Slavery in First Millennium Arabia: Epigraphy and the Qurʾān

1 Introduction

In the recent The Cambridge World History of Slavery. Volume 2. AD 500–AD 1420, Matthew S. Gordon argues that “Slavery was […] well-rooted in pre-Islamic Arab society.”^[1] Although Gordon recognizes that “it is difficult to reconstruct Arabian social and economic history more generally”, he upholds slavery as playing a crucial role in pre-Islamic times, which he defines as “a legacy of ancient Greek, Roman, and Iranian traditions”.^[2] In the Oxford Dictionary of Late Antiquity, Majied Robinson writes that “Slavery would have been familiar to pre-Islamic Arabians of the Hijaz”,^[3] and the assumption that slavery was a widespread practice in pre-Islamic Arabia is overall common. Although no mention of pre-Islamic Arabia is found in another recent publication on the topic, Slavery in the Late Antique World, 150 – 700 CE,^[4] many publications in the past have argued for the presence of enslaved people in pre-Islamic Arabia.

In a short paper published before his 2019 book Slavery and Islam, Jonathan Brown argued that “slavery was an integral part of economic and social life in the Late Antique world, and the Quranic revelation accepted it as a reality.”^[5] In a later passage, Brown claimed that “in the cases of polygamy and slavery, there is salient continuity between pre-Islamic and post-Islamic society in the Near East.”^[6] This is arguably not Brown’s most controversial statement, but it is nonetheless quite problematic.^[7] In the same year of the appearance of Slavery and Islam, Bernard K. Freamon published Possessed by the Right Hand. The Problem of Slavery in Islamic Law and Muslim Cultures. Freamon’s text devotes an entire chapter to “Slavery, Slave Trading, and the Law in the Pre-Islamic Middle East.”^[8] Less than two pages among the sixty-eight of the chapter are dedicated to an Arabian region, i. e., the northwest kingdom of the Nabateans. Freamon recognizes that the assertion that the Nabateans traded in enslaved people can appear as rank speculation.^[9] Still, he claims that “The pre-Islamic literature and lore from the region is replete with references to Africans functioning as slaves”.^[10] To support his claim, Freamon refers to B. Heller’s entry “Sīrat ʿAntar” in the Encyclopaedia of Islam^[11] and a brief passage on ‘Antar in Bernard Lewis’s 1992 Race and Slavery in the Middle East: An Historical Enquiry.^[12] Poetry and romance are far from authoritative sources on the matter. Nevertheless, relying on unreliable sources is still better than relying on no sources. For instance, Hend Gilli-Elewy has claimed that “Arab prisoners of war” constituted the main category of enslaved people in pre-Islamic times but has failed to mention a single primary source.^[13] Gilli-Elewy’s bold statement that the Meccans “dominated the international trade network” in the second half of the sixth century is instead based on secondary literature.^[14] Kecia Ali has also claimed that “patriarchal marriage and slaveholding coexisted in pre-Islamic Arabia” but failed to mention any pre-Islamic Arabian source.^[15]

The attestation of slavery in the Near East has led many scholars to assume that the practice of slavery also thrived in the Arabian Peninsula. In the section of Slave Soldiers and Islam devoted to the “pre-Islamic antecedents” of military slavery in Islam, Daniel Pipes wrote that the three “ancient cultures” of Rome-Byzantium, Iran, and “Arabia” “contributed to the sudden, new, and original Islamicate synthesis”.^[16] After briefly dealing with the Roman and Iranian antecedents, Pipes reserved a brief section to “Jahili Mecca”, arguing that “information on slaves fighting in pre-Islamic Arabia derives almost entirely from Mecca in the years preceding Muslim control”, and that “Mecca consistently used slaves in warfare”.^[17] Muslim sources are the only investigated material. By contrast, Noel Lenski’s excellent article Captivity and Slavery among the Saracens in Late Antiquity (ca. 250 – 630 CE) does not use Muslim literary accounts but literary sources produced outside Arabia though contemporary to the events narrated, such as the works of Ammianus Marcellinus, Malalas, Jerome, Evagrius, Theophanes, Cyril of Scythopolis, and Dionysius of Alexandria.^[18]

The following sections address the imbalance between the many secondary references to Arabian slavery and the present paucity of primary sources by focusing on the material written by Arabians in Arabia during pre-Islamic times.^[19] The analysis of the indigenous corpus of Arabian epigraphy will be followed by an inquiry into the first Arabic literary source, which offers some insight into the topic: the Qur’ān. The conclusion will attempt to situate Arabia in the broader context of Late Antiquity. In the article mentioned above, Lenski argued that the “Muslim conquerors” continued the pre-Islamic “habits of capturing, handling and dealing with slaves cultivated by their Saracen predecessors”.^[20] Assuming that it is possible to trace a direct line between the “Saracens” mentioned in the Greek and Syriac literary sources, was there continuity from pre-Islamic to early Islamic times regarding the practice of slavery? While attempting to answer these questions, I will provide an overview of the heated debate on “slave societies” and “societies with slaves”.^[21]

