Each of the four discussants of my paper has offered constructive comment on specific points of detail in the piece. I am grateful for their insights, and for their consensus that the time has come to break down the disciplinary restraints that inhibit interdisciplinary cooperation in the study of slavery (and much else). In terms of my central focus on the potentials and pitfalls of a comparative archaeology of Roman slavery, however, I gained most from placing these commentaries side by side and reflecting on them as a group. That is also what I plan to do here.
Whilst all the commentators share my interest in the history and archaeology of the unfree, they each work in very different fields. One of the four (David Mattingly) is an archaeologist who has championed the concept of ‘discrepant’ experience within the Roman provinces, one (Constantina Katsari) is an ancient-historian whose work explores the potentials of a diachronic approach to aspects of slavery in the Roman world, and two (Martin Hall and Paul Mullins) are historical archaeologists working in countries with histories of European colonial settlement and racial slavery. The former works on colonial South Africa, and has pioneered the use of synchronic comparative analysis in exploring the material worlds of 18th-century slave-owning societies, whilst the latter explores 19th-century and later consumption amongst marginalized groups (including African-Americans) in the USA. Each, of course, responds to my paper in the light of their own disciplinary backgrounds and interests, and in what follows I have attempted to highlight key points of difference, and common threads of shared interest, running through the four commentaries and my own contribution.
Analogy and comparative slavery
The epistemological dilemma at the heart of my paper is succinctly summarized by Hall: does the comparative approach take us back to the bad old days of modernist ‘laws of human behaviour’? None of the commentators feels ultimately that it does, yet the ambiguity that Hall correctly detects in my own thinking on this (and robustly sorts out for me) is also evident in some of the other commentaries, highlighting the spectrum of answers that can be offered both to the question of whether we should compare, and to its applied or methodological corollary: how should we compare?
Katsari is particularly concerned with methodologies of comparison, discussing a number of options and ultimately favouring the ‘contrast-of-contexts’ approach employed by Peter Kolchin (Reference Kolchin1987) in his synchronic study of slavery in the antebellum American south and serfdom in Russia. If I have read her correctly, she appears to imply that those undertaking synchronic comparison avoid the pitfalls of analogous reasoning which routinely snare those attempting diachronic comparison. All comparison has to start somewhere, and it is usually with the scholar's own principal or initial field of study. In that context, many would sympathize with an ancient-historian's plea that diachronic comparative study ‘should always start with analysis of the ancient data’. But few archaeologists would accept Katsari's implication that the point of comparative analysis is to find points of convergence rather than disjuncture. To compare is surely also to contrast. Drawing only on data that support or amplify ancient practice, we would simply fulfil our own prophecies about the nature, and boundaries, of slavery systems in the ancient world.
Mattingly has long championed a comparative approach to Roman imperialism (1997, 2006), but is concerned that when it comes to slavery, we do not have enough of what he calls the ‘right sort of data’ (p. 136) to allow us to compare. As he rightly suggests, there is an urgent need for new work on rural estates, mines, quarries and other sites with large slave populations, particularly within Italy. The recent signs of a new interest in the archaeology of Roman slavery noted in my paper offer some hope that this work will begin to appear in the coming years. But even without it, as I attempted to argue in my case study on graffiti, comparative insights might allow us to begin to penetrate, using what we already have to hand, what Bradley (Reference Bradley1994, 180) called the ‘psychological world of the Roman slave’. Put another way, comparison may help us to broaden our understanding of what the ‘right sort of data’ for exploring the lifeways of Roman slaves actually is. Our sense of where we need to look – and what we need to look for – may not be as well developed as we appear to think.
Is there a ‘material culture of slavery’?
Mattingly suggests we might draw a distinction between the imposed culture of restraint and the material culture of slave identity (‘slave culture’). I take his point, but agree wholeheartedly with Mullins's comment that in slave-owning societies the material traces of slavery are inscribed into everything; a point also brought out, of course, in Hall's work on South Africa and the Chesapeake (Hall Reference Hall2000). But it is important to emphasize the corollary to this point: that attempts to isolate forms of material culture unique to the unfree will usually prove fruitless. In this qualified sense, there is no material culture of the experience of enslavement. We might know this, but that does not mean we all like it.
Katsari appears to imply that classical archaeologists ought to be able to identify a slave culture that is divergent from that of the master class – that there will be a material distinctiveness, if we look hard enough for it, and in the right places. This is, I think, debatable. We have yet to identify any class of artefact made by Roman slaves entirely for their own use, for example. Matters are slightly better in the Caribbean and the USA where slave-made ceramics have been identified with some success. The best-known category here is colonoware, now believed by many scholars to have been manufactured by slaves in Virginia and the Carolinas, although debate rumbles on as to whether some – or all – colonowares were not in fact made by native Americans (Mouer et al. Reference Mouritsen, Mouritsen, Reynolds and Varone1999). The colonoware debate also highlights the extent to which – even in the USA plantation belt – slave-made artefacts are in fact a rarity: colonowares make up a tiny percentage of the artefact assemblages on most sites on which they occur. Thus, whilst slaves in the Americas may have made some of their ceramics, they more commonly made use of European manufactures. We will need to bear this point in mind when an archaeology of slavery finally does emerge in the Roman world, and we turn our attention to the range of sites Mattingly rightly asks us to explore.
