Introduction

This chapter discusses slavery in medieval Arabia1 by focusing on Yemen, the Arabian Peninsula’s southwestern part. Sources from the eleventh to fifteenth centuries CE feature slaves in diverse life situations: as human commodities sold on public slave markets, commanders of cities, mothers of their masters’ children, resistors to enslavement, laborers in kitchens and workshops, and as freed persons. The following case study will characterize the life trajectories of slaves in medieval Yemen and highlight three central aspects: their enslavement (entry into slavery), how they lived and worked as slaves (experiences of slavery), and the possibilities of altering their unfree status (exits from slavery). The evidence presented comes mostly from the Najahid (1021–1158) and Rasulid (1229–1454) eras of Yemen’s medieval history. While offering a detailed account of slavery in these two Yemeni polities, this chapter will also discuss the phenomenon of slavery throughout medieval Arabia more broadly, highlighting temporal and geographical continuities and differences.

Entry into Slavery

In the early 2000s, the Yemeni scholar Muḥammad Jāzim published a fascinating collection of administrative documents from late thirteenth-century Yemen that had survived in a private library in Sana’a. Known as Nūr al-ma ‘ārif (“The light of knowledge”),2 this source provides rare information on slave trading practices across the Red Sea to Yemen, including routes and procedures, as well as prices and taxes paid for different categories of slaves. The following passage from Nūr al-ma ‘ārif is part of a broader section on trade between Ethiopia and Yemen. Tellingly, slaves are listed here among many other “products” to be evaluated, sold, and taxed:

The good eunuch (al-khādim al-jayyid) is expensive up to a hundred wiqīya [a common weight measure] in coins, and that is precious. […] The eunuch of medium quality is for 50 or 60 wiqīya. The regular eunuch is for 40 wiqīya.

The uncastrated slaves (al-ʿabīd al-fuḥūl): the good slave is the pure Ethiopian slave boy (al-waṣīf al-ḥabashī al-ṣarīḥ), such as the jizlī and the amḥarī, or the saḥartī or any kind (jins) as long as he is flawless. His price in Ethiopia is 20 wiqīya, and the one of medium quality 15 or 14, and the one of lesser quality 12, 11 or 10.

As for female slaves (al-jawār): the good slave girl of excellent quality (al-jayyida al-waṣīfa al-ʿāl) is for 20 wiqīya, and of medium quality for 15 or 16, and of lesser quality for 12 or 10.3

This short passage highlights many of the complexities characterizing the medieval slave trade to Arabia and thus requires some context. The Rasulids at whose court this document was written ruled over large parts of lower Yemen from 1229 until 1454. Of Turkoman descent, they had come to the country with the Ayyubids, their predecessors, whom they served as high military and administrative officers. Yemen’s geographical location on both the Indian Ocean and the Red Sea meant that important maritime trade networks intersected there, linking up with caravan routes leading to Mecca and Medina, the holy sites of Islam. The Rasulids exploited this strategic advantage by actively supporting trade through the provision of administrative, infrastructural, and security services. The state in turn shared in the merchants’ wealth by taxing them heavily.

According to medieval Islamic law, slaves could be bought and sold like any other property, although their humanity was acknowledged. In the above passage, slaves are divided into categories—female, male, and castrated male slaves (eunuchs)—and ranked by their perceived “quality”—low, medium, and high. Eunuchs were up to five times more expensive than other slaves, matching accounts from the Roman, Byzantine and Ottoman empires, all of which considered eunuchs to be the ultimate luxury possession.4 The reasons for the high value placed on eunuchs were twofold. The castration procedure, despite being carried out in specialized centers in Ethiopia and elsewhere, was very dangerous, and many boys did not survive this mutilation. Furthermore, courts in the Islamic world had come to rely on the service of eunuchs, who were considered to be the perfect servants. In the Rasulid case, as we shall see below, eunuchs were important players at all levels of government and administration. From the perspective of slave traders, eunuchs were thus a high-risk commodity that potentially yielded high profits.

Ethnicity also impacted a slave’s price: the highest value was attributed to a “pure Ethiopian slave boy” who was “flawless.” The author of the above passage distinguished between three different Ethiopian ethnic affiliations—jizlī, amḥarī, and saḥartī—suggesting he had a fairly nuanced knowledge about the country’s inhabitants. Ethnonyms used for slaves in Arabic sources underline the finding that most slaves in medieval Arabia were of African origin. Yemeni authors writing between the eleventh and fifteenth centuries mainly used three terms to describe African slaves: habashī, denoting individuals from today’s Ethiopia and Eritrea, zanjī, roughly pointing to modern Somalia and coastal regions further south, and nūbī, indicating origins in today’s southern Egypt and Sudan. The geographical meaning of these terms was approximate and fluctuated over time. What is more, it appears that many individuals were captured from populations living further in the African heartland, rather than in the coastal areas from where they were later trafficked to Arabia.5 The exact origins of enslaved Africans who labored in medieval Arabia therefore remain unknown. A new body of evidence that has recently been introduced into slavery research are genetic studies. This type of data shows that female African contributions to the gene pool of the Arabian Peninsula are on average triple that of the male contributions, pointing to slavery and specifically to concubinage (the sexual exploitation of female slaves by their masters). Most African genetic input into Arabia is of Ethio-Somali and Nilo-Saharan origin, roughly validating the geographical information given by medieval Arabic authors.6

Nūr al-ma ‘ārif contains several other passages that offer striking insights into slave trading practices during the Rasulid period. The source shows that children were captured or bought from populations living in today’s Ethiopia and Eritrea, and then transported by Ethiopian traders along commercial routes to Zaylaʿ, a port city and slave trading hub commonly identified with today’s Saylac in Somalia. There, they were sold to Yemeni merchants with the help of an Ethiopian intermediary known as nazīl. Yemeni traders then transported the slaves on boats carrying mixed cargo to Aden, Yemen’s major port on the country’s southernmost tip, and to the ports of Zabīd, one of two Rasulid capitals. Upon arrival, some slaves were immediately selected for government service by state officials and taken overland to Ta’izz, the dynasty’s second capital, while all others were sold on the public slave market.

