This “injection” comes from the point of view of a historian of slavery turned historian of sexuality. The history of sexuality sits very uneasily with the history of enslavement, however, because of the problem of agency and its relationship to consent. Historians must not erase the personhood of the enslaved and their ability to make choices; but the realm within which they make those choices is so constrained that recognizing their agency can be akin to blaming the victim or implying consent when it cannot really be given. Can we really say that every sexual encounter between an enslaved person and their enslaver is an act of rape, if the enslaved person has been the one to initiate it in the hope that it could better their life in some way? Indeed, we can, by analogy with the laws about rape of minors or others who are incapable of consent. Even if a 16-year-old initiates a sexual encounter or agrees to one, sex with them is legally rape because their consent is not valid. Of course, sex with one’s own human property was not legally considered rape under Western and some other legal systems. The enslaved person had no ability to withhold consent from their enslaver, in the same way a wife until quite recently (1990 in Ireland; 1991 in England; and 1993 in some US states) could not be raped by her husband because her marital vows constituted permanent consent. But although sex between enslaved people and their enslavers would not be considered rape under the law of the time, it is entirely reasonable for us to call it that. Even if the encounter was not in proximate terms violent or coercive, even if the enslaved person participated enthusiastically, the coerciveness of the entire situation was such that they cannot be considered independent actors.

This essay focuses largely on the actions and choices of enslavers and treats enslaved people as victims. This does not deny their humanity, their choices, or their desires, but they were nevertheless victims of a violent system in which they had no redress. There are three factors in the sexual exploitation of the enslaved by their enslavers, although they overlap a good deal. One is, simply, desire. If enslavers wished to have a variety of sexual partners—or even one new sexual partner—systems of enslavement allowed them to choose from among their enslaved people, or to purchase additional enslaved people, with whom to have sex. Already the enslaver’s property, they did not need to be courted, or paid, or endowed. They were practically and legally available, and if there were no social sanctions there was no reason for the enslaver not to engage in sex with them if he so desired (usually he, because sex between free women and enslaved men was a serious offense in most cultures).

But, of course, there is no such thing as simple desire without issues of power being involved, and there is a significant overlap between the factors of desire and domination. The assumption that whenever one feels the wish to have a sexual encounter, that encounter is immediately available is a result of a huge power differential and unquestioned privilege. For men of the enslaving class to assert that privilege is an act of domination over people of other groups, notably people of whatever ethnic or racial group or social class was being enslaved in that particular culture. It may not feel that way to the one who holds the power, but privilege is not any less the case for being unexamined.

Sex as domination is both the result of unexamined power and privilege that manifests in an individual, and the result as well as the cause of larger group processes. Sex with one’s enslaved women or men could carry with it a message of dominance over other men: men who are raped; enslaved men who are less masculine because they are unable to protect the victims; men whose wives, daughters, sons, sisters, brothers have been captured and raped (even if the men in question are dead, their posthumous reputation may still be dominated); men within the enslaving society who are not wealthy or powerful enough to have slaves or perhaps even any sex partners, and are dominated by the assertion of a monopoly. These messages need not always be deliberate, but sometimes they are. And, of course, for a man to have multiple enslaved sex partners can also be an assertion of domination over other women as well, that is, his free sex partners, whether wives or concubines, whom he may be trying to make jealous or more eager to retain his favor.

There is a third factor in men’s having sexual relations with their enslaved women, which may not be obvious to those of us living in a culture where the fathering of children out of wedlock is often secret or shameful and where birth control is commonly practiced: precisely the wish to beget children. Even when law or religious teaching restricted the inheritance rights or the social standing of children born out of wedlock, the begetting of such children could still be a proof of masculinity. Children born of an enslaved woman and her enslaver could often be expected to be loyal and trusted supporters even if they could never take the full role of a child born in wedlock of a free woman. In many premodern societies such children were either automatically considered free or could be freed. Such a child, if recognized, might have full inheritance rights equivalent to those accruing to a child of formal marriage. Enslaved women were thus often valued for their reproductive labor.

For medieval Christian Europe the status of a child born of an enslaved woman impregnated by her enslaver was legally problematic. First, relationships between men and women other than within marriage were condemned by the church, regardless of whether there was a status differential between the parties. In practice, the definition of what relationships constituted marriage took a long time to work out. Historians have placed the triumph of church control over marriage, and the consequent definition of those born outside of it as illegitimate, anywhere from the ninth to the thirteenth century. Even when marriage came to be considered as within ecclesiastical jurisdiction, what constituted it was not necessarily a blessing by a priest (not required until the sixteenth century). Marriages were meant to be conducted publicly but were not invalid if they were not. By the twelfth century it was generally agreed that marriage was constituted by the consent of the parties, and that of their families was not required. Secular law might judge the validity of a marriage by the payment of a dowry or other marital assigns, which did require the consent of the family from which the money came. More importantly, however, whether a particular union was considered a marriage by the community, and sometimes by various legal systems, could depend on the relative status of the parties.