2 Pre-Islamic Epigraphy

Pre-Islamic Arabia is a particularly fertile site for a historical inquiry into literacy and linguistic habits in the ancient world. The present study is based on several epigraphic testimonies found in the region. Thousands of inscriptions written in various scripts and languages bear witness to several communities and geo-political entities. The already mentioned Nabateans developed their script from Aramaic in the second century BCE. After the fall of the Nabatean kingdom in 106 CE, the disappearance of monumental inscriptions gave way to the emergence of thousands of graffiti, some of which have been identified as “Nabateo-Arabic”, “Transitional”, and “Old Arabic”.^[22] Two additional corpora of the Arabian Peninsula are today labeled “Ancient North Arabian inscriptions” and “Ancient South Arabian inscriptions”. The first group is attested in the south of Mesopotamia and northwest Arabia from the first millennium BCE.^[23] One of the most notable subcategories is “Safaitic”, which comprises thousands of short graffiti in southern Syria and dates between the second century BCE and the fourth century CE.^[24] The South Arabian corpus is more accurately dated than its North Arabian counterpart based on its attestation of a local calendar and its mentions of ruling dynasties. The corpus dates between the early first millennium BCE and the second half of the sixth century CE, and it is written in four different languages, named after the four most powerful kingdoms of the area (Ḥaḍramawt, Qatabān, Ma‘īn and Saba’).^[25] Due to the average length of these texts, it is one of the most valuable sources for uncovering the history of pre-Islamic Arabia.

Several South Arabian inscriptions offer insight into the region’s society in pre-Islamic times. At least four lexemes indicate a person in a subordinate position in the corpus. ‘bd is the most generic way of referring to a person wholly subject to another, being attested more than five hundred times. The word is used in religious and mundane contexts: a man can be a ‘bd towards divinity or another human being.^[26] Similar blurry connotations are attested in other corpora, such as the Sogdian documents from Turfan.^[27] In a temple in wādī Shuẓayf, a ʾs¹dm son of Ḫḏym who was an ʿbd of Kbr Ḫll dedicated a bronze plaque to the god dhū-Samawī.^[28] On the same site, a smaller bronze plaque was dedicated by S¹ʿdn dh-Nṣrm ʿbd of ʿbṯ Hbs²nyn, (i. e., ʿbṯ the Aksūmite).^[29] In another context, the ʿbd of the family Nʿmbrl and Ḥbt dedicated two statues in bronze to the god ʾlmqh “when he returned safely from S²bwt and the sea”.^[30] An offering table contains an Early Sabaic inscription with a similar formula.^[31]

It is difficult to ascertain whether the people defined as ‘bd of other men were “enslaved people”. An Early Sabaic inscription mentions a “repair” carried out by a man who defined himself ʿbd of two rulers.^[32] In another case, a man who defined himself not only as a ʿbd but “property of the king” (qny mlkn) declares having surrounded a temple with a wall and having constructed an altar in a Southern Middle Sabaic dedicatory text.^[33] Elsewhere, a certain ʾs¹ʿd son of S¹ʿdn and ʿbd of the banū ʾḥwḍ constructed the funerary monument and tomb of his brother Yʿmr, son of ʿmrn and ʿbd of the banū ʾḥwḍ.^[34] Furthermore, a bronze bowl contains a carved inscription: “Mrkḍm ʿbd ʿnnn placed for [the god] S¹mʿ for his security and to protect his work”.^[35] This would have been an expensive object, as the bronze sculpture of a hand dedicated by a ʿbd of the family S¹ḫmn to his patron Tʾlb Rymn.^[36] Christian Robin translated ʿbd with the French lexeme “esclave”, a term corresponding to the English “slave” and not “servant”, stating that the writer’s “position of dependence seems to be without time limit”.^[37]

A unique insight into the relationship between ʿbd and overlord comes from a legal inscription datable to the end of the first millennium BCE. The inscriptions concern the granting of a portion of a tomb (ʾṣbʿ b-qbrn) to the family of a ʿbd ml[kn] (ʿbd of the king).^[38] Irene Rossi has recently translated the apposition as “servo del re”, claiming that the term “indicated the king’s employees or members of the court. In any case, they were certainly people who enjoyed a position of respect”.^[39] If so, the inscription would allow for the hypothesis that the ʿbd were not enslaved people but special attendants of the elites’ entourage. Although prisoners and privileged guests at royal courts were often one and the same, clear proprietorial rights over a ʿbd are mentioned in a wood stick inscription which reads:

