Slaves and Freedwomen

Slaves and Freedwomen

Although the Romans are attested as possessing slaves from an early period, slave-ownership on a large scale seems to have developed only with the expansion of Roman power, from about the latter half of the third century B.C.Slaves became an integral part of Roman society, and this is reflected in the range and volume of references to slavery in legal texts.1 ‘Outside the region of procedure, there are few branches of the law in which the slave does not prominently appear’ (Buckland 1908: v); this is an inevitable consequence of the slave’s being both an item of property and also a person, with human qualities, capable of certain actions, yet lacking full legal capacity.

Most of the law relating to slaves could apply to persons of either sex, and need not be separately discussed here, and in practice female slaves were less likely than men to be employed in certain areas of activity, e.g., as bailiffs or agents (vilicae, adores) in rural estates. Although male slave bailiffs (vilici) and adores are frequently attested in inscriptions, many along with their ‘wives’, the few vilicae mentioned are likely to have been the slave-consorts of male bailiffs rather than farm-managers themselves. One was actually a vicaria, that is, a ‘slave’s slave’, part of the peculium of a vilicus. The duties of the vilica, as described by Columella, appear to be essentially those of a kind of housekeeper—supervision of the cleaning, the cooking, the sickbay and, to some extent, the activities, such as weaving, of the female slaves. On the other hand, slave institores, i.e., managers of branches of a business, are rather harder to trace, since they belong to the category of slaves most likely to achieve freedom, but both Ulpian and Gaius take it for granted that institores might be either male or female, and Ulpian says that the sex makes no difference to the question of legal liability, any more than the servile or free status of the institor.²

Those areas of the law in which the sex of a slave was important concern mostly offspring, blood relationships and sexual relations between slaves or between slave and free, and their consequences.

Slaves as Breeders³

As a woman, an ancilla (female slave) would normally be capable of procreation, and slave-owners had an interest in the results. The R Roman equivalent of a Sale of Goods Act, the aedilician edict, laid on a seller the duty to inform a purchaser of any unsoundness or defect (morbus vitiumve) in the merchandise; in regard to slaves, this covered not only physical characteristics, but also mental and moral (e.g., whether a slave had a record of attempted escape, or theft). In Egypt, the most bureaucratic society of the ancient world, sale and transfer of slaves were tightly regulated, and slaves were, it appears, sold, as cars are nowadays, with the equivalent of a logbook, initiated at first sale and attesting successive changes of ownership. One slave girl called Euodia, herself the child of a slave born in Egypt, had, by the time she was 14 years old in A.D. 225, been sold five times. An Egyptian certificate of sale would contain guarantees of absence of defects; some slaves were specifically sold “without guarantee’. At Rome, according to Gellius, slaves sold without guarantee were marked at the point of sale by the wearing of a felt cap.4

Buyers were, to judge by the legal texts, interested in the breeding capacity of female slaves, although this was not the only motive for purchase. If a slave was actually pregnant at the time of sale, this might be held a defect. Vitruvius (de archit, 2.9.1), writing about the weakness of a tree’s wood in spring, draws a comparison with pregnant women, who, he says, are not judged sound until they have given birth, because the foetus is pre-empting the body’s nourishment. For this reason, he says, slaves, when pregnant, are not guaranteed sound. He seems to be thinking of the slave’s impaired capacity for work, rather than the possibility of her death in childbirth. According to Ulpian, (D. 5.3.27 pr.) non temere andllae eius rei causa comparantur ut pariant, ‘slave-girls are not generally acquired as breeders.’

This remark was made, however, in the context of a discussion of the ownership of offspring; as far as the law of sale was concerned, Ulpian saw no reason to insist on the buyer’s being informed if a slave was pregnant. ‘Everyone agrees,’ he says, ‘that she is healthy, for it is the chief and special function of females to conceive and to gestate.’ One who had just given birth could also be sold as sound, unless some additional complication had affected her health.⁵

However, even if breeding was not the only consideration, the amount of detailed attention paid to the matter by jurists suggests that buyers did regard a capacity for childbearing as forming at least part of the specifications of a sound and saleable slave. Jurists on the whole seem to have sided with the sellers rather than the buyers, in giving the opinion that barrenness was not a notifiable defect. Labeo thought that a woman found to be barren, like a barren sow or a male slave found to be a eunuch, could be returned to the seller, if the condition had not been made known before sale. In this he went against the general trend. Unless some physical abnormality was present, a woman who was barren, according to Trebatius, or one whose babies were always stillborn, according to Sabinus, was not to be regarded as unsound—presumably because they were still ‘of merchandisable quality’ for the other purposes for which slaves could be used. This applied, however, only if the woman was congenitally barren, according to Trebatius; if the barrenness arose as a result of some physical ailment she had developed, then she was to be regarded as defective, and declared as such by the seller. A woman ‘too narrow to become a woman’ was not regarded as sound, nor one whose periods were either too frequent or non-existent (unless, in the latter case, because of her age).⁶ Whether the aediles made provision in the market for medical examination or whether, as is more likely, they expected the sellers to provide information is not known.

