BMCR 2005.07.81

L’achat et la vente des esclaves dans l’Egypte romaine

, L'achat et la vente des esclaves dans l'Egypte romaine: contribution papyrologique à l'étude de l'esclavage dans une province orientale de l'Empire romain. Archiv für Papyrusforschung und verwandte Gebiete. Beiheft; 14. München/Leipzig: K.G. Saur, 2004. xliv, 373 pages, v pages of plates: facsimiles. ISBN 3598775490 €92.00.

Straus sets out to examine all (fewer than 200) Greek and Latin papyrus records of purchase and sale of slaves in Egypt from 30 BCE to the end of the fourth century CE.1 Instead of posing and defending a thesis, for fear of making the evidence fit the thesis, he instead asks a question: what do the papyrus documents bring to the knowledge of the sale of slaves in an eastern province of the Roman Empire? Straus is scrupulously careful, excluding any doubtful evidence; perhaps there should have been another section that presented the doubtful evidence. He also makes careful distinctions between documents of sale and tax value declarations, or, e.g., a slave offered as guaranty on a loan. Straus’s most frequent dialogue partner throughout the study in questions of uncertain interpretation is H. J. Wolff.

The original title of the study was to be “Commerce in Slaves,” but the title was changed mid-course because the documentation does not support the existence of a sustained business of slave-trading, rather there were discreet instances of purchase and sale by private persons or their authorized agents. Egypt was never a slave society, most of the labor being provided by indigenous peasants. Slavery was probably introduced by the Greeks of the Ptolemaic era and continued under predominantly Greek and Roman influence.

Part One deals with administrative and legal questions. Chapter One introduces the sources for the study, especially contracts of sale and certificates of registration of sale. But there are other documents as well, e.g., the ἀνάκρισις.2 Other relevant types of document include those of tax on sales, etc.

Chapter Two discusses legal formalities. Most buyers enter into their own contracts, but some documents give an agent authority to buy or sell a slave. Those chosen as agents or representatives in purchase or sale are most likely to be family members or tutors, especially when the buyer or seller is absent. Most often a woman designates a husband, brother, or other relative, who may also be her κύριος, to buy a slave for her. A number of legal terms are used for different documents, without clear difference. οἱ παρὰ + genitive designates everyone except the person named who is involved on one side of the transaction, including legal heirs, and Straus argues that οἱ περί τινα appearing in the same document is inclusive of the person named. A καταγραφή, is not the same as bill of sale, contract, or registration of contract. An ἀπογραφή is probably a declaration of ownership for a slave bought elsewhere and brought into Egypt. A δημιώσις and an ἐκμαρτύρησις are probably forms of public registration of a private transaction.

An ἀνάκρισις is a certification of the servile status of the person to be sold, required before a first sale in Egypt. This inquest takes place as close as possible to the sale, sometimes the same day, happens only once in the life of the slave, and suffices for further sale, apparently even outside Egypt. P. Herm.18 suggests that the basis is an interrogation of the slave as to the servile status of him/herself, parents, and siblings. The sale tax is payable by the seller and can be taken out of the sale price.

Chapter Three examines the typology of Egyptian sales contracts, placing them in the context of H. J. Wolff’s typology of Greek legal documents. Whatever the form, the contract contains date, place, duration, seller, buyer, description of the slave, payment, price, guarantees, and attestation of seller and perhaps others. Regarding date, the most common forms are year of the emperor, Egyptian month and day and occasionally the Macedonian month name, even well into the Roman period. The manner of describing slaves is identical to that describing free persons in other documents. If the slave is house born, the name of the mother is given. Receipt of payment is an essential part of the contract. The guarantee is offered against invalidation, sometimes with a co-guarantor named, though most often the seller is also the βεβαιώτης, testifying that the slave is indeed his/her own property, with no private or public claims by a third party. There is frequent guarantee against hidden defects, and often statement that the slave is not returnable except for epilepsy or ἐπάφη, probably leprosy. A very brief Chapter Four on co-ownership notes that the law allowed co-ownership of as little as one-sixth of a slave. In Greek but not Roman law, co-owners could and did manumit their part of a slave while the rest of the slave remained enslaved. How this worked in practice remains unknown.

Part Two takes up economic and social questions. Chapter Five on buyers and sellers of slaves begins with the sad fact that only 5.1% of this group is represented by more than one document. The three legal categories of persons in the Roman law of Egypt, Roman citizens, ciues peregrini, and peregrini Aegyptii, are represented in three fiscal categories: 1) those exempt from personal tax, which includes all in the first two legal categories and some in the third; 2) the more hellenized of the third legal class, with reduced tax; 3) all the rest. The decree of Caracalla in 212 had little effect on these distinctions. Barring explicit reference, one cannot be sure of Roman citizenship in the documents, even when the tria nomina are used. Some women have tutors “according to the Roman custom,” or do not use tutors “according to the law of children in Roman custom.” Of the four Greek cities in which citizenship can be held, Alexandria, Naucratis, Ptolemas, and (after 130) Antinooupolis, only Alexandria and Antinooupolis are represented here, with thirteen sure and eight probably Alexandrians and three from Antinooupolis. Others call themselves “of” a certain city, but may not be citizens.