Pope Leo, I had written to Rusticus of Narbonne in 458–59 that “Marriage is a legitimate agreement between freeborn and equal persons,” and suggested that someone who wished to marry his daughter to a man who had an enslaved concubine need not worry that the man could be considered already married.1 Leo’s opinion was based on Roman law, which did not consider any union involving an enslaved person as a marriage. The enslaved could have only contubernium and not coniugium which only existed between Roman citizens. By the twelfth century the church would come to reject Roman law on this point.2 The jurist Gratian held that a marriage between an enslaved and a free person or between two enslaved people was valid as long as there was no fraud and the status of the enslaved party was known to the free party.3 But the older attitude that a marriage could not be valid one unless the parties were of approximately equal status still persisted, particularly in areas whose legal systems were based on Roman law. Bartolo of Sassoferrato (1314–1357), one of the great legal authorities of the fourteenth century, enunciated a general principle that any son of an unmarried man who is acknowledged by his father should be considered legitimate unless the father calls the child “natural.” Children born to servants, however, were exceptions to this rule: “if some honorable and noble citizen should have children by some servant who served him or another, then by those words he cannot say he is legitimate because marriage cannot happen with that woman, at least honorably.”4

The Roman law principle for the children of the enslaved had been partus sequitur ventrem, or “the child follows the womb”: because there was no marriage of the enslaved, children born to them were by definition illegitimate, and illegitimate children did not take the status of their father. As an enslaved woman was legally property, her children were considered the property of her enslaver as well. This was the case whether or not her enslaver was the father; fatherhood brought with it no rights here, as indeed motherhood did not. The principle that the children of an enslaved woman and a free man were themselves enslaved did not hold throughout medieval Europe, however. In Scandinavia, for example, the laws held that children of the enslaved followed the status of the higher-status parent. And even in places where Roman law was generally followed, such as medieval Italy, by the later Middle Ages both legal commentators and judges were discussing whether a man could make legitimate the child he had with an enslaved woman, without referring to any preceding manumission. In other words, they were assuming that the acknowledged child of a free man was already free. Sally McKee has argued that this change originated in Venetian Crete, where enslavement was quite common, and was then imported to the Italian metropole where it was less so.5 Hannah Barker argues that by the later Middle Ages in many areas of Mediterranean Europe, enslaved women were valued precisely for their ability to produce offspring. Children, especially sons, born of an enslaved woman and her owner before his marriage could be his heirs if he did not eventually have children from a marriage.6

Late medieval Italy gives us examples of elite men’s children with enslaved women being treated as part of their marital family. Gregorio Dati (1362–1435), a Florentine merchant who left a famous ricordanza (journal/account book), had a son with his enslaved woman Margherita (a “Tartar”, possibly a Russian) in between his first and second marriages.7 The child, Tomasso (Maso), was born in Valencia where Dati spent 1391–92 on business. He sent Maso back to Florence at the age of two or three months. Dati subsequently had eight children with his second wife, eleven with his third, and six with his fourth. He records that Maso was still alive in 1422, calling him his eldest child, clearly acknowledged and considered part of the family.8 What he does not tell us, however, is anything further about Margherita. He had purchased her while in Valencia, but does not say whether he brought her back to Italy. Although we may guess that Maso was to be wet-nursed in Florence and that his mother may never have seen him again, it is possible that she went with him, and that she was simply not significant enough to be mentioned. This would not be out of character for this particular text, which is largely an account of Dati’s business; he mentions his wives’ dowries, the children they bear, and their deaths, but is not otherwise concerned with household matters.

What can we guess, then, about Margherita? Dati was single when he went to Valencia and may well have purchased Margherita as a domestic servant, but sexual service was often a part of this. Sexual involvement with her would be seen as only a minor offense, if an offense at all: she was his property, and he was not married at the time. Bringing her back to Florence with him could be seen as a risk, however, in that he was seeking to marry well to secure a large dowry to invest in his business. If Margherita did come to Italy with her son, she likely was no longer enslaved in Dati’s household.9 The fact that he had a subsequent 25 legitimate children (not counting stillbirths and miscarriages) but does not mention any further illegitimate ones makes it likely that he remained either faithful to his wives, or careful.

