John bodel introduction
Domestic slavery was imagined as existing at Rome from the beginning — it could hardly have been otherwise with a cultural institution so deeply embedded in daily life during the historical period — and its established presence by the middle of the fifth century bc can be presumed from references to manumission and the liability of slave-owners in the Twelve Tables. However, slave labour did not become a significant phenomenon in Roman culture before the fourth century bc, when its rise in importance coincided with a decline in the institution of debt-bondage (nexum), as foreign conquests brought captive manpower to Roman territory and sent citizen colonists abroad, while displaced peasants migrated to the city in search of new means of support. From then on, if not before, agricultural slavery in Italy and, eventually, throughout most of the western empire predominated over all other categories of slave labour in importance for as long as landholding remained the cornerstone of the socio-economic system and the ideal of self-sufficiency was aspired to by the elite. Anecdotal reports in our literary sources of domestic servants in the houses of the kings, like the legendarily servile origins of the sixth king of Rome, Servius Tullius, reflect the conventions of foundation myth-making more surely than they do any historical reality, about which the most that can be said is that the houses of the wealthy of the regal period might seem to require staffing. We simply do not know how and when slave labour (no doubt of war captives) was first exploited at Rome, nor can we reasonably infer its original character from the developed system of the late Republic and early Empire. If we restrict our focus to the periods of classical Roman history when contemporary written sources of suitable quantity and quality survive (roughly 200 bc to 300 ad), we can see that the labour of slaves was integral not only to the economic but to the cultural ideology of the Roman slave-owning classes, and that the two aspects were inextricably intertwined.1
1 Twelve Tables 7.12,12.2a, with Watson 1975: 81—97. Nexum: Finley 1980: 83; Cornell 1995: 280—3, 333, 393—4. Agriculture: App. B Civ. 1.1.7. Servius Tullius: Cic. Rep. 2.37; Val. Max. 1.6.1, 3.43.
31I slave labour
‘Slave labour’ as a concept would have seemed both redundant and paradoxical to Romans of the upper classes, who owned most of the slaves — redundant because labour (in the sense that we understand the term, which is different from what Romans in many contexts meant by labor) was by nature servile, and paradoxical because there were few types of work from which slaves were excluded and none which only they performed: there was no peculiarly slave labour. Certain jobs undertaken predominantly by slaves naturally came to be associated with them, but the Romans did not think of work as ‘free’ or ‘servile’ in the sense that they regarded certain tasks as ‘man’s work’ or ‘woman’s work’, nor did they consider any type of employment — with the important exceptions of politics, the law and military service — as inappropriate for slaves.448
Slaves of both genders worked from childhood to old age in jobs more or less suited to their physical condition and capabilities, and were on call, at least potentially, at all hours of the night and day. In principle there was no task that a slave might not be called upon to perform, although with time a convention developed that slaves specially trained for particular roles should not be assigned to labours for which their training ill-suited them. The idea of a slave performing a service beyond the call of duty might engage a philosopher for the sake of argument, but even philosophers acknowledged that such an idea was perverse. ‘There is no leisure for slaves,’ remarked Aristotle, quoting a proverb (Politics 1334a). By this he did not mean that a slave’s work was never done (although idleness in slaves was considered wasteful, since they required maintenance whether or not at work), but that for a slave, who in daily life enjoyed neither freedom nor the exercise of free will, work was whatever the master wanted done whenever he or she wanted it done; there was no protected leisure time. That is the sense also in which Chrysippus called a slave a ‘hired man for life’ (Seneca, On Benefits 3.22.1). Everything a slave did, except what was done at the master’s sufferance, was done for the master and thus constituted work.449
For the slave-owner, on the other hand, a slave was much more than labour. A slave was property, a commodity of independent, if not unrelated, value. Nothing in the institution of slavery, it has been rightly noted, requires that a slave be a worker, and for many Roman slave-owners, work was not the most important function a slave performed. As well as representing a source of labour, slaves were caretakers, companions, partners and advisers — persons whose roles were not defined merely by the work they performed and whose value to the owner could scarcely be quantified. Cicero’s amanuensis Tiro is perhaps the best-known example, but the complexity of interpersonal relations that governed Cicero’s attitudes towards his human property, which ran the gamut from the avuncular affection evinced towards Tiro to the callous pragmatism of his response to a runaway slave, is unusual only in being so clearly articulated through his private correspondence.
Often, for men and women of Cicero’s class, slaves were status symbols. When Mark Antony discovered from their differing accents that two comely slave-boys purchased as identical twins at a high price (HS 200,000) from the slave-dealer Toranius Flaccus were not in fact brothers, he complained ofthe fraud but was persuaded that the boys were the more valuable because of their similarity, despite their differing origins, and wound up considering no other possession more indicative of his wealth and position (Pliny, Natural History 7.56). What their job in Antony’s household may have been we are not told — attractive servants seem to have been particularly desirable in the dining room, especially in the Ganymede role, as cupbearers — but it is clear that their main service to Antony was simply to be seen together and to be admired. For slaves of this sort, known as capillati, because of the long hair on their heads, or glabri, because of their smoothness elsewhere, the principal ‘work’ was simply to be on hand and to look sexually attractive. Not all jobs for slaves involved much labour. Indeed, when it came to associating labour with status, it was the hired workman in ancient Rome, rather than the slave, who was stigmatised with the title that reduced identity to work — opera, which came to mean metonymically both ‘a day’s work’ and, more reductively, ‘workman’. A runaway slave, by contrast, was guilty of the crime not of dereliction but of‘theft of self, on the principle that he or she had stolen property.450
The viability of the system has been endlessly debated — without resolution, or possibility of resolution, since we, like the Romans, lack the information needed to calculate and compare loss and gain. Because slavery in antiquity was grounded in ideological rather than economic considerations, slave labour was endemic in Roman culture — and was bound to be so, regardless of its profitability — during all the periods when the ideology prevailed (from at least the third century bc through the third century ad).451 Put differently, Roman slave-owners were naturally interested in profiting from slave labour, but profitability was not their main interest in owning slaves.
