Download file in PDF format: TRAC 1991: Roman Peasant and rural Organisation in Central Italy: An Archaeological Perspective (pp. 167–186)
ROMAN PEASANTS AND RURAL ORGANISATION IN CENTRAL ITALY: AN ARCHAEOLOGICAL PERSPECTIVE
By Peter van Dommelen
Roman rural landscapes have long been associated with villas, that generally were regarded as its central feature; studies of Roman landscapes were consequently often limited to a villa and its immediate vicinity. This bias in most historical and archaeological research is now increasingly being redressed, mainly by intensive archaeological surveying. The resulting large numbers of small rural sites in particular point to a much more varied Roman countryside with considerably more complex organisation. Concentrating on the rural landscape outside the villas, therefore, the peasantry and their relations with the rural elite are the central issues in a case study of a small valley in Roman northern Etruria.
INTRODUCTION: STUDYING ROMAN RURAL LANDSCAPES
Landscape archaeology is a relatively new development in Mediterranean archaeology, and until recently most archaeological attention has been directed towards the city. Traditionally, in fact, the countryside has been considered to be of hardly any interest at all, because ‘the Mediterranean world is a world of town-dwellers’, as Collingwood forcefully stated. As the urban bias of Classical archaeology is now increasingly being redressed, and the awareness is growing that cities are not isolated features in an otherwise ’empty’ landscape, studying rural landscapes has become a valuable means of understanding wider socio-economic organisation in Antiquity (Snodgrass 1987,67-69; Barker and Uoyd 1991).
Roman landscapes, however, have been less overlooked, as some rural aspects have been studied intensively from a historical perspective. Because of the nature of available literary sources these studies have generally been rather one-sided: agrarian organisation, as part of the ancient economy, has constituted the main theme of the historical debate and other aspects of the rural landscape have been treated as secondary to it. A certain ‘urban bias’ has moreover affected these studies as well, as attention has been focused on the larger villa sites, and consequently no coherent notion of a continuous rural landscape has been developed. Concepts such as the ‘consumer-city’ have furthermore regarded the countryside as a secondary counterpart to the city.
The perspective I want to adopt in this paper is an archaeological one. An important reason for doing so is constituted by the increased prominence of Mediterranean survey archaeology and its achievements during the last decade. First of all, the regional approach of survey archaeology is of course particularly appropriate for studying rural landscapes. As such, it is well suited both to improve traditional archaeological practice and to complement historical research in a fruitful way (cf. Snodgrass 1987, 99-131). Secondly, intensive archaeological surveying has not only redirected archaeological attention to rural landscapes, but has also upset some traditional views: the substantial numbers of small sites that have been discovered everywhere have unequivocally refuted the notion of an ’empty’ landscape (cf. Barker 1991a, 6-7). An approach based on archaeological survey, therefore, may be able to compensate somewhat for the rather elitist perspective inherent in most literary sources (cf. Hodges 1989, 179-186). The theme I intend to pursue regards rural organisation in general rather than agrarian organisation in a more narrowly defined sense; this paper accordingly deals not only with economic aspects of rural organisation but with its social dimension as well. Taking into account small-scale settlement and organisation, I particularly intend to focus on those features of rural organisation that ‘have no history’ but nevertheless represent an essential part of it.
A VILLA LANDSCAPE
Roman landscapes are traditionally considered entirely in terms of villas. Such a view has largely been based on documentary sources dealing with agriculture from a landlord’s point of view and discussing how to run large estates (cf. Garnsey 1979, 1-3). Historical research has consequently treated villas as dominating and determining Roman agrarian organisation. In its wake, extensive topographical work has resulted in more or less ’empty’ landscapes dominated by relatively large villa sites (cf. Leveau 1983, 921-933).
Similarly, Roman agrarian organisation is generally regarded as being both characterised and dominated by the so-called ‘villa system’: large villabased estates using slave-labour to produce specialised crops and agricultural products such as olive oil or wine. The villa buildings, usually divided into residential and working parts, constituted the central and typical feature of the villa system (Carandini and Settis 1979, 46-49). Initially, this system was presented as being rather uniform over large parts of Italy although three variations were proposed. The first, an intensive villa system producing oil and wine, was thought to be typical of the coastal areas and large alluvial plains of western central Italy; the second system, an extensive pastoralist one practicing large scale transhumance, was found mostly in the mountainous regions of central Italy, Apulia and internal Sicily, and the third villa system, geared at the intensive production of cereals, was regarded as typical of Sicily (Carandini 1979, 198-199). All three types, however, are minor variants on an essentially uniform theme of large-scale slave-run estates and as such these have been contrasted with small-scale rural settlement (Carandini 1981,252-253).