A graphic description of sale procedures on the slave market of Aden was recorded by Ibn al-Mujāwir, a thirteenth-century traveler in Yemen:

Selling slave girls. The slave girl is fumigated with an aromatic smoke, perfumed, adorned and a waist-wrapper fastened round her middle. The seller takes her by the hand and walks around the souk [market] with her; he calls out that she is for sale. The wicked merchants appear, examining her hands, feet, calves, thighs, navel, chest and breasts. He examines her back and measures her buttocks in spans. He examines her tongue, teeth, hair and spares no effort. If she is wearing clothes, he takes them off; he examines and looks. Finally, he casts a direct eye over her vagina and anus, without her having on any covering or veil. When he has examined, expressed his approval and bought the slave girl, she remains with him for about ten days. When [the buyer] has taken care of her, had his fill, become bored and tired of her and got what he wanted from her, his lust is at an end. Zayd, the buyer, says to ʿAmr, the vendor, “Indeed, sir, we have a case to settle in court!” So they attend in front of the judge and one makes a claim against the other, [suggesting there is] a defect [in the slave girl].7

Ibn al-Mujāwir’s distressing account closely matches the advice given in medieval Arabic slave-buying manuals, which recommend that prospective buyers inspect slaves’ private parts before making their choice. Doing so in public went against the principles of Islamic law and morality, but nevertheless seems to have been quite common in the medieval era. The latter part of Ibn al-Mujāwir’s report reveals his penchant for scandal, while also addressing a specific provision of Islamic law pertaining to the slave trade: the buyer had a right to annul the acquisition of a slave if a defect was discovered that had not previously been disclosed by the seller.8 Zayd, the buyer in this story, makes a mockery of this right by sexually exploiting the enslaved girl he had purchased and then attempting to return her. The author does not reveal how this legal dispute ended, but recounts a similar case that was dismissed in court. What is most striking about this description, however, are not the insights into customs and legal provisions surrounding the sale of slaves. The main character of this vignette is the anonymous enslaved girl, who first suffers objectification and humiliation through the actions of the seller and the prospective buyer, in full view of the public, and is then sexually exploited by her new owner. As Hannah Barker has pointed out, slaves were considered objects to be bought and sold, but at the same time, their human capacity to think and act independently was recognized and exploited.9 The moment of sale marked the point in the life of enslaved persons when they were almost fully reduced to commodities and their humanity was suppressed through humiliation and coercion. Ibn al-Mujāwir’s account does not tell us anything we do not know from other sources—enslaved girls and women were sold to satisfy the sexual appetites of their masters—but the detail and crudity of his description make the human suffering of this particular girl palpable. In light of this source, the common scholarly assertion that Islamic slavery was of a relatively harmless nature becomes difficult to sustain.10

Combined evidence from Nūr al-ma ‘ārif and Ibn al-Mujāwir has allowed for a vivid reconstruction of slave trading practices between East Africa and Yemen during the Rasulid era. Looking at the medieval Arabian Peninsula more broadly, it is important to note that slave trading routes were not stable over time, but fluctuated according to political and economic developments. Overland slave trading networks in late antiquity connected Syria, Jordan, and Arabia. In fact, slave raiding practices of nomadic Arabs during that time probably served as a model to the early Muslim conquerors.11 In pre-Islamic times, African slaves were shipped to Yemen and then transported northwards to the Hijaz12 on caravan routes, rather than being “imported” directly through Red Sea ports in Arabia.13 Mecca was already a major commercial hub before the advent of Islam, and slaves were among the many commodities that changed hands on markets in and around the city. Most slaves were brought to the peninsula from East Africa via the Red Sea. As the Periplus Maris Erythraei attests, by the first century CE, the trafficking of humans across the Red Sea was an established practice.14 Slaves were usually carried on mixed-cargo ships in small numbers, and their acquisition was often arranged through personal networks rather than on larger slave markets. During the eighth and ninth centuries, two dynamics spurred the demand for African slaves in Arabia. A boom in mineral exploitations in today’s Asir region of Saudi Arabia increased the demand for cheap labor.15 At the same time, local rulers began assembling slave soldiers (mamluks), a practice first introduced by the Abbasid caliph al-Muʿtaṣim (r. 833–842). In Yemen, the Ziyadids (ca. 818–981) were the first dynasty to establish a mamluk army. According to the Yemeni historiographer ‘Umāra b. ʿAlī al-Ḥakamī (b.1120 or 1121), the Ziyadids imposed a tribute on the ruler of the Dahlak archipelago in the Red Sea that included 1000 male and female slaves identified as Ethiopian and Nubian.16 An eleventh-century traveler observed enslaved workers from Zanzibar and Ethiopia in the oasis of Al-Aḥsāʾ in today’s eastern Saudi Arabia, suggesting that African slaves were not only traded across the Red Sea, but also directly to the Persian Gulf.17 This evidence is also a rare glimpse into agricultural and rural slavery, both of which have likely been underestimated by modern scholars due to the urban and elite focus of most available sources.18 In the twelfth century, the Arab geographer Muḥammad al-Idrīsī describes how Arab traders lured children on the coast of present-day Kenya with dates, in order to capture and enslave them.19 Around the same time, letters from Jewish merchants attest that slaves were shipped from the Horn of Africa to Aden.20

Not all slaves in medieval Arabia were African or of African origin. In fact, the enslavement of Arabs by Arabs was not uncommon until Muslim jurists ruled that only persons born into slavery and enemies captured in warfare could legally be enslaved.21 Although the medieval slave trade from India to the Arabian Peninsula still awaits thorough scholarly exploration, the appearance of Indian slaves in sources on medieval Arabia proves its existence. The Yemeni evidence shows that female slaves from India were prized as entertainers and concubines.22 In the thirteenth century, the traveler Ibn al-Mujāwir reports that enslaved boys (ghilmān) were imported from India to the Yemeni ports of Aden and Al-Shiḥr, even specifying the customs due on slaves who were likely Goan.23 Aden also served as an entrepôt for enslaved Indians destined for Egypt.24 A fifteenth-century source describes the eunuchs guarding the Prophet Muhammad’s tomb in Medina as being mostly Indian, but also of East and West African as well as of Byzantine origins.25 Finally, Turkish slaves are occasionally mentioned in medieval Yemeni sources, and a handful of references to slaves from other parts of the world further complicate the picture.26