Sometimes, though, females may have been bought specifically with regard to their breeding potential. If a female slave was sold and the seller had allowed the purchaser mistakenly to believe that she was a virgin, the sale was not void, but the buyer had an action against the vendor. The reason for this ruling is not stated. The supposition was most likely to arise when young girls were sold. The buyer’s interest may in some cases have been primarily sexual (e.g., as a leno stocking up a brothel), but many buyers may simply have preferred to acquire a female as near the start of her reproductive life as possible.⁷ Unfortunately, little evidence has survived on the age of female slaves at sale. Bradley (1978) examined 29 examples from papyri of female slaves sold in the age range from four to 35 years. Of these, seven were under 14 and only four over 30, while the average age was 22 to 25. On the basis of this evidence, he suggested that the breeding of slaves was a prime consideration in the sale of females in the Roman empire. He did not, indeed, suggest that it was the only consideration; however, the sample is a very small one on which to base conclusions applying to the empire as a whole.8 In Egypt, it seems that the rearing of slaves from infancy, not merely those born at home but foundlings as well, was a frequent undertaking, and so presumably it was regarded as financially worthwhile. Biezunska-Malowist (1962:158 ff.) has suggested that, specially for lower-class owners seeking additional money-earners, it was preferable to rear slaves and have them trained in a skill, rather than to buy at greater expense slaves already trained. This spread the cost. In large urban familiae, such as are attested at Rome, it made sense to rear home-bred slaves and make use of the training facilities afforded within the household’s own ranks of skilled workers. The babies were sometimes sent by wealthy owners to their country estates to be reared in infancy by other slave women, while their mothers presumably continued with their duties in the urban household.⁹

On rural estates (where female slaves were perhaps scarcer than in urban familiae), writers such as Varro and Columella recommended that females should be encouraged to have children by the offer of rewards, such as exemption from work for the woman who had had three children, and even manumission for those who exceeded that number.¹⁰ It is suggestive that in the Delphic manumission contracts analysed by Hopkins and Roscoe (Hopkins 1978:133–71) 80 per cent of the slaves freed as children were freed separately from their mothers, and probably after them. Some manumission contracts for women specified that children born to them (either all or a specified number) during the term of service of their ‘conditional release’ (paramone) must remain the manumitter’s property, in order for the woman to secure full release; some relatively generous owners specified in the contract that the children were free, but whether they were the majority is not known.¹¹ Paramone was a Greek practice, not to be confused with the operae Roman owners were allowed to require from freed slaves. A closer parallel in Roman law is the situation of the statuliberae. These were slaves to whom freedom had been given by the will of their late owner, subject to the fulfilment of a stated condition; until the condition was fulfilled, they remained slaves, and so did any children to whom they gave birth. The latter became the property of the heir.¹² Sometimes, at any rate, the condition of liberty seems actually to have been the production of a specified number of children. Two jurists, Ulpian and Tryphoninus, gave their opinions on the difficult case (possibly a ‘textbook’ example) of a woman, called Arescusa, given freedom by a will, ‘if she gives birth to three children’. She had, according to Tryphoninus, one child, and then triplets; according to Ulpian two children, then twins. They agree that the last-born, the fourth, child should be free, since the condition was fulfilled before it actually left the mother’s body. The matter,’ says Ulpian, ‘is one of fact, not law.’ Tryphoninus hypothesises a further complication. The mother’s freedom being dependent on the payment of a sum of money, this was paid by a third party just as she was in the process of giving birth. The child, he says, should be deemed free. In some paramone contracts these were specified as alternatives—either children or a specified sum of money as the price of freedom.¹³

Julian took up the question of what should be done where the heir wanted to retain the woman (now called Arethusa), and so tried to prevent her giving birth, either by administering a contraceptive or by making her have an abortion. Julian’s opinion (D. 40.7.3.16) was that she should be given her freedom at once, since she could have had triplets at one birth. This may be merely an imaginary example, but we should not be justified in assuming that no Roman owner was capable of such physical violence against a slave; and in a case of such a kind there is quite likely to have been a third party, namely the prospective father, interested in securing the woman’s freedom.

Women slaves had no automatic right to freedom after the production of a given number of children; that was solely dependent on the generosity of the individual master, and such manumissions may have represented a small proportion of the whole. When a paramone contract specified that the woman supply either money or children, in effect, the freed slave was being required to furnish her owner with a replacement. One Delphic contract specifies that a replacement shall be purchased. In one curious example, the manumitter actually saw children as a drawback. His interest was in retaining the slave’s services to look after his mother as long as the latter lived. Any child born to the servant during that time was to be free; he did not want to keep it. She could rear it or kill it, as she liked, but she was not to sell it (and so, presumably, raise the means to buy her freedom).14

Slaves as Livestock—the Offspring

Like other kinds of property, slaves could produce ‘fruits’ (fructus). These might be, in the more abstract sense, ‘profits’, such as earnings from work; however, those which a female could produce (i.e., children) were more directly comparable with the yield of plants and the natural increase of animals. The latter were ‘fruits’ under the law of usufruct, and were assigned to the person who held the right of usufruct. Did this apply to the partus ancillae, the children of a slave woman? Roman law was, from at least as early as the second century B.C., certain that it did not, though the Romans were less certain why it did not.¹⁵

In the second century B.C., the jurist Brutus stated the view that there was no usufruct of the partus, on the grounds that they were human (homines) and ‘Man cannot be the fruit of man.’ Ulpian was puzzled since (a) as Sabinus and Cassius recognised, the offspring of animals could be the ‘fruit’ of animals and, indeed, had to be used to make good deaths in the herd held by the usufructuary (as—in the case of dotal property—by the husband) and (b) both slaves’ children, and the usufruct of such children, could be the subject of legacies. Ulpian’s own explanation was that female slaves were not normally purchased for breeding purposes and therefore their offspring were not ‘fruit’. Gaius’ explanation, which, as Kaser (1958:158–9) rightly observes, is mere philosophising, was that, as the fruits of the rest of nature were made to serve the interests of man, it was absurd that man himself should be a fruit.¹⁶

Kaser himself suggests the derivation of the rule, not from any humanitarian principles or ‘respect for human dignity’ (Buckland 1908:21), but from the general principle that it was desirable that a slave child should stay with its mother, for the practical reason that she was best able, in the child’s early life, to look after it. This explanation, while fitting well enough short-term grants, which would restore the mother to her and her child’s owner quickly, does not apply particularly well to usufructs lasting over long periods of time, such as those bequeathed to widows during their lifetime, or until their children came of age. Since the slave mother would not be due to return to her owner until the end of the usufructuary period, she and her child could be separated. In any case, as Kaser himself recognised, mother and child could always be, and not infrequently were, separated by sale.