Increasing evidence from intensive surveying of small-scale farm sites is resulting in a much more differentiated perception and representation of rural landscapes. Small and medium-sized sites are often generically referred to as smallholders’ farms and are usually assumed to represent small possessions of independent free-holders. Small-scale rural settlement has not only been found to precede the villa system, but has also turned out to have coexisted with it (cf. Potter 1979, 120-139; Celuzza and Regoli 1982, 37-41). Consequently, small-scale farms have often been incorporated in the villa system as a significant but secondary element: it has been suggested that free-holders supplied temporary labour at peaktimes (e.g. harvest), thus allowing the villas to maintain a smaller slave workforce and to avoid surplus labour (Corbier 1981, 428-434). Alternative interpretations play down the dominance of the slave mode of production and propose a wide variety of non-slave labour coexisting with slave-run villas (cf. Garnsey 1980; Foxhall 1990). From such a point of view the relationships between smaller rural sites and villas can be highly variable and the interpretation of small-scale rural settlement itself becomes more problematic.
The coexistence of small farms and large estates cannot be negated and it is now widely recognised that the villa system has never been the most widespread type of agrarian organisation (cf. Pucci 1985, 17). Their relative importance remains to be defined. A recent detailed examination of archaeological evidence in Italy has contributed significantly to these ends (Vallat 1987) and has led to some important refinements of current generalisations: differences in agrarian organisation between various regions sometimes depended on geomorphological characteristics of the area; in other cases historical traditions of land use proved to have had their influence. Even in one region several different villa systems or types of agrarian exploitation might coexist and result in a far more complex agrarian organisation and rural landscape (ibid., 212-213). As a result, the presupposed dominance of the villa system or slave mode of production in Roman agrarian organisation must be rejected; small farms and large estates must no longer be regarded as two different types of agrarian organisation that might coexist and even be interdependent but yet remain distinct and separate. Instead, rural settlement should be treated as one continuous settlement system made up of several elements. Variations in – for example – land use, land quality or pre-existing settlement may have given rise to specific forms of rural organisation in particular areas. Moreover, new ways of agrarian organisation, such as the villa system, may complement rather than replace older forms (ibid., 212).
ANALYSING ROMAN RURAL SETILEMENT
A means for examining rural settlement systems is provided by survey archaeology and regional analysis (cf. Lloyd 1991). Intensive surveying in the lower Biferno valley (Molise, southern Italy) has resulted in a complex rural settlement pattern, in which relatively large and probably slave-run estates are not lacking; these nevertheless are far less lavish than their central Italian counterparts (Lloyd and Barker 1981, 296-301). As small farms, hamlets and villages coexisted with Roman villas, the part played by the villa system must have been limited; the close association of villas with urban centres may suggest a certain importance of cash crops (Barker 1991 b, 46-51). With regard to central Italy and Roman Etruria in particular, archaeological evidence has revealed equally complex settlement patterns even within one single region: the occurrence of villas is highly variable; they show considerable variability in size and other features and are more often than not accompanied by small-scale rural settlement (Vallat 1987, 182-199). While the mere coexistence of several types of agrarian organisation and consequently of different modes of production may be monitored quite closely by archaeological survey and now in fact seems to be beyond doubt (cf. ibid., 199 and 212-213), establishing interdependence or perhaps dominance of any mode of production requires closer study and more detailed analysis.
Previous attempts to deal with the interpretation of rural settlement usually have set out by classifying the raw field survey data. An important criterion adopted has been site size but the relevance of other features such as durable or expensive building materials and fine wares was also soon realised (cf. Celuzza and Regoli 1982, 56-57). The problem of identifying various types of rural settlement has been difficult to resolve; different classes of rural settlement have been proposed but the applicability of each classification seems to be limited to each specific situation (cf. Vallat 1989, 113-114). Perhaps the only useful categorisation is the string villa, casa and tugurium (corresponding to large, medium and small farms: Potter 1979, 122) but even this generic division has only relative significance in specific local contexts, as a comparison of the ‘villas’ in the Biferno valley with those in South Etruria shows. As any fixed relationship between morphological features of rural sites and their function seems problematic, if not improbable (cf. Slofstra 1983, 84), sites should be interpreted in relation to the whole of rural settlement in an area, using various criteria including not only building materials or associated ceramics, but also location (cf. Vallat 1989,114-116).
Understanding rural settlement, therefore, should not be focused on any site in particular, whether a villa or a cottage, but should start analysing the functioning of an entire settlement system in a regional analysis (cf. Johnson 1977). An important criterion is site size, as this is related to ‘functional size’, i.e. the number of functions exerted by a certain site. A site possessing a large number of ‘functions’ is usually referred to as a central place and the larger a site is, the more central it is supposed to have been Oohnson 1977,495). The implication is that most sites are dependent on such places for one or more ‘functions’ in the form of services or items; the relationships between settlements might therefore be described in terms of ‘dependency’ and ‘independency’. By ranking all sites according to site size, such relationships can be represented graphically. A second basic criterion is site location. The significance of this reveals itself when both criteria are applied jointly by classifying the mapped sites after the rank size graph. Although such a classification remains essentially arbitrary, the combination of both features allows the identification of some basic characteristics of the settlement system (Fig. 27).
