Progetto Tiber Valley. La Sabina e le due rive a confronto

Abstract

An English translation of the section by R. Witcher: Central to the Tiber Valley Project has been the collation and integration of the results of over 100 years of landscape research and the development of new fieldwork projects to improve our understanding of specific themes and problems. The combination of archive work, restudy of existing collections and targeted or problem-oriented fieldwork is becoming more common as new methodological techniques and research questions arise for landscapes which already have long histories of archaeological study (see papers in Attema et al. 2002). The middle Tiber valley, through its proximity to Rome, has a particularly long and distinguished history of research and can be reasonably argued to be one of the most intensively surveyed landscapes in the Mediterranean. As such, the quantities of data are enormous; here we’ll concentrate on just one smaller area of the middle valley, the Sabina Tiberina, and the sequence of archaeological surveys conducted in the area over more than a century. In the Sabina, survey has tended to focus tightly along the line of the Tiber, rarely extending more than five or six kilometres to the east. Various reasons for this include archaeological interest in ‘historical’ centres which concentrate closer to the Tiber (e.g. Campo del Pozzo, Magliano Sabina, Cures Sabini) and differences in modern land use; the area closer to the river is dominated by arable and olive cultivation which is better suited to field survey techniques than the forest and exposed geology which dominate further east in the hilly interior. The first large-scale survey of the area formed part of a broader project to map archaeological evidence across southern Etruria and Sabina during the 1880s (Gamurrini et al. 1972). The emphasis was upon establishing the courses of ancient roads and listing standing or rock-cut monuments (e.g. mausolea, chamber tombs, villa platforms). Relatively few scatters of surface material were noted and overall site density was relatively low and strongly skewed towards funerary evidence and prominent structures. Nonetheless, the survey forms an essential framework for all subsequent fieldwork in the area. Ashby and other topographers worked particularly in the southern part of the area during the early twentieth century, again concentrating primarily on roads and associated sites, though with increasing attention to surface scatters (e.g. Ashby 1906). The basic field technique remained largely unchanged until the 1960s and a survey focussed around the pre-Roman centre of Eretum (Ogilvie 1965). The survey formed an ‘outpost’, on the east bank of the Tiber, of the British School at Rome’s South Etruria Survey commenced in the 1950s (Potter 1979). The Eretum survey, conducted by Robert Ogilvie, used the techniques successfully applied in South Etruria to identify a dense distribution of artefact scatters, many producing evidence to suggest they were both luxurious residential villas (e.g. marble veneers, architectural fragments) and centres of production (especially bricks and pottery). High site density, especially villas, was also recognized by Maria Pia Muzzioli’s survey of the area around Cures Sabini during the 1970s (Muzzioli 1980). As part of the Forma Italiae series, Muzzioli’s survey encompassed the entire Montopoli in Sabina 1: 25000 mapsheet (c.100km2), revisiting sites mapped by Gammurini et al. and adding significant numbers of new identified sites in between. As with Ogilvie’s work at Eretum, the emphasis had shifted to include not only structures but also artefact scatters. Such work provided important evidence with which to reassess the predominant text-based approach to the ancient countryside. However, there were methodological issues which limited the use of these results, mainly focused around ceramics. Firstly, only small samples were collected and the resulting site chronologies were divided into very broad periods, e.g. Republican and Imperial. More specifically, there was a heavy bias towards the better-known ceramics of the Roman period. The pre-Roman and especially late antique and early medieval periods were much less visible. In particular, this was a function of heavy dependence on finewares for the dating of site occupation. However, the continual development of ceramic typologies meant that coarsewares were increasingly identifiable (e.g. Patterson & Roberts 1998). To take full advantage of such ceramic typologies – and more general developments in surface survey methodologies – new survey is required. During the 1980s, another survey was launched – the Farfa survey, directed by John Moreland (Leggio & Moreland 1986; Moreland 1987). This explored a transect extending along the Farfa river valley from its confluence with the Tiber, c.13km up into the hilly interior, encompassing part of Muzzioli’s Forma Italiae survey area. This survey differed from earlier work in two important ways. Firstly, it adopted an ‘off-site’ strategy; the basic unit of record for earlier work was the ‘site’. In contrast, the Farfa survey recorded artefact densities across the entire landscape, and defined ‘sites’ at a later interpretative stage. Secondly, the survey collected and quantified all artefacts including coarsewares permitting the recognition of late antique and early medieval settlement long after African Red Slip had ceased to be widely available. The survey is not yet published, but promises to add much to our understanding of this landscape, especially in relation to the medieval abbey of Farfa and the medieval landscape (Moreland et al. forthcoming). The particular area chosen for survey by Muzzioli has also been the subject of two more recent and one current survey projects. As part of the Tiber valley project, a small area to the south-west of Cures Sabini was selected for resurvey during 2000 (Di Giuseppe et al. 2002). Bibliographic research and restudy of the South Etruria Survey material from the other side of the Tiber had raised several problems of interpretation. These included the enormous disparity in settlement density on either bank of the Tiber, the possibility of using new coarseware chronologies to extend the dating of known sites, and the opportunity to explore an hypothesized area of land division (ager quaestorius) (see Muzzioli 