Whether or not, in law, mother and child remained with the same person depended, in practice, on where the full legal title of ownership rested. Where there was merely usufruct or a conditional title, human offspring were distinguished from those of animals and from crops. A husband’s title of ownership of a dotal slave was conditional on the continuance of the marriage. If the dowry fell due for return, any children the slave had had must go back with her—unless, that is, he had held the dowry only ‘at valuation’, in which case he could, if he wished, hand over only the agreed sum. However, he need return only an equivalent number of animals to those originally received, and could keep any increase.17

Matters were more complicated in those situations in which the transfer of ownership was of doubtful validity, e.g., sale of a pledged or stolen slave, or sale in default of creditors. Two elements which jurists appear to have considered important and relevant were the time of conception, and the ‘good faith’ of the purchaser, i.e., whether at certain material times he had believed the purchase to be lawful and valid.

The offspring of animals were, in effect, treated like other products of the animals, such as wool or milk; so, for example, the offspring of stolen animals belonged to the bona fide purchaser, under the title of ‘fruits’, regardless of when they were conceived.¹⁸ If a slave had been stolen, legal title both to her and to any children subsequently born belonged to the rightful owner, regardless of when they had been conceived. If, however, the thief had sold her, then whether or not the bona fide purchaser could claim any children born while she was with him depended on when she had conceived. If she conceived before the theft, or after theft but before sale, the offspring belonged to the original owner; if after sale, she still belonged to her original owner, but the position regarding the child conceived after the sale is less clear. Most jurists seem to have held that the bona fide possessor could usucapt the child, providing that he still believed the purchase to have been genuine at the time the child was born.¹⁹

As we have already seen, the notion that a child had, in some sense, a legal existence from the time of conception, was of some importance in determining matters of status and of inheritance rights, and this was applied not only to the child of a free woman but of one who at some time during her pregnancy was, or should have been, free. It is less obvious that it could be held to have any relevance where what is involved is no question of status but merely of legal ownership, i.e., where the slave is being considered primarily as an item of property. Time of conception was not held material when a slave woman was sold in the ordinary way; children born after the sale belonged to the buyer. The principle was that prenatal existence should be considered only where that was to the expected child’s benefit; Buckland (1908:22) speaks of’a modification of this in favour of the owner of the ancilla at the time of conception’. This will not do; the bonafide possessor was not the legal owner of the ancilla at any time, yet he might be allowed to usucapt the child.

The time factor was important, but surely in a different way from that supposed by Buckland. A slave’s child was partus, noifructus, and belonged to the owner, not the possessor. However, one year’s uninterrupted bonafide possession of movable property gave a right of ownership.20 If a slave woman could be shown to have conceived after theft and sale, then by the time the child was born the best part of a year would have elapsed. The jurists were probably trying to find ways of ensuring that an innocent purchaser of stolen property should derive some benefit for his expenditure. Fructus he could have; that was easy. In allowing usucaption of children in the circumstances specified, they were stretching a point in the interests of equity. The interests of the child were not the issue; all that was being decided was the legal title to ownership.

An explanation is still needed of the distinction between Truits’ and partus ancillae. Buckland’s ‘respect for human dignity’ is not particularly apt. Both humans and animals could be bought, sold, bequeathed and pledged and, where there were no complicating factors, legal title to the offspring of either, born after the transaction was complete, vested in the new owner. The offspring of slaves, as well as the offspring and produce (e.g., wool) of animals, could be bought, sold and pledged in advance. Paul, commenting on the aedilician edict, describes one type of fraud practised (D. 19.1.21 pr.): ‘If the partus (i.e, prospective offspring) of a woman is sold, and the woman is barren or over fifty years old, and the buyer was not aware of this, he has an action against the seller.’

One might perhaps develop a hint from Ulpian’s remark already cited, that since slave women were not normally acquired for breeding purposes, their offspring were not regarded as ‘fruits’. Animals were bought or their usufruct (effectively, a kind of leasehold) retained for the sake of their bodily produce—meat, wool or hides, milk and young—while slaves were acquired primarily for their services. Animals were essentially consumables, and interchangeable. Slaves were not. Their qualities and abilities were many and various, not always predictable or susceptible of simple numerical quantification. Moreover, slaves were a great deal more durable than most farm animals, even the larger ones. No ancient jurist seems to have developed an argument along these lines, and we have no evidence what reasons, if any, were adduced when the distinction was first drawn. Nevertheless, something like this may be close to the truth. Slaves were simply too valuable to be treated like lambs.

Family Relationships among Slaves

Slaves could not contract legal marriages and have legal children; nevertheless, some, particularly in larger households, could and did form lasting paired relationships, have children and behave, for a time at least, as a family. Since their continuance was entirely dependent upon the slave’s owner, who might separate the couple from each other or from their children by sale or bequest, and grant or refuse manumission at will, the stability of these relationships was never to be relied upon. What proportion was stable cannot easily be determined. Different kinds of evidence present different facets of the situation.

In the Delphic manumission records of the last two centuries B.C. the husband-wife relationship is explicitly attested only once, as is that between father and child. The mother-child relationship appears more frequently; there are 29 examples of their joint release. At the same time, as we have seen, mothers might be severed from their children, surrendered in partial fulfilment of conditions for their release. Even allowing for the fact that relatives might sometimes be released together, without the relationship expressly being stated, the overall high ratio of adults to children, together with the preponderance of single over multiple releases, makes it likely that family members were often manumitted separately, and some families may never have succeeded in freeing both parents and children.21 Not all masters were so generous as the author of the will cited by Scaevola (D. 34.1.20 pr.) who manumitted Stichus, grandson of his old nurse, and bequeathed to him his contubernalis and their children (so there were at least four generations in that slave family). Stichus would then be able to free them, and he, and not his patron’s son or daughter, would have patronal rights over them.