The ideology of work: the view from the top
In order to understand the place of slave labour in Roman society, then, we must approach the question as primarily a cultural rather than an economic issue, and we must begin by recognising at the crux of it an intersection of two conceptual polarities — slave/free and work/leisure — that together informed much of the ideology and behaviour of the ruling class. Where the two axes met is well illustrated by a famous passage of Cicero’s treatise On Duties (150—i), in which the New Man surveys the illiberal arts — the types of work considered sordid and thus unbefitting a free man — in a sweeping catalogue that embraces almost all the jobs we would think of as constituting labour and ultimately leaves as ‘liberal’ only the non-occupation of land-ownership. The blacklist begins with occupations that incur ill-will (customs dues-collecting and usury are mentioned), and goes on to retail-merchandising, which involves misrepresentation, shop work (‘for no workshop can have anything freeborn about it’) and, most sweepingly, all work done for hire, on the grounds that ‘wages are pay for slavery’. Lowest of all are trades that cater to sensual pleasures — those of ‘fishmongers, butchers, cooks, poulterers, fishermen’ (Terence, Eunuchus 257), perfumers, dancers, cabaret players and the like. Professions requiring a higher degree of learning or especially benefiting society, such as architecture, medicine and teaching, are suitable for those whose station they befit. Trade is vulgar if petty, but, when conducted on a grand scale, not wholly to be despised and, if ultimately abandoned for land-ownership, rightly praiseworthy. Of all the profitable occupations, however, none is ‘better, richer, sweeter, and more worthy of a free man’ than farming.452
Two points only need be noted here. First, Cicero’s hierarchy of occupations associates the type of work performed with rank (ordo) and arranges the scale according to social rather than functional criteria, grouping a wide range of more or less servile employments in ill-defined groups beneath a single category comprising the one activity befitting a free man — agriculture. Secondly, as Cicero knew perfectly well, a large portion of the land in Italy under cultivation in his day was worked by slaves, and more slaves worked in agriculture than in any other activity. (These two conditions obtained in Italy for more than three hundred years, from before Cicero’s day through at least the early third century, a fact of fundamental importance for understanding the place of slave labour not only in the economy but in the ideology of the upper classes, who sustained the cult of the peasant farmer on the backs of their agricultural slaves.) No contradiction existed for Cicero for two reasons: first, because, for Cicero, the reputation of any particular type of work did not depend upon the status of those who performed it but was determined by the nature of the work itself, and, secondly, because slaves, as labourers, were socially invisible — the work they performed was performed vicariously, not only for but as if by the owner. It will be useful to consider this second perception first.
When Cicero praised agriculture, what he had in mind was not actual farming but the management of mid-sized rural estates, and when Cicero thought of the slave staff associated with properties of that sort, he thought of it as equipment rather than labour. Jurists debating testamentary law quibbled over the parameters of the servile instrumentum of a villa but agreed on the principle that the slave workers needed to run a farm went with the property; they were considered integral elements of the land, like the buildings built on it or the trees planted in it. According to Varro (On Agriculture 1.17.1), one authoritative classification of the things needed to cultivate fields grouped slave labourers with oxen and carts as types of tools: articulate, inarticulate and mute. On this understanding, an absentee landowner overseeing the operation of his rural estate by a slave staff was ‘farming’ as surely as the peasant who put hoe to earth: each was employing a tool of the trade appropriate to his position. Separation from the physical toil of sowing and reaping, often at several removes from the actual labour via a hierarchy of slave intermediaries culminating in the farm bailiff (vilicus), did not distance a landowner from farming any more than dictating to a slave stenographer removed him from writing. The mechanics were performed by human instruments, but authorship by the master was presumed in both cases. By the same logic, the slave of a slave (a legal fiction recognised and condoned by the jurists), although he or she might have a specific job or function, was not normally labelled with an occupational designation alone (ifat all), but rather as a ‘substitute’ (vicarius) performer of the same task, the credit for agency being deferred onto the fictitious ‘owner’ and thence up the line, ultimately, to the master, the one with a juridical persona and thus the property rights necessary to claim the work.453
It was the transparency of the slave as a surrogate that made slave labour essential to the ideology of self-sufficiency, since it allowed the slave-owner to reap the material and social benefits of labour (revenue, prestige, autonomy) without incurring the physical and social costs (fatigue, toil, dishonour). The advantages of this personal transparency can be seen throughout the system. At the opposite end of the scale from field workers in prestige and circumstances were the business agents and managers (actores or institores) who conducted much of the daily business of commerce and whose lack of juridical personae enabled them to represent their owners in contracts vicariously in a way that free persons, even freedmen, employed in the same capacities could not. Within a legal system that did not recognise a concept of direct representation, free persons, having their own juridical identities, were personally responsible for contracts they entered into and for work they performed. This handicap severely limited their usefulness as business agents: if an actor were not the principal’s slave, the principal to a contract had no action against a third party but could be sued by one. Dealing with free persons at the level at which most institores and actores operated meant engaging directly in commerce, whereas conducting business through slaves or freedmen enabled a slave-owner to maintain a respectable distance from the sordid business of trade while yet retaining full control over profits.8
The transparency of the slave as a tool, however, applied only to the world of business (negotium); in the activities of leisure (otium), servile representatives were decidedly out of place. Recognition by slave-owning Romans of the inconsistency inherent in such an ambivalent conception of the slave’s surrogacy explains the hostility that ex-slaves invariably provoked when they failed to comprehend the parameters of their own newly recognised identity. The wealthy Calvisius Sabinus, whom Seneca characterised as having both the patrimony and the mentality of a freedman, and whose social ambition led him to affect an education by training his slaves to memorise Greek poetry, on the theory that what his slaves knew, he knew, offended upper-class sensibilities by confusing labour with learning; but the underlying principle of vicarious capability on which he based his claim was not so much misconceived as misapplied: Sabinus erred merely
By Keith Bradley (12), Neville Morley and Walter Scheidel in this volume. Cicero on agriculture: Sen. 51-8, with Powell 1988: 205-7; Or. 1.249; Fam. 16.21.7; Rep. 5.5, with Carlsen 1995: 60-1; cf. De Robertis 1963: 87—93. Jurists: Dig. 33.7.19.1, 33.7.8; Cod. Theod. 6.35.1,12.1.6,10.8.4. Stenographers (notarii): Cic. Fam. 16.10 (Tiro); Mart. 5.51.2; Plin. Ep. 3.5.15, 9.36.2; CIL 6.9130, 6.10229.43 (testamentum Dasumii’)-; P Oxy. 4.724. Vicarii: Weaver 1972: 200—6. Legally the master’s property: cf. Dig. 33.7.12.44.
