This is a characteristically trenchant and persuasive critique from Jane Webster, who has done more than most to persuade Roman archaeologists to engage with comparative and postcolonial perspectives (Webster 1997; 2001; 2003; 2005 are required readings). Yet, as she argues, despite some notable theoretical advances in the last decade or so the study of the Roman past still remains hidebound within classificatory and descriptive categories. The result of this is that we remain far better at delineating our subject than at interpreting it – especially where we enter territory that is not well served by supplementary sources (such as texts or epigraphy). However, the problems that she has identified with regard to the archaeology of slavery apply with equal force in many other areas where the material culture is ambiguous or discrepant. Thus, while I find myself in strong agreement with much of what she says, I would like to pose a series of points for further reflection.
(Why) are Roman archaeologists failing to compare?
The accusation that Roman archaeologists are neglecting the comparative potential of more recent slave-using societies is in general a fair one, but in broader terms the picture may not be quite as bleak as presented. Postcolonial approaches in archaeology have engendered a considerable engagement with comparative approaches (Hingley 2005; Mattingly 2006) and, while this is acknowledged by Jane Webster, I think she is unduly pessimistic about the extent to which other scholars have picked up on her ideas. It seems to me that much of what she wants to see happen in relation to the slavery debate is already under way in broader theoretical discussion about identity and material culture (Mattingly 2004). More serious problems may, in fact, concern the quality of data available on which to deploy the new theoretical insights (see below).
It is also debatable whether, overall, ancient historians have been quite as dynamic in embracing the comparative perspective as she seems to suggest in places. With the honourable exceptions of the cited examples of Finley, Hopkins and Scheidel, and Dal Lago and Katsari's (2008c) recent book, the typical concerns of the historical study of Roman slavery are often narrower and more conventionally based on ancient sources.
But something has constrained the potential contribution of Roman archaeology to this debate. It is a telling indictment that Carandini's analysis of the Settefinestre villa remains the cardinal reference point for identifying slaves in the Italian landscape (Carandini 1984). It is perplexing that there has been so little interest in subjecting other Italian villas to the same level of scrutiny regarding their zoning and social correlates. Jane Webster (2005) has tried to show that there is an archaeology of slavery to be investigated even in provinces such as Britain – that were simply slave-using rather than representative of the intensive slave society of the core territories. She also disputes above that we need to limit our comparative study to the ‘five “genuine” slave societies’ (p. 113). However, it stands to reason that the best point to start an investigation of comparative slavery would be the zones of most intensive slave employment. The failure of Roman archaeology to capitalize on Carandini's initial exploration of slavery in Roman Italy thus inhibits the work in less intensive slave-using regions. To put it another way, the academic priorities of Roman archaeology have simply not generated the quality of data needed to facilitate comparative study of ancient and ‘modern’ slavery. One of the reasons we have failed to compare is that we do not have enough of the right sort of data on the Roman world.
Thus, while I agree with Jane Webster that we can use comparative slavery data to shed light on the Roman situation, I think this can only get us so far unless the broader agenda and methodologies of Roman archaeology are revised and greater attention focused below the level of elite and urban society. That agenda needs particular elaboration in Italy, if we are to make the most of the provincial evidence. Indeed, it is significant that Jane Webster's most interesting suggestions for future directions focus on Italian data (graffiti from Pompeii certainly have some potential). Identifying the signs both of constrained lives and of creolized identities of servile communities elsewhere in the Roman Empire will be a deal easier if we have a clearer understanding of what we are looking for by comparison with both modern examples and the evidence from the most intensive regions of Roman slave society.
I think there is also an issue to discuss concerning the appropriate scale of analysis of the ancient institution of slavery. It seems to me that as slavery was a key feature of ancient imperialism – especially in its diasporic effects – we need to consider the one in relation to the other. To put it simply, can we divorce the comparative study of slavery from the diachronic study of imperialism? In order to break down the ‘tiresome tradition’ among classicists of characterizing ancient slavery as milder than that of more modern ages, we also have to debunk the modern nostalgic and affectionate respect for Roman imperialism. This is surely the greater obstacle to exposing the flaws in conventional wisdom about a specific issue such as the idea of benign Roman slavery. My book on Britain as an ‘imperial possession’ (2006) was designed to spark just such a debate and the reviews have made interesting reading concerning both perceptions of my agenda and the preconceptions of some readers. We cannot expect to progress far with revising views on Roman slavery without exposing the nature of the modern discourse on Roman imperialism and setting in its place a new set of theoretical models.
Some additional thoughts on the archaeology of slavery and slave identities
What does the archaeology of slavery consist of? We might make a primary distinction between the imposed culture of restraint and the material culture of servile identity – with the former represented by, inter alia, slave chains, slave barracks or cells, representations of slaves commissioned by their owners; the latter by the material culture of suspected slave dwellings and burials, or, as suggested by Jane Webster, by graffiti and other possible forms of self-expression. But slave ‘culture’ did not exist in isolation, and to identify it we surely need to adopt a more nuanced approach to the question of the differences in the expression of identity at all levels of Roman society. This can best be achieved, I would argue, by rejecting the Romanization paradigm and embracing my agenda of ‘discrepant identities’ (Mattingly 2004; 2006). Comparative studies may help us identify slave material practices across time and space, but we also need to compare slaves with, and seek to understand them alongside, other group identities and behaviours in Roman society.