Mullins, like Katsari, points to the fact that captivity is not reflected in unique material patterns or goods. As a North American historical archaeologist familiar with material culture creolization and hybridity, he is far less worried about the archaeological implications of this point, regarding the fact that ‘there is no such thing as a Roman context untouched by captivity’ (p. 127) as the starting point for an analysis of slavery centred on everyday materiality. I endorse Mullins's view that it is in everyday (‘prosaic’) objects that at first sight appear unimpressed by slavery that we will actually see the material world Roman slaves made for themselves. Like Wilkie, whose work on the consumer choices of Bahamian slaves is discussed in my paper, I would argue that we are likely to have most success in seeing the material world of the unfree when we isolate materials according to the ways in which slaves actually used them – when we explore the cultural logic informing the selection and use of European (or ‘Romanized’) manufactures.
Categories of servitude
Varied forms of enforced servility have existed in most slave-owning societies, and Mullins prefers to speak of the ‘archaeology of captivity’ (p. 123) rather than of that of ‘slavery’ for this reason. Mattingly similarly locates slave identities within the broader spectrum of discrepant identities occasioned by the operation of Roman power. There were certainly many forms of servile status in the Roman world, where coerced labourers, semi-servile kin, coloni and other known or putative semi-free groups laboured alongside slaves, and where manumitted slaves (freedmen) were tied by bonds of patronage and dependency to their former masters. In asking whether it is possible to isolate a material culture of slavery, Katsari at the same time asks whether it is possible to distinguish materially between slaves, other unfree labourers and freedmen. She suggests, for example, that it might not be possible to distinguish between graffiti made by slaves and that made by other servile groups. She is no doubt right about this, but I would suggest that our focus – within and between specific contexts – should be on the material culture strategies developed and shared by all those who have experienced captive servility. Graffiti were not exclusively produced by slaves and former slaves, but in the Roman world (and beyond it) both of these groups appear to have favoured this form of discursive strategy, and that is surely the important point. Self-inscription by means of funerary epigraphy was equally important to freedmen (Mouritsen Reference Mouritsen2005), and (contra Mattingly, who regards epigraphy as a ‘distancing’ strategy by former slaves) clearly seems to me to fall within the same spectrum of activity. It may never be possible to say with certainty that a non-verbal graffito X was made by a slave (or a freedman), but closer attention to graffiti and dipinti, and a better understanding of comparative material from beyond the Roman world, may help us to unlock some of the non-verbal strategies through which Rome's servile communities – both before and after manumission – expressed their identity.
Slavery and the politics of inequality
As Mullins argues very forcibly in his commentary, the archaeology of slavery is necessarily politicized. In this context there is a broader agenda on offer both in his own contribution and (in a less consciously politicized form) in that by Mattingly.
Mullins envisages a cross-cultural, Pattersonesque ‘archaeology of global captivity’, a project systematically comparing slaveholding societies across time and space, and within which Rome (my own starting point) becomes simply one ‘player’ among many. If I retain any ambivalence about the comparative project, it lies here, in the tension between the local and the global that (for rather different reasons) troubles Katsari too. Ultimately, I work comparatively to understand more about what it meant to be a slave in the Roman world, and I do actually want to be doing the fine-grained contextual analysis that Mullins suggests that I would like to overturn. Yet it is clear from Hall's commentary that he sees no tension here, and I take comfort from that.
In apparent contrast to Mullins, Mattingly argues that slavery should not be separated out from the wider, diachronic study of imperialism ‘at large’, and locates Roman slave identities firmly within the broader spectrum of discrepant (non-egalitarian) identities occasioned by the operation of power in Roman society. With reference to the latter, Mattingly shares with Mullins and with Hall the understanding that to study slaves and other disadvantaged groups is to study the technologies of power: to explore, as Mullins puts it, ‘how complex societies discipline their subjects and the range of ways people are integrated into states’ (p. 125). But should the archaeology of chattel slavery be framed entirely within a broader archaeology of imperial/colonial inequality or discrepancy, as Mattingly appears to be suggesting in his piece? I remain uneasy about this, despite having emphasized above that ‘captives’ (and former captives) of all sorts can often share behavioural traits and material culture strategies. What would be lost by subsuming or diluting the archaeology of chattel slavery into a wider archaeology of power inequalities? Certainly, this route was avoided in the USA and the Caribbean, where vibrant archaeological subdisciplines devoted to the study of the ‘peculiar institution’ have developed in the last 30 years. Classical archaeologists might usefully ponder this, as slavery finally emerges from the shadows, and as we grapple with the consequences of an archaeology of the human being as property.