We do not know how Dati’s wives felt about Margherita, if she did come to Florence. Mark Meyerson gives the example of a baptized former Muslim named Maria who was enslaved in a household in Alicante but sent to Valencia in 1503 to be sold because the wife of her enslaver was jealous.10 But a child born prior to the marriage was less likely to threaten a wife. Francesco Datini, a merchant in Prato, had a daughter Ginevra with an enslaved woman; his wife Margherita, who had no children, raised the girl and wrote that she considered her own daughter.11 In order to be raised in her father’s household she had to be taken from her mother Lucia, who was freed and married to another servant of the family. It is possible that some enslaved women who had their children taken from them were subsequently sold as wet nurses, a task in high demand which was often performed by enslaved women.12

On the individual level, men in the European Christian Mediterranean who purchased enslaved women particularly (as opposed to purchasing both men and women for large-scale agricultural enterprises) did so to fulfill their own needs, whether for sex, domestic help, reproduction, or some combination. But on a societal level it was also about having a class of people who could be dominated, and the conspicuous “consumption” of a luxury “good.” Sally McKee has estimated that 80% of enslaved people sold in the Christian Mediterranean between 1360 and 1499 were women; a sharp rise in prices cannot have been justified by the economic cost of domestic service and must, she argues, be due to men’s purchasing of enslaved women for sexual services.13 A man might feel the responsibility to care for his children with enslaved women, might even welcome and love these children, but the availability of the enslaved facilitated the access of the elite men to these women without creating a responsibility to care for them.

We turn now to a different medieval culture. For the Jewish merchants whose business and family letters have come down to us from the Cairo Geniza, particularly from the tenth through thirteenth centuries, the possibility of jealousy between enslaved women and wives was made much more explicit. This Jewish community, like the Muslim majority community within which it was embedded, permitted men to have more than one wife, but the ketubah or marriage contract could be drawn up in such a way as to severely limit this possibility (it could not absolutely prevent the man from doing so but it could provide serious financial penalties if he did). Marriage contracts could also provide that the man could not force his wife to move abroad, or commit him to living with her family. The inclusion of provisions like this in marriage contracts, of course, varied and reflected the relative strength of the two sides in the marriage negotiation. But one of the most common clauses included in marriage contracts was that the husband would not purchase any enslaved woman of whom his wife disapproved. Sex by an enslaver with his enslaved woman was forbidden in Jewish law (unlike in Muslim law) but evidently was enough of a concern in practice that this type of stipulation was commonly found in marriage contracts along with the undertaking not to marry a second wife or a concubine. Typically, too, contracts stipulated that any enslaved woman had to be sold if the wife demanded it.14 This type of clause begins to appear in large numbers in the twelfth century; S.D. Goitein suggests that it reflects local custom that prohibited men from forming such secondary unions.

The Cairo Jewish community seems to have been stricter than Christian communities about unmarried men enslaving women for sexual purposes, and several cases ruled against men who placed their enslaved women under the control of their sisters to preserve appearances.15 The stipulations in marriage contracts that provided wives with a veto over enslaved women in the household were not restricted to sexual jealousy alone. Enslaved women generally did domestic work (as opposed to enslaved men who generally worked in the family enterprise), and would have been under the supervision of the woman of the household; thus, she had a major stake in making sure that they were capable workers. What men did while they were on long trips, however (and like Gregorio Dati, many of the men in the Cairo Jewish community did travel for business) would not be under their wives’ eyes. Jewish merchants of Cairo were involved in the Indian Ocean slave trade along with Muslim traders, and had ample opportunity for sexual exploitation of those they sold.16

Islamic law provided that an enslaver’s child with his own enslaved woman, assuming that he recognized the child as his, was free and legitimate; the mother could not be sold and was freed upon the father’s death. As Hannah Barker puts it, “[i]n an Islamic context, therefore, sex with slave women produced heirs, while in a Christian context, it produced property.” Barker suggests that the fact that by the late Middle Ages in some Christian jurisdictions these unions did not produce property but rather free children who could be legitimated and brought into the inheritance at the father’s discretion, may have been due to the influence of Islamic law.17 Within the Cairo Jewish community, even though it was located in a Muslim realm, these unions produced people who were treated as property. Girls born to enslaved women could be raised in the household and trained in domestic duties, and boys trained to work in the family business or sold for use as military slaves. The fact that enslaved women could not produce children who would be the heirs of their enslavers did not mean that they were not valued for their reproductive abilities, both in producing children and in serving as wet nurses.

