To refer to this please cite it in this way :

  • Keith Bradley, Frank M. Snowden, Jr. Lectures, Howard University, 'The Bitter Chain of Slavery': Reflections on Slavery in Ancient Rome, Center for Hellenic Studies, Washington, DC. November, 2005

‘The Bitter Chain of Slavery’: Reflections on Slavery in Ancient Rome

Towards the middle of the fifth century AD the Christian presbyter and moralist Salvian of Marseilles composed a highly polemical tract, On the Governance of God, in which he explained to the decadent Romans around him how it was that the destructive presence in their midst of barbarian invaders was the result not of God’s neglect of the world but of their own moral bankruptcy. In their general comportment the Romans, though Christians, were full of moral failings and were far more morally culpable than the slaves they owned. Their slaves committed crimes such as stealing, running away, and lying, but they did so under the comprehensible and forgivable compulsion of hunger or fear of physical chastisement, whereas the Romans were simply wicked and had forfeited all claims to forgiveness because of their terrible behaviour. Among other things the Christian slaveowners had completely desecrated the institution of marriage: regarding their female slaves as natural outlets for their sexual appetites and considering adultery unexceptional, they thought nothing of acting upon their impulses and of satisfying their desires. As a result, Salvian said in an ironic metaphor, they had become the bad slaves of a good Master, which meant that the barbarian invaders, while pagans, were in fact their moral superiors. In Salvian’s judgement it was this moral superiority that accounted for the barbarians’ stunning invasionary success (On the Governance of God 4.13-29; 6.92; 7.16-20; cf. 3.50; 8.14).

Despite his critical assault on Roman slaveowners, Salvian makes very clear the low esteem in which slaves were held in his society. Slaves were naturally inferior, criminous, and corrupt, they lived only to satisfy their base wishes, and they were expected to show unquestioning obedience to their owners, including sexual obedience. In recognising the motives that drove them to steal, lie, and run away, Salvian was notably sympathetic to them and he maintained that kindly treatment was a useful alternative to physical coercion in rendering slaves submissive. But he never questioned the reality of slavery, and he could proclaim without any sign of discomfort: ‘It is generally agreed that slaves are wicked and worthy of our contempt’ (4.29).

Such views were hardly new. Images of immoral and criminous slaves, appeals for adopting a carrot-and-stick approach to handling them, and statements that obedience should be expected of them can be found in any number of earlier Greek and Latin writers. The precise form of slavery Salvian knew in fifth-century Gaul is a matter of controversy, but the terms he used to describe it, and the conceptual attitudes underlying them, were those which Greek and Roman slaveowners had used and drawn on for centuries past.

His remarks nonetheless are striking. Salvian was writing at a very late date in classical history, and while directed to Romans in general his audience in the first instance was an entirely local body of men, the wealthy lords of southern Gaul—and both he and his local audience were of course Christian. Despite its conventional aspects, therefore, Salvian’s evidence brings into sharp focus two well-known but important facts. The first is that there was no time or place in Greco-Roman antiquity, even on the margins of time and place, that was altogether free from the presence or influence of slavery. Across the vast chronological interval from the Mycenaeans to the Roman Empire of the fifth century and beyond, and in all the regions where Greco-Roman culture took root—Europe, the Near East, North Africa—slavery in one form or other was an integral part of the social order. The second is that across time and space no one, not even Christians, ever seriously thought to question slavery and slaveowning. To moderns living in societies in whose democratic traditions the abolition of slavery in the nineteenth century is a landmark event, it may seem problematical that a call to end slavery never arose, especially in view of the appearance in late antiquity of a socially sensitive attitude like Salvian’s. But this is a modern not an ancient problem, and it is not the absence of an abolitionist movement in classical antiquity, even Christian antiquity, that is historically peculiar so much as the rise of abolitionism in post-Enlightenment Europe and North America. For most of human history, the enslavement of one group in society by another, or of one people by another, has been a quintessential element of normal social relations. Western liberalism cannot be allowed to obscure this fundamental truth, or to justify the assumption that the absence of slavery is in any sense socially normative, no matter how socially desirable. There is a world of difference between Salvian and, say, the nineteenth-century opponent of slavery Henri Wallon, who in his celebrated Histoire de l’esclavage dans l’antiquité (1847), wrote in a climate when slavery had come to be regarded in Christian ethics, in a mode of thought totally alien to classical antiquity, as a sin.

The forms of servitude known in the classical world varied across time and place. They included debt-bondage, helotage, temple slavery and serfdom, but also chattel slavery, an absolute form of unfreedom in which enslaved persons were assimilated to commodities, akin to livestock, over whom, or which, owners enjoyed complete mastery. Chattel slavery was not found in all times and places in antiquity, but it was especially evident in Italy during the central era of Roman history and it is with Roman chattel slavery that I am concerned here. I want to consider the nature of the master-slave relationship and the basic character of Roman chattel slavery, and to suggest from a cultural point of view why slavery at Rome, as I understand it, never could present itself as problematical. For the sake of convenience and because it is relatively well-attested, I concentrate particularly on Roman domestic slavery. My account is necessarily generalised, impressionistic, even superficial and schematic, and at every stage allowance must be made for the ambiguous and the exceptional.

I take as a starting-point the observation from the Roman Antiquities of the Greek author Dionysius of Halicarnassus (1.9.4; cf. 4.23.7), that when Romans manumitted their slaves they conferred on them not only freedom but citizenship as well. To Dionysius and the Greeks for whom in the age of Augustus he was writing this was an unusual and generous practice. And it has seemed unusual and generous to moderns as well, so much so that scholars have often concluded that Roman slavery was a mild institution, milder by implication at least than the race-based slavery systems of the New World. As an example let me quote a passage from another celebrated book, Jérôme Carcopino’s Daily Life in Ancient Rome, which was first published in the United States in 1940, a year after the French original, and which I select because it has always enjoyed enormous influence and is currently enjoying a new lease of life in re-edited versions. Carcopino is speaking of the age of the Antonines:

Everyone learned to speak and think in Latin, even the slaves, who in the second century raised their standard of living to the level of the ‘ingenui’. Legislation had grown more and more humane and had progressively lightened their chains and favoured their emancipation. The practical good sense of the Romans, no less than the fundamental humanity instinctive in their peasant hearts, had always kept them from showing cruelty towards the ‘servi’. They had always treated their slaves with consideration, as Cato had treated his plough oxen; however far back we go in history we find the Romans spurring their slaves to effort by offering them pay and bonuses which accumulated to form a nest egg that as a rule served ultimately to buy their freedom. With few exceptions, slavery in Rome was neither eternal nor, while it lasted, intolerable; but never had it been lighter or easier to escape from than under the Antonines.

