Neville morley introduction
The slave society of Roman Italy, characterised by the presence of large numbers of slaves (forming perhaps as much as 35 per cent of the population) in all kinds of activity from personal service to crafts to business to education, and in all regions and all levels of society from the depths of the countryside to the houses of the urban elite, developed over the course of the last two centuries bc. Over this same period, slave labour maintained a central role in agricultural production, in the market-orientated villa system of central Italy described by Cato and Varro; slaves were by no means the only people involved in productive activity, or even the majority, but they played a vital role in accumulating the marketable surpluses that sustained the lifestyles and ambitions of many of the elite. Their role in ensuring the social reproduction of the elite, both through apparently ‘unproductive’ personal services (which in fact were vital for their owners’ participation in the competition for status and the display of an ‘appropriate’ lifestyle) and through their dominant position in the process of educating and socialising the next generation of aristocrats, should also not be underestimated. Indeed, in a family environment that was characterised, as Bradley (1991a: 125—55) has argued, by a high risk of emotional uncertainty and dislocation, slave tutors and nurses offered some degree of continuity in personal relationships for young aristocrats, and so, perhaps, shaped the behaviour of generations of elite Romans.1
The key question for this chapter is how Roman slavery developed over the next few centuries, if indeed it developed at all. First, there is the question of its diffusion; how far Roman models and practices of slavery spread into the provinces of the empire, whether taken there by Roman officials, soldiers and other migrants, or adopted by the provincials themselves in imitation of their rulers. Secondly, there is the question of development and change, whether in the uses to which slaves were put, their treatment and place in Roman society, or attitudes towards them.
1 Cf. Vogt 1975c: 105.
265
Thirdly, there is the related question of whether slavery ‘declined’; the fate of the slave system in Italy itself during this period is often presented in terms of crisis.
The diffusion of roman slavery
Despite his provincial origins, Columella, like the other Roman writers on the management of villa estates, focuses almost entirely on Italy; his imagined reader is resident for most of the year in Rome, able to visit his estates in central Italy regularly and to contemplate different management strategies for ‘far distant’ estates that could not be properly supervised (Columella, On Agriculture 1.7.6; cf Pliny, Letters 9.37). There is no reason why Columella’s advice on the selection of an estate, the choice of crops or the disciplining of slaves could not be applied in the far north or south of Italy, or in more distant provinces; most of the precepts of the agronomists were readily (indeed, intentionally) transferable (cf. Columella, Rust. 3.13.1). It is simply that Columella, like the other agronomists, offers no evidence either way for the diffusion of the classic ‘villa mode of production’ — to be exact, the range of cultivation and management strategies that might be employed on an estate worked and managed by slaves — beyond the central regions of Etruria, Latium and Campania. Even when he discusses agricultural practices outside Italy, for example noting the different systems of training vines in the provinces, he gives no indication of the nature of the workforce that undertakes this work besides a few passing and vague references to agricolae and rustici (e. g. 4.1.5, 5.4.3). One might conclude that his failure to specify the status of the vine-dressers and other labourers implies that (in his view, at least) provincial estates were worked in the same way as Italian estates, by slaves, but this is no more provable than the idea that he is contrasting the methods of Gallic and Spanish peasants, developed in different conditions, with the longer-established and more ‘scientific’ approaches of Italy. Roman expansion had brought the vine to the western provinces; it remains to be seen whether Italy had also exported its system of agricultural exploitation.
Other literary sources are equally frustrating in their treatment of this topic, offering little more than passing comments; as Hopkins (1978: 99) put it, ‘the ancient evidence on slavery in the Roman empire outside Italy is so thin that it seems compatible with many theories’. It is clear that slaves could be found throughout the empire, sometimes in significant numbers — Galen’s suggestion (5.49K) that they constituted a third of the population of Pergamum may be no more than a guess, or an extrapolation from the elite households with which he was familiar, but it implies a considerable presence. In many regions, different forms of unfree labour long pre-dated Roman occupation; Roman writers noted the differences between native practices and their own but were not surprised to encounter some form of slavery in other societies (e. g. Tacitus, Germania 24—5; Bradley 1994: 22—3). There was a long tradition of slaveholding in the Greek East, with slaves involved in personal service, craft activity, trading and mining; in Pergamum there is evidence of slave managers on large estates, while many of the agricultural labourers in these regions were certainly ‘unfree’, if not necessarily chattel slaves.347 There was little, besides the specifics of the central-Italian villa system, that the Romans could teach the Greeks about slavery. In the West, on the other hand, slavery was marginal; war captives might be kept in the household for personal service or might be compelled to practise some craft, but increasingly in the last two centuries bc they were sold to slave-traders or merchants, to fuel the slave system of Italy.348
With the exception of a few special groups like the workers in the Spanish silver mines, many if not most of whom were unfree (Diodorus Siculus 5.38.1; Strabo 3.2.10; Edmondson 1987; Thompson 2003: 56—81), the role and importance of slaves in different sorts of production under the Roman Empire is unclear. It is sometimes assumed that the large numbers of slaves to be found in some elite familiae in Africa and Gaul (e. g. Apuleius, Apology 17.1) are explicable only if a significant proportion of them worked in agriculture, but that runs the risk of underestimating the ingenuity of Roman aristocrats in developing ever more complex demands for personal service (cf. ILS1514; Bradley 1994: 61—4). Even where we do find indications of the presence of slaves in the countryside, it is important to bear in mind the distinction between a slave employed in the highly organised villa system of the agronomists and a slave employed as an independent farmer, scarcely distinguishable in some respects from a tenant (as Tacitus described the slaves of the Germans: Germ. 25). Where rural slaves appear in ancient novels of the period, they generally appear in this latter role (cf. Apul., Metamorphoses 7.15—28; Longus, Daphnis and Chloe 4.19).
