The Roman Empire developed one of the largest and most economically and culturally integrated systems of slavery in world history. It thrived on a remarkably robust supply stream that included enslavement by birth, capture, sale from foreign and domestic sources, the reclaiming of exposed infants, and—in late antiquity—self-sale, child sale, and debt bondage. Enslavement was imposed upon people from all regions, inside and outside the empire, and was never inflicted exclusively on a particular racial or ethnic group. Those enslaved to Rome worked in agriculture, industry, service, and even knowledge production, allowing them to be the primary workforce behind the generation of elite wealth. Escape from slavery could at times involve resistance, including everything from open revolt to flight, but Roman society was also remarkably generous with manumission. This and many other features reflect a hybridity between ancient patterns of captive integration and modern habits of slave exclusion.

Entry into Slavery

Roman slaving was practiced in a Mediterranean context where captive taking and slaveholding were nearly universal. From the Semitic and Greek peoples that populated the East, to the Germanic and Celtic peoples in the North and West, to the Phoenicians, Africans, and Egyptians in the South, slaving was an ancient and entrenched tradition before, during, and after the flourishing of the Roman Empire between the third century BCE and the fifth CE. Aware of this, the Romans enshrined it in the legal principles by which they structured their own practices. When Roman jurists described the legal basis on which an individual might enter slavery, they did so with reference to a bipartite division between the “law of all peoples” (ius gentium) and their own “civil law” (ius civile). The former included two of the most common ways to enter servile status, birth to a slave and captivity to the Romans in enemy combat. For the Romans, birth into slave status depended entirely on one’s mother, for the Romans followed the principle of partus ventrem sequitur (offspring follow the womb): an enslaved mother bore children who were slaves, while a free woman birthed free children, regardless of the status of the father. We have no quantifiable data to help us determine the relative importance of various channels into the Roman slave supply stream, but modern historians agree that birth was the biggest.1

Nevertheless, captivity by all means played a crucial role in bolstering the enslaved population. Massive captive-taking events are recorded regularly in the sources, many of them so sizeable that they temporarily crashed prices on the slave market. This was largely a result of Rome’s repeated successes in large-scale combat against foreign enemies and the ruthless efficiency with which Roman armies disposed of enslaved persons into the market. Julius Caesar, one of Rome’s greatest conquerors, is said to have enslaved 1,000,000 Gauls in his sweeping campaigns in the territory of what is today France between 58 and 50 BCE; on a single day in 57 he sold 53,000 members of the Atuatuci tribe to slave dealers; and when he achieved victory at the Battle of Alesia in 52, he distributed one Gallic captive as plunder to each of his 80,000 soldiers.2

Indeed, the Roman Empire attracted captives from all across the Mediterranean—wherever Rome’s armies went, captives were generated for its markets. These could come in waves as individual territories or polities were integrated into the imperial machine: when Tiberius Sempronius Gracchus defeated the rebellious inhabitants of Sardinia in 177 BCE, he brought 80,000 captives into the markets, equal to as much as a ten percent of the Italian slave population at the time; Aemilius Paulus nearly doubled that number when he defeated the peoples of Epirus in north western Greece, bringing in 150,000 more.3 The trend continued as Rome shifted from a Republican democracy to an autocracy under the rule of an emperor in 31 BCE. To take just two examples, when Titus sacked Jerusalem in 70 CE, he carried off 97,000 Jews into slavery; and Trajan’s campaigns against the Dacians in what is today Romania probably generated at least 100,000 captives.4

Nor were all these captives put to work in menial labor, for the Romans were happy to employ victims of enslavement with skills in their respective fields of expertise. These included craftsmen, but also those trained in service professions such as medicine, architecture, or music. According to Pliny the Elder, the satyrist Publilius Syrus, the astronomer Manilius, and the grammarian Staberius Eros, each of whom had an important impact on Roman intellectual history in the first century BCE, all arrived in Rome on the same ship.5 If younger captives stood out for their aptitude, these might be trained in a profession rather than being sent directly to the fields. Two captive boys from the eastern kingdom of Parthia have left inscriptions confirming that one was trained as a treasurer, the other as a pedagogue to the imperial family.6

All of which is to say that Rome used slaving as part of a larger imperial apparatus for the integration of conquered peoples into its world system. As it expanded its territorial reach, it generated first captives from the territories it conquered and only later, through techniques of domination and pacification, provincial subjects. And neither provincials nor captives were entirely banished from the levers of power. They were instead absorbed—together with their skills and to some degree their customs. Roman captive-taking practices offer only one of the many examples we will see in this essay of the way in which Rome’s slave system sits squarely between the fluid and integrative captive-taking systems that predominate in earlier, less complex societies and the more compartmentalized systems of permanent racial slavery that characterize the highly stratified and bureaucratized societies of modernity.

If birth into and captivity constituted the two primary ways into slavery through the ius gentium, the Roman ius civile was also remarkably efficient at enslaving people or perpetuating their enslavement. The most obvious example is sale, which allowed one master to transfer the body of a previously enslaved person to the ownership claims of another. We have over sixty slave sale contracts preserved on papyri (ancient Egyptian paper) and other writing materials, such as wood and leather, to serve as concrete evidence of such transfers.7 Because most of these documents were found in Egypt, where Greek was the official language of the eastern Empire, most are written in Greek, but we also have sale contracts in Latin and Syriac from find spots stretching from Britain to the Euphrates River.8 These generally list the ethnic origin of a slave and thereby confirm the staggering variety of regions from which Roman slave traders drew—literally every corner to which the empire stretched and also well beyond it. The most common and indeed the preferred origin was, however, the “homebred slave” (Greek oikogenēs; Latin verna).9 Birth into slavery generally rendered a person more docile and manageable than foreigners or captives because homebred slaves were acculturated in local customs and languages and had never known any life outside of enslavement. All of this means that the “slaving zones” that characterize modern slavery did not exist for the Romans. The reach of their enslavements was coexistensive with their world.