More recently and more compellingly, the preeminent historian Susan Treggiari has shown how a relatively benign picture of Roman slavery like that of Carcopino might emerge. Exploiting two types of evidence, commemorative epitaphs and the writings of Roman jurists, Treggiari has investigated in a remarkable series of studies the personal lives of slaves and former slaves who worked as domestic servants in the resplendent households of the Roman elite under the early Principate, and she has proved that much can be learned about the world these people created for themselves. What emerges, first, is the vast range of highly specialised work-roles that helped slaves to establish individual identities for themselves, and, secondly, the formation of familial relationships, sometimes of long duration, that restored to slaves something of the human dignity of which slavery deprived them. The value found in their work as domestic servants becomes clear, and the manner in which despite their legal incapacity slaves constructed and memorialised familial ties is repeatedly made plain. Special attention is paid, moreover, to the roles played by women and what might be termed the female contribution to the infrastructure of Roman society is brought to the fore as evidence is compiled of the spinners and weavers (quasillariae, textrices), the clothes-makers and menders (uestificae, sarcinatrices), the dressers (ornatrices), nurses and midwives (nutrices, opstetrices) who populated the domestic establishments of the Roman elite. With the perceived development under the Principate of a more humane attitude to slaves—a view that goes back beyond Carcopino to at least Gibbon’s belief that a certain ‘progress of manners’ alleviated the hardships of slavery in the imperial age—a confident picture of life in slavery is presented. Here are three representative statements:


The Monumentum Liviae gives us the first full and vivid evidence both for the staff of a Roman woman and for the middle class of domestic servants, a class which enjoyed scope for a variety of talents and which displays esprit de corps and considerable satisfaction in being employed by the wife and mother of a ‘princeps’ and the daughter of a god.

In the large, hierarchical but closely-knit society of the rich household, with its records of births, deaths, manumissions, and ‘contubernia’, slave family life could often attain comparative security and dignity. Scraps of evidence, the commemoration of parents, brothers, sisters, and sometimes other relatives, friendships close enough to be honoured after death, ‘contubernia’ which lasted a lifetime, help to illustrate this.

We can see in the inscriptions evidence for a tightly-knit and supportive community, creating its own goals and work ethic and organizing its own social life, under the supervision of freed administrators and largely without the interference of the upper class masters whom the staff was bred or bought to serve.


There are other avenues of investigation which lead in the same incontestable direction. In an absorbing recent book, Art in the Lives of Ordinary Romans, John Clarke suggests that the response of slaves to certain works of Roman art was to encourage hopes of manumission among them that presuppose while they were waiting an unquestioning willingness to accept and conform to the values of established free society. The grain measurers (mensores) who can be seen in a mosaic from the Piazza delle Corporazione at Ostia will have communicated to slave viewers how they had to work hard to win their freedom (‘and perhaps an easier life’). A tomb such as that of the freedman C. Julius Apella at Ostia would allow slaves from the familia who waited at table at the dinners held there to delight in its decorations and to imagine eventually coming to rest there themselves as liberti. The household shrine (lararium) from the House of the Sarno in Pompeii, which apparently depicts a master and his slaves at work on the river, would foster among the viewing slaves in the garden where the shrine was located, as Clarke puts it, ‘a certain pride in seeing themselves with their owner every day as they stood behind him to sacrifice to his Genius and the Lares.’ The assumptions are clear that that the slavery system was benign, that within it the boundary between slavery and freedom was easily permeable, and that Roman slaves, always acquiescent to their masters’ demands, wanted nothing more than to cross the boundary and strove in every way to do so.

There is much of course to show that many Roman slaves did adopt the enticements of the free to conform to the dominant ideology, and also that they successfully make the transition from slavery to freedom and fully integrated themselves into the life of established society. One impressive illustration is the way in which slaves and former slaves willingly responded to the emperor Augustus’ division of Rome in 7 BC into 265 administrative districts (uici) and agreed to become officiants (magistri and ministri) in the renovated cult of the Lares Compitales, now the cult of the Lares Augusti, which was in effect a not so subtle form of emperor worship. Felix, slave of L. Crautanius, Florus, slave of Sex. Avienus, Eudoxsus, slave of C. Caesius, Polyclitus, slave of Sex. Ancharius (ILS 9250)—these names typify the many men who participated in the cult. Another illustration is the willingness of wealthy freedmen to enter the new status-category of Augustales created by Augustus and to use their wealth, much like free men higher in the social hierarchy, to provide games and shows and other public benefits in the towns and cities of Italy, as though they were citizens of longstanding devoted to the promotion of the established civic and social order. The freedman C. Lusius Storax, who died about AD 40, was acclaimed for his provision of gladiatorial games at Teate Marrucinorum and built a monument commemorating himself and his generosity in the enclosure of the burial society (collegium) to which he belonged. There can be no doubt that over time there were many like him, slaves who were set free and subsequently made vital contributions to the well-being of Roman society; and it can readily be granted that as slaves they aspired to achieving manumission and worked diligently to this end. The ‘optimistic’ view of Roman slavery has much to commend it.