Slave trading was not the only way in which people wound up as slaves in medieval Arabia. In pre- and early Islamic times, most slaves on the peninsula seem to have been Arab prisoners of intertribal warfare.27 Between the third and sixth centuries, the Ethiopian kingdom of Aksum suffered a number of military defeats in South Arabia, likely increasing the African slave population there.28 Military expansions during the early Islamic period (seventh to eighth century) provided many opportunities for the capture and subsequent enslavement of prisoners of war. Although figures are difficult to extrapolate from the available sources, the scale of these practices was undoubtedly significant. Once conquered populations had either embraced Islam or accepted the authority of their Muslim rulers, enslaving them became illegal. In this respect, it is instructive to briefly consider how the entry into slavery was regulated and restricted by Islamic law. Islam’s view of slavery rests on two main sources, namely the Quran and the ḥadīth. The principles stipulated in these sources are congruent and form the basis of Islamic law (fiqh), which was codified in the eighth and ninth centuries. Islamic law recognizes two ways of entering into slavery: only children born to enslaved parents and enemies captured in warfare could rightfully be slaves, meaning that previous practices such as debt bondage, enslavement as punishment for crimes, or self-sale were forbidden. Furthermore, as will be discussed below, medieval Islamic law promoted manumission and granted the children of concubines free status. These legal principles caused a steady reduction in the number of slaves but left open a loophole, since the supply of slaves from beyond the Islamic realm remained unregulated. Hence, while the normative framework established by Islamic law arguably sought to reduce the number of slaves in medieval Islamic societies, it ultimately led to the development of a commercialized slave trade.

Experiences of Slavery

We know that in medieval Yemen, enslaved persons belonged to royal courts in large numbers and were also found in the households of local elites, but there is only minimal evidence for slave ownership among the broader population. Due to a strong bias toward a historiography of the elites, only slaves who closely associated with influential individuals appear regularly in narrative sources from medieval Yemen. This imbalance of evidence means that the lives of eunuchs and concubines can be reconstructed in much greater detail than those of low-ranking slaves. Administrative documents partly compensate for this shortcoming, offering glimpses into the lives of enslaved menial workers. It is important to consider that slaves in medieval Yemen lived under vastly different conditions, but that they nevertheless shared the same state of unfreedom or partial freedom which set them apart from the free population. The study of slaves in relatively privileged positions thus also offers clues on the experience of enslaved persons in medieval Yemen more generally. Harnessing the complementary strengths of sources from the Najahid and Rasulid era allows us to gain important insights into the lived realities of slavery in medieval Yemen, illuminating a broad range of aspects of this phenomenon. For example, Rasulid works are rich in information on the lives of eunuchs serving at the Sultanic court. Furthermore, the administrative documents contained in Nūr al-maʿārif not only describe the slave trade across the Red Sea, but also offer rare information about slaves at the bottom of the social ladder. The surviving Najahid chronicle, on the other hand, offers a detailed account of the lives of (formerly) enslaved girls and women, a demographic that remains largely unmentioned in Rasulid sources.

Concubines

In the twelfth century, the Yemeni poet and scribe ʿUmāra b. ʿAlī al-Ḥakamī recorded a number of anecdotes featuring an enslaved singer known as Warda (Flower). Trained as an entertainer by her first owner, a well-known slave trader, she then became the concubine of a high-ranking military commander. Later, she attracted the attention of an influential vizier who devised an intricate plot in order to gain possession of her. In the following passage, Warda’s second owner describes the moment when she appears amid other enslaved girls to entertain guests at a banquet, and later informs her that she will be given to the vizier as a present:

Then we ordered [Warda’s] presence, as the tenth of ten [enslaved girls]. They kissed the hands of the vizier and began to sing in his presence, with uncovered faces. […]. Then [the vizier] ʿUthmān became drunk and slept, and the women became drunk, except Warda. I had wanted her to be alert. I went to the privy, called for Warda and informed her of the story. She said: “I don’t desire anything except my master.”29

After a number of scenes in which Warda merely appears as the voiceless object of men’s desire, she finally speaks. Although the above quote is at best a re-narration by the author, perhaps even mere fiction, it neatly encapsulates the precariousness inherent in the lives of concubines. Warda has just discovered that she will be handed over from her second owner to her third. Her statement “I don’t desire anything except my master” expresses her lack of choice and is masterfully diplomatic, in that it could be read as a statement of submission to any one of her previous or future owners. Medieval Islamic law legitimized and strictly regulated concubinage, the sexual exploitation of female slaves by their male masters. These relations were licit, and the children born out of them were free and their father’s legitimate offspring. A concubine who bore her master a child acquired the status of umm walad (literally “mother of the boy/child”), which meant that she could no longer be sold and automatically acquired her freedom upon her master’s death. Statistical analyses of prosopographic evidence have recently shown that concubinage, which was rare in pre-Islamic Arabia, expanded dramatically in the conquest era due to the ready availability of female captives.30 This dataset also shows that the number of concubines dropped significantly by the time of the Umayyad caliph Hisham’s reign (724–43), when the major Islamic military expansion came to an end.31 The practice of concubinage exposed enslaved girls and women to sexual exploitation, while offering some of them opportunities for social advancement that were unavailable to other slaves. Al-Ḥakamī’s account reveals that enslaved girls such as Warda were trained by slave traders and then sold to the elites as entertainers and concubines.32 This trajectory parallels practices in the Hijaz during the Abbasid era. Around the ninth century, Mecca and Medina became centers for the training of quiyān, enslaved girls and women who were then sold onto the Abbasid elites, whom they entertained with musical performances, poetry, and quick-witted banter.33 For instance, an early eleventh-century slave trader relates that he bought nine-year-old girls and trained them for three years each in Medina and Mecca, upon which they arrived in Iraq perfectly equipped to carve out careers as quiyān.34 What followed—both in Abbasid Iraq and in Najahid Yemen—was a life of precariousness in which concubines were frequently moved from one owner’s household to the next. In Warda’s case, she was owned by three men consecutively (that we know of) before finally gaining her freedom. After the vizier’s death, Warda is described as choosing her own marriage partner. It is therefore likely that she had born the vizier a child and was manumitted according to the umm walad laws. A careful analysis of the available evidence reveals that Warda displayed remarkable ingenuity enabling her to endure the coercion of enslavement and to navigate the complex dependencies tying her to her former and current owners, before she finally attained greater agency and social standing later in life.