Rural exploitation systems might similarly be distinguished, since perhaps one of the most distinctive features of the slave-based villa system is the virtual absence of any rural settlement in a considerable territory surrounding the villa site. That area is assumed to have been worked by slaves housed in the central villa (Celuzza and Regoli 1985, 51-53; cf. Carandini 1988, 121-129). This phenomenon is interpreted as ‘a drastic alienation of the producers from the means of production (land)’ (Carandini 1981, 250; cf. Carandini 1979, 151). A situation where, on the contrary, most of the area is occupied by numerous small sites may be interpreted as one in which the productive means are not alienated from the producers. If a surplus value is somehow extracted, this is usually achieved through imbalanced exchange between the producers and a non-producing elite: control of some critical resource – not necessarily land but possibly commodities such as imported objects or salt – enables the elite to obtain the producers’surplus in exchange for access to that scarce asset (cf. Amin 1983). The elite thus ensures their privileged economic position by control of exchange and distribution rather than production (Smith 1976, 311-312).
These make up distribution networks that, in the Roman case, are organised through primitive or peasant markets that are generally associated with central places. Typically, such central places are closely related with the rural elite (ibid., 313-321; cf. Hodges 1988, 127-131). The functioning of rural settlement, therefore, may be understood in terms of these distribution networks; the latter, furthermore, exhibit spatial characteristics that have been crudely summarised in a number of descriptive models (see Smith 1976, 333-338). An analysis, therefore, of a region’s settlement system and distribution networks may lead to a broad outline of regional economic organisation and overall socio-economic structures, in which these networks are firmly embedded.
PEASANT CULTIVATION AND ROMAN RURAL ORGANISATION
Taking the term ‘rural organisation’ as a reference to the socio-economic dimension of the Roman countryside, the anthropological concept of ‘peasantry’ seems to be particularly applicable: since Roman society was a complex state society with clear-cut social strata or ‘classes’, the vast majority of the rural population lived and worked in a position of economic and political subordination that neatly fits Eric Wolfs principal criteria of dependency and integration (1966, 11):
“Thus, it is only when a cultivator is integrated into a society with a state – that is, when the cultivator becomes subject to the demands and sanctions of power-holders outside his social stratum – that we can appropriately speak of peasantry.”
Peasantry as a concept is remarkably difficult to define, but as a first approximation peasants might be described as (Shanin 1987, 3):
“small agricultural producers, who, with the help of simple equipment and the labour of their families, produce mostly for their own consumption, direct or indirect, and for the fulfilment of obligations to the holders of political and economic power.”
Although this definition admittedly covers only part of peasant variability and heterogeneity, it does stress the fundamental interwovenness of peasant economic and social organisation (cf. Shanin 1987, 3-4): particularly in agrarian, non-industrial economies social rather than purely economic motives play a significant, if not decisive part in agrarian decision making. With regard to the Roman agrarian economy, the relevance of a ‘peasant economic’ rationality has recently been argued for, as opposed to the primitivist point of view denying any economic rationality in agrarian production and the ‘bi-sectorial’ perspective combining an ‘irrational’ subsistence sector with capitalist-like, maximising principles (Foxhall 1990, 98-104).
From this point of view, a fundamental theme regards the relationships between two basic social groups: how did peasants relate to landlords and vice-versa? This relationship essentially has been one of economic exploitation of peasants by a landlord in which peasants could occupy several different positions, ranging from owner-occupants, tenants, share-croppers to slaves l (Garnsey 1980, 36-41). The efficiency of the various exploitation systems has been much discussed and from a purely economic point of view each system has its own advantages, albeit mainly for the landlord. Nevertheless, comparative evidence tends to favour tenancy and sharecropping, as, at least in non-industrial economies, large possessions can be exploited most efficiently in small units (Martinez-Alier 1983). The main differences are non-economic and consist of social or legal obligations of peasants or landlords. Again, tenancy and share-cropping offer most advantages to both peasants and landlords as such relationships must often have coincided with debt-bondage or patronage (Foxhall 1990, 101-104 and 111-112); given the extensive patronage networks existing in Roman society, a close correlation of such relationships with economic ones seems very likely (cf. Slofstra 1983, 89-95). Nevertheless, exploitative systems obviously have coexisted, resulting in complex, varied and often fluid relationships (cf. Pitt-Rivers 1971, 36-46). In this way peasant cultivators and landlords presumably have often been linked into a single system of both economic and social interdependence.