Similar traces of separate manumission appear in the small number of inscriptions from Rome, noted by Flory (1984), in which the wife had achieved freedom while the husband was still a slave (though some of these were imperial slaves); of the children mentioned, the inscriptions mostly do not give their status at birth, but one-sixth were still slaves.²² Where, as sometimes happened, specially in larger establishments, freed staff continued work, and possibly residence, in the household, the effects would be less keenly felt. Bequest or sale was more likely to impose actual physical separation.

Some papyri reveal the existence of ‘families’ stretching over two or three, even four, generations in a single household. Usually, however, the relationship is specified only through the mothers. In bequests of slaves, also, where family relationships are stated, they tend to be between mother and child. Not infrequently the father will have been the woman’s owner (or her owner’s husband).²³

The largest single collection of evidence for slave marital couples is the funerary inscriptions from the city of Rome itself. Here, if free husband and wife have the same gentilicium, there is at least some presumption of a previous slave (or master-slave) link; the presumption is strengthened if the couple are expressly stated to be freed, and strengthened still further in the relatively small number of instances where some such word as colliherta or contubernalis is used. There are also some slave couples.²⁴

Slave ‘marriages’ probably had greater chance of stability in very large slave households, or where the ‘husband’ was living apart from his owner, as institor, vilicus or the like, and especially if he had slaves in his pecuiium. Parent-child relationships are harder to trace. The children appearing in the inscriptions of freed couples are mostly freeborn; the others may have been left behind, changed owners (and so have different names if manumitted) or lost contact in other ways.²⁵

Keeping in touch may have been easier in the large noble and imperial familiae, or in the countryside, than it was in towns generally, specially in Rome. We find, for example, a freedwoman of Domitian’s wife, Domitia Athenais, associated with her slave brother Januarius in burying their father Phoebus. The funerary collegia of the great families probably played an important part in maintaining contact between the freed and their slave relatives.²⁶

One mother belonged successively to three households, of the Modii, the Claudii and the Gavii, and left a son in each, before being bought by the Volusii. Freed, she became, as Volusia Justa, the wife of P.Volusius Renatus. Her three sons, all still slaves but now in public service, two with the priestly college of the septemviri and one with the fetiales, set up her epitaph (CIL VI. 2318, with Mommsen’s note).

Strictly speaking, these familial relationships between slaves had no existence in the eyes of the law. References in the legal texts to wives and children of slave bailiffs on estates are essentially terms of convenience.27 The whole concept of the family rested on conubium and potestas, and slaves had neither. Since, however, slaves on manumission moved into free society, some cognisance had from time to time to be taken of these relationships, at least among the freed.

The rules of dowry did not apply between slaves, since slaves could own no property. However, if a slave woman used her peculium as ‘dowry’, to give to a male slave, and the couple both obtained manumission and continued in the marital relation afterwards, then the law tacitly recognised its conversion into dotal property—always providing, of course, that the former owner had consented to the retention of the peculium. Shtaerman and Trofimova (1975:234) mistakenly take this as amounting to a recognition of slave marriage, and think it an innovation dating to the Antonines or Severi. Buckland (1908:199) rnisleadingly talks of ‘an effective transfer from one peculium to the other’. In fact, there was no marriage until both were free and continuing to cohabit; and if either had been freed without the other, the master was under no obligation at all to allow the ‘dowry’ to remain in the male slave’s peculium. The point is made even clearer by another reference in the Digest to slave ‘dowry’. If a slave woman, passing herself off as free, had ‘married’ a free man, the law was quite clear that the ‘dowry’ did not become his property but remained that of her owner. (More will be said later about some of the incidental legal consequences of slave-free relationships.)²⁸

The laws of inheritance did not apply between slaves. Twice, however, in the Digest we find the natural relationship being recognised in judgment of an inheritance case, both times apparently in the absence of any freeborn children. The more complicated case, possibly a textbook example, tells how Pamphilus was manumitted by his mistress’s will and left a half-share in a piece of land, with the instruction that the land be transmitted to his children. When he died, his daughter was still a slave, and he instructed his heir to pass on the land to her when she should become free. The question exercising Scaevola was this: under which will, when she became free, had she a claim to the property? If her father’s, she had to give the heir the ‘Falcidian fourth’ of the land’s value. Scaevola’s response was that the recognition of the natural relationship satisfied the intention of Pamphilus’ former owner.²⁹

This, however, was a matter of interpretation, rather than of legal right. Labeo thought that the provision of the praetor’s edict that the children should not bring suit against their parents should be interpreted as applying to those who had had children while in slavery, as well as to illegitimate children generally. Only mothers, it seems, not fathers, could benefit from this protection. Paul explains: the identity of the mother, even of an illegitimate child, was certain, whereas the pater was the person designated by marriage. The basis of this interpretation was the reverence due to all parents.³⁰

When it came to marriage by freed persons, however, blood relationships among slaves (servilis cognatio) were recognised in law. Marriage between freed persons was banned if, as slaves, they had stood in the prohibited degrees of relationship to each other, since, as Pomponius explains (D. 23.2.8), the ban on ‘incestuous’ marriages derived originally not from law but from custom (and so, presumably, legally non-existent relationships could be recognised).³¹ Nor could a man marry his son’s or his fathers former slave mistress, since she stood in the relation of daughter-in-law or stepmother, respectively. Her mother was excluded as well. This rule must go back at least as far as the possibility of conubium between freeborn and freed.

Certain blood relationships were also taken into account in the lex Aelia Sentia, which included among the ‘good cause’ exemptions from the prescribed lower age limits (20 for owner, 30 for slave) that the slave was the natural son or daughter, brother or sister or (according to Ulpian) cognate of the manumitter.32 Clearly, this could apply not only to natural children of the manumitting patron by his ancilla but also to slaves who, having been manumitted themselves, had succeeded in buying, or perhaps had in their peculium and were about to manumit, their own relatives.