Only in Egypt, where detailed records have survived of management practices on large estates such as that described in the Heroninus archive, and other papyri permit the detailed reconstruction of life in the countryside as well as in the cities, is it possible to offer a certain response to this question. Overall, approximately ii per cent of those recorded in Egyptian census returns were slaves, a figure that is often used — in the absence of alternative evidence — as the basis for an estimate of the slave population of the empire as a whole.4 Slaves were more common in the metropoleis (i3 per cent of people returned) than in the villages (8 per cent), and it is assumed that they must have been still more numerous in Alexandria. About one household in six listed slaves on its census return, usually just one or two; as far as their occupations are concerned, they appear in the papyri as scribes, cooks, barbers, other kinds of personal servants, craftsmen and ‘slaves without a trade’, men-of-all-work (e. g. P Oxy. 3197; 3510). It is striking that, in contrast with Italy, few seem to have been employed as business managers or agents for their owners.5 In agriculture, the large estates of the wealthy, who could have afforded to invest in the human and material capital involved in the Roman villa system, preferred to rely on various forms of peasant labour; tied to the land and dependent on the landowner to different degrees, but clearly distinguishable from chattel slaves.349
For the rest of the empire, arguments about the prevalence of slavery have developed in three main directions, drawing on different sorts of evidence. The first focuses on the epigraphic material.7 Inscriptions commemorating slaves and freedmen are known from every province of the empire; they are almost all found in urban contexts, which of course reflects the general pattern of epigraphic finds. Where the occupation is indicated, it is almost always urban; the majority of slaves whose role is recorded were employed as personal servants, to officials, soldiers or local dignitaries, with a few involved in the imperial administration and a few employed in craft production or business. Even in Spain and southern Gaul, where it is often assumed that slaves were extensively employed in agriculture, there are few inscriptions from rural contexts or relating to rural occupations; the majority of finds from the countryside refer to personal servants, with a few commemorating overseers (vilici). ‘Everywhere but in Italy and Sicily. .. a decided majority of slaves did nothing productive at all’ (Mac-Mullen 1987: 379); free labour, he argues, was the norm in agriculture and craft production.
It is worth noting that the epigraphic record from the provinces, limited though it is, offers little support for the idea of a decline in slave numbers from the second century. Manumission inscriptions from Thessaly go through an apparent trough in the period ad 50—74 but rise again by the mid-second century; their decline in frequency over time, like the decline in other inscriptions commemorating slaves and freedmen, parallels the decline in the total number of inscriptions over the period of the Principate, rather than showing a steeper fall. As MacMullen (1982; cf. Bodel 2001) has argued, this epigraphic record reflects social behaviour, the wish on the part of those who could afford it to advertise their status and social climbing in this particular way. The decline in attestation points to the development of new patterns of display and competition, not to a decline in the numbers of those groups who had previously erected such inscriptions; just as the relative absence of inscriptions commemorating the freeborn does not imply racial degeneration or the catastrophic decline of the free poor.
However, precisely this point can be used against MacMullen’s argument that the absence of epigraphic evidence for slaves in provincial agriculture indicates the actual absence of slaves.350 There are, in the first place, far more inscriptions of all kinds from urban contexts than from rural. The inscriptions that have survived offer a sample of those people who put up inscriptions, not a sample of all slaves and freedmen. It is generally agreed that slaves in the familia urbana, and especially those whose occupation brought them close to their owners, stood a better chance of being manumitted or, if they died as slaves, being commemorated; those working in crafts or business had the opportunity to build up their peculium in order to purchase their freedom or a tombstone.9 The vilicus on the rural estate, occupying a role of responsibility and interacting directly, if infrequently, with his master, might entertain some hope of freedom; ordinary rural workers had little chance. In other words, the epigraphic record offers an entirely believable picture of patterns of slave wealth and manumission; the absence of ordinary rural workers is wholly predictable. One might also argue against MacMullen’s assumption that domestic service is automatically to be considered ‘unproductive’, but the key point is that the epigraphic picture from the provinces mirrors that of Italy, where other evidence leaves little doubt that slaves were heavily involved in agriculture in at least some regions of the peninsula. As Samson (1989: 100) suggests, the argument from silence could equally be used to demonstrate that Gaul and Spain contained no coloni or peasants either.
The second line ofargument in this debate has focused on archaeological material, on the grounds that it is less susceptible to bias against both poor farmers and rural slaves. Excavated sites have been examined for evidence of slave quarters and related artefacts (iron fetters, in particular), while survey evidence offers an overview of changing patterns of settlement and exploitation in the countryside, which might indicate the effects of the intrusion of villa-based slave agriculture on the Italian model. Here too, however, there are significant problems of interpretation, which are familiar from debates on the archaeology of the villa in Republican Italy; archaeological evidence also needs source criticism.10 Depending on the methods used and the intensity of the sampling, field surveys tend to identify large, wealthy sites more easily than small, poor ones, especially if different sites did not have equal access to the fine-wares which usually provide the basis of dating. Large, distinct concentrations of finds are identified as ‘sites’, but attempts to establish typologies or hierarchies of such sites, let alone to identify them with particular types of rural settlement known from other (mainly literary) sources, are fraught with difficulty.351 It is important to remember that the ‘villa’ of archaeology and the ‘villa’ of the agronomists are not identical; the former is a more or less luxurious residential complex, often identified by the remains of bath complexes or mosaic tesserae, representing a particular way of displaying wealth in the countryside. Such a rural residence might indeed be the end-product of several decades of successful exploitation of slave labour in the way described by Cato, Varro and Columella, but it is equally possible that the source of the owner’s wealth might have little to do with agriculture or slavery.12 Some excavated villas contain areas that may be interpreted as the quarters of slave labourers, but by no means all of them do (see below, p. 272).