To Whr of the Ḥbrn from Yʿll, ʿbd of the ḏ-Grfm. As for you, take care of the ʿbd of the ḏ-Dwrm, who was sent with S¹bʿm, and do not shame this man. Yʿll.^[40]

Proprietorial rights are also attested in a monumental inscription on a statue aimed to thank the god ʾlmqh for having granted the writer “proprietorial rights over Ṯwbṯwn, son of Mrdm ʿbd of the ḏ-Ḥbb”.^[41] A dedicatory text records that a ʿbd was hit and killed (“Hwtrʿṯt, of the family ʾḥgrm, hit his ʿbd Ṯwbʾl, servant of the family ʾḥgrm, and just for that beating Ṯwbʾl died”).^[42] Another inscription records a ʿbd stealing from a temple (“ʿ[.]m ʿbd of the banū Mwqṣm stole from the temple of Bʿl ʾwʿln”).^[43] There is no mention of punishment. If we assume that the term ʿbd has a somewhat uniform meaning throughout the corpus, a lengthy Early Sabaic monumental stele appears in contradiction with Rossi’s thesis. The stele mentions ʿbd as war prisoners and thus unlikely to be “people who enjoyed a position of respect”.

He brought together the children of ʾlmqh and the people, its free men (ḥr) and its ʿbd from the territories of Awsān and its cities […] and the men of ʿAbadan, free men (ḥr) and ʿbd, with full ownership rights […] and Krbʾl took possession of all the qs¹ṭ of Kḥt, its free men (ḥr) and its ʿbd and their children and their property and all the men.^[44]

Since the word ʿbd is in contrast with ḥr in this passage, ʿbd seems to indicate enslaved people, who had been taken as captives. The same inscription also includes elsewhere the word s¹bym, which could be translated as “slavery” (“demanded them from him, in death and s¹bym”). This lexeme is attested around one hundred times in the corpus of South Arabia, from the first millennium BCE to the fall of the kingdom of Ḥimyar in the second half of the sixth century CE. The word indicates “captives” in Central Middle Sabaic and Southern Middle Sabaic inscriptions.^[45] Both men and women are considered captives,^[46] and they often occur with other elements of booty, such as animals.^[47] In a few cases, the raids are against the Aksumites^[48] (“they came back with spoils, captives, and booty from the Abyssinians (ʾḥbs²n)”).^[49]S¹bym also appear in a few Late Sabaic inscriptions.^[50] For instance, a dated commemorative text written by one of the qayl protagonists of the massacre of the Christians of Najrān (523) by the Jewish ruler Joseph mentions “twelve thousand deaths, eleven thousand captives (s¹bym), two hundred ninety thousand camels, cows and small animals”.^[51] Two contemporary inscriptions refer to “nine thousand five hundred captives”^[52] and “eleven thousand captives” after a series of military accomplishments against the South Arabian Christians and the Aksumites.^[53] The second commemorative text also mentions the word ḫmrtm, which means “enslaved people”.^[54] The widespread attestations of the term in various ages witness the deportation of enslaved people held as captives in all societies of South Arabia.

In contrast to most traded goods, enslaved people were not weighed, as was the case with animals. Not many “receipts” of their trade have thus survived. Two notable exceptions are constituted by a receipt on a palm leaf stalk plausibly mentioning the purchase of three enslaved women^[55] and a stick wood incision allowing the Banū S²hrm to “acquire, go into possession of, possess, command, sell and exchange them or exchange something for those two women as well as for their children and grandchildren as they are proprietary right”.^[56] Enslaved women frequently served as concubines in neighboring societies, and conspicuous attestations of enslaved women are found in the epigraphic corpus of South Arabia. Three Southern Middle Sabaic inscriptions dated to the dawn of the first millennium CE use the lexeme ḫmrtn to indicate enslaved women,^[57] which appears equivalent to ʾmt (plural ʾmh), attested in Minaic, Northern, and Central Middle Sabaic inscriptions.^[58] A stone inscription from Ṣirwāḥ similarly attests to a father’s selling off his daughter to the god ʿṯtr S²ymm, after having procreated her with his ʾmt whom he had purchased (qny).^[59]