However, recognition of servilis cognatio was accorded in classical law mainly where the interests of free, or freed, persons were concerned. Slaves could always be separated from their relatives by sale or bequest as the owner thought fit. Ancillae were sometimes bequeathed along with their children, but legacies of unborn slave children, as of future increase in livestock, or wine as yet unproduced at the time the will was made, were valid, and mother and child would not necessarily be left to the same person.³³

Paul and Ulpian both tell us that a legacy of an estate ‘with its equipment’ includes the contubernales or ‘wives’ and children of the male slaves; Paul adds, ‘and livestock kept to provide manure, with those who tend them, are included.’ This, however, is merely a matter of enumerating the persons and goods actually on the estate (in eadem villa agentes). Ulpian, to be sure, adds, ‘It is incredible that the testator wished to impose a cruel separation’, but mere presumption of intent such as this carried no binding legal force. According to Scaevola, the response was given in one instance that when a business manager (actor) living in town was made the object of a legacy, there was no reason to suppose that his wife and daughter were included.³⁴

Family groups were sometimes sold together. If the buyer wanted to return one of these as unhealthy, it was ‘usual’, says Ulpian, for the whole group to be returned, so as not to offend proper feeling (pietas); this ‘ought’ to apply to parents and children, brothers and contubernales.³⁵

This is not the language of compulsion. Perhaps a little more firmness is detectable when the relationship crosses the barrier between slave and free. When a man’s goods are seized and sold up for debt, says Paul, his concubine and natural children are not included. Even by the time of Constantine, recognition has gone no further than the emperor’s pronouncement that when an intestate estate is being divided, family groups ought to be kept together: Tor who would bear the separation of children from parent, brothers from sisters, husbands from wives?’³⁶

Clearly, by the time of later classical law there was a certain amount of sympathy, in theory at least, though we cannot say to what extent in practice, for the personal ties of slaves; but they had been accorded no legal rights in respect of these relationships against the specific dispositions of an owner.³⁷

Slaves sometimes cohabited with free men and women, not their owners. The consequences for the free woman’s children and, under the senatuscomultum Claudianum, for her own status, have already been discussed. The children of a slave woman belonged to her owner. The free partner might be unaware of the other person’s servile status—most likely because he or she was a runaway. The marriage was void, but if genuine mistake could be proved, some of the other consequences might be avoided. In general, however, the law did not go so far in avoiding penalising the innocent victims of mistake as it did in the case of liaisons of citizens with free nonRomans (G. I. 67–75). The free woman’s child had her status. Until the time of Vespasian, according to Gaius (G. I. 85), there was a law by which the child of a free man and an ancilla whom he had believed free was free if male, but if female belonged to the woman’s owner; though Gaius does not say so, the son probably had only peregrine status. Vespasian abolished the law; the children of both sexes were to be slaves. This is consistent with the judgment, attested for the second century A.D., that the runaway is a “thief of herself and of her offspring, and usucaption is excluded (D. 47.2.61); it is also consistent with the lex Minicia.

Money or property might have changed hands as ‘dowry’. It did not become the property of the free ‘husband’ of a slave woman passing herself off as free, but remained that of her owner. On the other hand, if a free woman knowingly ‘married’ a slave, one would expect the dowry to be irrecoverable. Since there was no legal marriage, there was no dowry, and she had in effect made a gift to the slave of another, which, in law, was a gift to his owner. There were ways round this. Paul cites the curious case of Lucius Titius, who gave his daughter in potestate, Seia, in ‘marriage’ to a slave. The dowry he gave in the form depositum sub cautione—that is, not making a transfer of the ownership, but merely entrusting it to the slave’s care, subject to return. The father died, and so did the slave husband, without his owner having challenged the arrangement. Paul gave judgement that, as no dowry could have been constituted, Seia as her father’s heir was entitled to an action for return of the deposit from the peculium. Honest mistake, however, may not have been penalised by loss of the dowry property, especially if there were children. The emperor Antoninus assured Hostilia, who had married Eros in ignorance of his slave status and had children by him, that she could recover the dowry from his peculium.38

Flight, Abduction and Subversion

It was, it seems, not easy to prevent sexual relations between slaves and outsiders. Relations with male slaves gave concern mostly when they interfered in some way with the well-being of another household or with someone else’s property, although the late Republican jurist Alfenus Varus reported that he was approached for advice by an owner whose steward’s accounts did not tally because he had been spending his master’s money on a ‘popsy’ (muliercula). Varus advised him that he could sue her both for corruption of a slave and for theft of the money. Relations between free women and slaves were punishable as adultery (if the woman was married), but it was not until the sctum. Claudianum that any issue, and sometimes the woman herself, were liable to become slaves.³⁹

Relations with female slaves (or, indeed, homosexual relations between outsiders and male slaves) were regarded most seriously when they represented a threat to the security of their owner’s familia. When considering a complaint of corruption of a slave, deflowering of a slave girl or homosexual relations with a male slave, a magistrate should, says Paul, exact very severe punishment ‘if the perpetrator is corrupt or such a man that his act seems to go beyond the immediate damage and involve the overthrow of the whole house’. What Paul may be talking about here is alienation of the slave’s affections and the danger this might sometimes represent.40 Sexual relations with an ancilla belonging to another man was sometimes called stuprum, but Paul remarked that it did not constitute an offence unless it was used as a means of access to her mistress, or unless it made the slave ‘worse’. Rape of a slave girl was, as we have seen, punishable as damage to an owner’s property, not as an injury to the girl herself.⁴¹

Runaway slaves, of both sexes, were a perennial headache for the Romans. Possibly the women without the responsibility of children were the more likely to run away, but they could, of course, have children while at liberty, and the law was firm that these belonged to their mother’s original owner.⁴²

Slaves were valuable property, and slaves of both sexes were liable to be stolen. As we have seen, the law had a good deal to say about the legal position when a slave woman was resold by the thief, and about the legal ownership of any children she might have while away from the owner. Whether female slaves were more likely to be stolen than male is a matter for speculation. Their potential as breeders, and perhaps in prostitution, makes it a possibility. lavolenus, a jurist of the time of Domitian, specified a female when considering the respective rights of action of an owner and a subsequent purchaser, both of whom had in turn suffered the theft of the same slave. There are in the Digest four references to theft of a pregnant slave.⁴³ Male slaves, being stronger and less easy to handle, might be a more difficult undertaking to steal, though their possession of specialised skills might sometimes make it worthwhile.