With these caveats in mind, we can consider the findings of a number of field surveys from Gaul, Britain and Spain, regions of the empire where slaves were apparently not present in large numbers before the Roman conquest, to see if any pattern is discernible that might point to the intrusion of slave agriculture. The most obvious finding is that there was significant regional variation, relating partly to the indigenous pattern of settlement and partly to the circumstances of the Roman arrival. In the south of Gaul, for example, there was widespread redistribution of land for colonies under Caesar, the Triumvirs and Augustus; the patterns of centuriation still discernible show the extent to which traditional landholding patterns were disrupted.13 The north of the country escaped any such disruption to the rural economy. During the same period, vine and olive cultivation spread through the southern region, something which may be ascribed in part to the colonists. There is no direct evidence, however, that Italian settlers brought with them new methods of organising cultivation; the typical colonial allotment was much smaller than the estate recommended by the agronomists, and better suited to conventional subsistence farming than to slave-run, market-orientated agriculture.
By the end of the last century bc, the earlier settlement pattern in Gallia Narbonensis of numerous small villages on hilltops had been replaced by a dispersed pattern of numerous small farms, a process which can be interpreted as a response to the increased security provided by Roman rule; the north had been characterised by a dispersed settlement pattern long before the Roman conquest. Over the course of the first century ad, ‘villas’, identified by their distinctive architectural features, began to dominate the landscape in all areas of Gaul, either constructed ex novo or, more commonly (especially in the north), developed from existing farm buildings.14 These villas tend to cluster around towns, though whether they were orienting themselves towards their market, as the agronomists recommended, or in accordance with the social and political focus of their owners, or both, it is impossible to say. These new forms of rural site never wholly displaced ‘native farms’, which has led Woolf (1998: 159) to argue that they cannot have represented a new form of production but rather a specific form of consumption. However, the coexistence of villas and peasant farms in Italy has long been recognised, with peasants providing essential labour for the villas at certain times ofyear.352 In the south of Gaul, therefore, where traditional landholding patterns were clearly disrupted and Italian colonies were established, the survey evidence does at least not exclude the possibility of the introduction of the villa mode of production.
Britain resembles the north of Gaul in its development; villas appeared soon after the conquest, for the most part representing the reconstruction of pre-Roman farms rather than new sites. There is no evidence of any widespread redistribution of land or of significant disruption to the traditional rural economy, and British villas seem to represent a new form of displaying wealth and status rather than a new source of profit.353 In Greece, in contrast, luxurious villas appeared in conjunction with a general decline in the numbers of rural sites.17 In some respects this pattern resembles that of Italy in the late Republic, but in the case of Greece the literary evidence makes no mention of the displacement of peasants by slaves; both contemporary accounts and the decline in the level of off-site finds (interpreted as evidence for a decline in manuring) suggest instead the effects of post-conquest poverty and debt on the poorer proprietors, leading them to adopt less intensive methods of cultivation and to farm only the best land. Land-ownership seems to have become less stable, offering opportunities for a few well-off families to build up extensive holdings, and the country began to export grain, olive oil, flax and other goods; however, there is little evidence for market-orientated specialisation and none for investment in new, slave-based forms of agriculture.18
Many parts of Spain follow the same pattern as Britain and northern Gaul, with little sign of change in the countryside. However, the southeastern coast and the lower reaches of the river valleys of the Ebro and Guadalquivir more closely resemble Gallia Narbonensis, both in the introduction of new crops and in the spread of villas.19 Some of these larger sites were clearly associated with production for the market, with kilns for amphora production on the estate, and some in the Toledo province are found in conjunction with large-scale irrigation works, both of which imply a level of investment in agriculture that goes beyond conventional subsistence farming.354 Again, these larger sites never wholly displaced smaller farms, and further inland there are few signs of any major changes in the agrarian economy.355 Certainly these regions of Spain became major exporters of olive oil and wine during this period, as seen in the archaeological record at Ostia, and the development of a market-oriented agriculture might have involved changes in the organisation of production as well as in the crops produced. As in other surveys, however, the status of the labourers on such estates cannot be identified.
For a better idea of the nature of the ‘villa’ sites identified by surveys, we need to consider those that have been fully excavated.356 Here too, however, the problem remains that the legal status of individuals is archaeologically invisible. Plenty of excavated villas contain series of small rooms that might be interpreted as slave quarters, but which might equally be the quarters of tied labourers or hired workmen, or personal servants, or even storerooms.357 This can be said of villas like Foz de Lumbier in Navarre, rebuilt in the fourth century with a courtyard extension flanked by forty-four more or less identical small rooms, as well as rooms identified as ‘slave quarters’ in the main building,24 or Chiragan in Narbonensis, with its line of rectangular, timber-framed huts, some of them with verandas.25 As Purcell noted of the villa of Settefinestre in Etruria, ‘an alien archaeologist would spot the menial status of the majority of the occupants, but would not be able to deduce the phenomenon of slavery from the remains’ (1988: 197). In regions where the presence of large-scale agricultural slavery is not clearly attested by other sources, we cannot be certain that the workforce housed in such quarters was made up of chattel slaves.
In a few cases, the identification is more certain. The villa of Els Munts near Tarragona included a semi-subterranean structure with extra-thick walls, containing iron fetters; surely, its excavators believe, the ergastulum in which disobedient slaves were locked up.26 Iron fetters have been found across northern Gaul, especially in the regions of the Seine and the Meuse, and passing comments by archaeologists suggest that they are sometimes found in the context of villas.27 On the other hand, iron fetters are also found along the Rhine frontier in military contexts, and in East Anglia, where few historians believe in extensive use of slavery; the law codes of the later empire — and many of these finds date to the fourth century and later — make it clear that slaves were not the only class of inferiors who might be forcibly restrained.
