‘We Would Never Have Thought to Go There’ – The Changing Definitions of a Site in Central Italian Archaeology

Essentially, a ‘site’ has to be distinguished from a ‘non-site’ and the boundary between the two has to be drawn (Carman 1999). These definitions made in the field are archaeological decisions, not observations (Dunnell and Dancey 1983). Our practice is a pragmatic act of constructing boundaries on the basis of the density of finds and/or features. The definition of a specific site is subject to evaluation of the archaeological criteria used to define it. However, it is clear that the concepts behind those criteria have changed over time. In central Italian archaeology, the scholars have moved from the topographical archaeology of the 19th century to the GIS-assisted landscape studies of the 21st century and their definitions have evolved similarly. Central Italy is relevant as an example since Italian studies have contributed fundamentally to the developments in field archaeology.


Introduction
A 'site' is one of the key concepts in archaeology, and is not specific to central Italian archaeology.Archaeologists have tried to define what constitutes a site and how it can be measured.One definition of a site is 'as places where significant traces of human activity are identified' (Renfrew & Bahn 1994: 42).
Essentially, a 'site' has to be distinguished from a 'non-site' and the boundary between the two has to be drawn (Carman 1999).These definitions made in the field are archaeological decisions, not observations (Dunnell and Dancey 1983).Our practice is a pragmatic act of constructing boundaries on the basis of the density of finds and/or features.The definition of a specific site is subject to evaluation of the archaeological criteria used to define it.However, it is clear that the concepts behind those criteria have changed over time.In central Italian archaeology, the scholars have moved from the topographical archaeology of the 19th century to the GIS-assisted landscape studies of the 21st century and their definitions have evolved similarly.Central Italy is relevant as an example since Italian studies have contributed fundamentally to the developments in field archaeology.
Italian archaeology as such has two different meanings.Firstly, it refers to the archaeology done by Italian scholars, and secondly, it refers generally to all archaeological research undertaken in Italy.The relationship between these two spheres is symbiotic and so it can be difficult to treat them separately.However, due to the shear volume of regional studies, the focus of this article is on central Italy -Lazio is, afterall, the core area of the ancient Latins, Etruscans and Romans.In addition, many of the most influential schools in Italian archaeology originate from Rome (cf.Guidi 1988Guidi , 2000;;Loney 2002;Peroni 1992;Vanzetti 2002).Many different traditions come together in Italian archaeology, a fact readily induced from recent Italian literature (cf.Guidi 1994Guidi , 2000;;Barbanera 1998), especially in landscape archaeology (e.g.Bernardi 1992;Cambi & Terrenato 1994).This exchange of ideas has continuously transformed the concept of the 'site' (cf.Attema et al. 2002;Francovich & Patterson 2000;Patterson 2004).

Places in the Sun -The Beginning of the Topographical Archaeology
Italy is one of the birth places of modern archaeology (cf.Barbanera 2000).Italy has been important not only for classical archaeology but also for the development of landscape archaeology, where the work of the early topographers discussed here, has an important place.The way Dennis, Canina, Gell and Nibby were working is not dissimilar to modern surveying.Their approach is now considered 'old-fashioned' but their discoveries are still counted as archaeological 'sites'.Like modern-day cultural heritage managers, Gell and Nibby were aspiring to create an archaeological map of Rome and its environs.Their project resulted with the books Analisi storico-topografico-antiquaria della carta de dintorni di Roma (Nibby 1837) and The Topography of Rome and Its Vicinity (Gell 1846).These volumes were closely followed by The Cities and Cemeteries of Etruria (Dennis 1848), although the latter was considered to be an antiquarian travel guide.
In these publications, 'sites' exist only in relation to a town, past or present.Nibby and Gell were cataloguing places in alphabetical order, in the tradition of the encyclopaedists.For Dennis, the defining principal was the order of twelve city states of ancient Etruria.All these sites presented were monuments connected to a central place.As monuments, they were clearly definable: visible ruins, with entities like walls, chamber tombs or ancient road cuttings.A typical site tended to be Etruscan or Roman (Rajala et al. 1999); prehistoric sites were not of interest to classical scholars in search of the patrimony of civilisations.This concept of a site was qualitative, defined by tangible structures.This concentration on centres and structures is understandable as the topographers were guided and inspired by ancient authors.Literature defined what was important, so much so that sometimes the sites of which nothing remained were counted.
These monuments lay in a well-defined landscape.The topographical locations and the geographical attributes of these places were perceived important enough to be described.This landscape was often reduced to a network of specific, value-laden points.The mental landscape was structured by the distance from Rome or the importance of the nearest major centre.Conceptually, sites were evaluated and appreciated on the basis of important artworks or magnificent structures.It was very seldom that a 'site' would have been defined as important because of the presence of fragmentary finds.When this happened, like in Capena, Municipio de Romani (Galletti 1756), the finds counted were perceived high status or high art pieces such as marbles and inscriptions.

