Slaves and Peasants: Approaches and Problems

One of the most influential attempts so far to make sense of the miscellaneous data referred to above has been Hopkins’ analysis of the causes and effects of the growth of a ‘‘slave society’’ in Italy during the last two centuries of the Republic.8 The starting point of Hopkins’ account was the observation that republican Rome was an uncompromisingly militaristic society in which a large proportion of the adult citizen population was recruited for military service. Initially most of the wars involving the Romans and their allies were fought on Italian soil, allowing peasant soldiers to return to their farms at the end of the fighting season. From about 200 onward, however, the theatre of war moved to Greece, Asia Minor, Africa, and Spain, taking over 100,000 Romans and other Italians overseas. The expansion of large estates was thus encouraged by what was effectively a form of peasant emigration.

Another factor propelling this expansion was the enormously increased income (in the form of booty, for example) generated by the acquisition of an empire. Most of this income flowed into the purses of the Roman elite and a significant part of it was used to buy Italian land, partly because there were few alternative opportunities for investment and partly because landowning conferred social status. Since much of the most desirable land in central and southern Italy had previously been cultivated by free peasants, the inevitable outcome of this development was a drastic decline in their numbers. During the second century free peasants who no longer had any land to work had two main options available to them: they could make a fresh start in a newly founded Roman or Latin colony (although few colonies were founded after 177), or they could try their luck in the expanding city of Rome. Here newly arrived peasant migrants became part of the urban proletariat that constituted the most important market for the surpluses produced on the slave-staffed villae of the rich. In this way the displacement of free peasants further stimulated the production of cash crops on large estates. When the number of impoverished peasants continued to grow during the first century, the Roman government devised a new solution: mass migration to newly established colonies in southern Gaul, Spain, Africa Proconsularis, and other parts of the Mediterranean. In economic terms the effect of this was to make even more land available for occupation by the elite.

One important merit of Hopkins’ account is that it attempts something more than an impressionistic sketch, interpreting the scattered literary data against the background of an overall reconstruction of the history of Italy’s population during the last two centuries of the Republic. Two pieces of ancient evidence are central to this reconstruction: Polybius’ survey of Roman manpower resources in 225 BC and the Augustan census figures. Following the conclusions reached in Brunt’s Italian Manpower, Hopkins assumed that central and southern Italy had some 3 million free inhabitants in 225 BC and that the free population of Cisalpine Gaul was of the order of 1.4 million. This gave the whole of Italy some 4.5 million free inhabitants, of whom more than 400,000 were likely to have been urban. His next step was to interpret the Augustan census figures as suggesting that Italy had roughly 4 million free inhabitants in 28 BC; of these he assigned 1.1 million to cities and towns. These figures, if correct, would mean that over the course of 200 years the rural free population declined from c.4.1 to c.2.9 million, a drop of nearly 30 percent. Meanwhile, Hopkins estimated Italy’s slave population to have increased from c.500,000 in 225 to about 2 million in 28. Of these 2 million early-imperial slaves he assigned 1.2 million to the countryside. These calculations, of course, support Hopkins’ central observation that ‘‘Roman peasant soldiers were fighting for their own displacement.’’9

In view of the elegant coherence of this account of the demographic and agrarian history of late republican Italy, it is not perhaps surprising that much research on this subject from the early 1980s onward could fairly be described as a series of attempts to refine and supplement a theory whose main tenets were generally accepted. One example of this approach is De Neeve’s attempt to reassess certain aspects of Italian agrarian history in the light of Von Thunen’s theory about the location of various types of market-orientated agricultural production.1 This theory relies upon the observation that transport costs are a crucial factor in choosing sites for market-orientated farms. In practice this means not only that higher transport costs mean lower profits, but also that products ‘‘compete’’ for the use of land close to given market locations. Building on these ideas, Von Thunen was able to establish a broad relationship between location and intensity of production in market-orientated farms producing certain cash crops. The geographical pattern that his analysis showed consisted of a series of concentric belts surrounding the central market. In the belt closest to the center, market-orientated farmers tended to specialize in horticulture and intensive dairy farming. The second belt was characterized by market-orientated forestry and the third by arable farming whose intensity decreased as distance from the market increased. In the outmost ring, market-oriented producers concentrated on extensive stock-breeding.