Insofar as the Roman Empire favored certain regions to supply it with slaves, these shifted over the seven centuries during which it remained at its zenith. Rome’s expansion was always limited by military, financial, and above all technological constraints, which meant that the Empire was surrounded by ethnic and political others (“barbarians”), whom the Romans treated as a potential source of slaves. In the early first century BCE, when the Romans had conquered only a small slice of southern Gaul, they are said to have traded for slaves with the Celtic peoples who controlled the remaining Gallic territory at the rate of one amphora of wine per human.10 After Caesar’s conquest of Gaul, these same Celtic people became Roman subjects, such that they were sending their own aristocrats to Rome to serve as senators by the mid-first century CE. The Roman frontier had simply expanded and continued to do so into the second century CE, allowing the Empire to draw on new peoples beyond its frontiers to supply it with slaves, usually by human trafficking among members of their own or neighboring tribes or polities: Scotti (Irish) and Picti (Scots) in Britain; Frisians and Alamanni along the Rhine; Sarmatians and Goths on the Danube; Armenians and Arabs on the eastern frontier; Nubians and Gaetulians in Africa. These were transported, usually on foot, shackled together at the neck or ankle, over the hundreds or even thousands of disorienting miles to their new life as slaves to a foreign empire (Fig. 5.1). We also have textual and archaeological evidence for slave markets where they were resold. The city of Rome had markets in the Saepta Julia and near the Temple of Castor, but markets are also attested in most major cities of the Empire. The Aegean island of Delos—to which Rome had granted free-port status—became the marketplace par excellence, leaving us abundant textual and archaeological testimonies to its role as a trafficker of some 10,000 humans per day.11

Fig. 5.1
Roman gravestone with relief from Nickenich, Germany. The sculpture exhibits the relief of a Roman trader of enslaved individuals.

Grabstein aus Nickenich bei Mayen (etwa 50 n. Chr.): Das Relief zeigt einen Sklavenhändler, der zwei Sklaven an einer Kette führt, um sie zu verkaufen. Es ist im LVR-Landesmuseum in Bonn zu sehen. In regard to the use of pictorial material: use of such material in this press release is remuneration-free, provided the source is named. The material may be used only in connection with the contents of this press release. For pictures of higher resolution or inquiries for any further use, please contact the Press office publishing this directly.

The Romans also felt comfortable enslaving at least some of those whom they regarded as provincial subjects. Here it is crucial to note that, as the Empire expanded and incorporated new territories, it did not instantly grant all subjects Roman citizenship. This was a privilege reserved for assimilated members of the local elite, while others were offered more demanding pathways to citizenship, such as service in the Roman army. This meant that, up to the year 212 CE, when the emperor Caracalla issued a sweeping law granting citizenship to most of Rome’s provincials, less than twenty percent of Rome’s subjects had attained citizenship. In the centuries before this, Rome allowed its non-citizen subjects to follow their own regional legal systems, and many of these permitted enslavement from within their own culture—especially through the sale of children or debt servitude, neither of which were permitted to Roman citizens. Once enslaved, such provincials could be trafficked throughout the empire.

Furthermore, although the Romans of the High Empire never legally permitted the full enslavement of those born with citizenship, de facto they allowed the enslavement of unwanted infants who were exposed by their parents after birth. In the absence of effective birth control and safe abortion, the Romans (whether citizens or provincials) dealt with unwanted pregnancies by carrying a fetus to term, then exposing the newborn—usually on town garbage heaps. Knowing this, profit-minded enslavers kept watch and snatched up such infants to raise to four or five years, then sell as slaves. We know this from multiple sources, but especially the more than forty surviving wetnurse contracts, which were written to pay lactating women to feed exposed infants for the first year or two of life and then surrender them as toddlers back to the enslaver so he could sell them. We have no statistical data, but it appears that child abandonment provided a major stream in the slave supply. Even the emperor Claudius ordered a child born to his wife Messalina exposed because he believed it to have been conceived by another man.12

Finally, it should be emphasized that the Roman Empire was incredibly long-lived and that its customs and laws changed over time. In the flourish years from the second century BCE till the second CE, legal protections for citizens were strong and prevented most from falling into slavery. But up to 326 BCE, the Romans had been happy to allow citizens to fall into debt bondage to one another, and this practice became common once again by the end of the fourth CE. Then too, the introduction of Christianity as the religion of the emperor after Constantine’s conversion in 312 CE caused a shift in attitudes toward infant exposure and child sale. Prior to this, the Romans had left exposure unregulated but forbade the sale of recognized citizen children. Constantine began a process that reversed this formula, forbidding exposure but allowing parents to sell their children into long-term indenture—effectively servitude.13 This he did to protect parents from breaking the commandment against killing (exposure often led to death) and to protect children from being raised into a potential life of immorality, especially prostitution. So too the practice of punishing those who committed major crimes by enslaving them to the emperor (servitus poenae) was normal for the first five centuries but was eliminated in the sixth century by emperor Justinian, who based his decision on Christian concerns—because penal enslavement had the effect of terminating marriages.14

The practice of enslavement thus changed over time, adapting to shifting territorial, legal, political, and religious landscapes. But the Roman slave supply was always fortified with multiple streams which ensured that, even as the circumstances and attitudes that governed slaveholding shifted, supply always met demand. This fact is confirmed by the stability of documented slave sale prices throughout the imperial centuries, usually amounting to the equivalent of about five years of wages for a day laborer.15

Experiences of Enslavement—Labor Extraction

Historians and sociologists have long debated what constitutes “slavery”—how might it be defined? No consensus has yet been achieved, but two schools have circled around the definition provided by the 1926 League of Nations Charter (“Slavery is the status or condition of a person over whom any or all the powers attaching to the right of ownership are exercised”) and that first developed by Orlando Patterson in his sweeping 1982 monograph Slavery and Social Death (“[Slavery is] the permanent, violent domination of natally alienated and generally dishonored persons”). I have argued that the fullest definition would combine both, for the former emphasizes the claims and commitments of the master class while the latter conveys a fuller understanding of the experience of the enslaved.16 Only in the dialectical space between these two actors—enslaver and enslaved—can the dynamic of slavery be understood.