In the archaeological record economic aspects of regional rural organisation can be examined in rural settlement systems and distribution networks as described above.2 Such an analysis, however, only regards the more generic, overall trends within a region; it may point to some relationship of dependency between certain areas or even sites, but it cannot ‘identify’ this particular relationship as a share-cropping agreement and that one as a case of slavery. This difficulty to interpret any specific relationship between sites largely arises from the fact that most of the variability of exploitative relationships is social in nature. Therefore, other features in the archaeological record must be taken into consideration as well: besides site size and location, the range and combinations of objects found at a site are particularly relevant in this respect. Thus it may be possible to argue for a more detailed interpretation of specific relationships between rural sites or areas: for example the contrast of high-quality masonry and agricultural equipment with poor, coarse pottery found in some small farm sites as well as their remote location have been proposed as arguments to identify these sites as habitations of share-croppers (Foxhall 1990, 109-111).
PEASANTS AND LANDLORDS IN ROMAN NORTHERN ETRURIA
A small coastal plain in northern Etruria (Figs 27-29), in the region nowadays referred to as the Maremma toscana, has provided the setting for studying a Roman rural landscape. Naturally defined by the watersheds of the lower river Pecora – sometimes quite steep hills heavily wooded with macchia – the area is well suited to examine regional and local peasant-landlord relationships. As the steep and high hills south of the plain extend westward into the sea to form a small promontory and thus effectively block all easy access from the south, the coastal plain can be reached most easily either from the north where the hills are separated from the sea by a narrow strip of dunes and lowland or from the east along the upstream Pecora. To the west, the coastal plain originally was delimited by a now drained lagoon, the extent of which in Roman times is known only approximately.
In the years 1980-82 an archaeological survey was carried out in the lower Pecora catchment as part of the Scarlino archaeological project of the University of Siena (Cucini 1985, 147-150). The survey was concentrated on the coastal plain, part of which was surveyed intensively, while other samples were surveyed more extensively; the intensively surveyed area, referred to as the Scarlino area, roughly coincides with the southern foot-hills and is more or less naturally bounded by the lagoon shore and the southern hills (Fig. 28, number 1). As far as the earlier Roman period is concerned, the Scarlino survey has registered 113 sites in the coastal plain of the Pecora valley; of these, 27 sites have been dated in the Republican phase (more or less 3rd-2nd century BC) and 26 in the late Republican/ early Imperial phase (more or less 1st century Be – 1st century AD), while the remaining 60 sites could be assigned only generically to the Roman period.3 Far from being useless, these ‘generically Roman’ sites represent a significant feature of the valley’s settlement system, as the rank size graph shows (Fig. 27): while the sites dated to either of the phases represent more or less the same but incomplete part of the rank size distributions, the category of generically Roman sites roughly supplements their missing lower parts. These sites, therefore, effectively represent the lower part of the settlement hierarchy, the ‘small-scale rural settlement’. The difficulty of dating these sites more precisely is a reflection of this, as such datings are based on fine wares that are usually lacking in small farm sites (cf. Uoyd and Barker 1981, 296). The generically Roman sites, therefore, can and must be used in an analysis of the settlement system, although none of the sites can be assigned to any specific phase.
Figure 27. Lower Pecora, northern Etruria, Italy survey area: sites mapped by size (after van Dommelen, drawn by Sandra Hooper).
Figure 28. Lower Pecora, northern Etruria, Italy (after van Dommeien, drawn by Sandra Hooper).
Figure 29. Lower Pecora, northern Elruria, Italy (after van Dommelen, drawn by Sandra Hooper).
Taking into account site size only, the sites may crudely be grouped into four classes, as the form of the rank size-graph suggests breaking points at 5,000, 1,000 and 400 square metres (Fig. 27). Because the largest sites that traditionally would be labelled ‘villas’ do not occur in the Pecora valley before the 1st century BC (cf. Cucini 1985, 290), and villas in general have not been attested in central Italy before that date (Vallat 1987, 198-199), this analysis will be confined to the late Republican/ early Imperial period (1st century BC – 1st century AD). It should be noted, furthermore, that extensive surveying has generally failed to register smaller sites, as can be observed on the distribution map; but this obviously does not imply that these sites were lacking in other parts of the Pecora valley. As the Scarlino area can be considered a representative part of the coastal plain, the site hierarchy constructed for this area may be taken as valid for the whole of the lower Pecora basin. Although the lacking lower end of the site hierarchy reduces a plot of site categories to a mere distribution of major sites, it still is informative, as these sites in particular playa central part in the area’s settlement system.
The settlement pattern in the lower Pecora catchment during the late Republican/ early Imperial phase is characterised by a remarkably biased distribution of the major ‘villa’ sites (class 1). This distribution can hardly be a consequence of uneven covering, as even extensive surveying is unlikely to overlook sites of this size. Two major settlement nuclei can be distinguished that are both located in the Scarlino area, one close to the sea and the other one more inland, while another ‘villa’ site has been found in a more secluded position on the far north-eastern foothills. Sites that might be termed ‘large farms’ (class 2) have been found allover the coastal plain and are generally situated at some distance from the larger sites. Only one of these can be associated with a ‘villa’ site. Four out of the ten ‘large farm’sites, however, cannot be assigned to the late Republican/early Imperial phase with certainty nor can the possibility be excluded that some of these sites have been missed by the extensive survey. The remainder of small sites have mainly been found in the Scarlino area for the reason mentioned; they usually seem to be associated with larger sites.