Two texts in the Digest mention the payment by a slave to his master of a stolen ancilla as part of the price of his freedom.⁴⁴ We should perhaps think here of an urban context, in which the male slave was living away from his master and able to amass a peculium, even possibly to acquire slaves of his own, so that he could pass the woman off as legally bought. She may have been, of course, so far as he knew.

A slave girl might be abducted for sexual purposes. The law took this very seriously —unless, indeed, the,girl was already a prostitute. In which case, Ulpian comments, ‘it is not theft, for it is not the act but the motive which is in question and the motive was lust, not theft… nor is he liable under the lex Fabia if he concealed a whore for sexual purposes.’⁴⁵

More about prostitutes later. If the slave girl was not a prostitute, lawyers advise that her owner might muster the resources of the lex Fabia, the lex Aquilia and an action for theft against the man involved.

The lex Aquilia was an action for damages; loss of virginity is particularly specified. The lex Fabia, of uncertain date, but possibly going back before the beginning of the second century B.C., created the crime of plagium, or ‘kidnapping’, either of free persons or of slaves. Anyone who persuaded a slave to flee or concealed, kept under restraint, or knowingly bought or sold a runaway, was liable to a penalty which by the time of Ulpian was fifty thousand sesterces. Later, offenders of lower rank were sentenced to punishment in the mines, the higher-ranking to the loss of half their property.46

What the Romans were worried about was perhaps not only the possibility of physical harm to their property or a reduction in its worth but also the subversion of a slave’s loyalty to his or her master (D. 1.18.21). After all, not all sexual activity was necessarily rape. The slave may have been a consenting party. As for abduction and concealment, we may suspect that in many cases the slave girl was not abducted but was a runaway, who may even have eloped in order to go and live with the man in question. This may be one reason for the distinction drawn by Ulpian. If the girl was a prostitute, the man’s intention may have been to detain her only temporarily for his short-term gratification; but if she was an ordinary slave girl, eloping and hoping to set up in conjugal life and pass as free, the absence was intended to be permanent.

Sexual Use of Slaves

Owners could and did have sexual relations with their own female slaves.⁴⁷ The latter were scarcely in a position to refuse, but it need not be supposed that they were always unwilling. Literary sources speak, not only of the debauchery of a lord of a harem, like Martial’s Quirinalis (I. 84), but of the famous slave or freed mistresses of emperors and nobles, and inscriptions attest not only to the freedwoman concubine, but to the freedwoman wife of her patron. As we have already seen, such liaisons were not necessarily illregarded (though marriage was banned, after Augustus, with the senatorial class).

A double standard operated. Women owners could be rather less free in acknowledging slave amours, specially if they themselves were married. Sex with a slave counted, for a married woman, as adultery under the terms of the Julian law.⁴⁸ For a man, sex with slave girls did not count.

Many prostitutes were slaves. Their owners suffered the penalty of infamia but otherwise were permitted to operate their business undisturbed by the law. A slave leno, prostituting the slaves of his peculium, would not be affected by infamia at all, until he was freed. Severus specified in a rescript that a woman who had given up prostitution on manumission was not liable to infamia. It is not until the end of the second century A.D. that we hear of any official intervention to protect the slave whose master compelled her to prostitution. Hadrian forbade the sale of slaves to bawds or gladiatorial schools without cause being shown—that is, he did permit it as a punishment. Under Pius, magistrates were given a duty of enquiring into any complaints of ill-treatment made by slaves who sought sanctuary in temples or at the emperor’s statue, and among the grounds was infamis irimria, a rather unspecific term, which may refer to various forms of sexual outrage. Severus laid on the urban prefect the duty of hearing any complaints by slave women who had been forced into prostitution by their masters. Under the Christian empire, the masters were, in addition, severely punished.⁴⁹

However, there seems to have been a feeling that the business of prostitution ought, at least, not to be encouraged. This feeling may partly have underlain Ulpian’s view that a lenient view should be taken of someone’s abducting or harbouring a slave prostitute. According to Callistratus, also active under the Severi, if a prostitute was manumitted, her patron should not expect to derive any profit from her, since only those services to patrons (operae) were considered due which were not morally undesirable.50

Some owners at least were opposed to the prostitution of slave women and when they sold them stipulated a penalty, usually a right of seizure, or freedom for the slave, if the new owner did prostitute them. Several emperors, particularly Vespasian and Hadrian, took an interest in the matter, and insisted on the enforceability of these clauses. Even if the woman had been sold several times before the condition was breached, it became effective whenever she was prostituted, and the original seller benefited, either by right of seizure and recovery, or as patron on manumission. Marcus Aurelius may even have allowed the woman her freedom by virtue of a nonprostitution clause, even if liberty was not specified as a penalty on the buyer. Alexander Severus ruled that the device of employing the woman ostensibly as a waitress was a mere evasion, and could not avert the operation of the penalty.⁵¹

The effectiveness of this protection for the woman would, of course, depend on the thoroughness with which her previous history was known to, or investigated by, subsequent purchasers. As conditions of sale were not uncommon, it would be sensible for any prospective buyer to make careful enquiry.