Mental Maps Created: Traditional Italian Archaeology
'Traditional' Italian archaeology is an ambiguous concept, because archaeology in Italy has not experienced the turbulent shifts of paradigm as archaeology in the Anglo-American world.With 'traditional' I mean the kind of cultural historical archaeology that did not critically evaluate the assumptions underlying archaeological thinking and gave limited consideration to the theoretical archaeological debate outside Italy.Generally, almost all Italian archaeology pre-dating the 1980s can be defined as 'traditional' (cf.Cuomo di Caprio 1986), although during the latter part of the 1970s certain archaeologists started to discuss the issues of 'New archaeology' (Guidi 2000).
The new unified Italy at the end of the 19th century came to the idea that archaeological heritage should be administered and protected.As a result, the Ministry of Public Education (Ministero dell'Istruzione Publica) ordered a systematical survey of south Etruria to be carried out.The material of this Carta Archaeologica d'Italia was collected mainly during the 1880s but published only in 1972 (Cozza 1972).The aim of the project was to map the remains of the ancient roads, and to collect new material, mostly from the Etruscan period (Barnabei 1894).This work stands as a testimony to the unchanged standards in defining 'sites'.For example, Etruscan or Faliscan pagi were often recognised on the basis of their geographical characteristics, rather than with archaeological finds.A resemblance of an important place with known ruins was enough to identify a 'site' -archaeological or literary evidence was not needed.
The views on real 'archaeological sites' were persistent.A 'site' was defined be the presence of high class finds such as marbles, art pieces, fine pottery and/or inscriptions.These were catalogued in the periodical Notizie degli Scavi.The outer looks of a site were firmly rooted in ideals: Roman sites had considerable structures (e.g.Lugli 1931Lugli -1940) ) and Etruscan centres lay on upper tufa plateaux surrounded by river valleys and cemeteries (e.g.Pallottino 1942).Prehistoric sites were Villanovan cemeteries or lake dwellings of terramarcoli.However, more than Etruscan or Roman archaeology, prehistoric archaeology relied on fragmentary evidence.Nevertheless, a typical prehistoric site was also a standardised concept, as Rellini's (1920) catalogue of caves with or without archaeological layers demonstrates.
As late as 1967 there was some astonishment when archaeologists found unexpectedly open Villanovan settlements, places described as 'luoghi ove mai avremmo pensato andare a cercare resti di insediamenti' ('places where we would never have thought to go to look for the remains of settlements'; Colonna 1967: 9).A 'site' was still a highly qualitative concept, defined by clearly visible definite attributes.A real, recognised archaeological place had certain expected qualities.Furthermore, the guidance of ancient literature and the primacy of historical periods was still important (cf.Barbanera 1998;Guidi 2000).

