Slavery After Rome, 500-1100

In this masterful monograph, Alice Rio revisits one of the central questions in the historiography of early medieval Western Europe: how did the transition from slavery to serfdom take place? While many earlier answers to this question have proposed a more or less linear trajectory from late Roman slavery to the serfdom of the central Middle Ages, Rio sets out a compelling and elegant argument for a rather less elegant state of affairs: instead of trying to contort the messy source base into a seamless theory, she sensibly and convincingly argues that ‘there is a fundamental problem with looking for a single linear story here: the line is far too broken up, too frayed with little individual threads making their own bids for escape, and often leading nowhere’ (p. 248). Rather than smoothing over the very wide range of variation in what unfreedom could mean in this period, Rio aims to provide an interpretation of that diversity that can explain an overall trajectory without seeking to pare off the divergent possibilities that the sources present us with. The first part of this book examines pathways into the condition of unfreedom, namely sale after being captured in war, self-sale, debt slavery, and penal enslavement; and ways out of that status through purchase and manumission. Rio next discusses the nature of two kinds of unfreedom: household slavery, and unfree status on large agricultural estates. Her last chapter provides an analysis of the legal and institutional framework for unfreedom, with a focus on the ways in which Church, state, and private lordships approached it through efforts to control marriage. In most of these chapters, Rio moves successively through different regions: Francia, Italy, the Iberian peninsula, Anglo-Saxon England, and Ireland, of which the first two tend to get the most coverage; in some chapters, Byzantium and the Slavic lands get a look in as well. In the introduction, she surveys some of the main landmarks of the debates on the issue of transition from slavery to serfdom, and sets out the main questions that drive her own analysis: why did the powerful seek to categorise some people as unfree, and not others, and what were the advantages that they felt accrued to them therefrom? And crucially – although this is less explicit in the introduction than in the following chapters – how and why did dependants themselves negotiate these categories, and what benefits could they have got not just from moving from being unfree to free, but also the other way round?

My principal criticism of this book is worth stating at the outset, since it is a matter of terminology: is this a study of ‘unfreedom’ or ‘slavery’? In Roman sources, servus means something that we can translate unambiguously as ‘slave’, but its connotations in medieval texts can range from ‘slave’ to ‘serf’, and even ‘servant’. On a number of occasions, the choice between ‘slavery’ or ‘unfreedom’ is a matter of interpretation, particularly when other terms like ‘debt slavery’ or ‘penal enslavement’ come into play, and there are times when what Rio is discussing looks to me more like ‘unfreedom’, even though she uses the term ‘slavery’. As she herself notes, ‘it is more or less impossible to find a working, unified definition or translation for servus and mancipium as either “slave” or “serf”: the whole spectrum of possible meanings could apply at the same time within the same early medieval society—and indeed within a single lordship’ (p. 247). It is a bit odd, then, that if the title is to be believed, this book is only about one end of the spectrum: a good deal of it is clearly about ‘unfreedom’, by Rio’s own definition, rather than ‘slavery’, and a more accurate title would have been Unfreedom in Western Europe, 500–1100. In what follows, I use the terms ‘unfree’ and ‘unfreedom’ almost exclusively, with ‘slave’ and ‘slavery’ being restricted to household dependants who were most akin to the slave in the Roman world.

Perhaps the most obvious way into unfreedom, and into an unfreedom that can most un-problematically be defined as ‘slavery’, was through capture and sale as chattel. This tended to occur in two ways: either large-scale enslavement and subsequent use or sale of populations in the course of a military conflict, or small-scale cross-border raids, which probably resulted less frequently in sale (and are to be found more commonly in Ireland than most other parts of western Europe). Rio suggests that although much writing on the slave trade emanated from Carolingian Francia, it is unlikely that the Franks benefited much from it: their profits were made from tolls and taxes paid by traders passing through, rather than directly from capturing and selling people themselves. The most significant evidence of large-scale abduction by the Franks of a conquered population comes from the Saxon wars, and it seems to be certain that the Saxon captives were not sold, but rather resettled on estates in Francia. (Perhaps unsurprisingly, given the lack of evidence, Rio does not discuss the status of these Saxon captives: household slaves, unfree tenants, something else, or both at different times?) Slavery was not, therefore, at the ‘origins of the [western] European economy’ (as Michael McCormick suggested in 2001), but rather quite marginal to most of it, in particular to the economic boom taking place in the Frankish lands. The British Isles were a different case, but here too, the slave trade was of relatively little economic importance. The main raiders were vikings, and however important the slave trade may have been to the economies of emerging kingdoms and states in Scandinavia (not something discussed by Rio), it seems to have had a marginal role in the British economies in this period. In southern Europe, with closer contact with the Muslim lands, raiding and trading had a greater impact: Christian polities were vulnerable to Muslim raids, and as they in turn started conquering Muslim lands, Christians also enslaved captives. The bulk of the slave trade in this period, however, seems to have been in people enslaved in the far north and north-east (captives from England, Ireland, and the Baltic) and conveyed through Francia, Italy, and Spain, to the Islamic Mediterranean. The reason why there was little demand within much of western Europe itself for slaves was, Rio argues, because here the elites ‘had greater possibilities to derive substantial resources from the exploitation of land through tenants, free or unfree’ (p. 38); selling these people would have resulted in losing labour needed to extract essentials and profits from the land.