The epigraphic corpus clearly shows that it was possible to buy ʾmt/h, suggesting that they were enslaved. Some of these people were captives of war, as indicated by a short commemorative text (“These has captured Gḥmt, of the house of Yṯʿʾmr, his ʾmt, at ʾrydy and Dnm: 20”).^[60] A bronze plaque found in a temple condemns the sexual violence carried out by a male servant/enslaved man (ʿbd) of an Aksumite man towards a female servant/enslaved woman (“he did violence to the ʾmt of his master”).^[61] The name of the female servant is missing, so we cannot infer if she was an Aksumite too. However, it is unlikely that enslaved people were only captives of war, as the onomastics of the inscriptions suggests that some indigenous people were also classified in this category. There is some uncertainty regarding the exact status of the son or daughter of an enslaved person, though children of enslaved people were usually considered to be the property of the owners of their parents. It is thus unclear if “the descendants of ʾmt S²[.]”, mentioned in a monumental text, need to be considered enslaved.^[62] Four Late Sabaic inscriptions bear the lexeme hgn, which points to a freedman but could potentially refer to enslaved mothers’ sons too.^[63] The hgn did not hold the exact position of the ʾḥrr, freemen,^[64] and ʾḥrrt free women,^[65] while free men were considered in opposition to s¹byw,^[66]ʿbd,^[67] and ʾdm.^[68] The latter, attested more than three hundred times, could be translated as “liegeman” or “subject to the king”.^[69] The word appears to a certain extent similar to ʿbd, as it was also used in religious contexts as “servant of a divinity”.^[70]

A preliminary inquiry into the Ancient South Arabian corpus has shown that a series of words referred to people in subordinate positions. While ʿbd and ʾdm occur in religious contexts to indicate subordination to a deity, the terms s¹bym are only employed to indicate war captives in subordinate position to elites. Ḫmrtm and ʾmt are used to indicate female servants. A freedman is a hgn, while freemen and free women are ʾḥrr and ʾḥrrt, often found in opposition to s¹byw, ʿbd, and ʾdm. I have already mentioned the case of a long Early Sabaic inscription.^[71] A Ḥaḍramitic inscription datable to the late second to late third centuries CE exemplifies the stratification of society in pre-Islamic South Arabia.

Ṯwybm and his sons

Mrṯdʿm and ʾbkrb,

of the family Wdm,

dedicated to their lord

Wrfw ʾmrʿm a statue

in bronze for their safety

and for the safety of their house,

their free men (ḥrr) and their subjects (ʾdm).

They committed to the protection of

Wrfw their persons, as a protection

and as goodwill which is favorable to

them and to all they possess in ḏt

Ġylm and its two valleys.^[72]

The inscription shows the close-knit between freemen and servants in pre-Islamic Arabia. People who lay in subordinate positions could carve inscriptions on bronze plaques. This could suggest that they were not enslaved but rather dignitaries and liegemen, but there is some degree of opposition between “freemen” and “subjects”. Either way, people in subordinate positions could commission the construction of expensive votive objects. Not differently from ‘bd, a ʾmt was able to dedicate steles to ask the gods for wellbeing,^[73] or ask for forgiveness (“because her daughter ʾbʿly drew water from the well of ʿdn when she was not pure”).^[74] While the material of several steles is stone, two enslaved women (plural ʾmh) dedicated a bronze statue to the god ʾlmqh,^[75] and two more were dedicated to the god Ns¹rm by a mother and a daughter who defined themselves as ʾmh ʾlt Wdnʿm (“enslaved women of the family Wdnʿm”).^[76] Similarly, a mother and a daughter in pilgrimage left a bronze statue in a temple, right before the kingdom adopted monotheism,^[77] and two more bronze statues celebrated ʾlmqh’s grant of a daughter to other ʾmh.^[78] Statues, likely representing the dedicants, were often the preferred material for divine offerings in South Arabia due to their material value.^[79]

At the dawn of the first millennium CE, a certain South Arabian Ḫwlyt, ʾmt of Sulaymim, “clothed herself with an unclean cloak and a woolen garment that had become unclean” and concealed this from “her masters” (ʾmrʾh).^[80] A bronze plaque found in a temple commemorates the event. She had to “confess and repent before ḏṭ-Samāwī, the Lord of the (temple) Bayyin” and now begs Dṭ-Samāwī to “restore her happiness” after she “had submitted and humbled herself”. Ḫwlyt “had sinned”, but she will “pay a fine” to receive ḏṭ-Samāwī’s forgiveness. The bronze plaque attests to the story of Ḫwlyt, the ʾmt, a “woman in a subordinate position”, “enslaved woman”, or “female servant”. In the remaining space of this section, I suggest that bronze plaques played a profound function in South Arabian society. Most of these objects were not mere votive objects but served as legal texts: the culprit begs for the deity’s forgiveness after confessing their sins. The bronze materiality made the plaques a sort of “divine covenant”. The objects were probably read aloud, and a ceremony involving all social strata plausibly took place. Religious and civil practices did not have firm boundaries, so the reasoning behind making these votive tables was both sacred and profane. Bronze plaques attested to the contrition and expiation of culprits; elements fundamental to establishing the correct functioning of the South Arabian social order.