With this as background, we can turn to experiences of enslavement as it was practiced by slaveholders, but also experienced by the enslaved. As we do so, we must once again emphasize that Roman society functioned in a way that simultaneously shared elements of antiquity and modernity. Like some ancient societies, Roman society was at times ready to overlook the productive labor potential of the enslaved in order to capitalize on the numinous or entertainment value of their destruction. Into the middle Republican period, the Romans occasionally used captives for human sacrifice (usually by live burial) and they regularly forced enslaved persons to fight to the death in the arena in spectacles that originated as a kind of funerary rite but took on the primary role of entertainment by the mid-first century BCE. This pure wastage of human life and labor never fully ceased even in the later empire, for a selection of enemy captives was still regularly culled for spectacular execution in public arenas.

Yet the Romans also developed a remarkably complex economic system that bore many of the hallmarks of modernity. The Roman economy was highly monetized; it was supported by a sophisticated institutional framework at the level of commerce (banks and credit) and law (contracts and corporations); it developed specialized agricultural markets in trade commodities (wine and olive oil); it mass produced manufactured goods (tableware, metal work, oil lamps); it supported specialized service industries (transportation, entertainment, sex work); complex supply networks (capable of maintaining over a thousand cities); complex labor differentiation and social stratification; and remarkably developed engineering and technology (paved roads, aqueducts, watermills). The combination of these features allowed the Roman empire to compete with early modern west European economies in terms of GDP and demographic growth, which also meant that its slave system could operate with many of the same complexities of modern Atlantic slave systems—in fact, arguably with more, for the openness of the Romans to the enslavement of all peoples and races and their readiness to deploy slaves in nearly every economic sector allowed the Romans to develop what was arguably the most socially integrated slaving system in world history.

First and foremost, slave labor was deployed in primary industries, food production. This sector commandeered as much as ninety percent of the labor market, and slaves constituted the major force of production for members of the elite in the high Roman centuries. Thus, while subsistence farming continued among the free population throughout the Roman period, slaves represented the major producers of marketable surplus for elite landholders between the second century BCE and the second century CE. This we know above all from the four agricultural manuals surviving from the Empire, those of Cato (c. 160 BCE), Varro (c. 35 BCE), Columella (c. 60 CE), and Palladius (c. 470 CE). The first three are filled with advice on the management of slave labor on the farmstead, where slaves were organized into teams and placed under a manager (vilicus), who was himself usually a slave. All three manuals emphasize the need to keep enslaved workers busy throughout the year, but they focus above all on viticulture, which was particularly capital and labor intensive. Cato’s manual also lays out the size of rations permitted to slaves: four and a half modii of wheat per month in summer as well as some wine and olive relish for an average of c. 3500 calories per day. Cato also recommends each slave be given a new tunic, a blanket, and pair of wooden shoes every other year.17 These were, in other words, the barest of living conditions. Animal herding (sheep, goats, cattle) was also labor intensive and was thus regularly assigned to slaves. Herding slaves were, of course, given freedom of movement, and because of the danger of rustlers and predators, they were generally armed. This added to slave agency without generally threatening the system.

We can catch a glimpse of the lives of agricultural slaves from the archaeology. Members of the elite tended to hold land as a collection of geographically diverse estates, each equipped with a large multi-purpose villa structure. These were divided into large living areas for the owners, stalls for animals, functional rooms for food processing and storage, and a series of much smaller quarters (cellae) measuring 4–9 square meters for the slaves—probably one cella per slave family. Some, like Villa 34 at Gragnano, also had metal hooks affixed to interior walls where some slaves could be chained during sleeping hours. We learn from written sources that many such villae were also equipped with prison cells (ergastula) in which to pen slave workers, although archaeological examples have not been found.

It is important to bear in mind that enslaved persons were never the only labor source for any given economic sector. The size of the enslaved labor force varied over time but also geographically, with high-imperial Italy being the period and place of most intensive slave use. Estimates of the percentage of enslaved persons on the peninsula in this period have ranged as high as 35 percent, but most recently these have been revised downward to 15–25 percent (about 1–1.5 million persons from a total of 6–7 million).18 The only part of the empire where we can begin to assemble firmer numbers is Egypt, where it has been shown from extant census declarations that about 11 percent of the population was enslaved.19 This was perhaps representative of other imperial provinces, which, as mentioned earlier, tended to structure their societies and economies as they had before the Romans arrived. Thus, in Asia Minor (modern Turkey) or Syria, where elites had tended to work their estates using free (or semi-dependent) tenants in Hellenistic times, this type of production prevailed. The same held of North Africa, which has been shown always to have relied more heavily on free tenant labor rather than slaves throughout the Roman period.20 This was also true because of variability in crop production in different regions. North Africa, for example, was climatically suited to oleiculture, which tended to be less labor intensive and relied mostly on the development of new tree stocks, which took more time than effort and was thus best suited to free tenants. Similarly, cereal cultivation was demanding, but only in the seasons of plowing and harvesting. This meant that grain crops could be put in by relatively small teams of enslaved plowmen and farmers, then taken off by hired teams of harvesters each fall. Enslaved persons were thus one source of labor for elite surplus production, surely the most important one in High Imperial Italy, but not necessarily so in other places or periods. Already in the second century CE we learn of grand Italian landholders using a mix of tenant and enslaved laborers, and by the fourth century the imperial government tilted the balance in favor of the latter. Beginning in the 320s, the emperor Constantine permitted landowners to bind their tenants (coloni) to the plots of land on which they had been born. This may initially have represented an effort to protect tax revenues by eliminating the mobility of taxpayers, but over the course of the fourth century the “bound colonate” came to represent a third status between free and slave and to supplant slave labor on larger estates.21