Household slavery nevertheless continued to exist in western Europe throughout this period. It was more common for unfree people tied to a household to be used in agricultural production in Ireland than elsewhere, largely because of the small scale of agriculture; freedom and unfreedom were ways of distinguishing between people who otherwise performed much the same tasks. On the continent, it seems to have been common enough at least in the earlier part of this period for even peasant households – and indeed even unfree peasant households – to include household members who are probably best understood as slaves; these might have been considered a part of the holding and were possibly granted to the tenants along with the land. In southern Europe, there was an increasing tendency towards slavery being restricted to religious or ethnic outsiders, whereas landed unfree tenants tended to belong to the lord’s own religious or ethnic community. Thus there came to be a more firm line between household slaves and unfree tenants not just with regard to legal status and function, but also in terms of their ethnic and religious identity. In contrast, although there seems to have been little to distinguish Francia and England from Italy or Iberia in the early part of this period, by the ninth or tenth century, household unfree labour blurred increasingly into the continuity between free tenancy and unfree labour, with the landless unfree labourers on estates not being hugely different in legal status from unfree tenants. The possession of household slaves does not here seem to have be an aspect of display of aristocratic status. Rio also convincingly argues that in Francia at least, many of the best-documented instances of household unfree labour could have been a matter of life-cycle dependence, with some younger members of the family being provided for household service (with arguably greater constraints on their freedom) by tenant families, who would then return to a tenancy later on that was still unfree, but probably less so than household service.

People subject to these sorts of unfreedom were almost certainly not people who had been captured and sold. A more complicated entry into unfreedom arose when people chose to sell their freedom or enter into unfree status as a means of paying off a debt; or when they were made unfree as a punishment. In early medieval Francia, this change of status seems to have become less a matter of enslavement and more an agreement entered into between two parties as a payment for something, which could always, in theory at least, lead eventually to a reversion to freedom. People used the status of legal unfreedom as a means of overcoming economic hardship for a while, which in turn led to a more flexible concept of unfreedom: something negotiable and temporary, as opposed to a fixed and clearly defined slavery of the Roman sort. Rio finds that similar processes occurred in Italy, though somewhat earlier than in Francia; and in Italy, there were more occasions for a slide into unfreedom as a punishment for various crimes. As in Francia in the post-Carolingian period, in Italy the voluntary entry into unfreedom seems to have been more a matter of ‘self-gift’ rather than ‘self-sale’: poverty tends not to be stressed in the Italian documents as a motive. Like the Frankish monastery of Marmoutier in the eleventh century, ninth-century Italian monasteries ‘seem to have used entry into unfree service to create a range of ties with the people who worked for them’ (p. 61), which ties were presented as voluntary and reciprocal earlier in Italian than in Frankish sources. Iberian laws, in contrast, appear to be focussed almost exclusively on unfreedom as a form of punishment, partly because they seem to have been much more closely modelled on Roman laws; however, here as well, Rio suggests that although the evidence is patchy, actual practice may have been closer to Italian and Frankish practice. In England, although the evidence is sparse, temporary unfreedom seems to have become more common over the centuries, and thus it seems likely that the condition was, for those entering it in a more or less voluntary fashion, here also increasingly flexible in practice. Hierarchical social relations of lordship and dependence – inherently less absolute than the stark dichotomies of free and unfree, and linked also to poverty and piety, service and reward – were thus as important, and possibly increasingly more important, than legal status per se; as a result, however rigid the differences between free and unfree might have appeared in law, in practice they were overlaid with these categories of social relationships, and thus tended to be less of a binary and more of a continuum.