8 Indirect agency: Aubert 1994: 40—116; see also Jane F. Gardner’s chapter in this volume. Actores: Aubert 1994: 186—96; Carlsen 1995: 121—42; Schafter 2001. Handicap: Dig. 5.1.19.3, 14.3.1.1, 14.19.3; cf. 14.5.8. For slaves in commerce and banking, see Kirschenbaum 1987: 89—121; Andreau 1999: 64—70.
In treating the products of a gentleman’s leisure like the products of his labour. For the Romans, less sensitive than the Greeks to conceptual impurities, it was this ambiguity alone that elicited the ambivalence identified by Nietzsche in the Greeks, who regarded both slavery and labour as ‘a necessary disgrace, of which one feels ashamed, as a disgrace and a necessity at the same time’.454
More broadly, Cicero saw no contradiction between slave labour and the ‘liberal’ occupation of farming because for Cicero (and for upper-class Romans generally) work was not characterised by the status of the persons who performed it but was inherently more or less ‘servile’ according to criteria derived from the purpose of the task (occupations catering to sensual pleasures were more disfavoured than others), and from a system of values that disparaged any work for pay as a failure to attain the aristocratic ideal of self-sufficiency (autarkeia), which alone afforded the leisure and autonomy needed to pursue politics. Commerce and wage-earning compromised independence because they entailed reliance on others for sustenance and thus limited personal liberty; as such they were an impediment to political life. Farming, on the other hand, contributed to self-sufficiency and was therefore a suitable pursuit for a free man. But just as the dignity of the enterprise did nothing to ennoble the gangs of slaves who toiled in the fields, so the lustre of the work they performed was little tarnished by their own debased status. The dirty business of digging and ploughing, when not on one’s own land, Cicero classified among the illiberal pursuits not because of the nature of the work, but because of the lack of independence it implied. Cultivating one’s own homestead was another thing. The inspiring image of the sturdy farmer-soldier-statesman called from his fields to protect them by serving the state evoked emulation in spirit, if not in practice, long after the time when Italian soil was imminently threatened or Roman senators worked their own land. Recognising that the focus of aristocratic prejudice against hired labour was centred on the relationship of dependency it signified, rather than on the status of those who performed it, helps to explain why there is so little evidence of slave and free labour being distinguished from one another when the two were employed together, as they frequently were, not only in small crafts and shops, where it is generally believed that slave and free worked side by side at the same tasks, but in large-scale agriculture, where the two were virtually interdependent.455
Cicero did not speak for most Romans, but his views on labour, calibrated with all the precision of the successful arriviste, reflect well the ethos of the ruling class during the late Republic and early Empire. Very different was the attitude towards crafts and labour expressed by slaves and ex-slaves in inscriptions and art, where repeatedly throughout the early imperial period we find the life of the honest workman held up as a source of pride and occupation emerging as an important, in some cases primary, marker of identity. The signature of a slave craftsman (‘Diogenes builder’) carved into a tufa block high in an exterior wall at Pompeii beneath a sculpted representation of a set of building tools may be less telling as a general selfevaluation than the declaration of an inscribed tombstone, but it shows that expressions of pride in a craft were more than a trope of funerary art. Epitaphs set up by slaves for each other, such as that dedicated to ‘Hilarus, goldsmith’ and depicting the tools of his craft (balance, compass, engraving tool) commissioned by a burial society (collegium) of fellow-slaves from the house of a senator’s daughter, show that occupational designations were not merely a convention of slave-owners, but that slaves themselves associated individual identity with work roles.11
These few examples may suffice to illustrate the complexity of Roman attitudes towards labour and status, which depended to a large extent upon the social perspective from which they were viewed. Juridical status affected the legal capacities and social ‘transparency’ of persons of different condition in certain types of work (slaves predominated in managerial functions and domestic service, for example), but work was not characterised in relation to status associatively, by the condition of the workers who performed it, but conceptually, as either (for those who subscribed to the aristocratic ethos) detracting from autonomy and independence, or (for the great majority who did not) as contributing to a sense of identity and place in the social order and a source of pride.456 457
The mlter of evidence
Five main categories of evidence provide most of the information we have for Roman practice, but each enables us to view it only selectively and through a particular lens. ‘Occupational’ representations from across the empire, mostly funerary and virtually all commissioned by persons (of freeborn, freed and servile status) from the great mass of Romans whose social and economic circumstances freed them from elite ideological constraints, illustrate a wide range of craft and mercantile activities practised by slaves but tell us virtually nothing about agricultural labour or domestic service. Inscriptions, mainly from Rome, attest a wide variety of crafts and shop work and provide a strikingly detailed picture of the domestic service jobs to which slaves might be more or less permanently assigned; but it is unclear to what extent the micro-specialisation of work roles attested in the epitaphs of slaves from the largest senatorial households in Rome reflects the patterns of domestic service elsewhere, and agricultural jobs are conspicuously absent. Papyri, a far from negligible source of information on slaves’ work roles in Egypt, exhibit the biases of the documents they preserve, mostly tax records and receipts, even if the social and economic conditions they reveal are more typical of those elsewhere than once thought.458
Literary sources, including legal texts, provide prescriptive guidelines for labour management and reveal much incidental information about slaves at work in various capacities. Together with the physical remains of rural properties in Italy and the provinces, they supply virtually all the direct information we have about agricultural slavery. Almost without exception, however, these sources represent the circumstances ofthe upper classes, and each has its own peculiar limitations. Some of the larger rural properties in central Italy perhaps preserve evidence of the sorts of large-scale barracks one might expect to find housing the squadrons (decuriae) of agricultural slaves described by the Roman agronomists, but slave-prisons (ergastula) for chained gangs of slaves of the sort deprecated by early imperial writers are nowhere recognisable, and the simple functional rooms (cellae) often identified as slave quarters because of their lack of distinctive appointments could for the same reason equally well have served a variety of different purposes. Most mid-sized and smaller rural homesteads reveal no traces whatsoever ofspecial arrangements for housing slaves. In the end, the most that can be said is that certain large villa properties in Italy of the first two centuries exhibit features consistent with the intensive exploitation of slave labour for agricultural production — a conclusion about which our literary sources in any case leave us in little doubt.459