It is thanks to ʿUmāra b. ʿAlī al-Ḥakamī’s twelfth-century chronicle Al-Mufīd fī akhbār Zabīd, the only surviving source from the Najahid period, that we learn about the experiences of concubines like Warda. This work is a striking description of a society deeply impacted on all levels by enslaved and freed persons. Concubines and former concubines figure prominently in this work, allowing for a reconstruction of their lives in some detail. The opportunities for social advancement open to some concubines such as Warda were harnessed in an even more impressive fashion by a woman named ʿAlam (d. 1150). Her story is tightly intertwined with that of the Najahid Sultans, a dynasty founded by Ethiopian slave soldiers who had usurped power from their masters in 1021 and ruled parts of Southern Yemen until around 1158. The Najahids routinely took concubines as wives, likely because local elites shunned them due to their African slave origins. In this respect, ʿAlam’s story is not uncommon. She was an enslaved singer, concubine, and later wife of the fifth Najahid Sultan Manṣūr (d. 1130), whom she bore a son. What sets ʿAlam apart was her strong political engagement, which continued even after the death of her husband and son had deprived her of any direct link to the throne. The early twelfth century was a period of vicious strife at the Najahid court, with different factions of enslaved viziers vying for influence while the sultan’s role slowly grew obsolete. As al-Ḥakamī notes, ʿAlam became one of the most powerful political figures of her time by placing her own male slaves in strategically important government and military positions:

Men from among the slaves (ʿabīd) of the free lady, queen Umm Fātik b. Manṣūr [i.e. ʿAlam] were raised in the palace of king Fātik b. Manṣūr. They were Ṣawāb, Rayḥān, Yumn, ʿAzz, and Rayḥān the Elder. These governors were dignitaries, important personalities. And among the non-castrated ones [were] Iqbāl, Masrūr, Bārih and Surūr. He [Surūr] was the amir of the two parties, given his capabilities and affluence. This group were those who spoke with the sultan’s tongue.35

The deployment of trusted slaves enabled ʿAlam to exercise political influence in spheres that were inaccessible to women, such as the army and government. Al-Ḥakamī mentions elsewhere that ʿAlam had raised some of these slaves in the royal palace, thereby nurturing their careers from the very beginning.36 ʿAlam’s social standing is further underlined by the fact that she became the patroness of the yearly pilgrimage caravan from Yemen to Mecca and Medina, ensuring its safety through her presence.37 The life stories of ʿAlam and Warda exemplify broader trends in the biographies of concubines in medieval Yemen. While they might have enjoyed a relatively privileged lifestyle compared to the conditions of low-ranking slaves, concubines had to endure sexual exploitation, usually by several consecutive owners, and their wellbeing was largely dependent on the goodwill of these men. At the same time, the biographies of individual concubines show that some of them were able to incrementally increase their status and independence by bearing their masters children, making strategic use of their intimate association with influential men, and using their own slaves to expand their political reach.

Eunuchs

At the end of the fourteenth century, the court historian ʿAlī b. Ḥasan al-Khazrajī recorded the death of Ahyaf, the most influential eunuch of the Rasulid era:

In this year [1385], the chief eunuch Amīn al-Dīn Ahyaf al-Mujāhidī died. He was resolute, courageous, tough, wayward, blood-shedding, murderous, crude, uncouth, resolute, determined, canny, haughty, of great prestige and of severe spirit. He was brave, fearless in battle, and a good advisor to the sultan.38

Ahyaf was the figurehead among a number of eunuchs belonging to Rasulid Sultans who made stellar careers in the Rasulid army and administration. He served four consecutive Sultans, rising to the very top of the Rasulid military apparatus. Al-Khazrajī portraits him as a skilled military commander who was often charged with heading punitive campaigns against rebellious tribes. Rasulid sources show that promising young eunuchs were selected by government officials as soon as they arrived in Yemen, and were then trained by senior eunuchs in the royal palaces. Eunuchs were thought to display undivided loyalty to their owners, since they had been torn from their families of origin and were physically unable to start ou their own families later in life. For this reason, they were considered the perfect servants worthy of their masters’ trust. In the Rasulid era, the importance of eunuchs in the political and military apparatus greatly increased. They supervised slave soldiers, led military campaigns, and acted as governors of cities, castles, and administrative regions. A number of them even attained the position of amir, the highest military rank in Rasulid Yemen. Sultans also occasionally sent eunuchs as diplomatic envoys to Egypt. It is remarkable that enslaved persons of foreign descent were trusted to return from a mission abroad, rather than simply disappearing. Eunuchs also performed key roles in relation to Rasulid women and children. Administrative and narrative sources reveal that the households of Rasulid noblewomen comprised dozens of eunuchs whose roles included safeguarding the women’s quarters, educating royal offspring, supervising sultanic kitchens and storehouses, and likely also training and supervising female slaves. Furthermore, each Rasulid noblewoman had a eunuch administrator (zimām) who was usually appointed by a male family member and managed her household. The importance of this relationship is epitomized by the fact that Rasulid royal women were known not by their given names, but by the name of their eunuch administrator. The wife of the seventh Rasulid sultan al-Ashraf Ismā‘īl, for instance, was known by the name of Jihat Muʿtab, after her zimām Jamāl al-Dīn Muʿtab al-Ashrafī. This naming practice powerfully illustrates the role that eunuchs played in representing their mistresses in male-dominated spheres beyond the royal residences. Rasulid sultans often appointed their former educators and tutors as their wives’ administrators. In this way, eunuchs accompanied their royal masters from childhood throughout their adult life, serving their wives and children as well. It is easy to imagine that these intimate responsibilities also enabled eunuchs to exercise a subtle influence on the ruling family. In sum, eunuchs were the only slaves serving in all spheres of Rasulid private and public life—managing the households of royal women, educating royal children, supervising male and female slaves, commanding mamluk armies, and occupying the highest political positions. These elite eunuchs must be seen as important players in the Rasulid system of government. Their potential success, however, came at the price of physical mutilation and was contingent on their absolute loyalty to their masters.

Slave Soldiers

Slave soldiers (also known as mamluks) are among the most commonly encountered slaves in medieval Yemeni sources. Their deployment by local rulers is attested from the ninth century onwards. The Najahid dynasty, itself established by Ethiopian slave soldiers who overthrew their masters, relied on Ethiopian and Nubian mamluks as well as on mercenaries from Central Asia known as Ghuzz to fend off rivals.39 During the Rasulid era, the mamluks were essential in fighting internal and external enemies, but they also destabilized their masters’ rule through frequent plots and revolts. Despite their centrality to the medieval history of Yemen, it is impossible to establish the basic facts about the lives of mamluks during that time. Not only is information on their training and living conditions in the military barracks virtually inexistent in the sources, their origins are also shrouded in mystery. The fact that they bore Turkic names prompts the assumption that they were of Central Asian descent, as was the case for most slave soldiers serving the Mamluk rulers in Egypt during the same time period. 40 However, it is also entirely possible that the Rasulids relied on African mamluks, just as preceding Yemeni dynasties had done, while giving them Turkic names to mark them off as military slaves.41 The mamluks of Rasulid Yemen not only elude scholarship, but they are also remarkably often described as revolting against their masters. For example, in 1322, a contingent of Rasulid mamluks rebelled, arrested the young sultan al-Mujāhid and attempted to install his uncle al-Manṣūr Ayyūb and later his cousin al-Ẓāhir on the throne. After these events, it took al-Mujāhid a decade to restore his sovereignty.42