Being unfree generally brought with it complications not restricted to constraints on movement and obligations to perform labour: marriage between people of different status tended to be either illegal, or compromised the status of the free person and normally also the couple’s children. Such marriages nevertheless occurred quite frequently. The Church was, on the one hand, increasingly keen to maintain the sanctity of marriage, but equally, it had no interest in going against the desires of the landlords to keep their unfree tenants subject (the Church was, after all, the largest of all landlords, and thus would have had more unfree dependants than anyone else); it got around the issue by increasingly stressing the possibility for salvation for the unfree, and thus neatly sidestepping the problem of how the subjugation of another human being squared with Christian ethics. States, in contrast to the Church, seem to have been more active in legislating with regard to unfree status, and laws uniformly provide for a diminution of status of the free partner in a marriage between free and unfree. However, as Rio shows, in practice this was not always followed, and the free partner could not only often remain free, the children of the union could also be free. Thus she argues for a less literal understanding of the nature and utility of early medieval laws – ‘the legislative equivalent of a Swiss army knife’ (p. 244), usable in many ways without being the best tool for any single function. The reason for this flexibility was because the state – like the Church, also a lord, and comprised of lords – was in effect granting lords a framework to use, while also allowing them a great deal of flexibility in using it. This flexibility could be used as a bargaining tool, as indeed status itself came to be used: as a means of reminding dependants of their subordinate status, and as a means of extracting fines or dues. Indeed, by the twelfth century (and I would suspect in many instances often even earlier), control over marriage seems to have become primarily a means of obtaining more cash from dependants.

If being unfree was increasingly something that was flexible and could mean a number of different things, this was true equally of the process of being freed. In the early Middle Ages, manumission often implied continued service and labour obligations that could be hereditary and look very much like the kinds of demands placed on those who were legally unfree. One influential strand of the scholarship has thus argued that the origins of serfdom lie in the freedmen of the early Middle Ages: from being slaves, these dependants were turned into serfs, rather than being really free. Although manumission could result in a status that had less constraints than those of a Roman freedman, this form of manumission has left fewer traces in the sources, and seems to have been most common in Italy, though also to be found in Merovingian Francia. When manumission was a way of granting people complete independence, it was most likely to occur when the person had some particular military or artisanal skill, or when it was felt that the person freed could help to maintain or strengthen a relationship between manumittor and another patron. More controversial is the form of manumission – in Francia often to be found in the form of mass manumission granted in a testament – that placed the freed person and her or his descendants in a continued position of dependence, and tended to imply either labour service for a certain number of days in a week, or the payment of dues at fixed times in a year. Both in Italy and Francia, mass manumissions appear often to have occurred in conjunction with gifts of land to an ecclesiastical institution, and the freed were expected to stay on the land (otherwise the gift of land would have had little practical value to the recipient) or to commemorate their former lords (or both). The grant of freedom did not, however, necessarily also confer better conditions on the recipient: peasants would still need to stay on the land in order to survive, and thus they could easily be subject to conditions arising from the requirements of their new landlord, which could often translate into quite onerous labour dues. In other words, we are already on the way to something like ‘serfdom’, in which the conditions of access to land could be in effect a form of bondage, even if it might in theory be possible to escape bondage by giving up that access. While in the later Middle Ages such a form of flight (to a city, for example) might indeed have been an option, given the low urbanization in of most of Europe in the earlier Middle Ages, manumission could not necessarily do much to improve, in practical terms, the constraints on a person’s life.

Rio suggests that there was no consistent worsening of the condition of being freed, but rather a multiplicity of usages for manumission. Some form of bondage or constraint tended to accompany manumission when this took place in conjunction with a grant of land (and thus generally is in evidence in the context of mass manumissions); while such manumissions tend to dominate the evidentiary base because large ecclesiastical landlords are our principal sources of evidence, there is no reason to believe that fuller manumissions were not also common. However, as Rio herself notes, when we have a record of the manumission, but no record of conditions of land tenure for the manumitted after obtaining freedom, it could very well be – and in some instances it demonstrably was – the case that the terms of tenure ended up being the same as those of unfree tenants, even if the legal status after manumission was fully free. In other words: the conditions of access to the means of social reproduction are a key factor in assessing the actual, practical effects of changes in legal status; and if those conditions ended up being the same for the free and unfree, it is quite likely that, on a day-to-day level, there was very little difference in the experiences of people of either status. The point is well taken, however, that we should not see in the dues (in labour or money) owed in the form of chevage or census a distinct legal status (what German historians have termed Zensualität): these are dues that indicate a relationship between the one who pays or performs them and the lord, which could take a multitude of forms, and thus ‘there is a fundamental problem with considering them [freedmen] as a coherent group at any point in time, and with assuming them all to have been any one thing unambiguously—whether free, unfree, or half-free’ (p. 113).