The literary sources, in turn, provide evidence for only a small part of slave agriculture. The four surviving Roman farming manuals (our most important sources by far) tell us only about the operation for profit of mid-sized properties in the heartland of peninsular Italy. About how slaves were used in agricultural work on other types of rural properties and outside Italy we are much less reliably informed, but there is reason to think that conditions varied considerably from region to region. In North Africa, for example, the Roman conquest seems to have resulted in a decline in the local use of slaves for agriculture, with large-scale slave labour largely replaced during the Roman period by tenant farming, even as the slave mode ofproduction became the dominant mode ofagriculture in southern Italy. A similar replacement of slave labour by free tenancy as the principal mode of farming in the western European provinces in late antiquity has been explained as a result of a shift in the relative balance between slave prices and wages for free labour, but the factors that determined whether rural estates were farmed by tenants or directly by a slave workforce were complex and varied not only regionally but over time. Nor did the rise of tenancy mean that slaves were not still employed in farming; tenants exploited slaves in the same way as absentee landowners, and certain slaves farmed more or less independently, as if tenants (servi quasi coloni), but slave labour no longer occupied the central position in agriculture it once held.460
A perceptible shift of focus in the four farming manuals, a reflection in part of their uneven distribution over time, reveals a progressive growth in the elaboration of both the managerial functions and the servile staff on rural estates in Italy between the middle of the second century bc and the middle of the first century ad; but four centuries later Palladius addressed his instructions primarily to those managing a workforce of free tenants (coloni) rather than slaves quartered at the farm. Jurists writing at the beginning of the third century imply that the villa system of slave agriculture continued to thrive, but at what point subsequently and before the time of Theodosius II (emperor ad 408—50) the decisive change occurred, and why, are much debated. It is in any case clear that by the middle ofthe fifth century the dominant place of slave labour in Italian agriculture and the most economically significant use of slave labour in classical antiquity had come to an end, even as the ideology of Roman slaveholding continued to flourish and exploitation of the land continued to provide the greatest source of wealth.461
Categories of work and occupational designations
The historian Tacitus (Germania 25.1) characterised the Roman system as one in which specific duties were distributed throughout the household. Modern historians have been similarly impressed by the variety and specificity of the jobs held by Roman slaves. The range of occupations recorded in ancient texts is indeed striking — one recent survey culled more than 160 specific titles from the law codes, agricultural handbooks and epitaphs — but lists of the named occupations convey only a partial and in some respects misleading impression of the work that slaves did. If one considers slave labour to be anything that slaves were regularly expected to do, then the range of slave-jobs cannot be confined simply to those to which we can attach specific Latin (or Greek) names — often merely because they happen to be attested in one or two inscriptions from the city of Rome. At the same time, so closely was the practice of occupational labelling in certain contexts associated with the identification of slaves that one may reasonably question whether studies of the phenomenon are not more revealing of the Roman mania for classifying property than of the varieties of tasks that Roman slaves actually performed.462
On the other hand, linking an occupational title with a name labelled a person with a specific work function and thus marked out both the job and its practitioner: a single inscription attesting a particular title may suggest an entire category of designations, even if it says nothing about the typicality of the position it attests, and any slave labelled with an occupational designation acquired a distinctive identity related to the world of work that distinguished him from his undifferentiated peers. That even one ex-slave from a private house is identified in an epitaph from Rome as a silentiarius, for example, a slave charged with keeping the household staff quiet, tells us something about not only the mentality of Roman slave-owners and the working conditions of domestic servants, but also the types of servile assignments considered worthy of specific designations. Any occupational title might seem to mark a slave as being in a superior position, in respect to placement within a household if not in actual training, to the vast majority of slaves who undoubtedly had none, but even so simple an inference may need to be qualified by consideration of the context from which the information derives. Slave-owners, for example, may have been less inclined than slaves to dignify the mundane work assignments of domestic staff with formal titles; slaves and ex-slaves of socially prominent masters were manifestly more prone to boast of their positions than those of humbler households; slaves living in or near towns tended to be identified with occupational role-names far more frequently than those from less densely populated regions; and so on. Furthermore, the written sources that provide our evidence harbour a host of individual biases: legal texts revel in specificity and categorisation for their own sake; tax documents conform to the requirements of fiscal reporting; epitaphs further self-promotion as well as commemoration. Geography, chronology and gender, too, play a role in the fragmentation of our evidence, to the point that our quest to characterise generally the phenomenon of naming the jobs for slaves runs the risk of mirroring that of the three blind men attempting to describe the elephant: we may form distinctly misleading impressions by latching on to different parts of the animal, while the overall shape of the beast remains obscure.463