As De Neeve pointed out, the ancient evidence concerning market-orientated production in the last two centuries of the Republic fits this model remarkably well. Cato and other sources tell us that labor-intensive horticulture and intensive breeding (pastio villatica) were practiced near Rome; there is also good evidence for a second belt characterized by slave-staffed ‘‘plantations’’ for the commercial production of wine and olive oil.11 Furthermore, extensive stock-raising took place in peripheral regions such as Apulia and Lucania, exactly as the model predicts. For De Neeve, all this demonstrated that the market-oriented farms of the elite could properly be called ‘‘capitalistic’’ in that their owners or managers were engaged in cash-crop production that took into account the cost of all three ‘‘classical’’ factors of production (land, labor and capital). While this attempt to reinterpret the ancient evidence in the light of modern economic theory was strikingly novel, its results were of course entirely compatible with Hopkins’ theories.

A second new line of inquiry that has enriched our understanding of slave-based agricultural production without altering the overall picture has to do with slave-staffed estates for the commercial production of grain, whose existence is implied by Von Thunen’s distinction between ‘‘intensive’’ and ‘‘extensive’’ ways of growing grain. A publication that has had considerable influence in this area is Spurr’s book on arable cultivation in Roman Italy, which demonstrates that slave-staffed villae cultivating grain could have been profitable.12 Since there is good evidence to suggest that slaves were used in commercial arable farming (e. g., Col. Rust. 2.12.7-9), there can be no doubt that agricultural slavery was not specifically linked to commercial arboriculture. On the other hand, the fact that the three republican and early imperial agronomists do not have much to say about slave-staffed villae whose principal commercial product was grain may perhaps be taken as an indication that agricultural slaves were typically used on estates whose principal cash crops were grapes and olives.

In any case, the existence of slave-staffed grain-growing estates can easily be fitted into the traditional picture of the rise of the slave-staffed villa.

A third interesting development has been the realization that at least some of the features that characterize the agrarian economy of late republican Italy may well have dated back to the third or fourth centuries. Although the literary sources for this period occasionally refer to slave-run estates, the main reason for thinking that their development must have antedated the Second Punic War has to do with the history of debt-bondage. In a penetrating analysis, Finley called attention to the fact that the abolishment of voluntary contracts of debt-bondage in 326 or 313 must have created a demand for an alternative supply of laborers to work the estates of the rich. This demand can only have been met by using slaves.13 At the same time, the sending out of a large number of impoverished peasants to colonies in various parts of Italy must have enabled the rich to build up larger holdings.14 It is therefore highly probable that Italian agriculture was partly dependent upon slave labor well before the first decades of the second century. This convincing theory is, of course, entirely compatible with the overall tenor of Hopkins’ analysis. The idea that the widespread use of slaves in Italian agriculture antedated the Second Punic War is, indeed, implied by his suggestion that there may have been around 500,000 slaves in Italy in 225.

Alongside these investigations into the history of agricultural slavery, another wave of recent publications has focused on the history of so-called transhumant pastoral-ism, a system in which herds or flocks of animals are kept in the lowlands during the cold season and taken to mountain pastures during the summer. Although the amount of ancient evidence relating to this type of pastoralism is not exactly overwhelming, it is striking to find that most of the surviving sources refer to the movement of flocks over long distances. Varro, for instance, turns out to have owned flocks that were driven all the way from northern Apulia to Reate, northeast of Rome (Rust. 2 praef. 6). Since there is good comparative evidence for flocks being driven over similarly long distances in early modern Italy and Spain, there has been a tendency to assume the existence of an almost timeless, specifically Mediterranean form of pastoralism characterized by the ‘‘horizontal’’ movement of animals between widely separated areas.15 In recent years, however, considerable doubt has been cast on the idea that this type of pastoralism represents a logical response to the geographical and climatic conditions prevailing in southern Europe. Garnsey, for example, has emphasized the importance of various political and economic factors such as the unification of Italy, the confiscations that followed the Second Punic War, and the enrichment of Roman magnates during the wars of the second century, while others have identified the growth of Rome as a further precondition.16 It is, in short, becoming increasingly clear that the rise of large-scale ‘‘horizontal’’ transhumance was the product of specific historical circumstances many of which happen to be identical with those behind the rise of slave-staffed villae.