In my view, however, this is only one, partial, aspect of the picture. Roman slavery also had a much darker side, and it is to this darker side that I now want to turn, asking especially how slavery might have been fully experienced within the domestic context about which so much seems knowable. My intent is not at all to dispute the views of the contemporary scholars to whom I have just referred, but to point to the complexities of Roman slavery as I understand them. The question raises methodological issues. There are no extended accounts from slaves themselves to allow the historian a direct view of life in slavery, and much simply has to be inferred from sources that represent (and perhaps continually influenced and shaped) the attitudes and ideology of the slaveowning classes. These sources are primarily literary —sometimes imaginative and sometimes anecdotal — of a sort that historians of modern slavery systems would often dismiss as of minimal value. But there is scarcely any alternative. Epitaphs and legal sources are immensely important, but they are by themselves insufficient; and epitaphs, especially, cannot be expected to reveal much that is critical of slavery when they celebrate for the most part individuals who found ways to achieve some sort of conventional success in life. In what follows, therefore, I offer a set of inferential observations from my reading of certain literary authors of the Principate, who allow, I believe, credible glimpses of life in slavery that stand in strong contrast to what has been seen so far.

The first point to emphasise is one that slavery historians, especially modern slavery historians, have always known, namely that slavery in Roman antiquity was not a soulless legal condition—a point of view common in legal studies of Roman slavery—but a human relationship in which slave and master were always inextricably bound together. The relationship was obviously asymmetrical, comparable according to the third-century Greek author Philostratus (Life of Apollonius of Tyana 7.42) to that between a tyrant and his subjects. But it was not completely one-sided. In theory the slave was powerless: No slave is really happy,’ the Hellenized Jew Philo wrote, ‘For what greater misery is there than to live with no power over anything, including oneself?’ (Every Good Man is Free 41), and the slave was always subject to constraint, so that the medical authority Celsus could write (On Medicine 3.21.2) that a slave habituated to a life of compulsion endured the harsh treatment needed to cure an illness more eaily than the free. Yet because slaves were a human form of property, human agency could and did manifest itself in the relationship from moment to moment. Unlike the animals to which they were often compared, slaves were not easily manipulable, but had to be managed with thought and discretion to make sure that they did what was required of them and to prevent ‘criminal’ acts of the sort to which Salvian was still sensitive in late antiquity. The relationship therefore was one that on both sides involved constant adjustment, refinement, and negotiation. Some slaves, sure enough, enjoyed a privileged status in their households. Those who were stewards or managers of estates or supervisors of lowlier slaves or chaperons of their masters’ wives and orphaned children held positions of authority because they were trustworthy and so resembled the free; and yet, as Philo said in the same work (35), whatever influence these slaves had they remained slaves regardless and to that extent they could never be free from the constricting tie to those who owned them and the struggle for power the tie involved.

My point is well illustrated by an anecdote from Plutarch (Moralia 511D-E) concerning the consul of 61 BC, Marcus Pupius Piso. As follows: The orator Piso, wishing to avoid being unnecessarily disturbed, ordered his slaves to answer his questions but not add anything to their answers. He then wanted to give a welcome to Clodius, who was holding office, and gave instructions that he should be invited to dinner. He set up a splendid feast. The time came, the other guests arrived, Clodius was expected. Piso kept sending the slave who was responsible for invitations to see if he was coming. Evening came; Clodius was despaired of. ‘Did you invite him?’ Piso asked his slave. ‘Yes.’ ‘Then why didn’t he come?’ ‘Because he declined.’ ‘Then why didn’t you tell me?’ ‘Because you didn’t ask.’ Such is the way of the Roman slave! An anecdote like this, as everyone will be aware, cannot be taken at face value, as if literally true. It is what the story symbolises that is important: the fact that at any time any slave at Rome had the potential to challenge the authority the slaveowner commanded, which means accordingly that the relationship between slave and master always implicated the energies of both sides in a never-ending struggle for supremacy, and clearly it was not always the master who won. Owners knew this (as the anecdote shows) and they had to reconcile themselves to it. Philostratus (VS 516-518) offers another illustration, telling how a scheming cook named Kytheros, the slave of the father of the sophist Scopelian, in league with the old man’s concubine, was once able to turn the father against his son, replace the son in the old man’s will, inherit his estate, and subsequently achieve public prominence (though his origins were not forgotten). Philostratus was aware that the slave’s behaviour resembled something from the plot of an ancient comedy, but he did not doubt it as fact. Similarly Philo had no doubts (Every Good Man is Free 38) that sex was a particularly useful commodity in the relational contest of wills: maidservants with pretty faces and charming words might well take the initiative and seduce their masters—which is to say that slave women could use sex to their advantage and were not always its victims.

From the establishment’s point of view, what was at stake in all this was the maintenance of social order and the defusing of threats to the exercise of power inherent in slaveownership. Fear of upheaval was never far away. At the turn of the third century the sophistic writer Aelian (Characteristics of Animals 7.15) wrote of a woman named Laenilla he had known as a boy whose infatuation with a slave had led to the criminal indictment of her completely innocent sons, young men of senatorial descent who were embarrassed by their mother’s behaviour and tried to point out its shamefulness to her. The woman had been unwilling to give up her slave lover, and falsely accused the sons before a magistrate. They were subsequently executed. The combination of slavery, sex and shame was a recipe for social disaster, a deeply disturbing prospect to be avoided at all costs.

How might a sense of the never-ending in the master-slave relationship be recovered? I want to suggest at this point that for answering this question the famous section of Petronius’ Satyricon known as ‘Trimalchio’s Dinner-Party’ (Cena Trimalchionis) is one literary source that can be very useful. The Cena is of course a piece of fiction, as is the Satyricon as a whole. But no one would question that it reflects social conditions of the first century and for present purposes its value lies, I believe, in the way its narrative nature opens up the possibility of observing continuous interaction between a slaveowner and various members of his domestic entourage over a certain interval of time.

Petronius’ great creation Trimalchio, a former slave who has made himself enormously rich in commerce after inheriting and investing money from his former owners, has an elaborate domestic staff with titles just like those found in the epitaphs from elite Roman households. There is a porter (ostiarius), a major-domo (atriensis), an accountant (dispensator), a steward (procurator), a record-keeper (actuarius), a name-announcer (nomenclator); there are cooks and carvers, doctors and masseurs, musicians, acrobats and readers, and any number of attractive boys from Alexandria and Ethiopia to wait at table and catch the eye of guests. As Petronius’ heroes Encolpius and Giton suffer through the long ordeal of the dinner-party, they see these members of staff benefiting from their owner’s generosity (Trimalchio’s ‘humanitas’ [Satyricon 65.1]: surely ironic!) in any number of ways. Trimalchio orders drinks for the slaves who sit and attend his guests (64.13); he summons a new contingent of waiters so that the previous battery may leave to have their own dinner (74.16); he has his will read out so that his household will know the kindnesses awaiting them on his death—grants of freedom and even the bequest to one dependant of his contubernalis (71); he sets free an acrobat who accidentally falls on him, to avoid the shame of having been injured by a slave (54.4-5). These are all acts of random, one might say quixotic, ‘humanity’. They do not have to be taken literally as evidence of what slaveowners did in real life. But they reveal how slaves on a daily basis might reap the rewards of being close to their owners at specific moments in time.