The archive of the temple Yġrw of ḏ-S¹mwy suggests that the temple was a site in which the pilgrimage of both kings and people in subordinated positions was the norm, in contrast to the above-mentioned temple of Jupiter in Capua.^[81] In 1997 François Bron published four inscriptions from the temple, including the one by S¹ʿdn ḏ-Nṣrm.^[82] All inscriptions are on bronze plaques. The first is written by a certain ʾs¹dm bn Ḫḏym, ʿbd of Kbr Ḫll, in the hope that “dhū-Samawī continues to grant to His servant (ʿbd) Asadum the prosperity and benevolence of his lords, the banū Kabīr Khalīl”.^[83] One of the four inscriptions is dedicated by an Aksumite.^[84] Accordingly, two ʿbd and a third man, supposedly a free Aksumite, were able to dedicate bronze plaques to ḏ-S¹mwy’s temple. While unfeatured in Bron’s article, six more bronze plaques relate to the same temple. To mention a few examples: a) Ḥllm, son of Dhyt, the Aksūmite (Hbs²nyn), confess to having worn an unclean belt in the temple and having had sexual intercourse with a woman inside the building;^[85] b) ʾlʿz, son of Nhyt, begs forgiveness for having “failed to perform for him [ḏ-Smwy] a necessary rite on the occasion of an expedition”^[86]; c) Rbbm, son of Rdʿm, wishes to receive a divine oracle before embarking on an expedition;^[87]d) ʿlym, son of Qys¹mnwtm, confesses to the god of ḏ-S¹mwy of having had sexual relations with a woman inside the temple.^[88] The last text is found on a bronze pendant displaying a man and a woman who had sexual intercourse. Further considerations of this object can help us better understand the role of bronze plaques.

Manfred Kropp has recently suggested that the pendant was worn by the penitent who processed around the site after his confession.^[89] The aim of expiatory plaquettes is not entirely dissimilar from inscribed bronze collar pendants worn by Roman enslaved people with deterring phrases such as noli me tenere, non tibi experet/expedet (“do not hold me, do not try it/it is not good for you”).^[90] In most cases, bronze plaques were plausibly nailed to the temple’s wall, but in others, such as the case of the pendant, they were to be worn by the culprit. The aforementioned bronze hand could have also played a role during the procession after or before the trial, possibly representing a replacement for the penalty of amputation.^[91] These objects need to be considered not as mere votive objects but as legal texts that formed part of a society’s official archive and regulated the life of the South Arabian stratified society. In doing so, we can better understand how rulers, freemen, and servants were able to leave their testimonies in the same temple. Bronze inscriptions were legal texts aimed at surveilling the functioning of the stratified society of pre-Islamic South Arabia. The literacy level was plausibly low, so the inscribed objects were read aloud when offered to the temple and later held a symbolic role, becoming a monumental display of social justice and a witness to the devotee’s faith, publicly displaying his/her contrition. It is possible that the dedication of the bronze plaques was to be celebrated annually. A decree by the city of Pergamum attests that offering frankincense and sacrifices took place after the temple’s dedication of bronze tablets.^[92] The day became a public holiday for children and enslaved people. A parade was part of the celebration, whose total cost was covered by the local administration.^[93] The non-perishable material of bronze tablets made dedications in Arabia and elsewhere (a widespread Roman practice) inviolable and eternal. In some cases, the tablets may have even been paid by masters and rulers, with whom people in subordinate positions had a strong relationship, being considered parts of their masters’ household.

Before moving to notions of slavery in the Qur’ān, I would like to spend a few words on the mention of people in subordinate positions in the Ancient North Arabian corpus as we do find several evidence of a stratified society in the upper Arabian region in the regional epigraphic corpus. While a freedman is a ʿtq,^[94] a “male servant/enslaved man” is often a fty and a “female servant/enslaved woman” a ʾmt; the words ġlm/ġlmt, generally meaning “boy/girl”, also feature to express the concept.^[95] The word ʿbd appears similar to ʿbd in the South Arabian corpus, being attested in religious contexts expressing subordination to a deity and in more mundane ones.^[96] A Dadanitic inscription from al-ʿUlā records the offer of a what was possibly an enslaved boy (ġlm) made by a priest of the Minaean god Wd and his two sons to Dhġbt, the main deity of the kingdom of Liḥyān.^[97] In some cases, a drawing is featured in the inscriptions. This is the case of a series of Safaitic graffiti mentioning ġlmt and which they likely represent visually.^[98] However, most of the Ancient North Arabian inscriptions are brief stone graffiti bearing a few words, such as the following Safaitic graffito from Jordan:

l ʿḏr bn ʿd bn ʿbdt h- dr w ṣyr b- ʾmt m- h- bʿlt

By ʿḏr son of ʿd son of ʿbdt was here and brought back a female servant from the wedding.^[99]