While these new approaches have deepened our understanding of the emergence of slave-staffed estates and large-scale transhumant pastoralism, the fate of the free Italian peasantry has also attracted a considerable amount of attention. Here too some interesting new interpretations have been developed. A notable example is Rathbone’s attempt to question the existence of any direct causal link between the rise of slave-staffed villae and the numerical decline of the free peasantry of central Italy. His reanalysis centered on the coastal town of Cosa, founded as a Latin colony in 273. From a handful of literary data it would appear that Cosa originally had between 3,500 and 5,000 colonists, with their families. Livy (33.24.8-9) reports that the colony was permitted to take on a further 1,000 families in 197, which suggests that by then its original population had declined by about 20 to 30 percent. According to Rathbone, a decline of this order can be accounted for by assuming that on average around 700 adult men from Cosa would have been doing military service at any one time, that the annual casualty rate was just over 7 percent and that when a smallholder was killed on campaign his family would then abandon his allotment in one out of four cases.17 Against this theory it has been pointed out that the annual call-up rate assumed by Rathbone is excessive and that he takes insufficient account of the Roman census figures.18 There are also some grounds for thinking that many rural households were perfectly capable of dealing with the negative consequences of heavy recruitment and high casualty rates. It has been pointed out, for instance, that much of the productive work normally carried out by men could equally well have been done by women, many of whom are likely to have become the de facto heads of their households during the absence of their husbands or sons.19 Moreover, many rural households are likely to have contained an extended multigenerational family or two coresident nuclear families. Such households would have been in a good position to adjust to temporary or permanent changes in their manpower. We cannot, in fact, rule out the possibility that in strictly economic terms the effects of conscription may have been largely positive, since many rural households suffered from a structural labor overcapacity.20 Finally, Sallares has recently argued that the free population of Cosa’s rural territory in particular may well have declined not because of military death rates, but because of malaria (see also Chapter 5).

An alternative version of the theory of population decline stresses the importance of economic factors. One of those to take this approach has been De Neeve, who pointed out that even peasant families whose produce was principally used to supply their own needs must have marketed some of their surpluses.22 In his view, the emergence and spread of market-orientated villae, most deliberately situated so as to permit cost-effective transport, meant increased competition, which led to a further contraction of the free rural population. An important weakness of this line of reasoning is that most of republican Italy’s peasants are likely to have operated their farms according to the principles of what the Russian agronomist A. V. Chayanov called a ‘‘peasant economy.’’ 3 The kind of agricultural production described by this term is carried on by a peasant family whose primary aim is to satisfy the consumption needs of its members. Unlike ‘‘capitalist’’ farmers, the ‘‘peasants’’ within such an economy do not take into account labor as a separate factor of production. The practical significance of this is that peasants are willing to expend ‘‘irrationally’’ large amounts of labor on cultivating their plots in order to secure the family’s basic sustenance. Since the same applies to the effort involved in getting their produce to market, peasants are normally able to undercut the prices charged by highly efficient commercial farmers.24 The lesson to be learned from this is that the free peasants of republican Italy may well have been more resilient than is often assumed, since they were already prepared to accept the huge expenditure of labor necessary to survive on seemingly inadequate plots of land.