A pair of inscriptions from Rome and a papyrus from Egypt may serve to illustrate the variety ofour information, as well as the range ofinterpretative difficulties we face. When a high-ranking imperial slave, a financial official of the emperor Tiberius in charge of the Gallic Treasury in Lyon, died while visiting Rome, sixteen under-slaves (vicarii) travelling with him erected an epitaph to him in a monument for members of the imperial household beside the Via Appia in which they inscribed their own names and titles along with his own: the entourage of Musicus Scurranus comprised two cooks, two chamberlains, two slaves in charge of silver, two attendants, three secretaries, a buying agent, a treasurer, a doctor, a slave in charge of wardrobe, and a woman (the only one in the group), Secunda, who is listed last and without any job title. Do we assume, from her lack of a formal job-designation, that Secunda enjoyed a more personal and privileged relationship with her owner, or that her services to him were more discreetly left unspecified? It is in any case she, rather than her male colleagues, who represents the normal situation, since the great majority of slaves commemorated in epitaphs have no occupation indicated, and only io per cent of all epitaphs from Rome include such information.464
A more or less contemporary inscription of Augustan or Julio-Claudian date found beside the Via Praenestina outside Rome records the names of members of a funerary collegium of freedmen and slaves of an unknown private person, perhaps one of the Statilii Tauri, whose extensive household had a columbarium outside the Porta Maggiore. It lists some eighty persons, of whom only twelve are identified by occupational designations: four carpenters, two bailiffs, two workmen, a builder, a bath attendant, a marble worker, and one man identified as a painter and temple attendant. The first two (of four) columns of names seem to record the names of those already deceased, whereas the last two include some who had died and some who were apparently still alive at the time the inscription was carved, but there is no clear pattern to the distribution of job titles among them. As in the case of the vicarii of Scurranus, the inscription seems to have been set up by and for slaves (or ex-slaves), but mere pride in an occupation seems unlikely to have inspired the identifications, since those without designations, even if they could claim no specific craft or office, might well have been identified as mediastini, like the two who were.465
From Rome to Alexandria. A private contract drawn up in Egypt in ad 111 and listing the slaveholdings of a wealthy Alexandrian, Ti. Julius Theon, identifies individually at least fifty-nine male slaves (no females are recorded) and specifies job titles for only eleven of them. Do we conclude from their number that the five stenographers named (along with two secretaries, a scribe, a cook, a barber, and a repairer) served a different function in Theon’s household from the others, perhaps as part of a commercial venture, or that the position was particularly prestigious and thus worthy always of being named? That other slaves listed in the catalogue are identified simply by physical features (‘snub-nosed’, ‘the tall one’), origin (‘from Cussae’, ‘from the lower toparchy’), circumstances (‘at Berky’, ‘formerly the property of X’), or kinship (‘brother/son of X’), suggests that the occupational labels here may be mere descriptors rather than formal titles. That interpretation is consistent with the view of the second-century jurist Gaius, for whom an occupational designation was merely one of several ways by which a slave destined for manumission by testament could be distinctively identified.466
What stands out among the named jobs for slaves is a tendency towards particularism and specialisation, especially (and naturally) at the top of the social ladder, in the imperial household, where the scale of the emperor’s dominions and the administrative responsibilities of his staff necessitated hierarchies of managerial function not required in private homes; but these features are characteristic of all sectors of employment, from domestic service to crafts and trades to agricultural work. As it happens, our most informative statement of the rationale behind this propensity to compartmentalise comes from the farming manual of Columella. According to Columella, slaves should be assigned to particular tasks in order to reduce the risk of shirkers evading responsibility and to ensure that productive workers feel that their industry is recognised. One might infer from the latter prescription an encouragement to owners to stimulate in slaves a sense of pride in their work, but the tenor of Columella’s advice is cautionary rather than hortatory, and his view of the motivations of slave labourers is decidedly negative. He goes on to recommend that field labour be performed by squads of no more than ten, since greater numbers could not easily be guarded and might prove intimidating to an overseer, nor in groups of fewer than three, since workers widely scattered were not easily watched. Such an arrangement not only stirred up rivalry but exposed sloth, ‘for when work is stimulated by competition, punishment inflicted on the lazy seems blameless and justified’. Elsewhere Columella advocates rewarding industrious workers with honours and other preferential treatment, and in general shows himself to be far more attentive to the efficacy not only of material rewards but of positive psychological reinforcements in motivating a servile workforce than Varro, who acknowledges the idea but gives it little emphasis, or Cato, who says not a word about incentives.467
Columella’s advice was designed to enable slave-owners to extract the greatest possible productivity from a workforce with no inherent incentive to perform, and to that extent may be thought to have applied to other uses of slave labour as well; but practicality cannot be the only explanation for the pervasiveness of particularism throughout the system. In fact, the most extreme levels of specificity seem not to be concentrated in areas where the need for specialised skills or close oversight of performance was greatest. Instead hyper-specialisation is most apparent in the work assignments that served the interests of show and self-representation. From the end of the first century bc, for example, the emperor’s domestic staff included separate slaves in charge of gold, dinner plate, gold plate embossed with gems, gold drinking vessels, silver drinking vessels, serving bowls, wines, beverages and napkins, but only a single freedman procurator oversaw the entire budget of the emperor’s private affairs. That position, the most prestigious in the emperor’s household, ranking higher, it seems, even than the procuratorships of certain imperial provinces, might seem to belong to a different world from that of the dining room, but we should not be too quick to segregate the work regime of table service from that of high financial administration. The first ex-slave known to have held the post, a certain Bucolas, under Domitian, began his career as taster and diningroom supervisor for the emperor Tiberius. We may remember also in this regard Tiberius’ procurator of the Gallic treasury, Musicus Scurranus, whose name suggests a background in entertainment.468