1980: 37-9). The survey adopted a similar approach to the Farfa survey, recording artefacts across a continuous landscape, but sites were also carefully defined whilst in the field. In particular, the survey demonstrated the importance of coarsewares for identifying pre-Roman and late antique/early medieval settlement and considerably higher overall settlement densities were identified. Work by the Regione Lazio in the comune of Nazzano during 2001 has similarly recognized a number of new sites near the Tiber ‘loop’ (AA.VV. 2002). Finally, there is the on-going Galantina project; this extends north from the river Farfa, the southern half coinciding with Muzzioli’s survey, whilst the northern half moves into a significant blank on the archaeological map (Gabrielli et al. 2003). Again, intensive collection techniques have identified a dense distribution of sites from all periods. To summarise, there has been a strong bias to a relatively narrow strip of land alongside the Tiber, though more recent work has begun to move into the hills to the east. In general, the intensity with which survey has been conducted – measurable as person days per square kilometre or distance between walkers, etc. – has increased and this has led to a shift of emphasis from structures to artefact scatters and to an overall rise in settlement densities. The collection of larger samples and improved coarseware typologies has also allowed a wider range of sites to be identified. Overall, such work demonstrates the continued need for, and value of, survey both in the face of increasing development pressures and in order to understand better existing datasets through targeted fieldwork. These surveys also demonstrate the continued potential of the archaeological record despite its on-going erosion through climate and agriculture. The integration of these different datasets as part of the Tiber valley project has presented a range of issues (for an initial overview in relation to the Sabina, see Di Giuseppe et al. 2002). For example, the precision and accuracy of spatial data is exceptionally varied. Gammurini et al. worked with a series of bearings on 1:50 000 maps; in contrast, the Corese survey made use of a hand-held Global Positioning System receiver with an accuracy of c.5m. Similarly, earlier surveys recorded sites as dots (x,y coordinates), whilst more recent surveys have preferred to delimit sites as areas. Simply deciding whether or not a site is that recorded by an earlier survey can therefore be problematical and can only be resolved by plotting sites on maps and taking the accuracy and precision of spatial data, site descriptions and other characteristics into consideration. As well as spatial issues of integration, there are also chronological issues. The revisiting of sites, even 25 years or more after their initial identification regularly produces new dating material which both extends site activity and allows greater resolution, for example, a mid or late republican date instead of generic republican date. However, the vast majority of sites have not been revisited – indeed, many no longer exist. Thus this variation in chronological resolution can present problems when studying the changing numbers of settlement over time – it is important to remember that such detail is often only available for a small percentage of sites. But can, for example, the greater chronological detail available for the Corese survey be extended to Muzzioli’s Forma Italiae results or even further afield? The initial results of the Galantina survey to the north suggest some caution may be necessary as there appears to be some local variation of settlement development, for example, note the sharp contrast in the numbers of mid republican sites identified by the Corese and Galantina surveys (Di Giuseppe et al. 2002; Guidi et al. forthcoming). A final consideration concerns interpretation. Within Italian survey, the definition of the ‘villa’ has been a particularly vexed question; if any thing, the definition of smaller sites is even more problematical (for an overview, see Cambi & Terrenato 1994). A couple of examples will suffice. Muzzioli made use of a settlement hierarchy drawing on the classical sources; this emphasized villas whilst smaller and/or poorer concentrations of material were left uninterpreted as “aree di frammenti fittili”. In contrast, the Corese survey sought to interpret this latter group as farms or outbuildings on the basis of specific categories of material or proximity to other sites. As part of the collation of thousands of sites across the wider Tiber valley project research area, each individual site interpretation was reviewed. All the relevant data for each site were collated and new interpretations derived on the basis of a standard set of criteria for site definition. In each case, a note of earlier interpretations of the site was kept in order to follow changing understandings of the archaeological record. In conclusion, the middle Tiber valley, but particularly the Sabina Tiberina, represents a particular rich historical landscape with a well-developed tradition of archaeological research. As outlined above, this wealth of studies can present problems of interpretation, but the very historical depth of work provides the opportunity to study the changing nature of the archaeological record over the last century and to explore the relationship between diverse survey methodologies and results. By studying the history of the research of this landscape we can begin to understand the meaning of our archaeological data, not only in areas that area well studied, but also in the areas which have not been subject to fieldwork or where the details of that fieldwork are poorly understood. This section has concentrated on rural settlement in the Sabina Tiberina. These methodological studies suggest that settlement density is much higher than many older surveys suggest, and more akin the levels found on the western bank of the Tiber in Etruria. However, this is not to argue that the settlement trends found in the two areas are identical. In fact, the most recent surveys suggest that there were important differences not only between Sabina and Etruria, but also within Sabina itself. In contrast, the next section moves from rural to urban settlement and from a small sub-region of the middle Tiber valley to a synthetic overview.