Female Slave Attendants and Domestics

The role of female slaves as attendants and domestics of elite women is amply attested, even though details on the work and life conditions of these girls and women remain unknown. As noted above, Islamic legal texts present domestic slavery as the norm, and it was likely the most common form of slavery in medieval Islamic societies.43 Because their presence was taken for granted and considered to be irrelevant to historiography, enslaved domestics hardly figure in written sources and have thus frequently been underrepresented by modern historians.44 On rare occasions, however, these women and girls are thrown into the limelight, allowing us to gain insights into their lives. The story of Nukhba is a case in point here. She was an enslaved girl (jāriya) who worked in the Rasulid castle of Taʿizz, which at the time was under siege. Nukhba had just brought the reigning Sultan Al-Mujāhid ʿAlī (r. 1321–1363) water to perform his ablutions when he was almost struck by a ballista and barely escaped death.45 It was merely by being the only eyewitness to this event that Nukhba’s existence was recorded, and that we learn about her service in the most intimate realms of the Sultanic residence. How many others like her toiled in the households of the elites and are forever lost to history? While Nukhba is described here as serving the sultan, she actually belonged to the sultan’s powerful mother, queen Jihat Ṣalāḥ. The households of elite women comprised large numbers of female slaves. The labor of these slaves enabled their mistresses to live the largely secluded lifestyle that was expected from women of high rank and honor. Women-owned domestics were not allowed to perform sexual work, but anecdotal evidence shows that they were at times gifted to men and thereby turned into concubines.46 It should also be noted that although sex with male slaves was forbidden, both male and female slaves were at risk of sexual exploitation.47

Other Types of Labor Performed by Slaves

It is likely that the labor of enslaved women and girls comprised many tasks in the household setting, as well as in agriculture and craftsmanship, that were simply not recorded by contemporary historians and scribes. Even male slaves who were engaged in menial work are rarely mentioned. Nevertheless, medieval Yemeni sources do offer occasional clues on those slaves toiling at the bottom of the social ladder. Scattered hints, most of them from Nūr al-maʿārif, give some indication of the breadth of slave labor at the time. This information largely survived because at least some categories of slaves owned by the Rasulids received salaries and were given gifts during religious festivals. However, while male slaves at the Rasulid court frequently appear in salary and gift lists in Nūr al-maʿārif, female slaves are virtually absent, suggesting that they likely were not paid for their work but were instead maintained within the royal court. A fascinating exception concerns enslaved girls who received payment for producing buttons made of silk thread, under the supervision of a male slave, likely a eunuch.48 A handful of references also attest to the presence of female cooks and bread-makers in the royal kitchens. The work of male slaves in these kitchens is much better documented, and the range of salaries given to them suggests a diversification and hierarchization of tasks performed by them. The Yemeni sources analyzed also provide meager evidence that slaves were deployed in farming, especially in cattle husbandry.49 It is likely that slaves also worked in the royal stables as well as in mining and hunting.50 Furthermore, evidence from Geniza sources, a corpus of documents from the medieval Jewish community in Cairo, attests to the fact that merchants active in the Red Sea and the wider Indian Ocean world owned male slave agents who supported their business endeavors, sometimes even traveling abroad to trade on their behalf. The same practice existed among the Kārimī, a group of merchants active in the trade between Egypt and the Indian Ocean during the Ayyubid and Mamluk periods.51 Due to a lack of sources, the level of exploitation and violence suffered by slaves working in kitchens and workshops, on fields and ships is impossible to gage. The available evidence only paints a partial picture that obscures almost every detail about the lives of these children, women, and men.

Exits from Slavery

According to medieval Islamic law, slavery was considered to be a legal state, not an innate characteristic of the slave, and could therefore be reversed through manumission. The freeing of slaves was encouraged as an act of piety and recognized as a way of atoning sins, such as perjury or the violation of the Ramadan fast. Detailed legal provisions distinguished different kinds of manumission, such as a slave owner’s testamentary provision that her slaves be freed upon her death, or contractual manumission, in which slaves bought their freedom for an agreed sum of money paid in installments. In the foundational texts of Islam, manumission was understood as a way for slaves to regain their freedom and become full members of the umma; for slave owners, it offered the opportunity to perform a good deed and atone for sins. However, the way manumission developed in Islamic legal thought and lived social practice effected slaves in ways that were more complex. Manumission in the medieval Islamic context was not simply the reversal of enslavement. Rather, it is better understood as a long journey on which the slave traveled toward greater freedom, gaining more agency and rights along the way. The promise of freedom could also be used by masters to command obedience from their slaves, thereby becoming a tool of suppression rather than liberation. Not all slaves who were on a path to manumission were eventually freed. For example, if a master’s testament stated that his slaves be freed after his death, his heirs could still sell off these slaves to pay outstanding debts. Even achieving the status of freedman or freedwoman did not constitute full freedom in any modern sense. Rather, freed slaves remained connected to their former masters through bonds of clientage (walāʾ). This relationship was passed on from generation to generation, binding the families of former masters and former slaves together through responsibilities and rights toward each other. Walāʾ is modeled after kinship relations and is based on protection in exchange for submission. At its best, this relationship ensures that former slaves and their descendants are integrated into mainstream society through a connection to powerful patrons; at its worst, it can perpetuate exploitation and dependency for generations to come. Surprisingly, manumission is rarely discussed in medieval Yemeni sources. Exceptions include the fourteenth-century report that the Rasulid queen Jihat Taghā manumitted upon her death large numbers of slaves,52 and a mention of the freeing of a slave to atone for a sin.53 Biographical dictionaries occasionally feature manumitted slaves who became saints or merchants. For example, Abū al-Ḍiyāʾ Jawhar ibn ʿAbd Allāh al-Ṣūfī was a slave who engaged in trade on behalf of his master, a merchant. He was later manumitted and became a famous Sufi saint.54 If one traces the lives of individual concubines through al-Ḥakamī’s chronicle of the Najahid era, it becomes clear that they were indeed manumitted—either after their master’s death, as prescribed by Islamic law, or earlier. Whether and when the high-ranking eunuchs that figure so prominently in medieval sources on Yemen were manumitted remains a mystery. Only one eunuch named Niẓām al- Dīn Mukhtaṣṣ (d. 1267) is known to have been freed by his master.55 This remarkable silence around manumission could mean that it was rarely practiced in medieval Yemen, which however would be unlikely given its centrality in Islamic legal texts. Alternatively, it might have been such a standard occurrence that it deserved little mention by court chroniclers and scribes.