Neither epigraphic nor archaeological evidence can rule out the possibility that slave labour was extensively employed in the provinces; some of the survey material suggests disruption in the rural economies of some areas of southern Gaul and Spain that might be associated with the introduction of a new way of organising production. However, the case can scarcely be proved either way. Historians have therefore turned to arguments from probability and plausibility: how reasonable is it to take an optimistic rather than a sceptical view of this evidence? Much depends on prior assumptions about the reasons why villa-based slave agriculture was originally adopted in Roman Italy. Whittaker (1980) follows Finley in assuming that this development was largely a response to a shortage of free labour in the peninsula as a result of almost unceasing war; in the newly conquered western provinces, in contrast to either Republican Italy or the modern colonial experience, there was no shortage of exploitable labour, and so no need to import slaves. Moreover, Roman colonists tended to be few in number, isolated, not especially prosperous and so more likely to follow local farming practices than to import new methods. Columella’s remarks about provincial methods of training vines, which were less labour-intensive and required less specialised skills, might suggest reliance on a workforce of peasants and hired labourers rather than highly trained slaves.28
Another sceptical argument could be developed on the basis of demographic factors, drawing on recent work on the Roman slave supply.358 The generally accepted (albeit highly speculative) figure for the slave population of the empire is roughly 6 million (with 2 to 3 million in Italy); approximately 10 per cent of the total population, following the proportions found in the Egyptian census data. If we assume that slaves were a significant force in agriculture in southern Gaul and some regions of Spain, comparable to their role in central Italy, this implies a significantly higher figure for the total slave population. Such numbers might be sustainable, given a high level of slave breeding and a low level of manumission throughout the empire. However, the volume of slave imports and/or rate of natural increase required to support the additional expansion of the servile population from the late Republic into the Principate might strain credibility.
The alternative approach to this question sees the recourse to slavery in Italy as having been motivated by thoughts of profit; villa agriculture enabled the owner to take a much higher proportion of the surplus than if the land was cultivated by tenants or peasants, even if slave labour was not actually more productive.30 The villa represented a significant investment, but one which paid off because of the demand for its products; wine being exported to Gaul in the late Republic, but all manner of goods being shipped to the city of Rome. The same opportunity presented itself to the western provinces. Roman imperialism fuelled the growth of centres of demand in the capital and at the frontiers; it provided the transport facilities, the security and the framework of law that allowed trade to flourish; it created incentives for producers to increase their surplus and market it, the negative one of taxation in coin and the positive one of introducing new forms of consumption and social display which required cash to participate.31 Of course, not all regions were equally well placed to take advantage of the growing demands of the capital and the army; the villas of Italy are largely found within easy reach of the coast or of navigable rivers. But given the relative costs and speed of sea and land transport, the coasts of Gaul, Spain and North Africa were effectively closer to Rome than many parts of the Italian peninsula.32 The archaeological evidence for goods from these regions at Ostia shows that at least some provincial producers responded to the challenge. It is possible, therefore, that these producers might have sought to increase their profits by adopting new methods of organising production, and introducing slaves.
This argument still does no more than identify conditions which might have promoted the adoption of slave-based agriculture; it remains equally possible that provincial landowners were able to make a sufficient profit by exploiting peasants in a more traditional manner, by using less labourintensive cultivation methods and by investing in equipment (the oilpressing complexes found in North Africa, for example) rather than slaves.33 However, it should not be forgotten that, even if the Romans neither introduced slavery into the conquered provinces nor transformed their agricultural systems along the lines of the Italian model, they nevertheless presided over an expansion of the number and roles of slaves and freedmen in most regions. A figure of io per cent of the population (if we take Egypt as typical of a province where slaves played a negligible role in agriculture) is by no means insignificant; slavery may not have dominated production outside Italy, but it permeated the society of the empire.