The occasional acts of kindness at the dinner-party catch the eye, but more arresting to my mind is the threatening atmosphere of violence that overhangs and permeates Trimalchio’s relationship with his slaves. Physical proximity of slave and master, it needs to be remembered, could expose domestics, even those of superior station, to the punishing consequences of random bouts of temper or irritability as much as to the benefits of random acts of kindness—evidently a common enough problem for moralists like Seneca and Plutarch, and even the medical authority Galen, to be found giving counsel about it from one generation to the next—and there is no shortage of such consequences here. Trimalchio has an over-zealous slave who picks up a fallen dish boxed on the ears (34.2). He threatens with demotion to the ranks of the uiatores a cook, left to him by will and unrecognised, if a pig is not properly served, making the slave well aware of the master’s naked authority, his potentia (47.3). When another pig is brought in, another cook is threatened with flogging because he seems to have forgotten to gut the animal. This is a charade of course. Trimalchio is playing a trick on his guests because he wants to impress them with the sausages and black puddings he knows will appear once the pig is cut open. But the significant point is that the charade is credible: Trimalchio has the cook stripped and handed over to torturers (tortores) he keeps on his staff, as slaveowners could, expressly for the purpose of physically punishing members of his household (49.1-50.1). Later he threatens with decapitation a slave who drops a cup, relenting only when his guests intercede on the unfortunate man’s behalf (52.4-6). He also threatens to burn alive a certain Stichus (a good slave name) if the slave fails to take proper care of his burial clothes (78.2). Violence, physical, psychological, or both, figures everywhere in the relationship between owner and owned, and the extended dinner narrative expresses this dynamic reality in a way, I think, that inscriptions and passages from the law cannot.

That slavery was an institution based on brute force and terror hardly needs to be demonstrated. Plutarch’s observation (Moralia 462A) that the first thing newly purchased slaves wanted to know about their owner was whether he was ill-tempered is just one indication of the psychological truth, the Christian Lactantius’ exhortation (Divine Institutions 4.4.1) that his readers should fear God like slaves another. But the degree to which violence might at any time appear in the slave’s life cannot to my mind be overstated. In the Satyricon it is normative for the master to resort to the whip when angered by slaves he regards as ‘delinquent’, as a new slave, a nouicius, makes clear (139.5). And it was not the slaveowner alone the slave had to fear. It made sense to Petronius to imagine that a slave accountant could have a slave underling beaten for having lost his clothes at the baths (30.7-11), and that a freedman guest could verbally abuse one of Trimalchio’s slaves and physically threaten him for being impudent (58.5). It even made sense that slaves themselves could be called upon to commit acts of violence: in a later episode of the Satyricon (132ff) an upper-class woman calls on her slave spinning-maid to spit on an enemy, and on her slave chamberlain to beat him. Violence, or the threat of violence, was everywhere.

The violence of sale was a variation on this theme. It may have been a testament to Trimalchio’s bad taste that he had a mural of a slave-market in his house, complete with price tags on the merchandise (29.3); but this too was a statement of how power was distributed between master and slave, reminding the slave viewer of the violent disruption that could enter life at any time. So what, I wonder, would a real-life character such as the freedman L. Volusius Heracla, who was commemorated as both capsarius and a cubiculo (ILS 7413), have thought when looking at a picture like this before he was set free? Did he think only of his own good fortune, as the optimistic view might have it, or did he think of those sections of Rome where beautiful slaves, of both sexes, were, as Plutarch (Moralia 520C) notes, always to be found for sale along with the freaks of the ‘monster market’? Had he once been in the slave-market? Might he be there again? Was he once a slave whose face a slave-dealer had plastered with bean-flour to remove his freckles and moles to make him more attractive to buyers—a trick of which Galen knew (6.530K)?

But this is not real life, you will say. It all comes from a work of the literary imagination, and a work which by definition demands outrageous comic exaggeration, so that a strictly literal reading of the text cannot be justified. Yet this does not mean that the text lacks all sense of realism: Trimalchio’s specialised domestics are proof enough of that—and this means, I think, that the arbitrariness of the masterslave relationship that the Satyricon conveys must also be taken as authentic and the conclusion reached that Roman domestic slaves lived under a tense psychological regime which guaranteed them nothing as far as security or stability in life were concerned. Incidents such as that witnessed by Galen (5.18-19K) in which a travelling companion in Greece, enraged over the whereabouts of a certain item of baggage, struck two of his attendants on their heads with a large knife and seriously injured them—Galen later saved them, not surprisingly—have to be taken seriously as evidence of what slaves always had to contend with, not as isolated instances of aberrant behaviour. To dismiss Trimalchio’s actions therefore simply as ‘bluster’ and to describe him as ‘a perfectly kind master’ is to fail completely to my mind to probe how an atmosphere of intimidation, no matter how comically drawn, affected those against whom it was directed. Trimalchio had an inscription written on the door of his house: ‘Any slave who leaves the house without the master’s permission will receive a hundred lashes’ (28.7). I wonder again whether slaves in the well-documented household of Augustus’ wife Livia, those such as Antiochus the atriensis, Calamus the dispensator, and Dorcas the ornatrix, ever saw anything like this, and how would they have responded to it (CIL 6.3942; 6.3965b; 6.8958=ILS 1784).