The settlement cluster on the coast near modern Puntone consists of two ‘villa’ sites that must have been separated from each other by the lagoon outlet. The southern site is accompanied by a ‘large farm’ as well as by several medium-sized and small sites. The northern one, on the contrary, is closely related to extensive metallurgical activities: huge slag heaps as well as the remains of at least one furnace, all positively dated in the late Republican/ early Imperial phase, illustrate the enormous quantities of iron are that were transported from the island of Elba and smelted near Puntone. This settlement cluster must have been situated on the ancient Via Aurelia, a major Roman through way along the Italian west coast, and is usually interpreted as a statio or road station, probably the one named Manliana.
The Via Aurelia must have continued its way northward either around the lagoon or, after crossing the lagoon outlet at Puntone, over the dunes of the coastal barrier (Cucini 1985, 298-300). More inland, where circumstances are most favourable for agriculture, a ‘villa’ site has been located at La Pieve. This site is by far the largest in the Pecora valley, having more than twice the size of other ‘villa’ sites. A more modest but still significant ‘villa’ site has been found at nearby Casa il Pino and in the immediate surroundings several smaller sites have been located. At a short distance to the south a number of medium-sized and small sites have been found to cluster on the lower hill slopes. The ‘villa’ site at Casa Valmora is the only one outside the Scarlino area; it may very well be accompanied by several smaller sites but presently only one has been found. The location in the north-eastern corner of the coastal plain is a strategic one, as both the northern foothills and the upper Pecora valley are dominated by this large site.
INTERPRETING A ROMAN RURAL LANDSCAPE
Late Republican/early Imperial settlement in the lower Pecora valley can be interpreted from a regional point of view in terms of central places and distribution networks (cf. Uoyd 1991, 233-236; Gualtieri and de Polignac 1991). Four central places can thus be distinguished in the study area but each with slightly different characteristics: given its location at the Via Aurelia and the likely presence of a port, as well as the metallurgical activities, the statio at Puntone undoubtedly represents a major first-order central place with substantial external contacts. The large villa at La Pieve must have been an equally important central place of significance to the whole of the lower Pecora basin. Although agriculture presumably had a considerable share in the villa’s economy, it seems likely that most of the local agrarian activities were administered through the villa at Casa il Pino. This must have functioned as a second-order central place for most of the fertile south-eastern area. The villa at Casa Valmora similarly must have been a local central place for the upper Pecora valley and perhaps part of the northern coastal plain, although it may have been less directly subordinate to the villa at La Pieve.
In terms of economic organisation, the lower Pecora valley in late Republican and early Imperial times was operated by a so-called ‘dendritic’ central place at Puntone (cf. Smith 1976, 338-345), whereas internal distribution in the valley was administered by a ‘solar’ central place at La Pieve (cf. ibid., 345-352). Both Casa il Pino and Casa Valmora presumably were subordinate local central places. All four must have represented focal points of both economic and socio-political organisation in the lower Pecora basin where peasant surpluses were – usually unevenly – exchanged for access to land or other assets. Such exchange could take the form of either direct payment of rent or indirect contribution through a controlled periodic market (ibid., 333-338). These acts simultaneously reconfirmed and thus reproduced the peasants’ subordination to the elite; these in turn stressed their dominating role at the central places and in the entire valley by ostentious display of wealth.
The proximity of a first and second-order central place in a small, relatively secluded part of the valley suggests an uneven economic development of the entire lower Pecora area. In the southern part of the coastal plain, roughly coinciding with the Scarlino area, a rigid site hierarchy developed, indicating an increasingly vertical structuring of socio-economic relations that cannot be paralleUed elsewhere in the Pecora valley. Although agriculture was practised in the entire basin, the southern part seems to have benefited most. Non-agrarian economic activities on the other hand, such as metallurgy or exchange, all seem to have been limited to the Scarlino area where nearly all indications found are concentrated. Signs of wealth such as marble fragments, mosaic tesserae or remains of baths are also practically limited to this area, while elsewhere only modest farms have been found. The southern part of the coastal plain must have occupied an exceptional position in the lower Pecora basin. Both the uneven distribution of wealth and the regional division of labour suggest that the south of the valley was linked to the surrounding areas in the basin by a local centreperiphery relationship (Rowlands 1987, 4-5).