While the author of another Safaitic inscription invokes the goddess Lt and describes himself as “fty of Gʿd, son of ʿbṯn” who served in the army, as “he waited for the rains”,^[100] a man in some form of subordinate position left his name on a rock while “pasturing the sheep and leading a troop astray here”.^[101] The widespread use of writing stimulated questions about the region's literacy. Michael Macdonald has suggested that nomads wrote the Ancient North Arabian inscriptions found in the Syro-Arabian desert to relieve boredom during their often-solidary vigils as “writing never usurped the functions of memory and oral communication”.^[102] Ahmad al-Jallad has highlighted the highly formulaic structure of the Safaitic inscriptions suggesting that their writers “did not simply learn how to write, but what to write as well”.^[103] Al-Jallad has rightly pointed out that if writing was used to pass the time, certain names would be attested more than once.^[104] As most writers produced only a few graffiti in their lifetime, al-Jallad has suggested that the graffiti aimed to be “aesthetic depictions of reality”.^[105] People who appear to have been servants were also taught to use the same formulae and writing traditions to convey their still-life portrayals, pointing to their strong connection with freemen in pre-Islamic times.

3 Qur’ān

Enslaved people are mentioned in more than twenty-five Qur’anic verses, mostly dated to the Medinan period.^[106] They are often mentioned as “that which your/their right hands own” (mā malakat aymānu-kum/-hum/-hunna/yamīnuka), though the word ʿabd is also attested in the generic sense of “servant” and “worshipper”.^[107] When referring to enslaved people, the word can appear in contrast to a free person (ḥurr)^[108] or found with the attribute mamlūk (possessed).^[109] While the Sabean ʾmt is attested as amah, pl. imāʾ,^[110] other words refer to enslaved people more ambiguously.^[111] A verse of sūrah al-Naḥl attests to the difference between free and enslaved people in Muhammad’s community:

And Allāh has favored some of you over others in the provision, but those favored would not share their provision with that which their right hands own, making them equal. Then do they reject Allāh’s favor?^[112]

The separation between freemen and enslaved people is considered part of God’s plan for humanity. At the same time, freemen and enslaved people were deemed spiritually equal, so it was permissible to marry an enslaved woman, if a free woman was not an option, given that she was a believer.^[113] Manumission as an act of piety is encouraged and can become an instrument to expiate sins.^[114] We do not have information on the process of manumission in pre-Islamic Arabia. However, the inscriptions written by freed people point to the fact that manumission was a common practice in the region before the rise of Islam and not an innovation of Muhammad’s preaching. Studies on Jewish^[115] and Christian^[116] slavery have shown that the practice remained widespread after the rise and spread of Scriptural monotheism. Being war captives, slave trade, and birth to enslaved people were conditions for becoming enslaved people at the time of Muhammad. There is continuity concerning the Islamic condonement of enslaved women as concubines too,^[117] although the prostitution of enslaved people is condemned.^[118] As previously mentioned, marrying a believing concubine was permissible.^[119] This constitutes a significant difference from Roman practices, which also encouraged emancipation but did not allow marriage between free and enslaved people.^[120] One fascinating aspect regarding the condition of enslaved people in Muhammad’s society is the emphasis on enslaved people as part of the household. They are mentioned in a similar fashion to wives (azwāj)^[121] and underage children (lit. “those who have not reached puberty”, lam yablughū al-ḥuluma).^[122] A significant verse on the matter is found in sūrah al-Nūr, a few verses before a statement allowing marriage with enslaved men and women of good character (“marry off the singles among you, as well as the righteous among your enslaved men and women”),^[123] and another on masters encouraging manumission and condemning forced prostitution.^[124] The verse precedes these statements and exhorts pious women to a modest demeanor. The central part of the verses declares:

[…] And let them draw their headcovers over their bosoms and not display their adornment except to their husbands, their fathers, fathers of their husbands, their sons, sons of their husbands, their brothers, sons of their brothers, sons of their sisters, their women, what possess their right hands, or male attendants with no physical desire or children who are not aware of private aspects of the women. […]^[125]

The verse suggests that enslaved people were considered close family members. They were also considered part of the community of believers (Ummah); their manumission was encouraged, and forced prostitution was strictly forbidden. The institution of slavery evolved following the Muslim conquests and the enlargement of the community of believers. However, it is significant to notice the close-knit social framework of household slavery at the dawn of Islam. This portrayal is in line with the pre-Islamic society of South Arabia, where believing enslaved people and free people (including kings) frequented the same temples and were plausibly deemed equal to the eyes of the god. Converted captives may have been welcomed as part of this close-knit society, though “enslaving the other” remained a powerful instrument of control and rhetoric. As Max Weber puts it, “the slave market depended for supply on successful wars”.^[126] It is plausible that ethnicity played a role in pre-Islamic and early Islamic slavery practices in a not-so-different fashion from ancient Greece.^[127] In the twelfth century, however, faith had become the sine qua non to decide which Africans were to be enslaved by Muslims.^[128] The society of pre-Islamic Arabia, with its at least apparent egalitarian access to temple votive objects, preceded the Ummah’s granting of citizenship (and the consequent creation of otherness) on the basis of faith.