Slavery in imperial italy
Studies of the Roman economy in the western Mediterranean in the first century ad have often set up a contrast between dynamic, developing provinces like Gaul, Spain and Africa and a stagnant, declining Italy. A range of evidence drawn from literary sources is offered to illustrate this picture: Columella complained about the state of Italian agriculture in his day; the younger Pliny complained that he was unable to find suitable
Tenants for his estates; Trajan introduced alimentary schemes to assist the children of poverty-stricken Italians; in 92, Domitian introduced an edict to prohibit the establishment of new vineyards in Italy and to order the destruction of half the vineyards in the provinces.359 In recent years, archaeological evidence has been added to support the notion of a crisis in Italian agriculture, and in particular in the slave-run villas: in some regions — the Ager Cosanus in Etruria, the Ager Falernus in Campania — ‘villa’ sites began to decline in number from the end of the first century, while finds of Italian amphorae in Ostia declined dramatically after the Augustan period, to be replaced by imports from Gaul and Spain.360
Several theories have been offered to account for (and, to some extent, to bolster) this picture of crisis. One assumes that the cessation of significant Roman expansion must have had serious consequences for the supply of slaves, affecting their availability and price;361 another, that the development of Stoic doctrines that questioned the apparently clear distinction between free and slave — arguing, for example, that no man was a slave by nature — must have had some effect on the institution.362 Neither of these needs lengthy consideration. A peacetime slave trade across the frontiers of the empire had long existed and certainly continued to exist through the Principate, and home breeding of slaves played a significant, though unquantifiable, role in the reproduction of the slave population.38 As noted above, the epigraphic record offers no support to the idea of a first - or second-century decline in slave numbers in the provinces or in Italy, and literary evidence (especially legal sources) shows that slavery was widespread in late antiquity. Stoic ideas, like later Christian views, focused on the spiritual wellbeing of the masters, frequently using slavery as a metaphor and having little to say about slavery as an institution; there is no evidence for any ancient abolitionist movement, let alone for its having any effect on social practices.39
Two other theories have focused on the economic aspects of Italian slave agriculture. The first stresses the effects of competition as the provinces started to produce wine: Italian wine-growers lost their market in Gaul and were confronted with cheaper imports even in Italy.40 This idea places an undue, and anachronistic, emphasis on the role of exports in determining the health of an economy. Certainly Italian wine was no longer shipped to Gaul in large quantities after the first century bc, but the city of Rome could consume everything the villas produced, and their
Proximity to that market could surely outweigh whatever advantage Gallic and Spanish growers obtained by using different cultivation methods. Provincial growing methods might be cheaper, but they generally produced a much lower quality wine; this represented competition in a particular area of the market, not in the market as a whole, and the appearance of provincial products at Ostia and Rome may simply reflect the insatiable demands of the urban populace.363 The contemporaneous disappearance of Italian amphorae from the archaeological record might be explained in part by the adoption of barrels, while much Italian wine by-passed Ostia altogether on its journey to the city.364
An equally ideological approach, allegedly Marxist but in fact largely based on Stalin’s dubious interpretations of Marxist theories of historical development, attributes the crisis to internal contradictions in the slave mode of production: Italian landowners expanded their holdings in search of ever greater profits, but the increased costs of supervising larger slave workforces proved uneconomical.365 There is some evidence for the development of ever larger landholdings under the Principate (Pliny claimed rhetorically that six men owned the whole of Africa: Natural History 18.35), especially with the growth of the imperial estates; however, there is no reason to suppose that the individual farms owned by such magnates were even contiguous, let alone that they were managed as a whole rather than as individual units.366 Only if one assumes the existence of a law of historical development that one mode ofproduction must undergo a structural crisis to give birth to its successor — something which, in the case of ancient slavery, would anyway be more plausibly located in the third or even the fifth century —will this argument seem remotely plausible; it does not carry the imprimatur of Marx.45
It is clear enough that the picture of crisis in Italy is overdrawn, and that much of the alleged evidence for it is actually quite irrelevant.46 Domitian’s edict on vines relates to concerns about the over-production of wine in the context of a shortage of cereals, rather than a protectionist measure in support of Italian agriculture;47 Trajan’s measure to feed children in Italian towns cannot be taken as evidence for crisis in the countryside; the comments of Pliny and Columella are scarcely specific to this period (especially if Columella is simultaneously taken as evidence for the development of slave-based villa agriculture). We are left, then, with the archaeological evidence for an apparent decline in the numbers of‘villa’ sites in some, but by no means all, of the regions of Italy in which they had been established in previous centuries. Such a limited phenomenon may seem to require no further discussion; but, in so far as there were changes in the Italian countryside, they were apparently connected with the slave-run, market-orientated villa. This can be seen clearly in the contrast between the Ager Cosanus, where two-thirds of the villa sites disappeared in the course of the Antonine period, and the interior of the nearby Albegna Valley: further from the sea, and so less integrated into the wider Mediterranean economy, this region enjoyed a notable continuity of settlement, with no sign of a ‘crisis’ in either the the late Republic or the first centuries of the Principate.367
Farming is always an uncertain business; in the capricious environment of the Mediterranean there was a significant chance of harvest failure in any given year, and a run of bad luck would bring any estate close to disaster.368 It seems possible, however, that villas were more vulnerable than other forms of estate, but not because of their use of slave labour. Specialisation in just a few crops offered the possibility of significant profits in good conditions, but increased the risk of a disastrous harvest, rather than just a poor one, if the weather was unfavourable; peasants typically spread their risks by growing a wide range of crops.369 Compared with self-sufficient peasant farming, where the bulk of produce was intended for subsistence and storage, the villa was subject to the vagaries of the market as well as the environment; if prices were lower than expected, the landowner might not make a sufficient return on his investment in slaves and equipment. Some growers sought to insure themselves against such risks by passing the costs onto merchants, by selling the grapes on the vine rather than the finished product; if the harvest failed, the landowner was under no legal obligation to recompense those who had speculated on its success.51 Others focused on reducing their costs as far as possible and extracting the maximum value from their workforce; the need for close supervision of every aspect of the management of the villa is a central theme in Columella, creating the impression that intensive viticulture at least was a high-risk and rather fragile enterprise, with a slightly dubious reputation.52 It is possible, therefore, that the disappearance of many villa sites represents a change in wealthy landowners’ strategies for exploiting their lands, from direct management (with the need for a luxurious residence to accommodate the owner when he visited) to some form of tenancy; legal discussions of the management of the estates of widows and minors make it clear that tenancy was regarded as a safe and reliable option requiring minimal effort.53 This change need not imply any decline in the use of slaves in agriculture, since an estate might be leased to someone who would then exploit it using slave labour.370 At most, therefore, we might see this as a decline in a particular means of exploiting slave labour, not in slave labour in general; second - and third-century jurists take the presence of slaves in the countryside entirely for granted.371
As for the place of slavery in other areas of economic life in Italy, there is no indication of decline under the Principate and little sign of development. Slaves continued to be employed in all areas of production, with the only significant difference between free and slave workers being the more frequent employment of the latter as part of a large group, in gangs (such as the slaves who maintained the aqueducts of Rome) or factories (such as the potteries at Arezzo).372 Slaves’ main occupation continued to be domestic and personal service, in the broadest sense, from doctors, secretaries and tutors to cooks, dressers and masseurs.57 The Principate brought more examples of the ingenuity displayed by the Roman elite in the use of slaves to impress visitors and enhance their own public presence, such as Livia’s ‘pet child’ (delicium) and the dwarfs and other curiosities that, according to Quintilian, fascinated Romans in the slave market (Institutes 2.5.10.12; cf. Plutarch, Moralia 520c on the ‘monster market’; Barton 1993: 86). Roman moralists naturally revelled in all such manifestations of luxuria — as Pliny (HN 29.19) put it, ‘we walk on the feet of others, we recognise our acquaintances with the eyes of others, rely on others’ memory to make our salutations, and put into the hands of others our very lives’ — and it would be risky to conclude that such practices had actually increased since the virtuous days of the Republic, simply on the basis of their comments.58
The same could be said of the role of slaves in business and commerce; it does not seem to be a new phenomenon, but a number of sources allow us to see it in much more detail under the Principate than in earlier centuries. Above all we can draw on the writings of the jurists, who were fascinated by the problems of legal liability and responsibility that arose when a slave acted as an agent on behalf of his owner or even conducted business on his own account through the ‘legal fiction’ of the peculium.'59 As ever with legal sources, the attention devoted to a topic may be a reflection as much of its theoretical interest as of its importance in the real world. However, Roman law customarily developed in response to the situations it was asked to resolve; the presence of slaves in business was clearly taken for granted, and the problems this raised were real.60
Todd 1992: 32; Cunliffe 1997: 220.