Although the presence of villas in the lower Pecora catchment is evident, the existence of a slave-based villa exploitation system needs more and better arguments. In this particular case, a predominantly slave-based agrarian and economic organisation seems doubtful for several reasons. Firstly, numerous medium-sized and small sites were scattered allover the Pecora valley, resulting in a densely settled landscape of which the villas represented just one aspect; the small farms must have had their own separate agricultural production. It is also significant that the villas had not replaced earlier dispersed settlement but seem to represent an intrusive element in the local economy (see Fig. 27; cf. Cambi and Fentress 1989, 81). Not all the villas fit the model of a mainly agricultural settlement either; as the above analysis indicates, distribution and exchange must have constituted important activities of these villas: those at Puntone in particular seem to have had little in common with a ‘typical’ slave-run villa (cf. Carandini 1988, 19-108). Slaves nevertheless must have played some part in these villas and the Pecora valley as a whole: the metallurgical activities in particular are likely to have been dependent on slave labour. At the villas too, slaves undoubtedly were used for agricultural and related ends; this practice, however, must have remained limited to the estates proper, since slavery never seems to have become so important or dominant as to influence the valley’S agrarian economic organisation significantly (cf. Cucini 1985, 297). Therefore, instead of the means of production being alienated (cf. above), it must rather have been a surplus that was extracted from a free but dependent peasantry. The valley’s settlement system suggests this was achieved both directly and indirectly, with the villas in some cases functioning as a (periodic) market-place and in others as a landlord’s estate with dependent tenants. In either case, peasants must have been farming primarily for subsistence.
The above regional analysis has resulted in a ‘macro-economic’ review that merely indicates more general dependencies. Taking this analysis as a point of departure, the rather processualist representation of the valley’s socio-economic organisation can be refined, if particular cases are examined in their specific contexts. One such a specific case is constituted by the settlement at Puntone, where the villas have been suggested to represent a peculiar case.
Interpreting the settlement cluster at Puntone as a first-order central place and a centre of metallurgical production that was agriculturally non-productive has several implications for the smaller sites near the villas. It should be noted that the immediate vicinity does not permit agriculture at any scale. The smaller sites that may be associated with Puntone all cluster in the southern end of the southern foothills, where the nearest good agricultural soils are found. These are likely to have satisfied the settlement’s agricultural needs, since its inhabitants – not only the villa-elite but the labourers and slaves as well – must have been dependent on agrarian production elsewhere. It seems improbable that these small farms were inhabited by slaves, as the attention of the villa-elite was focused on other activities; these farms should therefore be interpreted as peasant farms. The archaeological remains of these farm sites are invariably poor, consisting of fragments of amphorae, coarse wares and usually roof tiles; they are most of all in sharp contrast with the lavish pars urbana of the two villas. The exact nature of this dependent relationship is of course difficult to establish, but the relatively large number of sites in a limited area suggests one of share-Hopping, as this is the most common and effective exploitation system in such a situation (Martinez-Alier 1983). The presence of a slightly larger farm may also be significant as a collective processing place (Gill 1983, 146-148). The combination of roof tiles and coarse ceramics might also point at a share-cropping arrangement with the landlord providing the house ( cf. Foxhall 1990, 109-111)_ The proximity of a population concentration at Puntone might even have encouraged cash-cropping on the smaller plots of land (Wolf 1966, 36; Foxhall 1990, 105-106). Sharecropping arrangements would be the best way of providing the settlement with a constant production, without the necessity of close supervision. Although share-cropping arrangements usually are complex and variable, they are likely to have been complicated even more by day labourers who are hard to trace archaeologically.
Small rural sites largely similar to the ones just discussed have been found in the northern half of the coastal plain; if their relatively remote location is taken into consideration, however, a slightly different situation seems to emerge. Archaeologically, these smaller sites are also different from the ones near Puntone because roof tiles are generally lacking: these sites should therefore represent even poorer farms. As the area closer to the river was best suited for grazing, animal husbandry presumably was more important at these farms. Agriculture could be practised on the aUuvial fans immediately to the north. The area north of the river Pecora where these sites are located is relatively distant from any of the central places and direct exploitation of these farms from there seems somewhat implausible. The relationships between these small farms and presumably the large estate at La Pieve may therefore have consisted of tenancy arrangements, with rents paid in produce for example (cf. Pitt-Rivers 1971, 44-45). The peasants can also have been owner-occupants who were somehow tied to the large estates: such ties may be socio-political or commercial, the estate functioning as a periodic market (Wolf 1966, 49-59). In both cases, however, peasants were subsistence farmers dependent on a landlord. Although ownership may guarantee some form of economic independence, this often can be annulled by relationships of social dependence, such as patronage or debt-bondage (ibid., 86-87). The effective difference between ownership and tenancy can thus be reduced, since tenancy offers certain advantages to peasants in marginal situations: a combined tenancypatronage relationship guarantees both a ftxed price and a protector (Foxhall 1990, 111-112). Such arguments probably apply to the northern half of the Pecora basin that has been argued to be peripheral to the southern part. The scanty archaeological evidence of these farms ftts this interpretation also quite well, since peasants were responsible themselves for housing and agricultural equipment, although complicated arrangements of cooperation often exist (Wolf 1966, 85-86).