The theme of resilience leads us to another aspect of Italian agrarian history that has attracted a considerable amount of attention since the early 1980s. As may be gathered from the preceding pages, authors such as Appian and Plutarch are insistent as to the negative effects of the spread of slave-staffed estates, which they describe as pushing the free peasantry off the land. A close reading of the Roman agronomists reveals this picture to be overly pessimistic. As Peter Garnsey and Dominic Rathbone have demonstrated, Cato, Varro, and Columella are unanimous in assuming that slave-run villae specializing in wine or olives were structurally dependent on outside labor during the harvesting season, for the simple reason that they had to keep the permanent workforce as small as possible in order to turn a satisfactory profit. The main reason why the villa system was more economically efficient than traditional farming was, in fact, ‘‘because it carried no surplus labor... because it exploited the underemployment of the neighbouring free peasantry.’’ In other words, the mere existence of a successful villa system implies the survival of a large pool of peasant smallholders (see also Chapter 28).25

The new approaches that we have discussed have undoubtedly introduced a more realistic view of agrarian change during the second century, but few ancient historians have gone so far as to challenge the prevailing view that the post-Hannibalic period witnessed a rapid growth in the slave population and a corresponding decrease in the number of free country-dwellers.26 The revisionist conclusions reached by a group of archaeologists during the 1970s, however, did exactly that, as did Lo Cascio’s radical reinterpretation of the demographic history of late republican Italy.

The first attempts to use archaeological evidence to investigate the fate of Italy’s free peasantry were inspired by the results of the extensive fieldwalking campaigns carried out by the British School at Rome during the 1950s and 1960s. One of the basic assumptions underlying these campaigns was that the occupational history of entire landscapes could be recovered by identifying concentrations of pottery as the remains of ancient farm buildings and dating them to broad chronological periods. Thus sites containing sherds of grey bucchero were assigned to the fifth or fourth centuries BC, while sites with black-glaze or red-gloss pottery sherds were respectively dated to the last three centuries of the Republic (300-30 Bc) or to the early Empire.27 Applying this method to the ager Veientanus north of Rome, British archaeologists attributed far fewer sites to the late Etruscan period (127) than to the late republican (242), dating even more sites to the early Empire (327). Since the fieldwalking campaigns carried out in neighboring areas produced similar patterns, the inescapable conclusion seemed to be that, in southern Etruria at least, the last three centuries BC had witnessed not a drastic decline in numbers of free country-dwellers, but an unprecedented ‘‘population explosion.’’28

As archaeologists became more aware ofmethodological problems, however, it rapidly became obvious that this scenario of continuous demographic expansion required revision. A reexamination of 600 sherds collected during the fieldwalking campaigns of the 1950s and 1960s appeared to suggest that some 80 percent of the black-glaze pottery recovered in the south Etruria surveys belonged to the fourth to second centuries BC and a mere 20 percent to the period between 200 and 30 BC.29 Could this mean that there really had been a ‘‘crisis of the second century bc’’ after all?

Further difficulties appeared when American and Italian teams carried out surveys in the territory of Cosa. Although literary evidence suggests that between 3,500 and

5,000 families were sent out to this Latin colony (cf. above), only a handful of the rural sites discovered during these campaigns could be assigned to the third century. 0 About 100 appeared to belong to the second century, but even this suggested a very poor recovery rate. This has led some scholars to suggest that many peasants may have lived in farms that were too flimsily built to show up as modern plow-scatters, and others to venture the hypothesis that many smallholders may have been too poor to afford black-glaze pottery.31

Yet another problem has to do with the tendency to interpret the ‘‘small’’ republican sites as representing the farms of subsistence-orientated peasants and the ‘‘large’’ ones as the remains of slave-staffed villae. Here the main difficulty is that most of the ‘‘large’’ republican sites seem to date from the first century. Since this finding does not agree with the literary sources, Rathbone has suggested that many of the ‘‘small’’ sites discovered in the Italian countryside may actually represent modestly sized villae of the Catonian type. A corollary of this reinterpretation is that really big sites such as the villa at Settefinestre are seen as representing new consolidations of property by citizens who gained by the Social War and the Sullan proscriptions.32 In any case, there are strong indications that the history of the villa followed different trajectories in different parts of Italy. It has long been realized, for instance, that the republican villae whose remains were discovered in the ager Cosanus were much larger than those represented by the ‘‘large’’ sites of southern Etruria.33 There are also good grounds for believing that villae remained a phenomenon of limited importance in many inland districts such as the territories of Saturnia and Volaterra.34 Finally, archaeologists have now begun to distance themselves from the traditional assumption that ‘‘large sites’’ (i. e., villae) should invariably be linked to what Marxist archaeologists used to call ‘‘the slave mode of production’’. It is becoming increasingly clear that villae existed in a variety of cultural and chronological contexts within which they had completely different functions. The few villae that have been discovered in the territory of Etruscan Volaterra, for example, seem to have functioned mainly as status symbols and as expressions of elite control over the landscape (see also Chapters 24 and 28).35