In some cases, terminological ambiguities cloud the picture. Not only did the titles assigned to particular functions in domestic service change over time, but certain titles described different functions at different periods. The position of atriensis, for example, fell in a markedly downward spiral during the late Republic and early Empire. In the time of Plautus, the atriensis was the most important slave in the household, assigning work to other slaves and overseeing the master’s business with outsiders. By Cicero’s day, the managerial functions of the atriensis had devolved onto a dispensator, and the position had become associated with that of cooks, bakers and house-cleaners. By the middle of the first century, the atriensis is found guarding the atrium, and a hundred years later the position is grouped together with that of doorkeeper, among the most menial of the domestic assignments. The position of doorkeeper, in turn, was variously identified by the terms ianitor and ostiarius. Both designations seem to have described the same job and, once the latter term had been introduced toward the end ofthe Republic, could be used interchangeably. Ostiarius became standard in inscriptions and the legal codes during the Principate and eventually passed, via Christian writings, into the Romance languages, whereas ianitor remained the preferred term in classical literary texts, especially of the more elevated genres. The variable usage therefore seems less to reflect a purely chronological development than to be a stylistic choice, with the older term perhaps conveying a pejorative tone alien to the professed neutrality of juristic opinion and understandably eschewed by those for whom the job was a source of pride. To judge from the literary sources, even a comparatively modest household might boast a doorkeeper, since the job was not onerous and could be entrusted to an old man or woman, but Varro imagined a villa as potentially not having one, and the working conditions of those we happen to hear of suggest that the position was not one held by the more privileged or trusted household servants. The rhetorician M’. Otacilius Pitholaus, for example, who rose to become the teacher of Pompey the Great and was the first ex-slave to undertake the aristocratic activity of historiography, began his servile career as a doorkeeper chained to his post.469
The trajectory of Pitholaus’ servile career was noteworthy, but advancement from menial household duties to prestigious positions requiring education, though exceptional, was not unparalleled. Nor should we imagine that slaves with job titles always performed only the work associated with those titles, or indeed that any occupational designation fully defined the range of a slave’s work: Trimalchio’s imperious ostiarius shelling peas onto a silver platter may have been intended to illustrate a crass combination of stinginess and extravagance, but the productive exploitation of the idle hours of slaves assigned to simple tasks was undoubtedly common. A papyrus from Egypt happens to reveal to us a weaver by day who baked bread at night, and literary sources confirm the impression that domestic slaves in humbler households were often factotums, performing any and every task a master required. Common labourers (mediastini) are found both in urban and in rural contexts, and the prescriptions of the Roman farming manuals confirm that there was no time off for slaves: if slaves were prevented from performing their regular work assignments, other tasks should be found.470
Roman jurists were much concerned to distinguish occupational designations from the work actually performed by slaves and to determine, in the case of slaves who filled two different roles, which took precedence over the other. The most important general principle they maintained drew a distinction between crafts that required training (and which usually carried specific names) and unskilled work assignments, which might or might not be identified by titles. The first were invariably given precedence, on the grounds that slaves with a skill were more valuable than those without. Already in the first century bc, the jurist Alfenus expressed the view that a trained weaver who had been made into a doorkeeper should be included in a legacy that bequeathed all a testator’s weavers, since he had not been transferred to another craft but merely to another use. So too, conversely, Marcianus at the beginning of the third century ad confirmed that a litter-bearer who became a cook would not be included in a legacy of litter-bearers, because he had replaced an office with a craft, whereas if the situation were reversed (that is, if a slave with a skill were transferred to an unskilled office, the situation imagined by Alfenus), he would go along with the skilled workers. For Paul, on the other hand, if a legacy bequeathed litter-bearers and a slave were serving simultaneously as both litter-bearer and cook, the slave should pass with a legacy of litter-bearers, evidently because no separate disposition specifically concerned cooks. In listing the details that a slave-owner was required to specify for a census declaration, Ulpian accordingly distinguished separate categories of slaves with duties, that is, general work assignments, and those with trades, which required special training. The distinction is borne out in practice by tax documents from Egypt that identify explicitly not only slaves with training in particular trades but those ‘without skills’ (atechnoi). Elsewhere in describing the staff of a rural estate, Ulpian mentions a (male?) slave with the title of‘baker’ but refers generically in the same sentence to ‘the women who bake bread for the slaves’. In this instance, it is unclear whether the distinction depends upon a difference of gender in the workers or of status in the consumers; perhaps the slave who baked for the master earned a title, whereas those who supplied other members of the slave staff did not. A slave’s perspective might countenance no distinction between the slave of a slave and fellow-slave, but Roman jurists and the property-owners they served recognised not only the status of the individual under-slave (vicarius) but the collective concept of ‘the equipment of the equipment’ (instrumentum instruments), the slave staff needed to support the slave labourers on a rural estate: these too, like the field labourers, were inseparable from the property, since their services were necessary to maintain the workforce.471