5 Aubert 2001: 101—2. 7 MacMullen 1987.
28 Cf. Tchernia 1986: 172-84. 30 Cf. Morley 1996: 122-9.
58 Cf. Edwards 1993.
Slaves might work under direct supervision, but it was routine for them to work as semi-independent agents and managers (institores), as shipowners, salesmen and financiers, with considerable freedom to invest and exploit their owner’s resources (Digest 14.3; Johnston 1999: 99—104). Such activities might overlap with the slave’s use of the funds which were allocated to him as peculium; the Murecine tablets from the Bay of Naples show a slave making two loans to a merchant, one on behalf of his then master (an imperial freedman) and one on his own behalf (Dig. 33.8 on the peculium; AE 1972: 86—8; 1973: 138, 143; Andreau 1999: 71—9). The use of such agents allowed the senatorial elite to exploit their wealth (and share in the profits to be had) in financing maritime commerce, without crossing the bounds of social decorum by becoming directly involved. However, the elite were by no means the only free Romans to make use of slave agents, either on a permanent basis or for particular tasks.61 The institution also operated as a form of limited liability, since the owner could be held liable for no more than the original sum of the peculium, regardless of the size of the debt which his slave had run up (the original sum was also lost, of course). The Romans in general showed a preference for running business through their dependants (including family members) rather than through salaried employees (though the institor, even if a slave, might indeed receive a salary for his work). The slave, meanwhile, gained the opportunity to save enough money to buy his freedom, or even to buy slaves to work on his behalf (cf. Dig. 33.8.25 on the servus vicarius). Slave procuratores and institores, with privileged access to capital compared with the majority of the freeborn, might build up significant wealth and even, by virtue of their economic activities, a kind of status in Roman society. It is striking that when attempts were made to regulate the process of manumission, by stipulating minimum ages for the slave as well as for the owner, they almost immediately called forth exceptions for those slaves who played a significant role in the food supply of the city of Rome, either by owning ships or by managing a bakery (Ulpian, Rules iii. i; Buckland 1908:
533-51).
Slaves and freedmen in roman society
The slave who worked as an agent for a member of the elite, like the slave who worked as a confidential secretary, a doctor or a tutor, was both an insider and an outsider in Roman society; a trusted member of the familia, with privileged access to its wealth and connections, but regarded in law and ideology as utterly dependent, inferior and powerless.62 We may surmise that such an ambivalent social position may have been a problem
Kirschenbaum 1987: 89—121.
62
Bradley 1994: 76—80.
For the slave; it was certainly a source of anxiety for the slave-owning society. One recurring theme in the portrayal of slavery in Roman literature is a concern that slaves are not in fact as cowed, let alone as inferior, as the ideology suggested that they ought to be, and moreover that slave-owners are more dependent on their slaves than vice versa.373 The slave agent or manager, legally dependent and allegedly deficient in reason while at the same time often educated and literate and certainly expected to exercise his independent judgement in his master’s service, highlighted the problem; agricultural manuals forlornly repeated the advice that the vilicus, the estate manager, should be as near as possible to his master in intelligence, but not think so himself (e. g. Plin. HN 18.36).
Roman society was, in theory, founded on clear distinctions of status, which determined political rights and social standing. In practice, the distinctions were not always clear, let alone fixed; demographic factors, for example, undermined any hope of maintaining a closed, hereditary elite.374 By the early Principate, slavery seems to have become a focus for anxieties about wider changes in the social structure, involving the replacement of the fundamental distinction between citizen and non-citizen (even if the citizen body was then internally stratified) with a society stratified by wealth (where political rights sometimes but not always determined a hierarchy within a particular group). Thus, Juvenal’s character Umbricius complains that the decline of Rome can be seen in the fact that the son of freeborn parents has to give way to the slave of a rich man (Satires 3.131). Of course the slave remained inferior in law, but this seemed to matter less than his association with money and power. Equally confusing was the fact that, in the absence of clear physical differences between slaves and slave-owners, it was not necessarily possible to identify one’s supposed inferiors in the street. There was no form of dress that was specific to slaves — Seneca suggests (On Mercy 1.23.2—24.1; Bradley 1994: 95—9) that the idea was once proposed in the Senate but abandoned for fear that the slaves would realise how many they were. Only citizens could wear the toga, but that was only for formal occasions; slave clothing was generally expected to be drab, practical rather than fashionable — but that was not true of all slaves. Several of the jurists’ discussions focus on the likelihood of mistaking a free man for a slave or vice versa: the sale of a free man, like the sale of temple land, may be valid if the purchaser is ignorant of the object’s status, ‘because it can be difficult to distinguish a free man from a slave’ (Dig. 18.1.4—5). Runaway slaves were to be treated more harshly if they pretended to be free (Dig. 11.4.2); such comments point to the effort invested, with debatable success, in trying to maintain proper, visible, social distinctions.