Enslaved persons were also employed in secondary industries, the manufacture of durable and consumable goods. These included trade commodities such as metal wares, clay lamps, and table wares. The last is particularly well attested archaeologically, allowing us to trace both shifts in the centers of production and networks of distribution. Much of the Mediterranean in Roman times preferred to eat and drink from “red slipware” vessels, also known as terra sigillata. These were mass produced to high-standards using clay molds and then transported throughout the Mediterranean basin via water transport on seas or rivers. Production was initially concentrated in Italy, then moved to Gaul in the late first century CE, and shifted again in the third to North Africa. Workshops employed up to sixty slaves each, but normally tended to function with ten to twenty. The material excavated from La Grafesenque in Gaul is particularly well studied and offers rich testimony to the potters involved, who proudly signed their wares.22 Yet the material from La Graufesenque makes clear that free laborers were also involved in pottery production. Indeed, while some industries relied heavily on slave labor, mining for example, others appear to have been predominated by free workers, especially building construction.23 Enslaved persons were also heavily involved in the processing and production of consumables. Baking offers rich testimony. Because watermills only came into use in the second century CE, grain processing was highly labor intensive, involving the turning of heavy horizontally rotating stone mills that required tremendous effort on the part of animals or humans. The second century BCE comic playwright Plautus often has enslaved characters lament the possibility of being relegated to a bakery, and the second century CE novelist Apuleius famously depicts a troop of emaciated slaves pushing a grain mill with their whip-scarred bodies.24 But we also know that a fair number of enslaved bakers eventually attained freedom, and some went on to become owners and managers of their own bakeries and to amass giant fortunes. The star example is Marcus Vergilius Eurysaces, a freedman of first century BCE Rome who gained enough wealth as a baker to construct a gigantic (10 meters tall) monument for himself just outside the city gates fashioned in the shape of a bread oven.25 Baking was thus quintessentially Roman in its approach to labor, for it entailed the violent professional induction of enslaved laborers, some of whom then escaped from their enslavement to enrich themselves using the professional skills they had learned.

Slaves were also involved in “tertiary” or service industries. Elite self-construction in the Roman world hinged upon the control of enslaved bodies that could be deployed to assist, accompany, curate, and enhance the lives and prestige of slave owners. Many of the tasks moderns accomplish with technology were more labor intensive in antiquity, which meant that those with ambitions to “live well” chose to exploit enslaved persons to do so. Elite houses had enslaved cooks, housekeepers, chamber servants, doorkeepers, watchmen, table servants, silver polishers, lamplighters, etc. This we know from their funerary inscriptions, sometimes preserved in the family columbaria of Rome’s super elite. These include the empress Livia (59 BCE–29 CE), whose cadre of enslaved workers included at least 49 discreet kinds of operatives, among them hairdressers (ornatrices) and a pearl-keeper (margaritarius).26 Enslaved personnel also shouldered the burdens of child-rearing as wetnurses (nutrices), nannies (tattae), and educators (paedagogi).27 Household slaves also played an important role in the daily rituals of the elite. Each time they went to the public baths, wealthy slaveholders would bring a large entourage of slaves who could carry their bathing implements, or wash and massage them. On such trips the slaveowner regularly traveled in a litter (lectica) born upon the shoulders of as many as eight specialized carrier slaves (lecticarii).

Finally, in contrast with most modern slave systems, the Roman Empire regularly employed slaves in quaternary industries, that is to say knowledge production and curation. Slaves served regularly as doctors, midwives, architects, astrologers, secretaries, accountants, property managers, business managers, ship captains, etc. Some of this arose from the reality, already discussed, that the Romans gladly capitalized on the skills and training of enslaved persons in any economic sector. It also arose from the fact that Roman law—like most premodern legal systems—had difficulty conceptualizing the assignment of agency to third party actors: a person was responsible for their own dealings, and these responsibilities were not easily transferred to an agent in ways that left the principal in an organizational chain legally liable. But a workaround was provided by enslaved persons, who were considered an extension of their master’s legal personality.28 The same legal fiction also led to the deployment of an imperial bureaucracy staffed by the emperor’s own slaves (servi Caesaris), who performed most managerial and secretarial functions of state for the first century of the Empire before being replaced by freeborn bureaucrats.29 The same was true of individual cities, including Rome itself, which owned their own teams of slaves (servi publici) to perform tasks ranging from street sweeping and aqueduct maintenance to bookkeeping and accounting.30

Particularly prized among household and imperial staffs were eunuchs, males castrated in childhood to serve as household attendants. Castration itself was distasteful to the Romans such that more than one emperor sought to forbid it by law, but the Roman appetite for eunuch servants meant that the practice continued inside the empire, and where supplies fell short, eunuchs were imported from foreign markets.31 Eunuchs were especially prized as chamber servants since they could not pose a threat to the reproductive capacities of the household.

As this discussion indicates, the sexuality of enslaved persons was always a primary concern and will be discussed in greater detail below. In regard to labor output, however, enslaved persons were also regularly deployed as prostitutes. Evidence for prostitution abounds in all periods of Roman history, and particularly from the first century BCE through the third CE. The most vivid testimony comes from the city of Pompeii which had an estimated 35 brothels to service a town of 10,000 persons.32 The use of female sex workers by male clients was greatly encouraged by Roman marriage customs, which tended to favor early first marriage for females (ages 12–18) and later for males (ages 20–25). Male surplus sexual energy in the years of adolescence and early adulthood was thus channeled to sex workers, and the problem was only exacerbated by marriage regulations introduced by the emperor Augustus, which strictly punished illicit extra-marital sex. But there were no legal consequences for males fornicating with prostitutes, whether free or enslaved, and the profits to be made led masters regularly to exploit the bodies of their enslaved females—and males—as sources of income.33

Experiences of Enslavement—Violent Domination

If the discussion thus far has focused on the use of slave bodies to perform labor services, we must still explore the question of the experience of enslavement by those who endured it as a regime of physical, social, and psychological repression. This problem was omnipresent, for even if some of the enslaved persons discussed above were at times able to escape their fetters and sometimes to benefit from the training or status imparted to them while enslaved, there was never a slave in the Roman Empire who did not experience slavery as a relationship of violent domination, natal alienation, and general dishonor. The final example in the previous section offers excellent proof of this concept. While some enslaved female sex workers gained fame as professionals and were even rewarded with freedom and wealth, they were never able to escape the anguish of a life of serial assault and the stigma of enforced bodily exploitation imposed on them by the master class.