Yet another case of small rural sites has been found upstream the river Pecora at Campo Ruffaldino. Where the river valley narrows, ftve small farms are clustered closely together on a river terrace. In the valley itself farming is possible on a limited scale, and the surroundings hills allow growing olives and vines. Shepherding must also have been a profitable activity of the farms’ inhabitants. These farms differ from the previous ones because of the relatively rich ftnds: at each site several fragments of ftne wares – sigillata italica – and amphorae have been found as well as roof tiles. An interpretation as a small nucleated village or hamlet seems plausible; such hamlets are usually referred to as a virus (cf. Garnsey 1979, 6). The relative well-being of these peasants and their outlying residence suggest a certain degree of independence: they most likely owned the small plots of land along the Pecora that they farmed essentially for subsistence. Their autonomy, however, must have been limited by the villa at Casa Valmora as the local periodic market place on which these peasants were reliant.
CONCLUSION: RURAL ORGANISATION WITHOUT HISTORY
The speciftc cases just discussed constitute an illustration of the points made earlier in this paper, since the variability and, most of all, the complexity of small-scale rural settlement in the lower Pecora valley are obvious. With regard to the villa discussion, these examples conftrm the conclusions reached by Vallat in his review of Italian rural organisation: villas are highly complex and variable (Vall at 1987). In the case of the Pecora valley, the principal archaeological similarity between the ‘villa’ sites is the presence of a pars urbana: in all cases the inhabitants’ wealth was openly displayed, while the way in which this prosperity was achieved may have varied considerably. The Pecora case study furthermore offers a good illustration of the interdependence of villas and small-scale settlement: these relationships have been distinguished not only at a local level (e.g. at Puntone) but regionally as well. Whereas the part played by a villa in an agrarian economy can largely be studied by analysing regional archaeological distributions, understanding agrarian organisation in all its dimensions requires a more detailed examination: from a processualist point of view, small-scale rural settlement may seem to fulfil similar functions, but a consideration of specific contexts results in a more detailed and varied representation.
The adoption of such a perspective in Roman studies may focus attention on the peasantry that has turned out to constitute an integral and essential feature of rural landscapes. The examples from the Pecora valley demonstrate that Roman rural organisation cannot be well understood if the peasantry is not considered (cf. Hodges 1989, 180). An approach to rural landscapes that appreciates the part played by the peasantry must be an archaeological one of necessity; it will, however, rarely be a definitive one. Even this limited case study has already demonstrated that no easy criteria can be formulated for distinguishing various kinds of small-scale rural settlement, since most variability in agrarian organisation is of a socio-economic nature and often archaeologically difficult to trace; alternative explanations may be possible. But such a discussion already represents a contribution towards constructing the peasantry’s part in Roman history.
Acknowledgements. I would like to thank Dr Pieter van de Velde for his
helpful comments on an earlier draft of this paper.
1. For the vague and often fluid distinction between slaves and free but otherwise dependent tenants or share-croppers (see Garnsey 1980, 34).
2. This holds in particular for regions with dispersed rural settlement as in the following case study; in case of more nucleated settlement – e.g. agro-towns – more attention should be paid to intra-site patterns within these sites.
3. All data concerning sites and the ScarIino survey in general are based on the survey publication (Cucini 1985).
Amin, S. 1983 ‘Economic and cultural dependence’ in T. Asad and R. Owen (eds),54-60.
Asad, T. and R. Owen (eds), 1983 Sociology of ‘Developing Countries’: the Middle East. London.
Barker, G. 1991a ‘Approaches to archaeological survey’ in G. Barker and]. Lloyd (eds), 1-9.
Barker, G. 1991b ‘Two Italys, one valley: an annaliste perspective’ in J. Bintliff (ed.), 34-56.
Barker, G. and R. Hodges (eds) 1981 Archaeology and Roman Society. Oxford; British Archaeological Reports (International Series 102).
Barker, G. and J. Lloyd (eds) 1991 Roman landscapes. Archaeological survey in the Mediterranean region. London.
Bintliff, J. (ed.) 1991 The Annates School and Archaeology. Leicester.
Brandt, R. and J. Slofstra (eds) 1983 Roman and Native in the Low Countries. Oxford; British Archaeological Reports (International Series 184).
Cambi, F. and E. Fentress 1989 ‘Villas to castles: first millennium A.D. demography in the Albegna Valley’ in K. Randsborg (ed.), 74-86.
Carandini, A. 1979 L’anatomia della scimmia. Rome.
Carandini, A. 1981 ‘Svilupp e crisi delle manifatture rurali e urbane’ in A. Giardina and A. Schiavone (eds) stet., 249-260.
Carandini, A. 1988 Schiavi in Italia. Rome.
Carandini, A. and A. Ricci (eds) 1985 Settefinestre. Una villa schiavista neU’Etruria romana. Vol I. Modena.