Manumission was thus a means of doing various different things, just as unfreedom itself could mean various difference things in practice, which could result in greater or lesser levels of actual freedom for the person(s) concerned. There was little possibility of the state intervening in any way – as it would have in Rome – to protect the status of the freed person. Thus, absent a commercial economy and towns as a point of refuge, in practice, ‘for a great many freedmen, freedom therefore still came with a tenancy agreement’ (p. 129), and the conditions of tenancy might have resulted in differentiating the freed person’s life very little at all from that of the unfree. It is not difficult to agree with Rio’s conclusion that the practice of manumission should not be seen as having had a significant impact on the construction of ‘serfdom’, as much of the historiography before her asserts. However, insofar as conditions of land tenure were a crucial – possibly the crucial – factor determining both how the freed person lived, and what unfreedom was in practice, it must also be stated that while it is true that the freedman is a bit of a red herring, what really mattered was the form of access to the means of subsistence for both free and unfree. Thus we are actually quite close to something that looks like later ‘serfdom’ already by the Carolingian period, regardless of the legal status of the people involved: free or unfree, for people working the land, their ability to live depended on fulfilling a certain set of services that were negotiated – as Rio herself cogently demonstrates – between individuals and communities and landlords, regardless of legal status. As indeed was the case in later centuries, legal status already often looks rather like a tool used by landlords to try and extract the most they could from their tenants.

This point emerges very clearly from Rio’s chapter on unfree status in estate communities. The most well-known form of land management – well-known because it has left the most written evidence, but emphatically not because it was the most common – was the bipartite estate, in which tenants were granted use rights to land sufficient to nourish their families and realise a surplus to pay dues to the landlord, in return for both those dues and frequent labour services on the land directly managed by the landlord. Almost certainly more common, though less evident in the sources, were forms of land management in which the landlord granted rights of use in return for dues of some sort, occasionally coupled with some minimal labour services; in such cases, the landlords may have retained some land that was cultivated by unfree dependants who were housed and fed and constricted to the landlord’s directly managed property, with no use rights to any land of their own. Most of the medieval population would have been involved with one form of estate management or the other; forms of land tenure were thus arguably the key factor influencing their lives. Perhaps paradoxically, it was not necessarily the case that the tenants with the heaviest burdens were consistently always the (most) unfree, although on some Frankish estates there is a certain correlation between number of days of labour services owed and legal status. Most tenants, however, were not on bipartite estates, and in such cases – where there was no obligation to service on demesne land – it is hard to tell the difference between free and unfree from what they owed, though there is one significant distinction, namely that the unfree have restrictions placed on their mobility (the free would not, however, have had anywhere to go at this point anyway, so their mobility was restricted in fact, even if not in law). In the great estate surveys (polyptychs) of the Carolingian period, a number of different terms are used for tenants, and clearly the compilers of these records cared about whether or not the tenants were free. Rio is convincing in her argument, against some of the older scholarship, that the polyptychs represent not a fossilised form of record, but rather the result of a negotiating process, presenting something that the landlord and tenants had managed to agree on. Legal status, in other words, still mattered to both parties, even if there was no consistent relationship between a particular set of obligations and a particular legal status.

Status and terms of tenure could differ so widely among the tenant community because they kept being disputed and negotiated, and the result of the negotiations depended on the balance between the landlord’s power as enforcer, and the place of the individual within the tenant community, which determined how much support that individual might or might not get. Rio argues cogently that it was often in the interests of some parts of the community to bear witness to the unfreedom of other members of that same community: this meant that if heavier burdens were to be assigned, they would fall on those now designated as unfree. Thus ‘peasants too could have a vested interest in the unequal distribution of labour duties’, and legal status became a ‘proxy through which to fight over minute differences in internal estate hierarchy and the distribution of labour duties’ (p. 198). There was no strong sense of division within the community on the basis of legal status, and as Rio herself says, most of the time the relationship between status and obligations was very loose; it only became an issue when attempts were made to use it as a tool to squeeze more out of the tenant.