Manumission

Of the three principal methods of manumission under Roman law, one, inscription in the census list, applied only to men, and was probably little used after the Republic. The common formal method of manumission during the owner’s lifetime was vindicta, which was modelled on the procedure for allegations of wrongful detention in slavery. Since it was a fictitious lawsuit, the presence of a magistrate was necessary, and the outcome was a foregone conclusion; agreement had already been reached between slave and owner on such matters as payment, services, etc. The other method recognised as legally valid was manumission in an owner’s will. The most important feature of both forms was that a slave so freed became a Roman citizen. Any other type of informal manumission left slaves, under the Republic, in practice at liberty, under the protection of the praetor, but legally still slaves, without property or legal capacity.⁵²

The situation remained essentially the same under the empire, except for the modifications introduced by the Augustan legislation, the lex Fufia Caninia of 2 B.C. and the lex Aelia Sentia of A.D. 4.⁵³ The former limited the number of slaves who could be manumitted under a will; the latter imposed various restrictions, such as minimum age limits. The effect of both was to create new categories of people who were not in fact legally manumitted. Some improvement in their situation was afforded them by the lex Junia (or Junia Norbana), passed possibly not until the principate of Tiberius (?A.D. 19), which gave them free status, not as citizens but as Latins.⁵⁴

Whether such was the intention of the legislators or not, male slaves achieving ‘Junian’ latinity were in a better position than female. A male Latin taking a Roman or Latin woman as wife could obtain citizenship for himself, his Latin wife, and his child, by having a child who reached the age of one year (annicult probatio), and he and his wife now also became legally married in Roman law. The treatment of the whole subject by Gaius, the principal legal source, is unsystematic, but it does not appear that this option was available to the Latin woman who wanted to marry a Roman citizen, or to her child. Gaius, discussing conubium, says (I79) that the offspring of a Roman citizen and a Latin woman has Latin status, “in spite of the fact that the lex Minicia does not apply to those who at the present day are called Latins’ (it applied originally to peregrines). Earlier, in discussing exceptions to the rules about status of children, in cases of genuine mistake, he says (I. 67) that if a Roman citizen took a Latin or peregrine woman as wife, unless he could prove genuine mistake (i.e., that he thought she was a citizen), the child would be Latin or peregrine according to its mother’s status, ‘because a child does not take its father’s status unless there is conubium between father and mother’. Again (I. 71), if a Roman man was mistaken about his own status, thinking himself a Latin, and therefore took a Latin woman as wife, he could, on birth of a child, prove a case of mistake and receive the benefit of the lex Aelia Sentia.

The conclusion seems inescapable. A male slave, informally manu-mitted, could nevertheless achieve citizenship for himself and his child, regardless of whether its mother was Roman or Latin. A female could attain legal marriage, citizenship and citizenship for her child, only by marrying another Latin. If she cohabited with a Roman, she was not legally his wife and her child was not a citizen, nor in its father’s potestas; nor would she, as a Junian Latin, be able to transmit property by will to her children (G.I. 23). Whether this was in fact the situation of many freedwomen concubines we do not know, and their nomenclature alone would not necessarily tell us. Whether or not the legislator’s intention was in part to discourage Roman citizens from taking Latin freedwomen as wives, the law possibly gave owners an actual incentive to manumit informally, in order to secure a slave’s intestate inheritance.55

Two later measures may have ameliorated the position of at least some Latin freedwomen. According to Suetonius (Cl. 19), the emperor Claudius as part of his endeavours to ensure the Roman corn supply offered to those who would build large merchant vessels substantial rewards adapted to the condition of each, ‘to a citizen exemption from the lex Papia Poppaea, to a Latin the rights of Roman citizenship and to women the privilege of four children’. Four was the number of children freedwomen must produce to qualify in the ordinary way for the ius lib. The women whose generosity Claudius particularly aimed at tapping may have been wealthy freedwomen (restricted, as we have seen, by having a patron as tutor legitimus); if they happened also to be Latins, they could gain Roman citizenship. The other measure was a senatusconsultum of unknown date which specified that a Latin woman could gain citizenship by having three illegitimate children. This sctum. is unlikely to have been very early; it can scarcely have been passed before illegitimate children were allowed to count for the ius lib. generally.⁵⁶

Obviously, it might be important for a woman to be able to demonstrate either that she was freeborn or that she had been properly manumitted. This adds significance to the efforts of Petronia Justa, already mentioned, to prove herself freeborn. If she had in fact been manumitted, it was probably informally, otherwise Calatoria would surely have been able to produce some proof of it. As an informally manumitted Latina, Petronia might face considerable difficulty making any match with a Roman citizen, to whom she could offer neither legal marriage, legitimate children, nor an inheritance for the children. She brought her case to try to establish her free birth, possibly because there was a Roman in the offing whom she wanted to marry. Given the informality of Roman marriage procedures, some Junian Latin women may have been able, for a time at least, to conceal their true status and find Roman husbands. Discovery was most likely when a child was born and the father attempted to register it as legitimate. His own conscientiousness, or that of the recording magistrate, might require from the woman proof of her own status. All was not yet lost; the child’s legitimacy and citizen status could be saved, and the woman could become a citizen, if the father could prove ‘honest mistake’. But proof that he had not known something might be difficult to establish, and some magistrates might pertinently ask why he had never troubled to enquire.

Whether men or women stood a better chance of manumission is hard to determine, and statistics are lacking. Two substantial collections of evidence do seem to show some bias in favour of women, though it may not be justifiable to take these as representative samples. Alföldy (1972) analysed 1,201 recorded ages at death of ex-slaves of private owners from the western provinces. Three-fifths died before the age of thirty (and so, perhaps, tended to be over-represented in commemorations), and of these three-fifths were female. The other set of evidence is the Delphic manumission records, in which women form the majority.57

On one hand, male slaves tended to have higher-level jobs, and their value as workers might be an inducement to their owners to retain them as slaves; on the other, the prospect of freedom could serve as incentive to the slave to work harder for the master, while at the same time the more responsible and independent slave had a better chance of raising a peculium with which to buy his freedom.⁵⁸ With women, a commoner motive for seeking manumission, probably, was to allow marriage and/or the birth of free children. As the specific jobs (if any) on which women were employed tended to be less lucrative for their owners, less value may have been set on retaining them; moreover, there will quite often have been a man, either the patron himself wishing to marry her, or another wishing to buy her freedom, interested in securing her release, and this may have gone some way to offset her own relatively poorer prospects of acquiring a substantial peculium and the loss to the owner of anyfuture children.