A second, overarching principle in the legal categorisation of slaves was to distinguish in their descriptive characteristics between a particular type (species) and a general class (genus), and always to give precedence, in matters of controversy, to the former. On this principle, a slave left in a legacy who was both a messenger and a home-born slave would go with the messengers, even if the latter designation marked a more privileged status, because the former was more specific. If one considers the penchant among Roman property-owners for distributing legacies among multiple beneficiaries, one may surmise from examples of this sort a plausible practical explanation for the proliferation of job titles within slave households of the wealthy, independent of any other potential psychological motivations. More difficult to decide were cases in which slaves carried double job-designations of the same class (skilled or unskilled), such as ‘cook and weaver’, ‘book-carrier and chamberlain’, ‘cooper and baggage-handler’, or, in the fictitious case of the clever slave, Massa, of Trimalchio’s friend Habinnas, ‘cobbler, cook, and baker — a jack of all trades’. If the slave had two trained skills and there were conflicting legacies, Marcianus’ view was that the slave should go with the craftsmen with whose trade he was most familiar. Unskilled slaves evidently fell under the general rules for the division of inanimate property. Sometimes the two titles attached to a slave’s name did not mark two occupations, but an occupation and a role, such as ‘stenographer and business agent’ or ‘scene painter and contractor’, in which case the general principle of preferring the specific to the general no doubt applied. Similarly, an ex-slave styled ‘mamma and nurse’ did not perform two functions but one, the former being an affective term, the latter an occupational designation.472
In theory a slave-owner could assign to a slave any job he or she wished, but the law allowed a seller to impose a restrictive covenant prohibiting a new owner from employing a purchased slave for prostitution, and a general principle evolved during the early Empire that slaves should not be assigned to tasks for which their training ill suited them. Ulpian articulates the concept at the beginning of the third century in discussing possible abuses of property inherited in usufruct. The examples he chooses to illustrate the idea are instructive in indicating the parameters of certain types of employment, but more importantly in suggesting what sorts of change in servile assignment were viewed as demotions and would therefore be seen as diminishing the value of the commodity. A usufructuary would be regarded as abusing the property put at his disposal, ‘if, for example, he sends a record clerk into the country and makes him carry a basket of lime, or if he makes an actor do the work of a bath attendant or a musician perform the duties of a butler, or if he takes a man from the wrestling arena and sets him to clean out latrines’. We note a decided preference for jobs in entertainment, which as a class were evidently regarded as more prestigious than those of personal attendance, cleaning or physical labour, but no clear pattern emerges with regard to jobs with specific titles as opposed to those defined by periphrasis.473
In accordance with the fundamental Roman division of work into two categories, farming and other, Romans classified slaves as belonging to either a familia rustica or a familia urbana according to the nature of the tasks they performed, with the former comprising slaves involved in all manner of agricultural operations and the latter embracing all the rest — domestic servants principally but also craftsmen, business agents and clerks. The adjectives allude to the custom among wealthy Romans of maintaining both a townhouse (domus) and villa properties, each with a slave staffappropriate to its functions, but in practice the labels designated types of work rather than the places where work was performed, so that an atriensis, for example, belonged to the familia urbana whether he served in a country house or a town house, whereas an ornamental gardener (topiarius) was part of the familia rustica even if he lived and worked in the city.474
Ultimately, the distinction depended upon the intentions of the owner, as indicated principally by a roster of the urban household staff, but the jurists were much exercised by particulars, because slaves were regularly transferred temporarily from the city to the country, or vice versa, for periods of variable but sometimes considerable duration. A slave might be sent from the city to the country, for example, for training as domestic staff, for upbringing, or for punishment, or from the country to the city for education or an ad hoc assignment. Whether physical removal from an urban residence or a country property marked a step up or down or sideways depended upon the circumstances: relegation ‘to rustication’ marked a demotion in status, but an apprenticeship in the country could raise the skill, and thus the potential value, of an ‘urban’ slave on the way up. A rustic slave might be formally promoted to the familia urbana or temporarily hired out or assigned to non-agricultural work with no change of status. A distinction was further drawn in inheritance law between the household staff of a rural property, which normally went with the estate, and the familia urbana of a town house, which, in Papinian’s opinion, did not. Sabinus defined the servile instrumentum of a farm as comprising those slaves ‘who were accustomed to being quartered there to stay’, but in order to make it clear that slaves of both categories went with a particular property, a testator might need to specify ‘the slaves who will be assigned there, both rustic and urban’. In other words, though formally distinct, the familiae urbana and rustica were permeable categories, and whether a slave belonged to one or the other depended entirely upon the master, who might transfer his human property from one to the other either temporarily or permanently, with positive or negative intent.475
Although the familia rustica provided some slaves with significant managerial responsibilities and relative independence, there is no doubt that agricultural labour was normally more onerous and unpleasant than most duties of the familia urbana and was regularly performed by slaves held in the worst conditions. The deplorable use of chain gangs for farm labour and the housing of agricultural slaves in underground prisons (ergastula) are lamented so frequently by writers of the first century ad that one is tempted to suspect the exaggerations of a rhetorical topos; but the consistently expressed view that urban slaves were too soft for farm work supports the idea, which further accords with the widespread ancient disparagement of country ways, that consignment to the familia rustica was regarded as an inferior assignment. On the other hand, jobs in the familia urbana brought slaves into direct contact with those whose ownership rights imposed no restraints on capricious behaviour and thus exposed them more readily to beatings and other physical humiliations, whereas rural slaves might live and work with relative independence. There is no simple answer to the question of which form of service was preferable.476
Training and career patterns