Even more problematic for Roman society was the ex-slave, the Freedman. As Greek commentators remarked (Dionysius of Halicarnassus 4.22.4—23.7), Roman slavery was remarkable for the (relative) frequency with which slaves were manumitted, and the fact that properly manumitted slaves were accepted into the citizen body with remarkably few restrictions on their rights.375 The ubiquity of freedmen in the epigraphic record gives an unrealistic impression of the actual frequency of manumission; consideration of the demography of the slave population suggests that its numbers could realistically have been sustained only if manumission rates were low.376 However, it certainly existed as a possibility for a number of slaves, above all those who held responsible positions or were intimate with their owners, and thus offered a clear incentive for loyalty and hard work — even if, at the same time, it reinforced the dependence of the slave on the owner’s whims and complicated their relationship still further.377 ‘Masters could afford to be generous with liberty, because they benefited from giving it’ (Hopkins 1978: 132). If the slave bought his freedom with money he had earned from a salary or from the use of the peculium, the master profited from his endeavours and then received the cost of replacing him; as patronus of a freedman, the former owner enjoyed the social prestige of a reputation for generosity and a following of dependants and could continue to call on the services of his ex-slaves.68 Many freedmen then continued to work as agents for their former masters, with access to their wealth and resources; others entered business on their own account, sometimes (as was the case with the fictional Trimalchio) deploying significant capital which they had received as legacies.69 Many freedmen became prosperous enough to leave a physical record of their manumission; in many cases they also proudly commemorated their occupation and social successes, such as a position in a collegium or membership of the Augustales7°
Particularly prominent in the first century of the Principate are the emperors’ slaves and freedmen, the familia Caesaris.71 As the imperial house expanded its wealth and power, so some of its dependants enjoyed vastly expanded opportunities. Freedmen such as Narcissus and Pallas under Claudius, like emperors’ wives and other family members, wielded influence because of their access to the emperor (cf. Plin. Ep. 7.29). They could not aspire to the throne themselves, and so (unlike senators) could supposedly be trusted to offer disinterested counsel and to act on his behalf; by managing his correspondence and audiences (both essential tasks), they could control the flow of information to the emperor and influence who would be able to have their petition considered.378 By the end of the first century, an ability to keep one’s freedmen in their proper place and to listen instead to the counsel of the Senate was, at least in the eyes of senatorial writers, a prerequisite for a successful emperor (Plin., Panegyric 88; Tac., Histories 1.49).
Roman sources naturally focused on such powerful, controversial figures, and their accounts have fairly been described as ‘prejudiced, sensational, repetitive and depressing’ (Weaver 1972: 9). The familia Caesaris contained a large cohort of less prominent, and certainly less politicised, figures, known above all from epigraphy, whose activities were far more important and influential in the management of the empire. Slaves and freedmen were to be found at all levels of the developing imperial bureaucracy in Rome and in the provinces, from lowly clerical workers to influential administrators. Until the reign of Domitian, freedmen headed all the great Palatine bureaux; even after those posts were reserved instead for equestrians, a few imperial freedmen who had been raised to equestrian status reached the top of the administrative ladder.379 In part, this can be seen as an extension on a far greater scale of the Roman tradition of using dependants as agents, along with the fact that financial posts such as dispensator could be held only by someone with no legal personality separate from his master, but there was the additional advantage that the administration was kept out of the hands of potential rivals to the emperor’s position. Imperial service offered some slaves, who might be born into the familia, acquired from another family or recruited specifically, the possibility of a kind of career structure, progressing through a series of posts of increasing importance and influence; the practice developed, particularly in finance, whereby the official would acquire another slave as a subordinate (the servus vicarius), to be trained as his eventual replacement when he was promoted.74 Such officials seem to have had a good chance of manumission around the age of thirty, after which they could progress to higher positions. They had opportunities, both legitimate and illegitimate, for amassing wealth, and the most successful moved in high social circles; whereas most freedmen married within their own group, to judge from the epigraphic record, the majority of imperial freedmen married freeborn women, the daughters of earlier generations of imperial freedmen or of municipal families from Italy and the provinces. The career of the father of Claudius Etruscus recorded by Statius, which culminated in the grant of equestrian status by Vespasian, is exceptional, if only because we can follow it in such detail, but it shows what was possible for an imperial slave with sufficient luck and ability.380
The wealth of rich freedmen, whether acquired from commerce or imperial service, presented another challenge to the traditional structures of Roman society. Access to political power and acceptance in elite circles had always been determined by wealth as much as by birth — the later European phenomenon of the ‘poor noble’ would be a contradiction in terms — and under the Principate birth seems to have become less significant than imperial connections. Freedmen were excluded from standing for office (emperors were very sparing in granting the gold ring that marked freeborn status), but their sons suffered no such restriction, and by the end of the first century it could be alleged that many equites and even some senators were descended from slaves (Tac., Annals 13.27). Some freedmen were richer than any member of the Roman elite and expected to be treated with appropriate deference; Roman sources complained at length about such presumption, including Seneca’s horror-struck account of the master who called on his former slave and was turned away (MoralEpistles 47.9; cf Juv. 1.24—30,102—ii). It became necessary to seek alternative means of social discrimination, through ‘taste’: in Petronius’ novel, the parvenu Trimalchio is shown to reveal his vulgar origins in every failed attempt at imitating the elite way of life, while Seneca refers disparagingly to someone having ‘the wealth and spirit of a freedman’ (Ep. 27.5; Veyne 1961; Edwards
I993).