4 Conclusion

Sometime in the second century CE, a tribeswoman of the Catuvellauni named Regina died in South Shields, a British coastal town just south of the River Tyne. Her husband was a wealthy Palmyrene man named Barates, who commissioned a lavish tomb in Palmyrene style for his deceased wife. Beneath the woman’s portrait, a bilingual inscription in Latin and Palmyrene reveals that she was a liberta/hry, a freedwoman.^[129] Two centuries later, an elaborate Roman villa was being built on the opposite side of Europe, at the foot of a hill in central Sicily. The unknown dominus commissioned marbles, frescoes, and mosaic pavements. A high-ranking officer beating enslaved people with a stick appears as one of the decoration subjects of the mansion. Elsewhere in the villa, a fresco shows two enslaved men carrying a wild boar on their shoulders.^[130] Enslaved and freed people populated Roman citizens’ physical and symbolic worlds, as suggested by the author of the Periplus of the Erythraean Sea, who mentions that enslaved people were traded in Roman Egypt, East Africa, India, and the Arabian Peninsula.^[131] A comparison between ancient South Arabia and Rome can elucidate some of the questions posited by this paper.

According to Anne Kolb, even enslaved people could afford funerary monuments in ancient Rome.^[132] Inscriptions mentioning enslaved people mainly consist of epitaphs (masters were required by law to provide for the burial of a deceased enslaved person)^[133] and, more sporadically, legal texts.^[134] Most of the “enslaved people” mentioned in the inscriptions were not servus/a, but formerly enslaved people, i. e., libertus/a (freed people), eager to erect memorials to celebrate their manumission.^[135] Enslaved people usually set up Roman votive inscriptions according to their master’s orders.^[136] Only one marble statue appears to offer a dedication by an enslaved man to his master. The statue must have been set up within private property as enslaved people were not allowed to erect a public monument on prominent sites.^[137] Although “no honorific representation of slaves is known from late antiquity”,^[138] a base from Isernia (Italy), plausibly part of an altar, bears the inscription Attalus Noni M(arci) S(ervus) and is likely an honorary monument dedicated to Imperator Nonius by his servant Attalus to thank him for receiving him into his familia.^[139] The base represents a thanksgiving sacrifice made by the Attalus to the goddesses Victoria, Fortuna, and Nemesis. John Scheid has recently suggested that the depiction portrays a suovetaurilia and the founding of an altar to Fortuna Augusta in the early first century CE.^[140] The sacrifice hints at a communal context, possibly that of a “collegium of slaves or ministers” of which Attalus was possibly the head.^[141] Scheid firmly states that “Attalus was no ordinary personage”, being the servant of the great family of Aesernia, the Nonii Gauls.^[142] Therefore, the ministers of the cult of Fortuna Augusta recruited from the family of a large local house, which included freedmen and enslaved people.^[143] This was not a typical situation, as enslaved people were usually excluded from public rites.^[144]

Although Attalus’ inscriptions play with religious tropes, the inscription’s recipient is another human being, i. e., his own master. A fair number of Roman votive dedications by enslaved people were dedicated to Lares Familiares (household deities).^[145] Enslaved people or freedmen were responsible for worshipping them, as in the case of Silvanus’ cult.^[146] A few bronze inscriptions appear to have been produced by enslaved people in the Roman world. Two bronze inscriptions from Sardinia appear to have been part of votive offerings by two enslaved men (donum dedit). However, due to their brief texts, any historical reconstruction is speculation. Two Latin bronze inscriptions, one discovered in ancient Summus Poeninus (Switzerland)^[147] and one in Monte Tifata (Campania, Italy)^[148] containing the widely attested abbreviation v.s.m.l. (v(otum) s(olvit) m(erito) l(ibens), i. e., “he/she has fulfilled his vow deservedly”), have been attributed to enslaved people. The one from Monte Tifata bears the abbreviation ser. for servus after the name Aprilis.^[149] Aprilis asks for the wellness of his master Publius Campanius Dexianus to the local god Jupiter who was venerated on site.^[150] The devotees mentioned on the Capuan site are of “low social prominence”, and it does not seem that diverse social strata simultaneously venerated the god.^[151] Strict lines of demarcation divided Roman social classes. Epigraphy reflects them. This paper has shown that South Arabians employed various terms to refer to people in subordinate positions, but the lines of demarcation between Arabian social classes are more blurred than their Roman counterparts. For instance, the terms ʿbd and ʾdm were used to indicate subordination to a deity or a human being. In contrast, s¹bym indicated war captives, which were often mentioned along with other elements of booty, such as animals. They called women in subordinate positions ḫmrtm and ʾmt. A freedman was a hgn, while freemen and free women were ʾḥrr and ʾḥrrt. As wood sticks and monumental inscriptions mention proprietorial rights over ʿbd and list the ʿbd as war prisoners, I argued that it is unlikely that all people forming this group enjoyed a “position of respect”, other than, perhaps, being conferred upon them as the property of a powerful owner. The attestations witness the deportation of enslaved people held as captives in both the polytheistic and monotheistic past of South Arabia. As in neighboring societies, enslaved women plausibly served as concubines. Although some were probably war captives, servants were also recruited among the local population, as suggested by the inscriptions’ onomastics.