Indeed, sexual assault was a regular experience for Roman slaves, both male and female, whether or not they were exploited in the sex trade. Owners of enslaved persons could and did have sex with them as they wished with no legal consequences. Romans of the means were thus less likely to use public prostitutes than simply to purchase sex slaves for their own exclusive exploitation. This had the consequence that male masters fairly commonly freed and married their slaves, a phenomenon well attested in funerary epigraphy and one that further points to the integrative role enslavement sometimes played.34 Sex between female masters and male slaves was generally stigmatized, though it too is attested, and some classes of elite male slaves, especially slaves owned by the emperor, are known to have been sought as marriage partners by free women. Sexual abuse was also common with same sex partners, particularly male masters who sexually abused their male slaves. This was especially common with male youths, who were often expected to grow their hair long and depilate body hair, sometimes well after the onset of puberty, in order to be ogled and sexually assaulted by their master and his friends (Fig. 5.2). The slave body, male and female, was thus a target for masters, who feared no consequences for what would today be considered felony behavior.

Fig. 5.2
The bronze statue of a nude ephebe from Xanten. The lower part of the right hand of the boy in the statue is missing.

Bronze image of a nude ephebe from Xanten. The boy, who would have carried an actual tray, is shown long-haired and garlanded with his nude body right at the end of prepubescence. Berlin, Staatliche Museen, Antikensammlung, Sk. 4. Bildarchiv Preussischer Kulturbesitz/Art Resource, NY

The bodies of enslaved persons suffered domination not just as objects of sexual exploitation but also in myriad other ways. The most basic of these was hunger, for food was always used as a tool of control. We have already seen that Cato prescribed set rations for his agricultural slaves, and other sources report that starvation was used to punish uncooperative slaves.35 The sources also confirm that the food given to slaves was distinctly inferior to that eaten by their masters (coarse bread, sour wine, table scraps). The hyper-frugal Cato recommended selling older slaves or otherwise getting them off the books since they could no longer earn their keep with their labor power, and some owners are known to have abandoned their slaves on the Tiber Island if they regarded them as too sick to survive without the expense of a doctor’s visit—leading the emperor Claudius to allow any such abandoned slave to go free if they survived their illness and abandonment.36 Slaves who were inclined to flee or whose behavior threatened or displeased a master could be bound with shackles and chains, a regular feature of the archaeological record. And those who fled habitually were often tattooed on the face so they could be identified at a glance and returned. After Constantine forbade this practice, the Romans turned instead to slave collars, permanently bonded to the neck with dog-tags reading “Hold me for I am in flight.”37

Above all, however, slave bodies were tortured and physically abused, even unto death, with no consequences for masters. Plautus’ second century BCE plays regularly feature slaves terrified over an impending whipping, a trope that was meant to elicit laughs from the audience. Similarly disturbing insouciance about physical abuse is found in the epigrams of the first century CE poet Martial: “You think me cruel and too fond of my stomach, Rusticus, because I beat my [enslaved] cook on account of a dinner. If that seems to you a trivial reason for lashes, for what reason then do you want a cook to be flogged?”38 And assaults were often much worse than a beating. The physician Galen speaks of his experience of masters, including his own mother, biting their slaves or gouging out their eye with a writing stylus.39 Ultimately, the master could even kill his slaves with impunity. This he sometimes did by contract, especially through the brutal punishment of crucifixion. An inscription of Puteoli (modern Pozzuoli) lays out prices set by a company that specialized in torturing and crucifying slaves on contract, allowing the master to hire out this messy and physically demanding affair to specialized professionals.40 Here again Constantine became uneasy with this level of violence and issued a law forbidding the deliberate killing of slaves in 319 CE, but in a subsequent law he granted tremendous leeway for masters who happened to kill a slave in the course of “corrective punishment.”41

Even when slaves were not openly abused, they lived in constant fear of violence. They also lived in a world of “natal alienation,” which meant that they were permanent outsiders, excluded from civic or political rights and privileges, excluded from control over their own birth families and offspring, and excluded from final control over their very bodies and personhood. Their names could be assigned to them by a master and could be changed at any time, particularly when they were sold to a new master. Their children could be exposed or sold by their master at will. And they themselves could be liquidated for their cash value at any moment. We have evidence of this process from multiple sources which reveal enslaved persons intended for sale were usually stripped down to a loincloth, displayed on a raised platform (catasta), made to wear a garland if they were war captives and/or marked with chalk on their feet if they were imported from overseas, their “defects” (disabilities, diseases, habits) were publicly proclaimed on placards hung round their necks, and they were subject to humiliating physical inspections by potential buyers (Fig. 5.3).42 They were, in other words, treated in the manner of livestock at market, with all of the attendant dehumanization and degradation.

Fig. 5.3
The tombstone of Capua exhibits slave trading in the Roman Empire.

