behave badly – and less a matter of moral rules. Particular rituals and festivals might be open to slaves or not, but Greek religion had no position on slavery and we would be led astray by “common sense” if we tried to find one. For this reason, careful historians try to learn as much as possible about the practices, feelings, and thoughts of ancient Greeks and Romans and their slaves, so that their inferences are based on ancient habits of thought and action rather than imposing a “common sense” that may be modern and thus inappropriate. Our understanding will never be as intuitive or as dependable as our common sense about the times and milieu we actually live in, but it can allow plausible conclusions to be drawn.
One problem with this procedure is that we often do not have enough information about ancient habits of thought and action, especially when it comes to slaves, who were often foreigners and cannot be assumed to resemble closely their Greek or Roman masters. As a result, historians of ancient slavery sometimes turn to comparative history.4 The basic assumption of this method is that slave societies are similar, and so what holds for one should hold for another. Comparative historians apply what we know about the better known slave societies of the modern world to the poorly known institutions of ancient slavery. For example, tens of thousands of epitaphs from Roman ex-slaves have survived. Historians used to infer from this that most Roman slaves gained their freedom. But nineteenth-century Brazil also possessed a vibrant and conspicuous urban population of ex-slaves. And, in the case of Brazil, historians have the actual statistics, which show that, while some skilled, urban, or household slaves had opportunities to obtain and save money and eventually buy their freedom, most agricultural or mine slaves, the vast majority of the slave population, had little chance of manumission (Garrigus 2011, 237). It is thus certain that conspicuous evidence of ex-slaves does not prove a high overall rate of manumission. Our evidence for ancient slaves regularly obtaining their freedom may only apply to a subset of the slave population.
The comparative method is far from foolproof, as you may have guessed from its dubious basic assumption. It may be more decisive in undermining than in establishing facts – as in the case of manumission rates. Often, when historians know something about both an ancient and a modern system of slavery, they find that the two are different: the existence of states without slaves and of abolitionists in the modern world provides an obvious contrast with the ancient world. Consequently, it is risky to fill in the blank areas in our map of ancient slavery by drawing on parallels from modern slavery.
But this is not the only way to use comparative history. Admittedly, it is difficult to prove something by the comparative method; the objection, “but ancient slavery might not have been like modern slavery in that respect,” is always possible and often persuasive. But, after all, other historical methods do not always result in indisputable conclusions. And almost all scholars of ancient slavery find it worth their while also to learn about other systems of slavery; often the benefit is simply to stimulate new ways of looking at things and new questions to ask. For instance, the retention of African identity by slaves has long been a focus of scholarship on New World slavery (Childs 2011, 171–173). Historians and archaeologists studying ancient Greek slavery, who knew that slaves in Athens were likewise imported from other cultures, were thus prompted to ask, “Can we tell whether slaves at Athens too retained any sense of their cultures of origins?” – a question we’ll explore in Chapter 6. Even when comparative history does not provide a conclusive answer, it can suggest new questions to ask.
Finally, it is always better to make an explicit comparison than to let an unstated comparison sneak in – sometimes under the guise of “common sense.” For example, one might unreflectively assume that Roman ex-slaves and their descendants would be subject to racial prejudice. This would be an unconscious use of the comparative method based on our knowledge that ex-slaves and their descendants in the United States were and are subject to racial prejudice. In fact, Roman attitudes towards ex-slaves, for example, could be hostile and were tinged with ethnic chauvinism. They were not linked to skin color nor were they as systematic as modern racism. As a result, ex-slaves and their children had a much easier time joining the mainstream of Roman society, especially in cities. And occasionally rich and powerful freedmen or their children and grandchildren were even able to join the Roman elite, as we’ll see in Chapters 5 and 8.
Although parallels with modern slave societies can stimulate new areas of inquiry and new approaches to our ancient evidence, modern concerns sometimes distort the interpretation of ancient slavery. In part this is because the evidence about ancient slavery is often ambiguous; the choice of one from several possible interpretations of this evidence may depend on which is more appealing in a modern political or social context.
For example, the study of ancient slavery first became a focus of intense modern interest in the context of debates about the abolition of the slave trade in the early nineteenth century. Pro-slavery writers could point to the prestigious high culture of the classical world and argue that the development of this culture depended on slavery. Anti-slavery writers argued that the alienation of the elite from practical matters stunted technological innovation and led to the fall of Rome. Or they claimed that the moral corruption attendant on this brutal system of oppression weakened the Roman Empire and left it vulnerable to invasion from outside. The attitude of the Christian church to Roman slavery became a particular bone of contention. Some abolitionist scholars tried to claim that the conversion of Rome to Christianity necessarily meant the end of classical slavery and therefore that slavery was intrinsically un-Christian. Christians today feel strongly that this last statement is manifestly true and current sympathies are, of course, with the abolitionists. Unfortunately, their historical claim about Christianity is untenable: centuries of Christian accommodation with slavery need to be explained away, no Christian abolitionist movement in antiquity can be detected, and no evidence exists that the decline of classical slavery was due to anything other than economic and social factors, as we’ll examine in Chapter 12 and 13. In fact, in parts of medieval Europe, “through the accumulation of power and real estate, the church found itself in possession of large numbers of slaves,” which it had no intention of letting go (Phillips 1985, 49).
The admirable qualities of Greek and Roman culture and their importance to Western civilization led to another distortion. Classical culture is liable to the charge, mainly accurate, of being based on slavery. Rather than saying that we admire the cultures of ancient Greece and Rome, but we deplore their use of slaves, some scholars have tried to palliate ancient slavery by exaggerating the differences between ancient and modern slavery. As we shall see, ancient slavery was plenty brutal: there is little to choose between an Athenian silver mine and a Brazilian sugar plantation. And, though not racist in the modern sense, ancient slavery was often justified by a sense of ethnic superiority.
It is easier to see the flaws in older historians’ viewpoints than your own biases, but historians today are also unlikely to be immune to the influence of politics and ideology. We probably also prefer topics or angles of argumentation that resonate with present political concerns. For example, historians today are acutely sensitive to multiculturalism and perhaps too eager to find it in the ancient world, an outgrowth of the topic’s political relevance. It is not possible to attain perfect objectivity and to escape the various influences of your own political views, and your cultural and social milieu. Indeed, it can be hard enough just to notice them. Despite all this, careful historians strive to and often succeed in writing accounts of history that are accurate and convincing regardless of where, when, and by whom they are read.
In addition to his influential redefinition of slavery, Patterson 1982 draws on evidence from a wide range of societies to explore key issues in the sociology and even the philosophy of slavery. For contrasting appraisals of his impact on the study of ancient Greek slavery, see Hunt 2016b and Lewis 2016. Vlassopoulos 2016b argues against “essentialist” conceptions of slavery altogether. Lenski and Cameron 2018 collect essays evaluating or applying the concept of a slave society. Some contributors hold that the concept itself is flawed while others argue that it needs to be applied more widely, that there were more slave societies in world history than the canonical five. David Lewis presents strong arguments that in the ancient Mediterranean Carthage too was a slave society (forthcoming, Chapter 13). Hopkins 1993 and Webster 2008 examine methodological issues in the study of classical slavery. Scott 1990, a brilliant work of comparative history, explores the systematic biases in our evidence that complicate historians’ understanding of slaves and other groups subject to intimidation and consequently unable to express their true opinions publicly and directly. McKeown 2007 focuses on several key controversies in the study of ancient slavery. He argues that the approach scholars take is determined in part by the traditions of the discipline of history in their countries and by varied ideological commitments.