Bryn Mawr Classical Review

Bryn Mawr Classical Review 2010.07.02

Rotman on Lenski on Youval Rotman, Byzantine Slavery and the Mediterranean World.   Response to BMCR 2010.05.14

Response by Youval Rotman, Tel Aviv University (

Noel Lenski's review of my book raises a number of objections to some of my arguments, which I would like to address. Furthermore, his review overlooks some of the most important arguments and misrepresents others.

The main subject of the first chapter, "Theoretical Approaches," is the dichotomy between slavery and freedom. Contrary to Lenski's assertion, I specifically show that this dichotomy is the only definition of slavery that really makes sense.1 The entire chapter's objective was to distinguish between the Roman/Byzantine juridical definition and modern concepts of freedom and slavery. Contrary to Lenski's assertion, I did not argue that the slavery-freedom dichotomy had meaning only in a republican democracy (in which case there would have been no sense in writing a book on Byzantine slavery). Instead I argued that if freedom is understood as political freedom, the dichotomy between slavery and freedom would be meaningful only in a republican democracy. Contrary to Lenski's assertion, I did not invalidate the definition of slavery as an economic institution, but argued that historians have interpreted ancient and medieval slavery according to economic models of modern slavery. These models were not necessarily in agreement with the definitions of the societies that were the subject of their research. I thus concluded that slavery should not be exclusively defined as an economic institution.

The second chapter, "Medieval Slavery in a New Geopolitical Space," concerns slaves as victims of war and trade. Although Lenski claims that natural reproduction was also a major source of slaves, medieval sources do not offer evidence to support this. Indeed, most of the slaves mentioned in Byzantine testaments were manumitted either during their master's lifetime or upon his/her death. The importance of both war and trade as source of slavery must be thus explained against the background of the new political and economic constellation in the early medieval Mediterranean, which changed dramatically the relationship between freedom and slavery. In a word, the status of free persons in the period under examination became unstable. This caused an incompatibility between being a slave de jure and de facto, a development that also provided the medieval world with slaves. As for the slave trade, Lenski asserts that the Mediterranean slave trade was oriented towards the South. I completely agree. In fact, this was one of my arguments. There was little need for slave traders to pass through the Mediterranean when importingslaves from Eastern Europe to Constantinople. . Constantinople's slave market was in competition with Arab slave markets, but contrary to Lenski's view, a competition does not necessarily mean that the competitors must share the same trade routes, as my map (pp. 60-61) clearly shows.2

In his review of chapter three, "Slavery, a Component of a Medieval Society," Lenski challenges my interpretation of the evidence for the use of slaves in the Byzantine countryside. There are indeed only five documents for a period of four centuries. However, they include both imperial and private documents and are complemented by descriptions from Byzantine literature. Lenski did not object to my analysis of the Book of the Eparch as evidence to the use of slaves in the urban economy. The five documents I use to show the existence of rural slavery are not different. As for "the imperial slaves" which, contrary to Lenski's assertion, I did not equate with servi Caesaris, the chapter shows their use throughout the Byzantine period, and demonstrates the difference between them and the "slaves of punishment," whom Justinian's legislation abolished.3 However, the main purpose of this chapter was to show that in every economic position slaves shared the same functions with persons of free status, which meant that they did not form an exclusive economic stratum.

Lenski's review of chapter four, "Evolution of the Concept of Unfreedom," implies, misleadingly, that this chapter was confined to the development in the concept of slave as agent in Byzantine literature. I agree with Lenski's point that the captive was discussed by various authors throughout Antiquity. However, captivity did not acquire religious meaning until the political enemy became the religious opponent with the arrival of Islam. Captivity thus became an important literary subject in the 9th-11th centuries because of the religious difficulties that Byzantines encountered in that period. As for the perception of the slave as a person, Lenski cites the homily by Gregory of Nyssa. One can go back indeed to Seneca's De Beneficiis. Intellectual arguments for the personhood of the slave are important, but take second place to arguments for the justification of slavery. Byzantine hagiographic literature became innovative not only in presenting as heroes slaves of unfree origin who did not choose their destiny, but also in confronting the slave with his master in order to undermine his loyalty to him. Thus, contrary to Lenski's review, the main subject of this chapter ("Evolution of the Concept of Unfreedom") was not the personhood of the slave, but the transformation of the relationship between master and slave into a tripartite relationship: master-slave-emperor, in which the slave's loyalty to the emperor took precedence over his loyalty to his master. The origins of this transformation, I argued, are already perceptible in imperial Roman law. The references Lenski brings from the Justinian Code in order to challenge my analysis actually provide supplementary support of this argument.