Within the imperial household we can discern a hierarchy of servile functions and can sometimes trace the course of an individual career, but we can speak of career patterns only in the most general terms and not of anything approaching a regular cursus officiorum. Even in the sub-clerical grades, where the inscriptional record is most full, most of the evidence concerns the post-servile careers of the emperor’s freedmen. Within a private familia, Ulpian could speak high-mindedly of ‘rank and station’ (ordo et dignitas) in discussing the quality of clothing and nourishment a usufructuary was obliged to provide to slaves in his care, but outside the imperial service the evidence for the categorisation of labour reveals more clearly the simultaneous influence of two conflicting principles in determining the careers of slaves in private households: specialisation and job-specification on the one hand, and regular, at times capricious, reassignment to different duties on the other. Within the realm of officia, a slave might move from one job to another over the course of a lifetime for a variety of reasons — for instance as a consequence of age: Seneca’s childhood delicium wound up a superannuated doorkeeper at one of his rural villas. More generally, certain light farm tasks, such as tending fowl or picking clover, were regarded as especially appropriate for old women and children; and in domestic service, the comely dining-room attendants known as glabri were less desirable once their youthful bloom had begun to fade.477
It is reasonable to suppose that the fluidity of transference from one job to another may have been linked to the level of training and education a slave received; but the miscellaneous and anecdotal information we gather from literary and documentary sources suggests that job mobility was possible throughout the system. In the case of litter-bearers and cooks, for example, although the former position was classified as a function and the latter as a craft, there seems to be no clear pattern of progression from one to the other: cooks could become litter-bearers and litter-bearers cooks without apparent system — despite the fact that slaves with trained skills were generally categorised separately, as being more valuable than those without. Training, however acquired, raised the value of a slave and was thus an economic as well as a practical investment. Indeed, the issue was of sufficient importance to educated slave-owners that in the time of Hadrian, one ex-slave, Hermippus of Berytus, wrote a treatise on the subject.478
The methods of training slaves ranged from simple apprenticeships, both formal and informal, to elaborate schooling in the higher arts, as well as in practical trades and crafts. Some slaves, such as the grammarian Remmius Palaemon, acquired marketable skills incidentally on the job, but most were systematically taught to perform the duties to which they were (or would be) assigned. Despite the prescriptions of agricultural writers such as Columella, however, who had much to say about the particular physical and mental qualities best suited for different types of farm work, the quality and level of training slaves received may have had less to do with their perceived abilities and native characteristics than with the investment strategies and educational horizons of their owners. Within the fictitious freedman’s world of Trimalchio’s banquet, the versatile slave Massa — the jack of all trades — had learned his skills from itinerant street-vendors, to whom Habinnas had sent him for schooling; another freedman, Echion, vacillates indifferently between educating his son in law or in a trade — barbering or hawking, or pleading — careers he views as equally rewarding. At the other extreme, the elder Cato, who advocated the sale of old and sick slaves, adopted a characteristically entrepreneurial approach by lending money to his own slaves to buy others, whom they would train for a year and then resell at a profit; Cicero’s friend Atticus maintained a squadron of copyists, each of whom was born and educated in his own household; and M. Licinius Crassus, reputedly the wealthiest man of late Republican Rome, is said to have derived his greatest profits from his household, personally supervised the training of his slaves and even taught them directly himself.479
Apart from simple on-the-job training, and often differing only nominally from it, the most common form of educating slaves, both male and female, was by apprenticeship, which might be undertaken on the initiative of the owner, a buyer, a usufructuary or a creditor. The training itself, which normally had a fixed duration but which sometimes lasted until the skill was mastered, could be contracted either for pay or for work in kind or for a combination of both. Legal sources imply that the hiring out of such contracts was common and attest apprenticeships for artists and secretaries, charioteers and pantomimes, but the training most frequently mentioned as being contracted outside the familia was for artisans and merchants: within that group the most frequently cited craftsmen are weavers (tex-tores), but the jurists also mention slave brickmakers, potters, blacksmiths, millers, bakers, pastry-chefs, innkeepers, tavern masters, clothiers, purple-dyers, wool-spinners, linen-weavers, seamstresses, tailors, textile workers, fullers, launderers, carpenters, construction workers, shipbuilders, plasterers, painters, erectors of statues, cobblers, cup-makers, barbers, morticians, money-changers, and all manner of merchants, including those who sold bread, olive oil, clothing, cloaks and slaves. Not all of the passages specify apprenticeship arrangements, but most imply it; documentary papyri from Egypt confirm the regularity of the practice for weavers and attest individual contracts for a flute-player, a tachygraphist and two hairdressers, and inscriptions from Rome reveal students learning the trades ofmirror-maker and goldsmith.480
Slaves born into wealthier households sometimes received special training in home-schools, paedagogia, that provided formal instruction in reading, writing and arithmetic, and some of which evidently specialised in the arts of seasoning and serving food. In the imperial household of the first two centuries ad, the paedagogia were large, elaborately organised institutions, with instructors and administrative staffs that might include, in addition to teachers, paedagogi, some of whom picked up basic literacy skills from accompanying their freeborn charges to schools, subpaedagogi, master masseurs, slaves in charge of furniture, anointers and hairdresssers. In the year 198, one paedagogium on the Caelian hill at Rome known as ‘The Head of Africa’ (Caput Africae) included twenty-four freedmen instructors and hundreds of pupils. In smaller households slaves might be sent out of the household to acquire litterae serviles — the basic clerical skills needed to conduct business — at informal street schools, and, for more advanced training, to specialist institutions for tachygraphy, accounting and medicine. Physically imposing or athletic slaves might be sent for special training in schools for gladiators or charioteers, or, occasionally, in the more refined activities of the gymnasium. Despite the ecumenical practices of many schools, a rescript of the emperor Domitian (if the text is correctly restored) inscribed near a gymnasium at Pergamum in late ad 93 or 94 shows that medicine and teaching were professions thought by some to be more appropriately reserved for free persons: those who taught them to slaves for pay forfeited their tax-exempt privileges.481
World History