Tombstone of Capua depicting the sale of an enslaved person, stripped and standing on a catasta with auctioneer (winged, to the right) and buyer, late first century CE. G. Fittschen Neg. D-DAI-Rom 1983VW1305)

Despite the repression inherent in this system, many enslaved persons managed to salvage a remarkable degree of agency in their own lives. Much of this occurred within the context of the paternalistic framework established by the master class, which offered a clear if not always reliable set of pathways to recover personal subjectivity. Roman masters considered slaves part of their familia, a word which included not just the nuclear family but also its enslaved dependents. As such the Romans associated their slaves with children, often referring to them as “boys” and “girls” (pueri, puellae), granting rights of agency (noted above) similar to those enjoyed by children, and offering both slaves and children the right to control some property in quasi-ownership—peculium, which was wealth over which they could dispose even if the father/master claimed bare ownership of it. This last in particular encouraged productivity, self-expression, and sometimes even freedom, albeit on terms ultimately controlled by the master. Here again, we see a system poised between the assimilative practices of captive-taking societies and the capitalistic practices of modern chattel slavery, for masters did allow slaves to amass fortunes large enough to buy their own freedom and even the freedom of their loved ones—spouses, children, parents, friends. And yet, even these freedpersons (liberti) were bound to offer obedience and often labor services to their former owner—their patronus, a word derived from the word for father. Within this framework, we find numerous instances of enslaved persons who embraced the profession assigned to them wholeheartedly and made it part of their identity, so much so that, apart from their name, their profession often constitutes their main or only identifier in the funerary inscriptions they have left.43 Slaves also formed family groups of their own which, while not recognized by law, represented a reality of their humanity that masters permitted and even encouraged.

But it is by no means the case that all slaves chose to operate within the guardrails established by the master class. This was in part because the actual living and working circumstances created for slaves only sometimes conformed to the ideal presented in the previous paragraph. Much depended on the temperament of any given master, but also on the structure of the work regime, which sometimes favored coercion, sometimes incentives. In less skilled professions, chief among them agriculture, coercion was the default, which meant that work regimes could be incredibly harsh and could foster resistance and even revolt. Three large-scale slave uprisings erupted in the late Roman Republic, the first two overrunning Sicily in 135–132 and 104–101 and the third rampaging through Italy in 73–71, the famous Spartacus revolt.44 All three necessitated the commissioning of full-scale Roman armies, thus showcasing the fact that Roman slavery was anything but a benign institution. Nor was revolt the only option for slaves to resist the hegemony of Roman slave society. Maroons of escaped slave brigands are attested across the Mediterranean, most famous among them that of Bulla Felix, who is reported to have sent a message to the emperor Septimius Severus, “Feed your slaves so that they may not turn to brigandage.”45 Roman slaves are also known to have murdered individual masters, a reality which led the Romans to issue a law ordering the execution of all slaves present in a household if any one of them murdered the master.46 At most times, however, slaves resisted their masters in quieter ways: shirking their work, wandering off for long periods, and damaging or stealing the master’s property.47

Roman slavery was thus fundamentally a system of domination and natal alienation, equipped with guardrails and safety valves, but never far from raw aggression. Chief among the mechanisms for relieving tension was Rome’s practice of relatively frequent manumission, but there were also structures of paternalism that granted enslaved persons some degree of agency in their lives even without achieving freedom. But the violence at the core of the system was always evident and made it such that Rome’s slave system could never be characterized as humane. It was the enslaved themselves who had to assert, uphold, defend, and recapture their humanity, at times through calculated cooperation, at others through resistance and open revolt.

Exits from Slavery

It is a sad reality that the most common exit from bondage for those enslaved to the Romans was death. For most this meant natural death, often due to overwork or abuse, but some enslaved persons chose to hasten death’s approach by taking their own lives. Seneca reports that one captive chose to avoid being forced to fight in the arena by suffocating himself with a toilet sponge and another by inserting his head into the spokes of the oxcart on which he was being transported.48 Another escape route from slavery was flight. This was common in Roman society and was assisted by the relatively undeveloped systems of communication available in antiquity. In the absence of printing presses and a public postal system (let alone wireless communication networks), it was more complicated for masters to track down escaped slaves, but this did not stop them from trying, often with the support of the state. We have, for example, a series of letters from the politician Cicero begging friends to help him find and return a person named Dionysius over whom he claimed ownership and who had fled his household, stealing a number of Cicero’s books on his way.49 In the imperial period, most cities had some sort of public police force one of whose major tasks was to track down and capture runaway slaves (fugitivi).50 We have already seen that tattooing and collaring were also common safeguards, but the collective weight of the evidence indicates that flight was common, and many surely managed to escape enslavement permanently.

The only legitimate means out of slavery was, however, manumission. We have already indicated that this was common in Roman society, especially when compared to other slave cultures, like those of the Atlantic world. Attempting to assign numbers and percentages is difficult given the nature of our evidence, but careful analysis of the sources (especially inscriptions) points to remarkably high rates of manumission. Perhaps more than 30 percent of urban slaves above the age of 25 could have expected to be given freedom, and many have argued the numbers were even higher.51

The openness of the system to manumission can also be measured in the variety of ways it could be achieved, ways which were adapted and shifted over the course of the Roman centuries. Manumission is perhaps best described as a formal ending to natal alienation and thus integration of the formerly enslaved into the enslaving society as a member with subjective rights. In early Roman society, this happened by literally enrolling a slave into the formal census roles and thereby rendering them a Roman citizen. Of itself this is remarkable since most slave societies never permit this level of integration, but here too we find Rome’s liminal position between earlier captive-taking societies, which were strongly inclined to the full integration of the formerly enslaved, and modern societies, which have tended to resist it. The Romans also permitted a state official or magistrate to manumit a slave in a formal ceremony called vindicta, which was based on transactions of sale and thus emphasized the property aspect of slavery, and which also conferred citizen status. Over time, the Romans came to practice “informal” modes of manumission such as freeing a slave before a group of friends or by a letter, attesting to a desire for streamlining of the formal processes—and for public display. But the emperor Augustus cracked down on informal modes in the first century BCE and imposed rules excluding informally manumitted persons from full citizenship. In the fourth century CE, a new mode arose, “manumission in church,” which was instituted by Constantine and also granted full citizen status but incorporated manumission into Christian ceremonial, thereby elevating it to the level of a “charitable act.”52