For Muslims, there was nothing legally wrong with a man having sex with his enslaved woman, and indeed if an enslaved woman were raped by someone else, any compensation would be owed to her “owner.”18 Michael Gomez has suggested that an entire political system—that of the Songhay Empire of West Africa—was based on slavery and in particular the enslavement of women who bore children to the ruler; these women were an integral part of the court and actively supported their sons in the struggle over the throne.19

Because sexual contact with enslaved women was licit in Muslim law, opportunities thus arose for these women within the household or court. The same thing might be true of eunuchs. But, like the eunuch, the enslaved woman who bore her enslaver’s children had to give up something in order to gain power and influence. The material conditions of enslavement could differ from place to place, and material conditions might not have been any worse than for many free people, and in some instances better if an enslaved woman became a favorite of a monarch. But good treatment in material terms, and especially the acquisition of political influence, were only possible because slavery was normalized and accepted. Margherita may have been separated from her son by Gregorio Dati, whereas in a Muslim society she might not have been; but the society that would not separate her from her child did not do so because such exploitation of enslaved women was routine. And whereas, in societies where married men’s children with slaves were considered heirs, this could improve the status of enslaved women while at the same time harming that of free women in relation to their husbands.

It is the Christian set of practices that should be treated as the historical anomaly, not the Muslim. Some overlapping of different categories within a polygynous system is very common in global history: even when a man is limited to only one official wife, or several, who are married with a particular ritual, it is common for him to have sexual access to enslaved women in his household, and for the line between an enslaved and a non-enslaved concubine to be blurred. For example, in the Tang Code dating from seventh-century China, the law acknowledged both a man’s right to sexual access to the servile women in his household, and the importance of the production of descendants in the Chinese family system, by adding this critical qualification to the statement that slaves and bondmaids were of “base” status and could not be wives or concubines: “If the bondmaid has a child and has already been manumitted, it is permitted for her to be a concubine.”20 She could even become a concubine if she had a child but was not manumitted (or vice versa). In all these cases, it was entirely up to the man. Her status was in no way guaranteed. This is, of course analogous to the Islamic law in which a woman who bore a child to her enslaver became free upon his death, but only if he had acknowledged that child.

The distinction between household women of low status and concubines eroded further in the Song period as more of the former were indentured servants rather than enslaved. In practical terms, whether someone was sold on a contract for a period of years or as a permanent possession may not have mattered very much to her treatment.21 The senior (non-concubine) wife was legally considered the mother of all her husband’s children, regardless of who their birth mother was. Whether they were considered heirs depended on the situation of the family and how many children the senior wife had, but courts supported their entitlement to a share.22 Under the Yuan dynasty the government attempted to freeze the fluidity that had previously existed between enslaved women, indentured servants, and women of higher status, to treat slavery as a permanent and degraded category, and to forbid the indenture of other people. In practice, however, the fluidity continued as it was not uncommon for people of high status to become enslaved.23

Christian Europe, as mentioned, developed by the later Middle Ages a somewhat analogous system in which the children of enslaved women with their enslavers could be considered free and where it was not uncommon to free the mothers as well. The issue of whether the child could inherit from the father is more complicated. This distinction between free and unfree in this context is a legal question and not one of religious outlook. The fact that Christian Europe for the most part did not legally recognize the relationships between enslavers and enslaved women does not mean that they did not force enslaved women into sexual relationships. These statuses existed, but without legal protection. Rather, the lack of legal recognition meant that everything depended even more on the enslaver; as in other systems he could grant or withhold status, but in the legal systems of Christian western Europe that status was more likely to go unrecognized by his heirs or the community, unless he took both legal and extra-legal steps to remedy it.

The fact that it was the enslaver had the upper hand legally and socially, in Christian Europe and elsewhere, did not mean that sex could not be an avenue of upward mobility for an enslaved woman, or that such women did not fight for their rights and those of their children through the legal system (and undoubtedly in other ways as well, but it is the records of the legal system that have survived). But even if these women did make choices—if what Steven Epstein refers to as “the motives behind the behavior of becoming pregnant” is something we can really consider24—the system constrained their choices within a context of rape and exploitation. These were not simply status-imbalanced unions but coerced ones, coerced on a systemic if not an individual level.

FormalPara Notes
  1. 1.

    “Ad Rusticem Narbonensum episcopum,” Sancti Leonis magni Romani pontifices epistolae, 167: 4, PL 54:1204a, quoted at greater length and discussed in Ruth Mazo Karras, Unmarriages: Women, Men, and Sexual Unions in the Middle Ages (Philadelphia: University of Pennsylvania Press, 2012), 31–32.