In line with Roman practices, I have highlighted that all social classes of ancient South Arabia appear to have been able to carve inscriptions on bronze, the preferred material for votive offerings. I have suggested that in some cases, the making of votive objects may have been facilitated by masters who conceived of their servants as part of their household. If a servant committed an illegal action, the master would subsidize the cost of a bronze tablet to beg for the god’s forgiveness. Religious practices and civil practices did not have firm boundaries in ancient times. Most of these objects should not be considered mere votive objects but legal texts with a deep ceremonial meaning, bounding a person to his/her mundane society and the divine. As such, they attested to the deep faith of believers and pilgrims and the culprits’ expiation of sins in front of a god, showing their proper position in the social order of South Arabia. An exemplary case is offered by a bronze pendant containing a public confession and displaying a man and a woman having sexual intercourse, stemming from the archive of the temple Yġrw of ḏ-S¹mwyw, frequented by both kings end enslaved people. The literacy level was plausibly low, so bronze inscriptions were read aloud and later became monumental displays of social justice.

Before moving to study the Qur’ān, I briefly addressed the question of literacy prompted by the study of graffiti written by enslaved people in the Ancient North Arabian corpus. I have concluded that people in subordinate positions were taught the same formulae and writing traditions as freemen, hinting at the strong relationship between various social classes in pre-Islamic Arabia. The Qur’ān instead attests to the difference between free and enslaved people in Muhammad’s community. However, at the same time, it mentions freemen and enslaved people as spiritually equal, which allows for the marriage between a free man and an enslaved woman. Although we do not have pre-Islamic inscriptions mentioning manumission, I have noted that the presence of inscriptions written by freed people suggests that the incentive to manumission was not necessarily an innovation of Muhammad’s preaching. In order to better understand the continuity between pre-Islamic to early Islamic times, I have mentioned the Qur’ān’s emphasis on considering enslaved people as part of the master’s household. Enslaved people were considered part of the community of believers (Ummah). Although views on slavery evolved following the Muslim conquests, we witness a close-knit social framework of household slavery at the dawn of Islam, in continuity with the pre-Islamic record of South Arabia. Assertions such as “the Qurʾan and the Prophet Muhammad encouraged emancipation to a much greater extent than any provisions in the pre-Islamic law had” simply do not find support in the available sources.^[152]

According to Moses Finley, there are three necessary conditions that society must display for which it can be defined as a “slave society”: 1) enslaved people need to be in large numbers; 2) they must play a fundamental role in surplus production; 3) slavery is an important cultural institution.^[153] According to Finley, only classical Athens and Roman Italy satisfy these requirements in antiquity. Based on the findings of this study, first-millennium Arabia would appear to be what Finley calls a “society with slaves” rather than a “slave society” as it is unclear whether slavery in the period could be considered to have been an “institution”. Commenting on Finley’s model, Bernard Freamon has admitted that he “cannot say whether Mecca was a ‘genuine slave society’ in the Finleyan sense”, though affirming that “there is no doubt that slavery was an extremely important aspect of Arabian society at the time of the rise of Islam”.^[154] More recently, the debate around whether a society satisfies Finley’s criteria has been superseded by the abovementioned What is a Slave Society? The Practice of Slavery in Global Perspective, which addressed several societies which Finley had not previously studied. In the volume, Noel Lenski criticizes Finley's model, magisterially showing its one-sidedness and its ethnocentrism,^[155] and advancing the following, more flexible definition of “slavery”: “Slavery is the enduring, violent domination of natally alienated and inherently dishonored individuals (slaves) that are controlled by owners (masters) who are permitted in their social context to use and enjoy, sell and exchange, and abuse and destroy them as property.”^[156] Elements such as bronze objects dedicated by people in subordinate positions to gods in temples opened to them as rulers; the perception of servants as members of their master’s household; the literacy level of Safaitic men in subordinate positions; and the fluid relationship between the two parties in the Qur’ān, do not suggest that Lenski’s perceptive definition of slavery fits perfectly well with first-millennium Arabia. Islam was not a rupture in this regard. The Qu’rān exhorts masters to behave beneficently towards their subjects. A preliminary inquiry into South Arabian inscriptions has shown the close-knit milieu of the region already in pre-Islamic times, suggesting that Muhammad’s preaching was not a watershed. Another datum which suggests that Islam was neither a rupture nor an alien product in the region.