As a control on Petronius some evidence from the Moral Epistles of his contemporary Seneca might be considered. You will think immediately of the famous forty-seventh epistle, which has often been held up as an example of Seneca’s humanitarian attitude and as a sign of an increasing compassion towards slaves under the Principate, a view I do not myself share. But I am more interested in the Moral Epistles’ casual allusions to slaves and slavery, which I think are especially revealing of Roman elite views because of their offhandedness and which consequently form a priceless guide to the conditions under which domestic slaves lived. Seneca himself was a slaveowner, on the evidence of the Moral Epistles alone (83.4; 123.1-2, 4) used to having around him a cook, a baker, masseurs, a bath attendant, a personal trainer, a major-domo—the constituents of what he terms the aristocrat’s formonsa familia (41.7). It is what he takes as normal or uncontroversial about slaveowning that is surely significant.

In the ordinary course of events Seneca expects elite Romans to have a mass of slaves attending upon them, litter-bearers to transport them, door-keepers to control access to their houses, masseurs to take care of their bodies (17.3; 31.10; 43.4). But contact with the slave is essentially degrading if, for instance, you have to take orders from the man who works as your trainer and so invert the ‘natural’ hierarchy of power (15.3). And slaves are a burden to the owner: they have to be fed and maintained, and they have a tendency to run away (17.3; 107.7). Seneca values the edifying story of the Spartan boy who killed himself rather than submit to slavery for what the story says about the need to secure freedom of the spirit; but when he tells it to his interlocutor Lucilius he shows no sympathy for or interest in the slave as a slave (77.14). It causes him no distress that a slave criminal should be burned alive (86.10). A master’s right to beat his slave when going over his accounts is not questioned (122.15). No problem that a slave might jump from a roof and kill himself to avoid the taunts of a dyspeptic owner or fall on the sword in order to avoid capture after running away (4.4). Slaves are essentially enemies, always involved in plots to kill their owners, creatures who, quite simply, like animals, have to be ruled (18.14; 4.8; 77.6; 80.9; 94.1).

Slavery itself Seneca regards as a state characterised principally by subjection to compulsion—this indeed is what he calls the bitterest part of slavery (61.3)—a condition in which the slave might be forced for ever to eat no more than meagre rations of poor food, ordered to tiptoe around the house in silence to avoid disturbing an insomniac master, required even to help a master kill himself (18.8; 56.7; 77.7). Or else it is a kind of living death, from which the slave will do anything to escape, saving money by going hungry so that freedom can eventually be purchased and slavery set aside (77.18; 80.4). When Seneca makes his grand Stoic statements about the brotherhood of man, claiming for example that the labels of elite Roman, freedman, and slave are no more than inconsequential words (31.11), it is difficult not to recoil in horror.

The poet Martial opens another window into the world of the master-slave relationship, and he is the last author I want to consider. Martial’s poetry belongs of course to a completely different genre from the genres represented by Petronius’ novel and Seneca’s sermonettes, and again as works of the literary imagination I stress that Martial’s Epigrams are not to be read as statements of literal fact. Once more, however, the poems can be read as statements that make assumptions about social norms in Rome of the first century, and it is this sense of the normative, and its consistency with what is evident in Petronius and Seneca, that I think is valuable. The evidence of different literary authors, I believe, cannot all be dismissed as simply ‘literary’ when consensus about the normative is so clear. And Martial, keep in mind, notably claimed that there was a direct correspondence between what he wrote and the life he knew around him: ‘let life recognize and read of her ways’ (8.3).

Martial’s poems contain any number of references to the occupations of domestic slaves, but if anything it is the humbler levels of the household hierarchy that predominate. There are stewards, pedagogues and nurses, musicians, cooks and bakers, and the freak (morio) who was kept as an object of amusement (e,g. 1.49; 8.44; 10.62; 11.39; 12.49; 11.78; 9.77; 3.94; 8.23; 16.39; 11.31; 8.13). But doorkeepers and litter-bearers are equally in evidence, and personal or body-servants seem to be everywhere: the woman’s hairdresser, the man’s barber, the bath-assistant, the personal trainer, the slaves who attended their masters at dinner—including those who took off their shoes and those who carried the lantern when they returned home in the dark—and the slaves who at the snap of a finger came running with the chamber-pot (e.g. 9.2; 6.52; 8.52; 11.58; 12.70; 3.23; 12.87; 14.65; 8.75; 6.89, 14.119).

Some of the most affecting of Martial’s poems commemorate the untimely deaths of young or former slaves: the boy Alcimus, who died as a teenager; the secretary (amanuensis) Demetrius, dead from disease at a similar age; the personal favourite (deliciae) Erotion, dead at only five and fondly remembered in three poems; the ex-slave Glaucias, dead at twelve and the subject of two poems; and the skilled barber, the completely good Pantagathus, who was taken while a boy (1.8; 1.101; 5.34; 5.37; 10.61; 6.28; 6.29; 6.52). Both the grief caused by death and the sense of intimacy in life between master and slave conveyed by these poems seem to me genuine, and it is difficult not to take them as evidence of the close personal bond between the two that might develop despite the enormous differences of status involved. Demetrius the secretary was even set free so that he might avoid the stigma of dying in slavery, a remarkable testament to the gulf between slave and free that existed in Roman society and also of a slaveowner’s sensitivity to it. In this context, a reference (9.87) to how at any moment a man might be called to witness an act of manumission suggests a slave world of relative ease in which once more the prospects of crossing the permeable boundary were rather good. The slave who was once in shackles, Martial says, might one day find himself wearing the ring of elite privilege.