The laws that Lenski cites to refute my analysis of the ius postliminii are taken from the Justinianic Code. My argument was that Roman captive had the status of slave while in captivity and, if ransomed by a third, was bound to him by a ius pignoris until the ransom was reimbursed. This is by no means in contradiction with the Code of Justinian.4 CJ 8.50: De postliminio et de redemptis ab hostibus, in fact proves my point that during the 3rd- 5th centuries the ius postliminii was not applied automatically to every Roman of free status ransomed from the enemy.5 In fact there is a confusion here between laws dealing with free Romans captured by the Roman army (in civil wars, or recaptured from the enemy) and laws dealing with ransoming of Roman captives by private persons through commercial exchange. The two laws Lenski cites to argue that there was a prohibition on trade in captives (CJ 7.14.9 and CJ 8.50.12) did not forbid trading captives of free origin back into the empire, but were enacted when there was no commercial exchange involved.6

Lenski ignores one of the most important issues: the laws that the Code cites apply only to persons who were sufficiently wealthy prior to their captivity, and who could then reclaim their property from the governor of the province, provided that they were living in the same province prior to their captivity, and that their buyer let them establish contact with their family. This is precisely what the ius postliminii was all about. I thank Lenski for his references, which actually show how, in assembling these regulations together in the Code, Justinian established the juridical setting for his policy of letting individuals and Church representatives deal with ransoming of captives, rather than having captives be ransomed by the state.

Against this background, and contrary to Lenski's assertion, the Byzantine innovation of the 8th century cannot be overestimated. This innovation lay not in attributing a free status to the captives returning to the Empire, but in considering them free Byzantines while in captivity. Contrary to Lenski's assertion, I did show (pp. 30-31) that the innovation of the Ekloga drew on the previous constitution of Honorius (CJ 8.50.20 = CTh 5.7.2), which determined a period of five years during which the redeemer will pay the captive a wage. Lenski's analysis fails to understand the great innovation that the Ekloga introduced: the wage of the ransomed captive was to be determined by state officials and not by the redeemer/buyer. If the Roman legislation only enabled the reimbursement of captives who were wealthy enough prior to their captivity, it would have left out most of the Roman inhabitants who were captured and ransomed, but had no means to buy their freedom back, especially if they had lost all their property during the war. In light of the Justinianic Code, the Byzantine imperial legislation of the 8th century thus appears even more innovative than I had argued previously. It eliminates all trade in Byzantine captives. The acts of private ransoming are thus replaced by the public acts of the state, as the historiography attests. In fact, the Ekloga (8.4.1) goes even further than that. Contrary to Lenski's reading, the issue here is not a ransomed free Byzantine but a Byzantine slave,7 who will be manumitted if he has behaved well towards the state by inflicting damage on the enemy. This was unthinkable under the Roman ius postliminii. This development was due partly to the religious identity of the captive, but also to the increasing intervention of the imperial authority in the private relationship between owners and their property.

As for coloni, enapographoi, paroikoi and other "semi-servile" groups as they are often termed in modern research, I agree that they must be differentiated from slaves, but not because they differ in the degree of dependency which their situation entails. This misunderstanding derives from the confusion of the different natures of these dependencies. Such groups cannot be put on the same scale of "freedom" as slaves. This is not because their degree of freedom is immeasurable, but because such a definition of freedom is fictitious and has nothing to do with slavery. This is precisely the main argument of the book and was, in fact, the reason for writing it.

In view of all this, I regret that Lenski's review is by and large inaccurate and ignores the book's main arguments.


1.   I did not argue against such a dichotomy, but relied on Institutes I.3 (p. 19; 211), the same definition Lenski brings up in order to refute my argument.
2.   Lenski wrongly writes that the map was not available in the French original. The map is the exact translation of the French original (pp. 340-341)
3.   Contrary to Lenski's assertion, I did not imply that there was no difference between servi Caesaris and servi poenae (bibliography: pp. 102-103; 172-173; 240; 261-262).
4.   See bibliography on pp. 214-215.
5.   Contrary to Lenski's assertion, the book does cite the Code on this discussion, but omits the laws dated to the 3rd-5th from the table of 6th-11th c. imperial regulations.
6.   According to CJ 8.50.12 Roman captives regain their previous status if they were liberated by Roman soldiers and were not ransomed by a commercial transaction. CJ 8.50.11 and CJ 8.50.13 both state the possibility of trade in captives. Contrary to Lenski's assertion, the book did emphasize that the property of Roman captives stayed in suspense while in captivity. The relation between usucapio (the acquisition of ownership by possession of property under certain conditions) and the ius postliminii, which Lenski invalidates, is in fact discussed in CJ 8.50.18. Unlike Justinian's Novellae, the Code sometimes includes laws which one cannot interpret as applying to the 6th century. This is the case with CJ 7.14.4, which contrary to Lenski's assertion did not forbid trade in Roman captives, but is a specific case of a Roman captured during the Zenobia's revolt, on whose behalf Diocletian and Maximian intervened.
7.   Lenski confuses here Ekloga 8.2 with Ekloga 8.4.1.

Comment on this review in the BMCR blog
Read Latest
Index for 2010
Change Greek Display
Books Available for Review
Bryn Mawr Classical Commentaries

BMCR, Bryn Mawr College, 101 N. Merion Ave., Bryn Mawr, PA 19010

HTML generated at 10:40:31, Friday, 02 July 2010