Masters were motivated to manumit slaves for a variety of reasons. These included social pressure, for Roman society had a stated “preference for freedom” (favor libertatis) which played itself out in law and practice. It was thus a matter of prestige to manumit slaves, particularly in one’s will—a practice so common that it too was restricted under Augustus, to no more than 100 slaves per decedent. Female slaves were often manumitted if they produced at least four viable offspring, thereby reproducing their labor power since all such children would have been born to slavery. Often loyal and productive slaves could expect manumission by a certain age, c. 25–30, although this was never guaranteed, and some made agreements with their masters to purchase their freedom using money they had saved in their peculium. In every instance these freedmen remained connected to their former masters through their family names, which were always those of the patronus¸ and they were also bound to their former enslavers in other ways. We have already seen that they owed the patronus obedience and labor services, and most also owed part of their inheritance. By the generation following, however, the children of freedmen had no further restrictions or obligations—even if they too retained the patronus’ family name. This meant that the children of freedmen could and did sometimes attain quite a lofty status—at least two children of freedmen became emperors, Macrinus and Diocletian.53

Exits from slavery were thus in part controlled by slaves, when these fled or otherwise resisted through revolt or suicide. More commonly, however, they were governed by the master class, which doled out manumission liberally but also in calculating fashion to serve as an incentive and a mechanism of control which could drive enslaved persons toward cooperation and eventually assimilation into their hegemonic system.


Roman slavery was thus a peculiar institution. Because Rome was an ancient society but also a remarkably precocious one, its slaving practices show traces both of much older patterns and of quite modern ones. Modern are Rome’s highly specialized and economically integrated uses of enslaved workers across labor sectors and its ability to convert slave power into money capital through rationalist investment and profit accumulation. Ancient are its tendencies to integrate enslaved persons into family and social structures and its readiness to assimilate the enslaved (almost) fully into the dominant culture. Pervasive throughout the system were the unmistakable hallmarks of all slave systems, violent domination, and natal alienation. These realities were crucial to the maintenance of Rome’s truly colossal slaving apparatus. Growing out of the Empire’s inherently militaristic ethos, Roman slaving flourished in an environment where violent power could be refined, controlled, and apportioned into quanta of domination that enmeshed not just the enslaved but also the enslavers in a nearly unbreakable social cage.


  1. 1.

    Walter Scheidel, “The Roman Slave Supply,” in The Cambridge World History of Slavery, vol. 2, eds. K.R. Bradley and Paul Cartledge (Cambridge: Cambridge University Press, 2011), 287–310.

  2. 2.

    Plutarch, Life of Caesar, 15; Caesar, Gallic Wars, 2.33; 7.89.

  3. 3.

    Livy 41.28.9; 45.34.3–7; Plutarch, Life of Aemilius, 29.4–5.

  4. 4.

    Josephus, Jewish Wars, 6.410; cf. Keith R. Bradley, “On Captives under the Principate,” Phoenix 58 (2004): 298–318.

  5. 5.

    Pliny, Natural Histories, 35.199–201.

  6. 6.

    H. Dessau, Inscriptiones Latinae Selectae, nos. 1836; 1980.

  7. 7.

    J.A. Straus, L’Achat et la vente des esclaves dans l’Égypte romaine: Contribution papyrologique à l’étude de l’esclavage dans une province orientale de l’empire romain (München: K.G. Saur, 2004).

  8. 8.

    Syriac: Dura Papyrus no. 28. Latin: S. Riccobono, Fontes Iuris Romani Antejustiniani, vol. III, nos. 87–89, 132, 134; R.S.O. Tomlin, “‘The Girl in Question’: A New Text from Roman London,” Britannia 34 (2003): 41–51.

  9. 9.

    Elisabeth Herrmann-Otto, Ex ancilla natus: Untersuchungen zu den “hausgeborenen” Sklaven und Sklavinnen im Westen des Römischen Kaiserreiches (Stuttgart: F. Steiner, 1994).

  10. 10.

    Diodorus Siculus, 5.26.2–3.

  11. 11.

    John Bodel, “Caveat Emptor: Towards a Study of Roman Slave-Traders,” Journal of Roman Archaeology 18 (2005): 181–95.

  12. 12.

    Wetnurse contracts: Mariadele Manca Masciadri and Orsolina Montevecchi, I contratti di baliatico (Milano: Tipolitografia Tibiletti, 1984). Suetonius, Life of Claudius, 27.1. More on the importance of expositi in the supply stream at William V. Harris, “Demography, Geography and the Sources of Roman Slaves,” Journal of Roman Studies 89 (1999): 62–75.

  13. 13.

    Judith Evans Grubbs, “Church, State, and Children: Christian and Imperial Attitudes toward Infant Exposure in Late Antiquity,” in The Power of Religion in Late Antiquity, eds. Andrew Cain and Noel Lenski (Farnham: Ashgate, 2009), 119–31.

  14. 14.

    Novels of Justinian, 22.8.

  15. 15.

    Kyle Harper, “Slave Prices in Late Antiquity (and in the Very Long Term),” Historia 59 (2010): 206–38.

  16. 16.

    Noel Lenski, “Framing the Question: What Is a Slave Society?” in What Is a Slave Society? The Practice of Slavery in Global Perspective, eds. Noel Lenski and Catherine M. Cameron (Cambridge: Cambridge University Press, 2018), 15–57.

  17. 17.

    Ulrike Roth, Thinking Tools: Agricultural Slavery Between Evidence and Models (London: Institute of Classical Studies, 2007), ch. 2.

  18. 18.

    Walter Scheidel, “Human Mobility in Roman Italy, II: The Slave Population,” Journal of Roman Studies 95 (2005): 64–79.

  19. 19.

    R.S. Bagnall and B.W. Frier, The Demography of Roman Egypt (Cambridge: Cambridge University Press, 1994), 70–71.

  20. 20.