Other poems, however, offer a starker set of images. First there is the commodity that can be loaned by a slaveowner to a friend, a transaction which might cause the owner difficulties of recovery but which hardly takes any account of the object of the loan (2.32). Secondly there is the commodity that can be bought and sold—sold on a whim to raise the price of a fancy dinner, or, with more calculation, as a result of a cash-flow problem—and bought especially, if you have the money, for sex, of any kind, boy-commodities in the Saepta and girl-commodities in the Subura (10.31; 9.59; 6.66). For Martial (and presumably his audience) the commodity’s sexual availability is simply taken for granted: slaveowning men and women are free to indulge any appetites they have, and slaves are to submit and to accommodate them (e.g. 1.84; 3.71; 3.73; 4.66; 6.39; 9.25; 12.58). The results might be literarily amusing—one man is utterly unaware that his apparent ‘children’ have all been fathered by different members of his household staff (6.39), and another is lampooned because he sells but then buys back a slave girl with whom he is infatuated (6.71): what a disgrace! But the assumption that the slaveowner is sexually sovereign is unmistakable. (The inference might be drawn that the slave, as seen earlier from Philo, was sometimes a willing sexual partner [a male slave on the run for instance who was shacking up with a discharged soldier (3.91)], but the servile perspective on sexual access is obviously hopelessly beyond reach.) Then, thirdly, there is the object (once more) of random violence, the object whose body is taken as a natural site on which to inflict physical pain and suffering. A woman distressed that one ringlet of her elaborate coiffure has not been properly pinned strikes her dresser with a mirror; a man annoyed that his dinner is not properly prepared flogs his cook; another punishes an errant slave by hitting him in the mouth (2.66; 3.94; 8.23). Martial repeatedly associates the slave with the whip, the cross, the shackle and the brand (e.g. 14.79; 2.82; 10.82; 3.29, 9.57; 3.21; 10.56). So no tenderness here of the kind that appears in the laments for the untimely dead, and the slave now is always anonymous.

Two of Martial’s poems recall the story of Marcus Piso and the theme of the contest of minds, offering further evidence of a master-slave relationship that was subject to constant negotiation. First, the poem (6.39) to which I just alluded in which a certain Cinna is derided because his seven children are the fruits of his wife’s liaisons with seven of the household slaves—a visible truth because the children all physically resemble their respective fathers. I can scarcely believe that Martial or anyone else knew of such a situation in real life. But what I find plausible is that the poem gives expression to a genuine, double-edged fear on the part of the Roman male slaveowner, first that despite her social subordination his wife’s sexual behaviour was beyond his control; and secondly that through the exercise of power that derived from their capacity to make human decisions and take human actions, his slaves were equally capable of challenging the authority the slaveowner commanded. Secondly, a poem (11.58) in which the poet contemplates a scene where a slave barber shaving his master, his razor at the master’s throat, demands his freedom and a small fortune besides. What is the master to do? In fear for his life he agrees to the slave’s demands and saves himself. But once the razor is safely out of the way he can immediately take his revenge by having the slave’s hands and legs broken as the ‘normal’ balance of power is restored. Here again I know of no real incidents like this. But the poem again plausibly expresses a slaveowner’s perhaps often latent fear that when instructing his barber to shave him he temporarily exposed himself to serious danger and literally placed his life in his slave’s hands, bestowing on the slave a power that the slave-commodity was never supposed to have. The psychology of the situation—could the slave be trusted?—can only be imagined.

The evidence I have described suggests that the meaning slaves sometimes found in their work, the family ties that they were sometimes able to create, and the freedom that they were sometimes able to win were remarkable successes gained in the teeth of an unspeakably difficult physical and psychological regime. It also suggests that there can be no justification for assuming that pride in their work was a natural and generic response, or that security within the slave household (familia) was automatically guaranteed, or that slaves easily and with benign encouragement from their owners always pursued a straightforward path to freedom. Roman slavery was a complex institution, full of paradox and contrast, allowing a poet such as Martial (as I noted) to speak almost simultaneously of certain slaves by name, as individual persons, but of most indifferently as nameless instruments. At times, due to contigency and temperament, human interaction between owner and owned led to favourable results for the slave. But as far s I can see none of this was predictable or all-embracing. Moreover, as I indicated at the outset, there was never any moment in the history of Roman slavery when individual acts of generosity developed into a society-wide call for ending the institution, even with the rise in late antiquity of the new ideology espoused by men like Salvian. The Christian bishop Ambrose of Milan (On the Duties of Christian Ecclesiastics 2.138-143) saw the redemption of enslaved prisoners of war as a Christian duty, the bishop Caesarius of Arles could limit the number of times a slave might be beaten (no more than thirty-nine lashes a day [Life of Caesarius 1.25]), and a pope like Leo (Ep. 4.1) might save the priesthood from the contagion of slavery by forbidding slaves to be priests. But slavery itself never raised any serious objection. The convert Lactantius believed that everyone was a fellow-slave of God (Divine Institutions 5.15.3), so a preoccupation with justice did not involve a problem with slavery.

Why was this so? Peter Garnsey has brought forward two possible explanatory factors: the absence of any rival social and economic system to offer competition to slavery, and the structural embeddedness of slavery within the classical household that made possible pursuit of the good life by those with the resources to carry it out. A third factor may have been the absence of any emphatic equation between slavery and race. To a degree, however, the question of ‘Why not?’ is specious, because it involves trying to explain why something did not happen that you think should have happened when in fact under contemporary conditions it could not have happened. Slavery was never considered a moral evil at Rome, and without that precondition there could be no moral impulse to eradicate it. A society without slavery was not of course beyond the bounds of imagination in the Roman world, but as Lucian (Satires 7) and Macrobius (Saturnalia 1.7.26) among others knew, a society without slavery lay only in a distant, mythical Saturnalian past. It was not a realisable goal for the present or future. The problem is ours rather than theirs, therefore, and to try to provide any wider explanation for the absence of an abolitionist programme is an exercise in virtual futility. Something, however, might be said by way of a reflective gloss that helps provide perspective, and to bring matters to a close I want accordingly to stress the connections between slavery and three aspects of Roman culture as I understand it.

The first is the fundamental and all-pervasive violence of Roman life, which appeared, most obviously, in Rome’s intense militarism (unique in the ancient world), a form of institutionalised violence that in the era of Republican expansion permitted Rome to field armies that fought with a level of engagement contemporary Greeks like Polybius found astonishing (6.39.11, 6.52.6-7; cf. 6.37.13), and that under the Principate allowed emperors to celebrate military victories with monuments such as the Column of Trajan that portrayed their enemies’ grisly deaths in very graphic terms. It appeared, too, in the ritualised killing of gladiatorial contests, a conventional type of entertainment that attracted mass audiences, that Rome exported to its provinces, and that the wealthy saw as a suitable subject with which in painting and mosaic to adorn the walls and floors of their houses. (The arena was a token of Roman identity.) There was also the violence of physical retribution laid down by Roman law. From the Twelve Tables to Justininian’s Digest, physical punishment was always assumed to be a natural and desirable way of exacting redress for criminal activity, especially for those of low social status, for whom under the Principate burning alive, exposure to wild animals in the amphitheatre, and crucifixion were standard, and unchallenged, forms of capital penalty. Judicial savagery moreover may well have increased in the later imperial age: Constantine’s order that molten lead was to be poured down the throats of nurses who betrayed their mistresses’ confidences (Codex Theodosianus is just one notorious example of what can be seen as a trend. Violence was endemic and inescapable in Roman culture. It was not sanitised or minimised, but stood front and centre as an elemental feature of a society in which life was lived in the raw and close to the edge. Traditions that Rome was founded in an act of fratricide and that the Republic was born in an act of political conspiracy occasioned by an act of rape cause no surprise.