The impressive methodological advances that have been achieved in survey archaeology have thus done much to undermine the credibility of earlier claims concerning the spread of slave-staffed estates and the survival or otherwise of subsistence-orientated smallholders. This loss ofinnocence cannot be regarded as a negative development. In fact, the realization that both the literary and the archaeological data are open to various interpretations may well prepare the ground for new inquiries that do not privilege one type of evidence over the other (see also Chapter 28).

Perhaps predictably, the abandonment of earlier attempts to use archaeological evidence to undermine the testimony of the literary sources has not ended the debate over the exact nature of the ‘‘crisis’’ that supposedly prompted the Gracchan land reforms. Just as new forms of archaeological ‘‘source criticism’’ were beginning to gain ground in survey archaeology, a new challenge to the prevailing orthodoxy appeared in the form of the theories of the Italian ancient historian Lo Cascio (see Chapter 14). One of the building blocks of Lo Cascio’s radical reinterpretation oflate republican history is the theory that the Augustan census figures included only adult males. If this is correct, Italy must have had some 12 million free inhabitants in 28 BC, in other words three times as many as has traditionally been believed.36 If there were only 4.5 million free Italians in 225 (cf. above), however, the theory requires an implausible rate of population growth. According to Lo Cascio, the solution to this problem may well be that the Polybian manpower figures included only those adult males aged between 17 and 45 (the so-called iuniores). The population of central and southern Italy would thus have been around 3.5 million and Italy as a whole would have had about 5 million inhabitants.37 If these figures are accepted, the Augustan figure can be accounted for by assuming that the free Italian population expanded at an annual rate of 0.5 percent during the last two centuries of the Republic.

For the purposes of this chapter, the most interesting aspect of this alternative reconstruction is that it is incompatible with the traditional view that the Gracchan land reforms were prompted by worries concerning the number of Roman citizens eligible for the call-up. 8 This explains why Lo Cascio reinterprets Appian’s statement that Tiberius Gracchus was trying to avert the threat of dysandria (usually translated as ‘‘lack of men’’) among the Italian population as referring specifically to the numbers of well-fed men who were physically fit for military service.39 Since a shortage of such men can be interpreted as the inevitable result of population growth without any corresponding increase in the amount of land under cultivation, the Gracchan reforms might be seen as intended to solve a social and military problem rather than to halt a demographic downturn in the free Italian population.40

Although this ingenious theory has impressed some ancient historians, it has also met with severe criticism. It has been pointed out that Lo Cascio’s estimates imply that the population of central and southern Italy (including slaves) was at least 13 million in 28 BC, even though the peninsula had only about 6.3 million inhabitants in 1600. Furthermore, if only 40 percent of Italy’s land surface was cultivated in early imperial times, Lo Cascio’s figures imply an average population density equal to that of the famously fertile Nile valley. Finally, the theory can only be maintained by assuming that the Roman censuses of the second and first centuries were utterly unsuccessful, with no more than one third of adult males being registered by the censors in 70/69. This has led one critic to comment that had Lo Cascio’s figures been correct, ‘‘the census of 70/69 BC would have been a joke rather than a census.’’41 For all these reasons, Lo Cascio’s theory cannot be regarded as a convincing alternative to earlier theories about the nature ofthe agrarian ‘‘crisis’’ that gave rise to the lex Sempronia agraria of 133 (see also Chapter 14).