Conclusions

Slavery has long constituted a blind spot in the study of medieval Islamic societies, but a number of important works published since the 2000s have greatly contributed to our understanding of the phenomenon. Yet, few of these works focus on slavery in Arabia, where the birth of Islam in the early seventh century brought about changes in the practice of slavery that continued throughout the medieval period. Our understanding of slavery in medieval Arabia largely rests on the work of a few scholars, which are listed in the suggestions for further reading below. Elizabeth Urban has studied the changing roles of Muslims of slave origin in the early Islamic community. Majied Robinson has fruitfully applied statistical approaches to Arab genealogical literature, thereby illuminating the role of concubinage in the tribe of Muhammad from the sixth to the mid-eighth century. Hend Gilli-Elewy has analyzed a wealth of Arabic source material to uncover the origins of slaves living in Hijaz during the Prophet Muhammad’s lifetime. Chase Robinson has examined practices of enslavement during the Islamic conquest period. Shaun Marmon’s study of the eunuchs who guarded the Prophet Muhammad’s tomb in Medina from the twelfth century onwards offers insights into slavery practices in both Egypt and the Hijaz. My own research focuses on the roles played by slaves at the Najahid and Rasulid courts of Yemen. Two recent works by Craig Perry and Jonathan Miran offer the larger transregional context for slave trading to Arabia in the medieval period. Additionally, scholars studying related topics such as historical commercial connections across the Red Sea and the Indian Ocean, the history of slavery in Africa, and slavery in Islamic law also offer important insights into the subject.

The story of how slavery developed in medieval Islamic societies begins in Arabia. It was here that the religious principles related to slavery were revealed through the Quran and first put to practice by the Prophet Muhammad and his disciples. The way in which the early community of Muslims (umma) related to slaves was both influenced by the region’s previous history of slavery and constituted a significant break from it. The early umma consisted not only of free Arabs, but also included slaves and former slaves. A prominent example among them was Abu Bakra, an Ethiopian slave who joined the nascent Islamic community as an equal member after having been freed by the Prophet Muhammad. The ḥadīth relates that the Prophet Muhammad did not reject slavery outright, but urged his supporters to treat their slaves kindly and to manumit them. Urban has argued that the early umma was a “radically inclusive, faith-based community,” a character that was lost by the time the Islamic scriptural tradition became codified in the eighth to ninth centuries.56 Slaves were first unconditionally welcomed to the early community of believers, but as the growing umma sought to define its boundaries, it became more difficult for slaves, former slaves, and their descendants to be considered full and equal members.

The conquest period of the seventh and eighth centuries saw a dramatic increase in the number of slaves owned by Muslim elites, due to the capture of large numbers of enemies in warfare. The proliferation of concubinage especially, which had been a modest phenomenon in pre-Islamic Arabia, would have far-reaching consequences for the nature of Islamic slavery, as well as for the composition of Muslim families and households. Islamic law, which was codified in the eighth and ninth centuries, reflects these developments, as jurists sought to translate the religious principles pertaining to slavery defined in the Quran and the ḥadīth into a legal apparatus that would regulate lived practice. This apparatus established who could be a slave—only children born into slavery or enemies captured in warfare—and urged masters to treat their slaves kindly. Slave owners were obliged to secure the upkeep of their slaves and provide them with medical treatment. However, no legal sanctions punishing abusive slave owners were introduced. Islamic legal texts presuppose a setting of domestic slavery in which masters possessed a limited number of slaves whom they interacted with personally. This scenario was likely the standard in the formative period of Islam and throughout the medieval period, although a lack of sources precludes us from gaining a full picture of all aspects of slavery. The normative framework around concubinage and manumission in particular was elaborated in minute detail, while other aspects of slavery were largely disregarded in the legal literature. The legal texts failed to account for the fact that the reality of slavery expanded and changed over time and depending on the geographical context, resulting in a growing mismatch between norms and practice.57

Yet, in the case of medieval Yemen, the sources prove that people were aware of the legal framework around slavery and largely sought to follow it. Around the mid-eighth century, after the great conquests were completed and Islamic polities had stabilized, slave trading replaced slave raiding as the main strategy for the acquisition of new slaves. Most slaves reached medieval Arabia from East Africa, having been trafficked as children first via overland routes that connected the African hinterland to the coast, and then across the Red Sea on mixed-cargo ships. Slaves from India appear regularly in sources on medieval Arabia, and enslaved individuals from Byzantium and other parts of the world are also occasionally mentioned. Evidence on the experiences of slavery in medieval Arabia is scattered and fragmentary. A heavy source bias toward male elite perspectives means that the lives of most slaves remained unrecorded. What is however clear is that slaves sustained the lavish lifestyle of Yemen’s upper classes for centuries. As in other historical and geographical contexts, slavery in medieval Yemen was highly gendered. This fact is most starkly on display in the case of concubines and eunuchs, who were valued for their sexuality and presumed lack thereof respectively. While concubines’ bodies were exploited to generate pleasure and offspring for their masters, eunuchs were mutilated to increase their versatility and perceived trustworthiness as servants. Female slaves cooked and cleaned, worked in agriculture and crafts, provided entertainment and sexual gratification to their masters, and bore them children. Male slaves also performed menial, domestic, and agricultural duties, fought and died for their masters, traded on their behalf, and represented their interests as high army and government officials. Most slaves lived within their masters’ households, a fact that should not automatically be considered as limiting the level of exploitation and violence they endured. The legal avenues to freedom were many, but the extent to which they were available to individual slaves depended on their status and on the benevolence of their masters.

Research for this article was conducted within the program “Visions of Community - Comparative Approaches to Ethnicity, Region and Empire in Christianity, Islam and Buddhism,” funded by the Austrian Science Fund (SFB 35). I am grateful to my colleagues at the Austrian Academy of Sciences’ Institute for Social Anthropology for their support throughout this project.  I thank Hannah Barker and Craig Perry for their valuable feedback on the draft chapter. Special thanks to Juliane Schiel and Damian Pargas for bringing together this volume and for organizing the writers’ workshop.

Notes

  1. 1.