Of course, the freedmen remained rich and powerful, regardless of this social judgement, and clearly felt no shame in displaying their status as former slaves on their tombstones. Most, we may imagine, were addressing themselves to their fellow liberti, just as they married (and probably socialised) within that group; imperial freedmen, however, moved in wider and higher circles, and it may be that the connection with the imperial house was far more important than the fact that it was a result of enslavement. Certainly this is the approach of Statius in his account (Silvae 3.3) of the career of the father of Claudius Etruscus. The unnamed man’s lack of noble lineage is noted but assumed to have been eclipsed by his success; he gains status from the fact that his masters were the rulers of the world — and there is no shame there since even kings have to obey the emperors (43—53). His constant closeness to Caesar is emphasised (63—6); this is clearly the basis of his claim to respect. This may indeed offer a partial explanation for the fact that those educated freedmen who were in a position to give an account of the life of a slave, admittedly an exceptionally privileged slave, chose not to do so; not only was there no ready audience for such a memoir, but they thought of themselves less as former slaves than as members of the imperial household — even if rivals and jealous aristocrats preferred to emphasise their servile origins.
Conclusion
Bradley (1994: 12—14) has identified three approaches to the definition of a ‘slave society’: the demographic test (how many slaves), the location test (where slaves are employed), and a more general emphasis on dependent labour. Either of the first two approaches would clearly restrict the term to Roman Italy alone (and perhaps just to central Italy), and only to the period from the third or second century bc to the third or fourth century ad. In the rest of the empire, slaves constituted perhaps 10 per cent of the population, if we take the Egyptian evidence as typical, a figure well below the 20 per cent of a ‘true’ slave society according to the demographic criterion, and were largely absent from the productive spheres of agriculture, commerce and industry. The third approach would allow the label to be applied to the entire empire under the Principate, but it has been justly criticised for failing entirely to distinguish the special characteristics of Roman slavery, whether as a mode of production or a form of consumption, from the myriad other forms of dependent labour that could be found throughout antiquity.381
Without wishing to downplay the special characteristics of Roman Italy during this period, with its high numbers of slaves and the particular ways in which they were employed in villas, it seems strange that a society in which slaves could be encountered in all areas of life and at all levels of social interaction, in which the ownership of slaves was one of the most important markers of social status and in which discussions of the state of society were dominated by the problems created by the presence of successful slaves and freedmen should not be described as a ‘slave society’. As Bradley puts it, ‘from a cultural point of view.. . slavery was at no time an incidental feature of Roman social organisation and at no time an inconsequential element of Roman mentality’ (1994: 29). The culture he describes was, by the end of the first century ad if not before, not confined to the capital or even to Italy. The Greek-speaking provinces had owned large numbers of slaves long before the Romans arrived, and the free/slave distinction was one of the fundamental determinants of social identity.77 In the West, however, slavery had been marginal before the conquest; the Romans established new rules for social competition, in which the display of one’s dominance over others took on a particular importance. Rome did not export the villa mode of production to any great degree, but it did export its beliefs, habits, practices and anxieties; the provinces were confronted with, and clearly influenced by, a culture that was permeated by slavery.
In the Italian heartland, the Principate is best seen in terms of the consolidation of the institutions of Roman slavery rather than their development — let alone their decline. Some of the slave-run villas underwent changes of management or fortune, and we have the evidence to explore the activities of slave managers and agents in more detail, but there was no qualitative (and little sign of quantitative) change in the ways that slaves were employed. The powerful slave and the wealthy freedman, and the social anxieties they aroused, were not new phenomena either (cf. Plin. HN 35.199), although the establishment of the Principate created new opportunities for a few, and consequent overreactions from the self-appointed guardians of traditional social values.
Having abandoned the idea that Roman slavery developed during the Principate under the influence of Stoicism, historians have tended to assume that the literary sources from Republic and Principate alike should be put together to delineate a composite ‘Roman attitude to slavery’ — even if this is characterised as a complex, even self-contradictory attitude (e. g. Fitzgerald 2000). The disadvantage of this approach is that it is difficult to decide if the apparent differences between earlier and later sources mark actual changes in Roman attitudes or are due simply to the accident of survival or the perception of historians. One might detect a certain increase in the fearfulness of Roman slave-owners in the first century ad, in Columella’s obsessive concern with surveillance and discipline, in Seneca’s anxiety about slave numbers and Pliny’s sense, on the basis of a single incident where a master was murdered by a slave, that he and his contemporaries were exposed to ‘indignities, outrages and dangers’, regardless of whether they treated their slaves well (Ep. 3.14). One senator, according to Tacitus, claimed that ‘you will never coerce such a medley of humanity except by terror’ (Ann. 14.44). Conversely, there are limited signs of interest, in the fictions of Apuleius and Longus, in imagining the experience of being enslaved — something which did, after all, happen occasionally to the freeborn — if not the experience of being a slave, which is not quite the same thing.382 For the most part, however, Romans remained fixed in their attitudes, even as their society, and the societies of the territories they had conquered, were transformed by the institution of slavery. Of the slaves themselves, the lucky few who gained their freedom or who enjoyed material prosperity were happy to maintain the institution to their own benefit (cf. ILS 1514); the vast majority lacked even the hope of eventual manumission.
World History