Ireland and Wales seem to be somewhat outliers in the sense that there is less evidence for large estates and thus for the subjugation of people working them, and independent peasant production with relatively low obligations in return for access to lands were more common. Anglo-Saxon England, in contrast, presents a different trajectory of the growth of large estates, with increasing amounts of land being classified after the Conquest as ‘inland’ (demesne land cultivated by dependants), without this leading in any way to a greater number of people being classified as unfree. The reason for this, Rio argues, was that the Norman lords would have needed to assure continuity of revenue and thus would have found it more beneficial to control directly the land, while avoiding peasant resistance by not insisting on unfreedom; the return of the yoke on the English peasantry had to wait another century or so, with the expansion of the manor and associated dues and unfreedoms expanding there paradoxically at the same time as they receded on the continent. In contrast to the British Isles, Iberia, at least in the tenth and eleventh centuries and beyond, seems to correspond reasonably well with the picture I have sketched above from Rio’s discussion of Francia; and northern Italy from the ninth century seems even closer to it. Large estates in Italy seem, however, to have had a larger proportion of prebendaries, or unfree labourers living on the estate with no land of their own, rather than tenants who had their own land and in return performed labour services on the demesne. As with Francia, the rich record of disputes is used by Rio to demonstrate cogently that legal status in Italy as well became ‘instrumentalized as part of a wide range of different strategies by different sorts of people, all of whom used it to achieve quite fine levels of differentiation’ (p. 211) with regard not solely or even primarily to their free or unfree status, but rather with regard to what they had to do to maintain access to the means of production (land), and where they ranked, as a result, within their communities.

Rio’s book is not about land tenure and the management of estates and labour forces (on which there is another, quite discrete, and equally massive and fractious historiography), but about legal status. She demonstrates convincingly that legal status mattered, both to those with power, and those without. Nevertheless, in a society in which most people lived off the land and were dependent on access to it in order to survive, conditions of tenure were in fact arguably the single most important socio-legal factor conditioning their lives – and indeed, Rio herself states that personal legal status became ‘a low background hum drowned out by existing terms of land tenure’ (p. 198). Rio does not discuss conditions of tenure in any detail – which is fair enough, since it would have taken another ten years for her to produce a book discussing them with as much care and cogency as she does legal status. But when a closer consideration of the nature of tenancy is brought into the picture, legal status, while still clearly important, does indeed seem to matter less (at least in Francia and I suspect parts of Italy) than conditions of tenure already by the ninth century. Rio’s statement that ‘we cannot distinguish between churches as patrons and churches as landlords’ (p. 111) leads inexorably to the insight that ‘without knowing the terms under which […] dependants held their land, it is impossible to judge what their likely level of subjection or freedom may have been’ (p. 113). What has happened between the Roman period and the Carolingian age, in other words, is a shift from concern about the nature of the person (free or slave) to concern about the best means of obtaining, for the landlord, labour, and for the tenant, access to land. For both parties, an agreement specifying some sort of unfreedom for the tenant might very well have been most favourable: this would give the landlord more labour or rents, and the tenant – in many, perhaps most cases – more land. While Rio’s point is well taken that disputes over legal status were, among other things, means of sorting out hierarchies within the community of tenants, for their day-to-day existence what mattered rather more would have been how much land they had access to, and what they were expected to do in order to access their lands; thus ‘debates over status seem to have been a way of discussing or re-evaluating terms of tenure’ (p. 195).

Rio has done a splendid job of analysing an extremely intractable source base and providing an interpretation that, without in any way trying to smooth over the inconsistencies and messiness of the evidence, nevertheless makes sense. Her book forces us to rethink narratives regarding the importance of the slave trade for the early European economy; the nature and indeed very existence of ‘half-free’ people and the German concept of Zensualität; and of course the problem of the so-called mutation féodale of the millennium in which the free and freed were, according to one influential theory, turned into serfs. The period covered by this book was one of slow but fundamental transformation, in which changes in legal status and conditions of land tenure, the nature and extent of the private power of landlords, and a slowly increasing level of economic complexity and commercialisation, converged to result in the social, legal, and economic structures we find in the twelfth century (whether or not this should be called ‘feudal’ is a matter best left aside for the present). While an interpretation of the trajectory of early medieval social history would need consideration of all of these issues and how they intersect, that was not the purpose of Rio’s monograph. She has done a tremendous service by giving us a new basis for discussion of one of those elements of social transformation. This study will be an indispensable addition to any and every library (and course bibliography) that includes early medieval history in its remit.

If I may be permitted, however, to end by pointing out an odd feature of this excellent book: it is fundamentally about the flexibility with which rules, often very rigid, are applied, so perhaps it is a homage (whether inadvertent or not) to the cogency of the author’s argument that her publisher has chosen to be flexible about one of its own rules. Although otherwise produced to a very high standard, it is striking that, despite the admonitions in the Press’s own indispensable guides, every single entry in the very comprehensive and impressively polyglot bibliography ends completely emancipated from the enslavement of punctuation.