    For the purposes of this chapter, the term “Arabia” will be used to denote the geographical area of the Arabian Peninsula, and the medieval period is defined broadly as the time between 500 and 1500 CE. Daniel M. Varisco and Thomas Bauer have convincingly challenged the term “medieval” in reference to Islamic history, but I consider the term to be useful in a comparative context such as this handbook (Daniel M. Varisco, “Making ‘Medieval’ Islam Meaningful,” Medieval Encounters 13, no. 3 (Sept. 2007): 385–412. Thomas Bauer, Warum es kein islamisches Mittelalter gab - Das Erbe der Antike und der Orient (München: C. H. Beck, 2018).

  2. 2.

    Anonymous, Nūr al-maʿārif fī nuẓum wa-qawānīn wa-aʿrāf al-Yaman fī al-ʿahd al-muẓaffarī al-wārif (Lumière de la connaissance. Règles, lois et coutumes du Yémen sous le règne du sultan rasoulide al-Muẓaffar), ed. Muḥammad ʿAbd al-Raḥīm Jāzim, 2 vols. (Sana’a: Centre français d’archéologie et de sciences sociales, 2003).

  3. 3.

    Anonymous, Nūr al-ma ‘ārif, I.362. Medieval Yemeni sources often give the plural of jāriya as jawār instead of the standard plural form jawārī.

  4. 4.

    David Ayalon, Eunuchs, Caliphs and Sultans: A Study in Power Relationships (Jerusalem: Magnes Press, 1999), 63; Ehud Toledano, “The Imperial Eunuchs of Istanbul: From Africa to the Heart of Islam,” Middle Eastern Studies 20, no. 3 (1984): 380.

  5. 5.

    Ayda Bouanga, “Gold, Slaves, and Trading Routes in Southern Blue Nile (Abbay) Societies, Ethiopia, 13th–16th Centuries,” Northeast African Studies 17, no. 2 (2017): 31–60.

  6. 6.

    For an overview of this research, see Craig Perry, “Slavery and the Slave Trade in the Western Indian Ocean World,” in The Cambridge World History of Slavery, eds. Craig Perry et al. (Cambridge: Cambridge University Press, 2021), 149.

  7. 7.

    Yūsuf b. Yaʻqūb Ibn al-Mujāwir, A Traveller in Thirteenth-Century Arabia: Ibn al-Mujāwir’s Tārīkh al-mustabṣir, ed. G. Rex Smith (Burlington, UK: Ashgate, 2008), 162.

  8. 8.

    Yaḥyā b. Sharaf Al-Nawawī, Minhāj al-ṭālibīn wa ʿumdat al-muftīn, ed. Muḥammad Ṭāhir Shaʿbān (Beirut: Dār al-Minhāj, 2005), 220–1.

  9. 9.

    Hannah Barker, That Most Precious Merchandise: The Mediterranean Trade in Black Sea Slaves, 1260–1500 (University of Pennsylvania Press, 2020), 98.

  10. 10.

    E.g. Jerzy Zdanowski, Speaking with their Own Voices. The Stories of Slaves in the Persian Gulf in the 20th Century (Cambridge: Cambridge Scholars Publishing, 2014), 50.

  11. 11.

    Noel Lenski, “Captivity and Slavery Among the Saracens in Late Antiquity (ca. 250–630 CE),” AnTard 19 (2011): 237–66.

  12. 12.

    The Hijaz (also spelled Hejaz) is a region of Arabia located in the northwestern part of what is today Saudi Arabia. The term usually designates the coastal areas on the Red Sea, roughly between the Gulf of Aqaba to the North and ʿAsīr to the South.

  13. 13.

    Hend Gilli-Elewy, “On the Provenance of Slaves in Mecca during the Time of the Prophet Muhammad.” International Journal of Middle East Studies 49 (2017): 166.

  14. 14.

    Richard Pankhurst, “Across the Red Sea and Gulf of Aden: Ethiopia’s Historic Ties with Yaman,” Africa: rivista trimestrale di Studi e documentazione dell’istituto italiano per l’Africa e l’Oriente 57, no. 3 (2002): 398.

  15. 15.

    Timothy Power, The Red Sea from Byzantium to the Caliphate: AD 500–1000 (Cairo: American University Cairo Press, 2012), 127–32.

  16. 16.

    ʿUmāra b. ʿAlī al-Ḥakamī, Al-Mufīd fī akhbār Zabīd, ed. Henry Cassels Kay (London: Arnold, 1892), 6.

  17. 17.

    Nasir-i Khusraw, Nasir-i Khusraw’s Book of Travels (Costa Mesa: Mazda Publishers, 2001), 112.

  18. 18.

    Chase Robinson, “Slavery in the Conquest Period,” International Journal of Middle East Studies 49 (2017): 159.

  19. 19.

    Abū ‘Abd Allāh Muḥammad b. Muḥammad al-Idrīsī. Kitāb nuzhat al-mushtāq fī ikhtirāq al-āfāq, ed. Alessio Bombacci et al., vol. 3 (Napoli: Istituto universitario orientale di Napoli, 1970), 58–66.

  20. 20.

    Shelomo D. Goitein and Mordechai A. Friedman. India Traders of the Middle Ages: Documents from the Cairo Geniza (India Book) (Leiden: Brill, 2007), 453.

  21. 21.

    Robinson, “Slavery in the Conquest Period,” 158; Gilli-Elewy, “On the Provenance of Slaves,” 164–8.

  22. 22.

    Al-Ḥakamī, Al-Mufīd fī akhbār Zabīd, 65; Goitein and Friedman, India Traders, 10, 481.

  23. 23.

    Ibn Al-Mujāwir, Tārīkh al-mustabṣir, 140–3.

  24. 24.

    Perry, The Daily Life of Slaves, 63; Ibn al-Mujāwir, Tārīkh al-mustabṣir, 146.

  25. 25.

    Al-Sakhāwī in Shaun E. Marmon, Eunuchs and Sacred Boundaries in Islamic Society (Oxford University Press, 1995), 39.

  26. 26.

    E.g. Badr al-Dīn Ibn Ḥātim, Al-simṭ al-ghālī al-thaman fī akhbār al-mulūk min al-ghuzz bi-l-Yaman, ed. G. Rex Smith (London: Luzac and Co., 1974), 179; ʿAlī b. Ḥasan al-Khazrajī, Al-ʿUqūd al-luʾluʾiyya fī taʾrīkh al-dāwla al-rasūliyya (The Pearl-Strings; a history of the Resúliyy Dynasty of Yemen), ed. James W. Redhouse and Alexander Rogers, 5 vols. (Leyden: Brill and London: Luzac, 1906), II.294.