The second aspect concerns Rome’s attitude to civic rights and privileges. It was a distinctive feature of Rome’s history, a secret of empire revealed long ago by A. N. Sherwin-White, that those who had once been Rome’s enemies might upon defeat not be killed or enslaved but, at least selectively, be incorporated within the body politic and given citizenship. The paradigmatic example is Attius Clausus (not to mention his 5,000 followers), the legendary Sabine chieftain who was admitted to the Roman community and became the eponymous ancestor of one of historical Rome’s most storied families, the Claudii. This feature was predicated on the principle that privilege, in this case the privilege of Roman citizenship, was selective not universal, as might be anticipated of a society that was deeply patriarchal and steeply hierarchical. By analogy the condition of liberty that some enjoyed was also a privilege, and one whose integrity depended on a very real absence of liberty for others. The sharp antithesis between freedom and slavery that is often metaphorically drawn in modern political discourse was a literal and necessary reality within the world of Rome, where the concept of universal rights so prevalent in contemporary western societies was completely unknown. Moralists like Seneca were not indulging in fantasy when they observed that a sudden shift of fortune could transport any free man into a life of bondage and coercion.

The third aspect concerns the moral code Rome’s leaders invented to justify their positions of leadership, according to which both under the Republic and during the Principate continuous military success and the continuous expansion of Roman military power were glorified and made the source of all prestige and honour. The capacity of the individual member of the ruling elite to excel depended on the acquisition of uirtus and gloria and on his ability to display clementia, attributes that in turn depended on the promotion of continuous warfare. It was in claiming and maintaining these attributes that the elite, the leaders of a ‘warrior state’ as Keith Hopkins put it (writing of the Republic), found its raison d’être. The vehicle by which individual virtue and glory were best symbolised was the military triumph, a spectacle in which the conquering general (the triumphator), elevated for a day to the status of a god, dramatised his accomplishments in battle by parading through the streets of Rome his victorious legions and, in chains, the captives he had enslaved, and by showing to throngs of spectators paintings and placards depicting the grand events which had led to the moment of rapturous celebration in which all became implicated. If under the Principate it was the emperor who monopolised military success, the ideology of the ruling class and of Roman culture at large lost none of its force: an empire without end was an article of faith for all. (The ideology was vigorously reinforced moreover by the literary genre of history-writing, which drew its inspiration from the record of wars waged and victories won and of leadership that for good or ill set models for new generations of the Roman elite to emulate or avoid.)

Under these conditions slavery, by nature an extreme and perhaps the most extreme form of violence, cannot have seemed to Romans an aspect of life worth interrogation, any more than warfare was to be interrogated, or the glories of military power and imperial expansion, or the contests of the amphitheatre, or the penalties of the law, or the selective tenure of rights and privileges, or the code of elite values. Instead it was vital that the power that freedom gave be exhibited, and one form in which to do so was through the grand entourages of domestic slaves with which, even in late antiquity, Roman magnates purveyed themselves to the world. Ammianus Marcellinus in the fourth century tells (14.17) how upper-class Roman women travelling through the city in their litters were escorted by squadrons of domestic servants in quasi-military formation: first came the weavers, then the smoke-blackened cooks, then the rest of the household with elements of the urban plebs mixed in, and finally a battery of eunuchs, all under the direction of the chiefs of staff, men easily distinguishable, as though military officers, by the rods of office they carried in their right hands. It could have been true of many earlier ages. The ‘work’ domestic slaves did was not defined by job title alone; it also involved the manner in which they could expect to be put on display, at any time, as human symbols of conspicuous consumption. Their lot was hardly enviable.

There were always those of course, though I believe not the majority, who managed to leave slavery behind. What is striking about many who did is that in their wish to conform to the values of free society—to render themselves, like Trimalchio, as ‘Roman’ as possible—they frequently became slaveowners. One of the most common formulas to be found in Latin epitaphs is the phrase ‘for my own freedmen and freedwomen and their descendants’ (‘libertis libertabusque meis posterisque eorum’), a phrase that appears when someone has made burial provisions not only for immediate relatives but also for the freed members of the household and has wanted to memorialise his or her wishes. Often the formula is used to express the will of former slaves who had made their way in the world, and who had demonstrated their upward social mobility and their acquisition of wealth by becoming slaveowners and setting some of their slaves free. The phenomenon, as Russell Meiggs observed, is well-attested for example at Ostia. Slaveowners who had once been slaves themselves might seem the most incongruous of slaveowning categories, and their ‘success’ only renders the degradation of slavery more peculiar. But every time this epigraphic formula is seen in a former slave’s epitaph, an instance has to be recognised of how manumission had made the Roman slavery system work from one generation to another and remain unchallenged. The ideology of slaveowning had been successfully transmitted to those who had once been its victims.