    Noel Lenski, “Peasant and Slave in Late Antique North Africa, c. 100–600 CE,” in Late Antiquity in Contemporary Debate, ed. R. Lizzi Testa (Newcastle: Cambridge Scholars, 2017), 113–55.

  21. 21.

    See the studies in Domenico Vera, I doni di Cerere: storie della terra nella tarda antichità (strutture, società, economia) (Turnhout: Brepols, 2020).

  22. 22.

    Luciano Lazzaro, Esclaves et affranchis en Belgique et Germanies Romaines, d’après les sources épigraphiques (Paris: Belles Lettres, 1993), 419–25.

  23. 23.

    On mining: A.M. Hirt, Imperial Mines and Quarries in the Roman World: Organizational Aspects 27 BC–AD 235 (Oxford: Oxford University Press, 2010) 222–28, 271–78, 314–18. On building: P.A. Brunt, “Free Labour and Public Works at Rome,” 70 (1980): 81–100.

  24. 24.

    Plautus: Amy Richlin, Slave Theater in the Roman Republic: Plautus and Popular Comedy (Cambridge: Cambridge University Press, 2017) 90–104. Apuleius, Metamorphoses, 9.12.

  25. 25.

    On baking see Jared T. Benton, The Bread Makers: The Social and Professional Lives of Bakers in the Western Roman Empire (Cham: Palgrave Macmillan, 2020), ch. 4.

  26. 26.

    Susan Treggiari, “Jobs in the Household of Livia,” Papers of the British School at Rome 43 (1975): 48–77.

  27. 27.

    K.R. Bradley, Discovering the Roman Family: Studies in Roman Social History (New York: Oxford University Press, 1991), 37–102.

  28. 28.

    Jean-Jacques Aubert, Business Managers in Ancient Rome: A Social and Economic Study of Institores, 200 B.C.-A.D. 250 (Leiden: Brill, 1994).

  29. 29.

    P. R. C. Weaver, Familia Caesaris: A Social Study of the Emperor’s Freedmen and Slaves (Cambridge: Cambridge University Press, 1972).

  30. 30.

    Alexander Weiss, Sklave der Stadt: Untersuchungen zur öffentlichen Sklaverei in den Städten des römischen Reiches (Stuttgart: Franz Steiner, 2004).

  31. 31.

    Helga Scholten, Der Eunuch in Kaisernähe: zur politischen und sozialen Bedeutung des “praepositus sacri cubiculi” im 4. und 5. Jahrhundert n. Chr. (Frankfurt: P. Lang, 1995).

  32. 32.

    Thomas A.J. McGinn, The Economy of Prostitution: A Study of Social History and the Brothel (Ann Arbor: University of Michigan Press, 2004).

  33. 33.

    Thomas A.J. McGinn, Prostitution, Sexuality, and the Law in Ancient Rome (Oxford: Oxford University Press, 1998).

  34. 34.

    Matthew J. Perry, Gender, Manumission, and the Roman Freedwoman (Cambridge: Cambridge University Press, 2014); Katharine P.D. Huemoeller, “Freedom in Marriage? Manumission for Marriage in the Roman World,” Journal of Roman Studies 110 (2020): 123–29.

  35. 35.

    Richlin, Slave Theater, 126–36; Kyle Harper, Slavery in the Late Roman World, AD 275–425: An Economic, Social, and Institutional Study (Cambridge: Cambridge University Press, 2011), 236.

  36. 36.

    Plutarch, Life of Cato, 4.4–5; Suetonius, Life of Claudius, 25.2; Dio Cassius, 60.29.7.

  37. 37.

    Hugh F. Thompson, The Archaeology of Greek and Roman Slavery (London: Duckworth, 2003), 217–40.

  38. 38.

    Martial, Epigram, 8.23.

  39. 39.

    Galen, De propriorum animi cuiuslibet affectuum dignatione et curatione [Kühn V. 17, 40–41].

  40. 40.

    François Hinard and Jean Christian Dumont, eds. Libitina: pompes funèbres et supplices en Campanie à l’époque d’Auguste (Paris: De Boccard, 2003).

  41. 41.

    Theodosian Code, 9.12.1–2.

  42. 42.

    Pliny, Natural History, 35.200; Aulus Gellius, 6.4.

  43. 43.

    Sandra R. Joshel, Work, Identity, and Legal Status at Rome: A Study of the Occupational Inscriptions (Norman: University of Oklahoma Press, 1992).

  44. 44.

    Theresa Urbainczyk, Slave Revolts in Antiquity (Stocksfield: Acumen, 2008).

  45. 45.

    Dio Cassius, 76.10.

  46. 46.

    Noel Lenski, “Violence and the Roman Slave,” in The Topography of Violence in the Greco-Roman World, eds. Werner Riess and Garrett G. Fagan (Ann Arbor: University of Michigan Press, 2016), 275–98.

  47. 47.

    K.R. Bradley, Slavery and Society at Rome (Cambridge: Cambridge University Press, 1994), 107–31.

  48. 48.

    Seneca, Letter, 70.20–26.

  49. 49.

    Cicero, Letters to Friends, 5.9, 5.11.3, 13.77.3, and Letter to his Brother Quintus, 1.2.14.

  50. 50.

    Christopher J. Fuhrmann, Policing the Roman Empire: Soldiers, Administration, and Public Order (New York: Oxford University Press, 2014), ch. 2.

  51. 51.

    Henrik Mouritsen, The Freedman in the Roman World (Cambridge: Cambridge University Press, 2011), ch. 5.

  52. 52.

    Justinianic Code, 1.13.1–2; Sozomen, Ecclesiastical History, 1.9; cf. Noel Lenski, “Constantine and Slavery: Libertas and the Fusion of Roman and Christian Values,” Atti dell’Accademia Romanistica Costantiniana 18 (2011): 235–60.

  53. 53.

    Macrinus: Historia Augusta, Life of Macrinus, 4.3. Diocletian: Epitome de Caesaribus, 39.1; Eutropius, 9.20.