  27. 27.

    Gilli-Elewy, “On the Provenance of Slaves”, 164.

  28. 28.

    Power, The Red Sea from Byzantium to the Caliphate, 95.

  29. 29.

    Al-Ḥakamī, Al-Mufīd fī akhbār Zabīd, 81.

  30. 30.

    Majied Robinson, “Statistical approaches to the rise of concubinage in Islam”, In Concubines and Courtesans: Women and Slavery in Islamic History, eds. Matthew S. Gordon and Kathryn A. Hain (Oxford University Press India, 2017), 17.

  31. 31.

    Elizabeth Urban, Conquered Populations in Early Islam: Non-Arabs, Slaves and the Sons of Slave Mothers (Edinburgh: Edinburgh University Press, 2020), 112.

  32. 32.

    E.g. Al-Ḥakamī, Al-Mufīd fī akhbār Zabīd, 38, 72, 78, 88.

  33. 33.

    E.g. Julia Bray, “Men, Women and Slaves in Abbasid Society,” In Gender in the Early Medieval World. East and West, 300–900, eds. Leslie Brubaker and Julia M. H. Smith (Cambridge: Cambridge University Press, 2004); Matthew S. Gordon, “The Place of Competition: The Careers of ‘Arīb al-Maʾmūniya and ‘Ulayya Bint al-Mahdī, Sisters in Song,” in Occasional Papers of the School of ‘Abbasid Studies, vol. 135 (Cambridge: Orientalia Lovaniensia Analecta, 2004), 62–81.

  34. 34.

    Ibn Buṭlān in Antonella Ghersetti, Trattato generale sull’acquisto e l’esame degli schiavi (Catanzaro: Abramo, 2001), 76–7.

  35. 35.

    Al-Ḥakamī, Al-Mufīd fī akhbār Zabīd, 82.

  36. 36.

    Al-Ḥakamī, Al-Mufīd fī akhbār Zabīd, 86.

  37. 37.

    Al-Ḥakamī, Al-Mufīd fī akhbār Zabīd, 71.

  38. 38.

    Al-Khazrajī, Al-ʿUqūd al-luʾluʾiyya, II.183.

  39. 39.

    Al-Ḥakamī, Al-Mufīd fī akhbār Zabīd, 77.

  40. 40.

    In scholarship on Islamic history, slave soldiers are usually referred to as mamluks, while the eponymous Sultanate (648‒922/1250‒1517) is capitalized (Mamluks).

  41. 41.

    Similarly, non-Turkic mamluks in medieval Egypt were given Turkic names as markers of group identity. David Ayalon, “Names, Titles and Nisbas of the Mamluks,” in The Mamluk Military Society, Collected Studies Series 104 (London: Variorum reprints, 1979), 194–95.

  42. 42.

    Al-Khazrajī, Al-ʿUqūd al-luʾluʾiyya, 1906, II.5 ff.

  43. 43.

    E.g. Al-Nawawī, Minhāj al-ṭālibīn.

  44. 44.

    Craig Perry, The Daily Life of Slaves and Global Reach of Slavery in Medieval Egypt, 969–1250 (PhD diss., Emory University, 2014), 68.

  45. 45.

    Al-Khazrajī, Al-ʿUqūd al-luʾluʾiyya, II.20.

  46. 46.

    For example, the daughter of a Najahid prince tried to evade the sexual advances of a powerful vizier by handing over to him “forty virgins from among her jawārī.” (Al-Ḥakamī, 73).

  47. 47.

    Shaun Marmon, “Intersections of Gender, Sex, and Slavery: Female Sexual Slavery”, in The Cambridge World History of Slavery, Volume 2: AD 500–AD 1420, ed. Craig Perry, David Eltis, Stanley L. Engermann, and David Richardson (Cambridge University Press, 2021).

  48. 48.

    Anonymous, Nūr al-maʿārif, I.104.

  49. 49.

    Ibn Ḥātim, Al-simṭ al-ghālī, 110–2; Anonymous, Nūr al-maʿārif, I.373; Al-Khazrajī, Al-ʿUqūd al-luʾluʾiyya, II.217; and Ibn al-Mujāwir, Tārīkh al-mustabṣir, 186.

  50. 50.

    Ibn al-Mujāwir, Tārīkh al-mustabṣir, 145; Muḥammad aṭ-Ṭayyib b. ʿAbd Allāh Bā Makhrama, “Tārīkh taghr ‘Adan”, in Arabische Texte zur Kenntnis der Stadt Aden im Mittelalter, ed. Oscar Löfgren (Leipzig: Harrassowitz, 1936), A.9; and Anonymous, Nūr al-maʿārif, I.294. II.353, I.126, I.175.

  51. 51.

    Goitein and Friedman, India Traders, 66 ff. Shelomo D. Goitein, Letters of Medieval Jewish Traders (Princeton: Princeton University Press, 1973), 13; Ṣubḥī Labīb, Handelsgeschichte Ägyptens im Spätmittelalter: 1171–1517 (Wiesbaden: Steiner, 1965), 112, 287, 489.

  52. 52.

    ʿAlī b. Ḥasan al-Khazrajī, Al-ʿIqd al-fākhir al-ḥasan fī ṭabaqāt akābir ahl al-Yaman, wa huwa ṭirāz aʻlām al-zaman fī ṭabaqāt aʻyān al-Yaman, ed. ʻAbd Allāh b. Qāʼid ʻAbbādī et al. (Sana’a: Maktabat al-jīl al-jadīd, 2009), 2502.

  53. 53.

    Al-Khazrajī, Al-ʿUqūd al-luʾluʾiyya, II.175.

  54. 54.

    Abū l-ʻAbbās Aḥmad b. Aḥmad al-Sharjī al-Zabīdī, Ṭabaqāt al-khawāṣṣ ahl al-ṣidq wa-l-ikhlāṣ, ed. ‘Abd Allāh al-Ḥibshī (Ṣanʿāʾ: Al-dār al-yamaniya li-l-nashr wa al-tawzi ‘, 1992), 120–21.

  55. 55.

    Al-Khazrajī, Al-ʿIqd al-fākhir, 1688.

  56. 56.

    Urban, Conquered Populations, 50.

  57. 57.

    Kurt Franz, “Slavery in Islam: Legal Norms and Social Practice”, in Slavery and the Slave Trade in the Eastern Mediterranean (c. 1000–1500 CE), eds. Reuven Amitai and Christoph Cluse (Turnhout: Brepols Publishers, 2018), 51–141.