How that process of ideological transmission took place—how the free learned say as children the habits of command and how slaves who were born into slavery and those who were enslaved as adults learned the habits of submission—are questions to which ancient historians have devoted remarkably little attention, despite M. I. Finley’s pressing declaration, in 1979, ‘That is the area—the psychology and ideology of slavery—which seems…most urgently in need of continued inquiry.’ ‘Psychology’ and ‘ideology’ are terms that I have used sporadically throughout this address in an effort to come to terms with the character of Roman slavery. Unquestionably much more could be done. Surveying the history of world empires in the later Roman period, Augustine (City of God 18.2) wrote that it had always been natural, and remained so, for peoples vanquished in war to accommodate themselves to servitude when the alternative was the far less attractive option of death, a suggestive statement for ancient mentality at large because it implies a deep and widespread understanding that slavery was always, everywhere, a condition or a fate that anyone might have to confront. It presents an opportunity (still) for further investigation. Meantime, whatever the dangers of reductionism (and I am conscious of them), the essential character of slavery seems clear enough: it was in the phrase of the early imperial writer Valerius Maximus (2.9.5) a ‘bitter chain’, a chain that no one, not even the sympathetic Salvian, ever thought possible to break.


This is the written version of an address first given in 2004 as the Frank M. Snowden Jr. Lecture at Howard University (April), then subsequently, and in slightly different forms, at the Gerder Lehrman Center for the Study of Slavery, Resistance and Abolition at Yale University (May) and the Pruit Memorial Symposium on Slavery, Oppression and Prejudice: Ancient Roots and Modern Implications at Baylor University (October); also in 2005 at the Hite Art Museum, University of Louisville (February) and Miami University, Ohio under the auspices of the John W. Altman Humanities Scholar-in-Residence Program (March). I am grateful to audience members who on each occasion provided stimulating discussion, and to the following individuals who invited me to speak: Rudolph Hock (Howard), David Brion Davis (Yale), John Nordling (Baylor), Linda Gigante (Louisville), and Peter Rose (Miami). The address was prepared for oral delivery, and in an effort to retain its original flavour I have avoided interrupting the text with footnotes, though I have included references to sources. Those who wish to pursue details may may read on. The relevant sections of Salvian may be read in Georges Lagarrigue ed., Salvien de Marseille, Oeuvres Tome II, Du gouvernment de Dieu, Sources chrétiennes no. 220 (Paris 1975); for English translation, see Eva M. Sanford, Salvian: On the Government of God (New York 1930). For the contemporary historical context, see R. Samson, ‘Slavery, the Roman Legacy,’ in John Drinkwater and Hugh Elton eds., Fifth-Century Gaul: A Crisis of Identity (Cambridge 1992), 218-227. On slavery and sin see the classic work of David Brion Davis, The Problem of Slavery in Western Culture (Ithaca, NY 1966). There are two new editions of Jérôme Carcopino’s Daily Life in Ancient Rome: The People and the City at the Height of the Empire (edited by Henry T. Rowell, translated from the French by E. O. Lorimer): one ‘With a New Introduction and Bibliographic Essay by Mary Beard’ (New Haven 2003), the other ‘Introduced by Keith Hopkins’ (London: The Folio Society 2004 [quoted 69]). The important series of articles to which I refer by Susan Treggiari is as follows: ‘Domestic Staff in the Julio-Claudian Period,’ Histoire sociale/Social History 6 (1973) 241-255 (quoted 250); ‘Jobs in the Household of Livia,’ PBSR 43 (1975) 48-77 (quoted 64); ‘Family Life among the Staff of the Volusii,’ TAPA 105 (1975) 393-401; ‘Jobs for Women,’ AJAH 1 (1976) 76-104; ‘Lower-Class Women in the Roman Economy,’ Florilegium 1 (1979) 65-86; ‘Questions on Women Domestics in the Roman West,’ in Schiavitù, manomissione e classi dipendenti nel mondo antico (Università degli Studi di Padova, Pubblicazioni del Istituto di Storia antica 13, Rome 1979, 185-201); ‘Contubernales in CIL 6,’ Phoenix 35 (1981) 42-69 (quoted 61). On slaves and work see also Sandra R. Joshel, Work, Identity and Legal Status at Rome: A Study of the Occupational Inscriptions (Norman 1992). See John R. Clarke, Art in the Lives of Ordinary Romans: Visual Representation and Non-Elite Viewers in Italy, 100 B.C.-A.D. 315 (Berkeley 2003), 78-81, 125-129, 203-206 for the items resumed (81 and 129 for the quotations), and 145-152 for Storax. Edward Courtney, A Companion to Petronius (Oxford 2001), 10 (quoted), 75 speaks of Trimalchio’s bluster. On Seneca and slavery the fundamental work is Miriam T. Griffin, Seneca: A Philosopher in Politics (Oxford 1976), 256-285. On Martial see in detail M. Garrido-Hory, Martial et l’esclavage (Paris 1981). See Peter Garnsey, Ideas of Slavery from Aristotle to Augustine (Cambridge 1996) for his explanation of the lack of abolitionism in antiquity, and complementarily, for the view that Stoic providentialism may have played a role, P. A. Brunt, ‘Marcus Aurelius and Slavery,’ in M. Austin, J. Harries, C. Smith eds., Modus Operandi: Essays in Honour of Geoffrey Rickman, BICS Supplement 71 (London 1998), 139-150. On judicial violence see Ramsey MacMullen, ‘Judicial Savagery in the Roman Empire,’ in Changes in the Roman Empire: Essays in the Ordinary (Princeton 1990), 204-217. A. N. Sherwin-White, The Roman Citizenship (2 &supnd; nd edit. Oxford 1973, 1 &supst; st edit. 1939) is the classic work on the secret of empire to which I refer. The essential works on freedom are Ch. Wirszubski, Libertas as a Political Idea at Rome During the Late Republic and Early Principate (Cambridge 1950) and P. A. Brunt, ‘Libertas in the Republic’ in The Fall of the Roman Republic (Oxford 1988), 281-350. On elite ideology see D. C. Earl, The Moral and Political Tradition of Rome (London 1967). For the ‘warrior state’, and much more, see Keith Hopkins, Conquerors and Slaves: Sociological Studies in Roman History Volume 1 (Cambridge 1978), 32. For Ostia see Russell Meiggs, Roman Ostia (2 &supnd; nd edit. Oxford 1973), 226. For Finley’s statement see M. I. Finley, ‘Slavery and the Historians,’ Histoire sociale/Social History 12 (1979) 247-261, quoted from Moses I. Finley, Ancient Slavery and Modern Ideology, expanded version, edited by Brent D. Shaw (Princeton 1998), 308.

This work is licensed under a Creative Commons License.