Carandini, A. and S. Settis 1979 Schiavi e padroni nell’Etruria romana. Bari.
Celuzza, M. and E. Regoli 1982 ‘La valle d’Oro nel territorio di Cosa; Ager Cosanus eAger Veientanus a confronto’ Dialoghi di Archeologi 4,31-62.
Celuzza, M. and E. Regoli 1985 ‘Gli insediamenti della Valle d’Oro e il fondo di Settefinestre’ in A. Carandini and A. Ricci (eds), 50-61.
Corbier, M. 1981 ‘Proprieta e gestione della terra: grande proprieta fondiaria ed economia contadina’ in A. Giardina and A. Schiavone (eds) 1981a, 427-444.
Cucini, C. 1985 ‘Topografia del territorio delle valli del Pecora e dell ‘Alma’in R. Francovich (ed.), 147-335.
Foxhall, L. 1990 ‘The dependent tenant: land leasing and labour in Italy and Greece’Journal of Roman Studies 80, 97-114.
Francovitch, R. (ed.) 1985 Scarlino I, storia e territorio. Florence.
Garnsey, P. 1979 ‘Where did Italian peasants live?’ Proceedings of the Cambridge Philological Society 205, 1-25.
Garnsey, P. (ed.) 1980 Non-Slave Labour in the Greco-Roman World. Cambridge Philological Society Supplementary Volume 6. Cambridge.
Garnsey, P. 1980 ‘Non-slave labour in the Greco-Roman world’ in P. Garnsey (ed.), 34-47.
Giardina, A. and A. Schiavone (eds) 1981a L’Italia: insediamenti e forme economiche. Societa romana e produzeione schiavista vol 1 Rome.
Giardina, A. and A. Schiavone (eds) 1981b Merci, mercate e scambi nel Mediterraneo. Societa romana e produzeione schiavista vol 2 Rome.
Gualtieri, M. and F. de Polignac 1991 ‘A rural landscape in western Lucania’in G. Barker andJ. Lloyd (eds), 194-203.
Hodges, R. 1988 Primitive and Peasant Markets. Oxford.
Hodges, R. 1989 ‘Archaeology and the class struggle in the first millennium AD’ in K. Randsborg (ed.), 62-73.
Johnson, G. 1977 ‘Aspects of regional analysis in archaeology’ Annual Review of Anthropolgy 6, 479-508.
Leveau, P. 1983 ‘La ville antique et l’organisation de l’espace rurale: villa, ville, village’ Annales ESC 38 (4),920-942.
Lloyd, J. 1991 ‘Forms of rural settlement in the early Roman empire’ in G. Barker andJ. Lloyd (eds), 233-240.
Lloyd, J. and G. Barker 1981 ‘Rural settlement in Roman Molise: problems of archaeological survey’ in G. Barker and R. Hodges (eds), 289-304.
Martinez-Alier, J. 1983 ‘Sharecropping: some illustrations’ Journal of Peasant Studies 10 (2-3), 94-106.
Pailler,J.-M. (ed.) 1989 Actualitl de l’antiquite. Toulouse.
Pitt-Rivers, J. 1971 The People of the Sierra (2nd edition). Chicago.
Potter, T. 1979 The Changing Landscape of South Etruria. London.
Pucci, G. 1985 ‘Schiavitu romana nelle campagne’ in A. Carandini and A. Ricci (eds) vol I, 15-21.
Randsborg, K. (ed.) 1989 The Birth of Europe. Archaeology and Social Development in the First Millennium AD Analecta Romana Instituti Danici, Supplementum 16. Rome.
Rowlands, M. 1987 ‘Centre and periphery. A review of a concept’ in M. Rowlands, M. Larsen and K. Kristiansen (eds), 1-11.
Rowlands, M. , M. Larsen and K. Kristiansen (eds) 1987 Centre and Periphery in the Ancient World. Cambridge.
Shanin, T. 1987 ‘Introduction: peasantry as a concept’ in T. Shan in (ed.), 1-11.
Shanin, T. (ed.) 1987 Peasants and Peasant Societies (2nd edition). Oxford.
Slofstra, J. 1983 ‘An anthropological approach to the study of Romanisation processes’ in R. Brandt andJ. Slofstra (eds), 71-104.
Smith, C. 1976 ‘Exchange systems and the spatial distribution of elites: the organisation of agrarian societies’ in C. Smith (ed.), 309-74.
Smith, C. (ed.) 1976 Regional Analysis vol II New York.
Snodgrass, A. 1987 An Archaeology of Greece. Berkeley.
Vallat, J.-P. 1987 ‘Les structures agraires de l’Italie re publicaine’ Annales ESC 42 (1),181-218.
Vallat, J.-P. 1989 ‘De la prospection a la synthese d’histoire rurale’ in J.-M. Pailler (ed.), 101-127.
Wolf, E. 1966 Peasants. Englewood Cliffs.