Marriage and Patrons

A freedwoman did not need her patron’s consent to marry, any more than a freeborn woman required the consent of her tutor (unless manus or a dowry was involved). This is shown by the fact that from time to time patrons tried to extract from their freedwomen an oath not to marry.59

The right to operae (services) was lost⁶⁰ if the freedwoman married with her patron’s consent, since her prime duty was considered to be to her husband. A patroness, however, could always require the operae, since her claims were not felt to compete with a wife’s duty. Clearly, the burden of proof would be on a patron, trying to exact operae, that he had not consented to the marriage; the oaths, even if not themselves legally enforceable, may have been meant to try to establish this in advance. If he had not consented to the marriage, it was not void, but he might claim that in opposing his wishes she had failed in the ex-slave’s duty of obsequium (‘respectful compliance’); this was not an offence that normally attracted a very severe penalty.⁶¹ Under the empire, jurists condemned these oaths as contrary to the lex Aelia Sentia; the penalty for the patron was loss of his patronal rights. The practice, however, was probably one that had been going on since the Republic. Patrons had much to lose under the empire, when freedwomen could acquire the ius liberorum and so control over their property and at least partial powers of testamentary disposition in favour of their children. The right to services (operae) might still be worth having, although these were, after the reforms of the praetor Rutilius Rufus (c. 118 B.C.), limited by law and created only by voluntary agreement; before that, patrons had had an almost limitless right to demand the time and services of their ex-slaves.

Whenever her marriage ended, whether by divorce or widowhood, a freedwoman became liable again for any operae that had been agreed. This gave her an incentive to fulfil the intention of the authors of the lex Julia et Papia and remarry as soon as possible. Since the marriage-laws applied only within certain age limits, once she reached the age of 50 she was not compelled to render the services; this was probably from an interpretation of the law, rather than one of its original clauses. If a patron had chosen to keep a freedwoman as his concubine rather than marry her, it was agreed that she ought not to be penalised, but should be exempt from any demand for operae, just as if she were his wife. Some freedwomen might have gone up in the world. They were not expected to go on rendering services once that became unsuited to the dignity of their position.

Marriage to her own patron placed the freedwoman under a certain restriction. If she divorced him without his consent, she was violating her duty to a patron. The lex Julia on marriage recognised the divorce but refused her conubium with anyone else, and she could not bring an action for recovery of dowry.⁶²

If she had been freed expressly in order to marry her patron, he could compel her to do so, even against her will. Legal opinion sided with the patron; if she married anyone else first, without the patron having renounced his right, the condition of liberty was not fulfilled and she was not free. However, he was not allowed to use this as a device for preventing her marriage to anyone else. If he freed her for the purposes of marriage under the lex Aelia Sentia (i.e., when either was under the prescribed minimum age) and did not himself marry her within six months, the manumission was invalid. Julian, indeed, seems to have taken an extreme view, and one not helpful to anyone; even if she was divorced, he said, by her first husband within six months and then her patron married her, she did not get her freedom. The law does not seem to have troubled itself about an alternative possibility—marriage to a complaisant patron followed by a quick divorce with his consent.63

None of this applied if the patron had freed her merely in fulfilment of afideicommissum, i.e., a testamentary instruction from a previous owner, since she was held to have been granted her freedom of necessity, and not under a condition of marriage imposed by himself. She was free, whether he married her or not, and she could divorce and marry again at will.⁶⁴

The special position of the patron-husband also explains an apparent anomaly in the law of divorce. As we have seen, a marriage was normally ended by the captivity of the husband, and postliminium did not apply. Lawyers differed, however over the patronhusband. Ulpian thought that, if he was captured, the wife was free to marry again, as if he had died. Julian, however, thought that the freedwoman-wife did not have conubium with anyone else, ‘because of reverence for her patron’. This may mean no more than that Julian found it slightly shocking that a patron could be divorced without his taking the initiative. Nevertheless, Julian may be right, and for good legal reasons.

It has been suggested that the reason that postliminium did not normally apply to the marriage of free persons was that it conflicted with the liberty of either to end the marriage at will. Since the freedwoman wife did not have this liberty, postliminium could apply, and so, although the marriage itself could not be held to exist while he was in captivity, she was in the same position as one who had divorced her patron-husband without his consent.⁶⁵

Not all patrons, however, married their slave-girl sweethearts, and slave men were, it seems, even less likely to marry up the scale. As we saw in Chapter 3 above, some marriages between patron and ex-slave, or between freeborn and freed, incurred at first social disapproval and later legal prohibition. Augustan legislation banned such marriages to senators. The ban on a freedman marrying his patroness possibly came later (after the extension of the earlier ban to senators’ descendants of both sexes). It is understandable from the point of view of a pater, especially if the family was rich; part of his patrimony might, as we saw in the previous chapter, end up in the hands of his son-in-law’s patron. What was the objection to a freedwoman’s manumitting and marrying a slave (not her former fellow-slave) whom she had acquired after manumission? The relationship may have begun before she became his owner. As with the sctum. Claudianum, the purpose of the ban may have been to try to ensure that a male slave’s sexual energies were directed to producing more children in the household of his original owner. An exception was made, if we may trust the relevant text (D. 40.2.14.1) for a freedwoman who managed to acquire and then manumit her own former fellow-slave.