An unusual case is that of first-century CE P. Valerius, a soldier who buys and manumits his sister and her children, she having been enslaved thirty-eight years earlier at the age of four (SEG XXXIX 1711= SB XX 15005). The reason for her enslavement is not given, and this is probably more a case of ransom paid than purchase. Though profession is seldom mentioned, there are soldiers, priests and athletes who appear in the documents. Attestation that a buyer or seller cannot read and write refers only to ability in Greek; one example has a secretary write the Greek document, but a different hand adds further text in Demotic. Many different levels of literacy are represented, and occasionally a specific statement in this regard: “ι, ____, wrote this for _____ who is illiterate.” Not every case of literary agency with no reason given, however, must be construed as illiteracy; it may often be a case of convenience, especially when there are several persons involved on one side of the transaction. Buyers and sellers include Romans, descendants of Greek colonizers, or hellenized Egyptians, and it is not always possible to tell them apart.

Means of acquiring slaves represented here include: the οἰκογένης, child abandonment, prior purchase, inheritance, and capture in war. The term οἰκογένης does not mean house-born in the house of the present owner, since one example so designates a slave who is at the same time Syrian. Rather, argues Straus, it means house-born in the house of the original owner. Only one example is represented of an abandoned child raised as a slave, a girl nursed by the woman who later sold her. Multiple sales of the same slave are common: one seven-year-old boy was being sold for the fourth time, and a girl for the sixth time. Slaves were often acquired by inheritance. There are two questionable cases of enslavement as prisoners of war, both brought into Egypt from elsewhere.

Part Six, finally, takes up the study of the slaves themselves. A brief Chapter One discusses terminology for slaves, the most common being the usual δοῦλος, παῖς, παιδάριον, οἰκέτης, and also combinations using σώματα, the familiar ἀνδράποδον being less frequent. Chapter Two examines names of slaves, comparing the collection to studies by Lambertz, Masson, Reilly, Fragiadakis, and Solin. Straus classifies the names in seven categories: geographical, foreign, derived from religion or mythology, named after historical persons, nicknames, names connected with luck, good or bad, and a few “other.” Some names are more associated with slavery than others, but there is no definite typology, no clear dividing line between slave and freeborn with regard to nomenclature. Chapter Three examines sex and age. There are slightly more women represented than men. Age at sale ranges from less than one year to fifty-one for men, from three years to thirty-five for women. Some were bought as infants, and Straus raises the question why anyone would buy an infant slave rather than acquiring an abandoned one free. He supposes the answer to be that the infant purchased has passed through the first dangerous months or year of life, and therefore has a better change of surviving. The seeming predominance of females contradicts known evidence elsewhere of a majority of male slaves in the population. But deeds of sale do not, of course, accurately represent the whole of slave activity. Females were undoubtedly of higher value between the ages of fourteen and thrifty-five for their reproductive potential, and in fact, there is no sale contract for a woman beyond childbearing years. Chapter Four discusses family ties. Slaves were no doubt encouraged to propagate. The multiple sales of very young children attested in the documents usually, though not necessarily, separated children from their mothers. Property interests seem to have dominated: one man sells first his concubine, then his own son because of financial problems. A woman sells a slave child she had nursed at her own breast. The trauma of rupture of family ties behind the contracts is inaccessible to us. Chapter Five examines country of origin. The majority of slaves represented in the contracts are indigenous, though a few are imported, sometimes ἀπὸ καταπλόου, imported by sea. Chapter Six on geographical mobility of the slaves traces some of the long journeys they endured up and down Egypt because of sale.

Part Seven takes up modes of payment, money used, and prices of slaves. Chapters One and Two discuss payment and money. Payment is almost always in drachmai, except in cases of slaves bought elsewhere, in denarii. Payment is often made on the installment plan, with an intermediary bank. Slaves did not legally change ownership until full payment was made. Chapter Three on prices of slaves gives five pages of tables, pp. 295-99, listing date, place, sex and number of slaves, age, price, and document. Prices vary considerably beyond the obvious factors of date and age of the slave. No consistent explanations for this are known.

Part Eight discusses briefly the evidence for commerce in slaves. The majority are indigenous, but those imported are mostly from the closest places in the Eastern Empire. Whereas Biezunska has argued that in Egypt the major source of slaves was foundlings and the homeborn, W. V. Harris has contended the opposite, that these sources of slaves were not adequate to fill the needs in most of the Empire. Straus suggests that perhaps Egypt was the exception here, as well as in the larger number of female to male slaves. There is no firm evidence in any of the documents of professional slave dealers or markets, even though such markets must have existed in Egypt like everywhere else.

A brief conclusion notes the trauma of uprooting implied in the sale of slaves, especially children, and the way in which the system of ancient slavery reduced human beings to property. Four appendices give: 1) several documents not yet edited or only partially edited that form part of the study; 2) a list of sales from Roman Egypt; 3) the sex of slaves for sale (59 males, 79 females); 4) new textual suggestions for a few of the documents.

This study is loaded with detailed information, about which little interpretation can be offered. Where questions can be raised, Straus raises them and gives judicious answers, erring, if anywhere, on the side of caution. Unfortunately, the question we would most like answered lies hidden: Why? Why did these persons decide to sell their slaves, especially young children, one’s concubine, son, or nursling? Were economic necessity and profit the only motives? Again, in spite of all the detailed knowledge contained in these ancient documents, the reality of slavery remains an enigma.


1. Only one Latin document of sale, P. Oxy. XLI 2951 = ChLA XLVII 1415, from the Roman military camp at Nicopolis.

2. For the reader not an expert in legal terms, the order of discussion here is problematic: the relevant documents are discussed on pp. 10-11, but the term is not defined until p. 62.