There are some powerful tools available for tackling the problem of the archaeological invisibility of slaves. If we accept the diasporic impact of slavery as an institution, one obvious approach is to seek to identify the presence of lower-status individuals who appear to have originated outside established local communities. There is now a battery of new and improved physical techniques for identifying migrants within a given population. One of the most rapidly evolving techniques involves the study of diet through the analysis of suites of isotopes in teeth or bone (Katzenberg 2000; Prowse et al. 2007; Schweissing and Grupe 2003). As these seem to develop characteristic signatures that can reflect differences in diet at various life stages, it is in principle possible to say that Individual A differed from other members of the community among whom they died in terms of having markers of a radically different childhood diet. Or that Individual B, while sharing early childhood patterns with others in the group, seems to have lived much of her life away from other members of the group, experiencing a different diet. The quality of the results can of course be dramatically enhanced by the addition of DNA sampling and of morphological analysis of skeletal differences.
However, before we get too carried away about our ability to recognize migrants (and following Jane Webster's suggestion to identify some of these with forced migration and slavery), we need to consider some limitations in the techniques. First, the science is still quite new and the sophistication of interpretation will increase as more results are published (we are in the equivalent of the first radiocarbon age, where every new set of data throw up as many questions as answers). Second, the extent to which slave populations were sustained across the centuries by deliberate breeding in the household or on the estate is a crucial variable. If many slaves were born into slavery at a specific locale, possibly with ethnic mixing, and did not move in their lifetimes, the migrant test will be less clear-cut and less relevant to assessing the potential presence of slaves. Third, these techniques need to be applied in funerary contexts where we have good stratigraphic control of where skeletal elements were found and of associated artefactual and material traces (the approach is being followed in my current fieldwork on the Garamantes of the Sahara, with the hope that the composition and varied identity markers (and possible slave component) of this ethnically diverse people may be better understood – see Mattingly et al. 2007; Mattingly, Dore and Lahr 2008 forthcoming). This will involve a more integrated approach to cemetery excavations and funerary archaeology and is arguably better suited to inhuming societies than to cremating ones. The implications of this for the western Roman Empire are that we are likely to have better data for the late Roman period than for the earlier principate, whereas it was precisely the earlier period that saw the more dramatic wars of conquest and therefore potentially the larger scale of forced migration.
Another underexploited area to look for sharp social distinctions in lifeways might be imperial mines and quarries where forced labour co-existed with salaried staff. At Mons Claudianus in Egypt the expected evidence of a largely servile labour force has been effectively displaced by a new model (based both on ostraca and on considerations of the quality of the diet revealed by archaeobotany) emphasizing a free labour force (Van der Veen 1998). Yet there are references to people being condemned to metalla in the Egyptian desert and I suspect that a component of the quarry population was effectively servile. Larger-scale excavations and careful attention to differences between separate habitation, cemetery and rubbish-disposal areas could produce a much more nuanced picture. The copper-mining centre at Khirbat Faynan (Phaino) in Jordan was another imperial metalla and here the condemnation of Christian communities during (and after) the Great Persecution is securely attested by Eusebius (Ecc. Hist. 8.13.5; Mart. Pal. 7.1–2; 8.1; 13.1–10). Survey has revealed the extent to which the landscape of the Wadi Faynan was placed under close surveillance and control (Barker, Gilbertson and Mattingly 2007, 305–48). This again would be an excellent location to investigate through follow-up excavation the varied pattern of material identities. There must be a clear presumption at such sites of a strong behavioural and material divide between slave/forced and free elements.
In conclusion, while applauding Jane Webster's argument and general advocacy of a comparative approach, I would caution against separating the study of slavery from that of imperialism at large, or the pursuit of slave identities from the exploration of a range of discrepant identities occasioned by the operation of power in Roman society. Slavery was a key feature of Roman imperialism and what for me makes the comparative case most compelling is to focus on cases where slavery has similarly been part of a colonial situation. Before we overstretch the potential of comparative analogies between more modern slave societies and evidence from regions of the Roman Empire where the scale of the slave presence is uncertain, I am also arguing that we need to create a much more substantial body of archaeological data on the physical and social loci within the Roman world where slavery was large-scale and likely to have left particularly distinctive markers. We need an agenda that will encourage much more work (and at a much more sophisticated level in terms of techniques, material typologies and spatial control) on slave sites – rural estates in Italy, imperial mines and quarries, service areas of elite houses in Italian towns, brothels and bars, pottery production sites at places like Arezzo, and so on. We have a long way to go before we can start to understand the commonalities and diversities of slave identities in Roman society, or the powerful transformative effects on the identities of the manumitted minority. Slavery was a deeply unpleasant and degrading experience to undergo, reflecting colonial power inequalities and exploitative strategies. Whatever coping mechanisms slave communities created through their uses of material culture and their development of a sense of identity, one of the great (and thought-provoking) paradoxes of the behaviours of freed slaves was the extent to which they exaggeratedly embraced practices and social expressions of the freeborn, distancing themselves from their former status group.