  2. 2.

    See Rachel Stone, “Overlooked Husbands: The Paradox of Unfree Marriage in the Carolingian World,” Gender and History 33, no. 2 (2021): 408–26 for how this developed.

  3. 3.

    Anders Winroth, “Neither Slave nor Free: Theology and Law in Gratian’s Thoughts on the Definition of Marriage and Unfree Persons,” in Medieval Law and the Origins of the Western Legal Tradition: A Tribute to Kenneth Pennington, ed. Wolfgang P. Müller and Mary E. Sommar (Washington: Catholic University of America Press, 2006), 187–218.

  4. 4.

    Transl. in Thomas Kuehn, Illegitimacy in Renaissance Florence (Ann Arbor: University of Michigan Press, 2002), 254. Further examples and discussion in Unmarriages, 97.

  5. 5.

    Sally McKee, “Inherited Status and Slavery in Late Medieval Italy and Venetian Crete,” Past & Present 182 (2004): 31–53.

  6. 6.

    Hannah Barker, “The Risk of Birth: Life Insurance for Enslaved Pregnant Women in Fifteenth-Century Genoa,” Journal of Global Slavery (2021): 1–31, esp. 11–15.

  7. 7.

    Dati’s ricordanza has been translated: Two Memoirs of Renaissance Florence: The Diaries of Buonaccorso Pitti and Gregorio Dati, trans. Julia Martines, ed. Gene Brucker (New York: Harper Torchbooks, 1967), 107–41.

  8. 8.

    Ibid., 134.

  9. 9.

    Cf a 1461 case cited in Debra Blumenthal, Enemies and Familiars: Slavery and Mastery in Fifteenth-Century Valencia (Ithaca: Cornell University Press, 2009), 175.

  10. 10.

    Mark D. Meyerson, “Slavery and Solidarity: Mudejars and Foreign Muslim Captives in the Kingdom of Valencia,” Medieval Encounters 2 (1996): 286–343, here 302.

  11. 11.

    Origo, “Domestic Enemy,” 344, 345.

  12. 12.

    Rebecca Lynn Winer, “Conscripting the Breast: Lactation, Slavery and Salvation in the Realms of Aragon and Kingdom of Majorca, c. 1250–300,” Journal of Medieval History 34 (2008): 164–84.

  13. 13.

    Sally McKee, “The Implications of Slave Women’s Sexual Service in Late Medieval Italy,” in Unfreie Arbeit: Ökonomische und Kulturgeschichtliche Perspektiven, ed. M. Erdem Kabadayi and Tobias Reichardt (Hildesheim: Georg Olms, 2007), 101–14, here 103.

  14. 14.

    S.D. Goitein, A Mediterranean Society: The Jewish Communities of the World as Portrayed in the Documents of the Cairo Geniza, Vol. 3: The Family (Berkeley: University of California Pres, 1978), 147–48.

  15. 15.

    Goitein, A Mediterranean Society, 24.

  16. 16.

    Craig Perry, “Slavery and the Slave Trade in the Western Indian Ocean World,” in The Cambridge World History of Slavery 2: AD 500-AD 1420, ed. Craig Perry, et al. (Cambridge: Cambridge University Press, 2021), 123–52.

  17. 17.

    Hannah Barker, That Most Precious Merchandise: The Mediterranean Trade in Black Sea Slaves, 1260–1500 (Philadelphia: University of Pennsylvania Press, 2019), 62 and 82.

  18. 18.

    Hina Azam, Sexual Violation in Islamic Law: Substance, Evidence, and Procedure (Cambridge: Cambridge University Press, 2015), 131.

  19. 19.

    Michael A. Gomez, African Dominion: A New History of Empire in Early and Medieval West Africa (Princeton: Princeton University Press, 2018), 299–303.

  20. 20.

    Beverly Bossler, Courtesans, Concubines, and the Cult of Female Fidelity (Cambridge, MA: Harvard University Press, 2013), 60.

  21. 21.

    Patricia Buckley Ebrey, The Inner Quarters: Marriage and the Lives of Chinese Women in the Sung Period (Berkeley: University of California Press, 1993), 217–34.

  22. 22.

    Bossler, Courtesans, Concubines, and the Cult, 246.

  23. 23.

    Bossler, Courtesans, Concubines, and the Cult, 333.

  24. 24.

    Steven A. Epstein, Speaking of Slavery: Color, Ethnicity, and Human Bondage in Italy (Ithaca: Cornell University Press, 2001), 129.