The South Etruria Survey
The British influence on Italian topographical studies has long traditions dating back to Gell and Dennis.At the beginning of the 19th century Thomas Ashby started his work of cataloguing monuments along the Roman roads radiating out of Rome (Ashby 1927).After the Second World War, when the Italian agricultural landscape was taken under deep plough, John Ward-Perkins and the archaeologists of the British School at Rome surveied south Etruria (Potter 1979).The methods of field walking was practised and propagated.'Sites' were now recognised as buildings from concentrations of pottery sherds and building material and were recorded on maps as points with a grid reference (Kahane et al. 1968).
Pottery scatters were systematically searched for in order to study rural settlement.Furthermore, it was acknowledged that a multidisciplinary approach was needed since the landscape had been continuously altered by natural and artificial processes.Therefore, it was understood that not all past sites were to be found.Among the fieldworkers the differing find distributions and the existence of eroded colluvial material brought along fierce postdepositional discussions (pers.comm.Michael Craven).Much attention was given to the classification of finds, and qualitative definitions were formulated for predominantly Roman sites.The poorest sites were classified as huts whereas richer ones with high quality finds like marble and mosaic tesserae were either farms or villas (Kahane et al. 1968: 154).The size of a 'site' was systematically estimated only when Potter brought new methodological ideas from Britain (cf.Potter 1992).At that point in the late 1960s a 'site' became a quantitative entity.
The South Etruria survey gave stimulus to multiperiod surveys in 1970s and 1980s (Barker 1995b: 2).Later surveys encountered larger and vaguer scatters of surface finds in lower densities (Wightman 1981: 278).Although qualitative classifications were still applied (e.g.Whitehead 1994; see also Mattingly 2000), fieldworkers began to move towards quantitative measures.The classifications were based on the ranked sizes of spreads as well as the quality of finds.For example, in Molise (Lloyd & Barker 1981) the sites were ranked from small to very large.However, the definition of the boundaries of different size classes and comparison with other surveys remains difficult.

New Archaeology, New Ideas, New Strategies
After the South Etruria survey it became a common understanding that a site exhibits definable limits, represents recurring human activity and has a density above a certain limit (Cherry 1983: 394-395).However, the American theoretical debate after 'New Archaeology' presented new ways to think about sites.New definitions were formulated rigorously and this led to the creation of well-developed site hierarchies (Flannery 1976).Moreover, Thomas (1975) was promoting an idea of siteless survey, which introduced the idea of recording find densities systematically across a grid.This approach was applied in the Boeotia project in Greece (Bintliff & Snodgrass 1985) but not in Italy.Results allow presenting concentrations around site cores and declining pottery counts in site peripheries (Bintliff 1992: 89-97).Nevertheless, this was seen as a vehicle to define sites since the idea of confined sites and spatially limited living areas stayed.Thus, gridded survey was applied in Italy only to measure changing distributions at definite sites whereas the general collection unit was pragmatically defined as 'a field' (cf.Barker 1995a;Mattingly 2000: table 2.1.).
The concepts of 'halo' and 'non-sites' were also developed in Italian archaeology (Coccia & Mattingly 1992).'Halo' were defined as less concentrated post-depositional spreading and 'non-sites' as archaeological 'background noise'.Distinguishing sites from background noise was based on required boundary values of densities, which were deduced through statistical analyses of fragment counts and weights of different material groups.As a result, a quantitative series of find density classes, running from 'light' to 'heavy', was created.The concept of 'background noise' was due to continuous human and natural interference, but it was also suggested that it could hide more temporary activities in the landscape (cf.Thomas 1975;Barker 1991).
The methodological refinement of the Mediterranean surveys has contributed to the understanding of the limitations of observing sites.It was soon realised that a site may be lost in the future because of site degradation and changes in visibility and land use (e.g.Cherry 1983: 398-399;Barker 1991).Although the concept of a site was seen in relation to natural and human activities in the study area, the whole idea in itself could be found unsatisfactory.When Paleolithic artifacts were collected in the Agro Pontino survey, idealized 'sites' made no sense.Thus, point-plotting of single finds was applied (Loving & Kamermans 1991).The resulting spots were considered in their totality (Koot 1991).The efforts to draw any boundaries were made only after a total pick-up and recording.

The Foreign Impact
Foreign projects have had a profound impact on the way the Italian archaeologists have carried out their fieldwork.The processual theoretical basis was presented in the seminar Economia e organizzazione del territorio nelle società protostoriche held in 1979 (cf.Cardarelli 1992).In addition, collaborations between the Italians and the British gave practical influence (e.g.Montarrenti survey, Barker et al. 1986).
Most importantly, the building boom of the 1970s forced local students and amateurs to survey in the areas of expansion following the British example.Around Rome, many eastern suburbs were visited before and during the building works (cf.Quilici ND).This resulted in a series of interventions from the part of the heritage management authorities (cf.Carboni & Ragni 1984: 47-48).The value of pottery scatters was acknowledged and projects such as the survey of the pre-and protohistoric sites in the administrative territory of Rome (Bietti Sestieri et al. 1984) were undertaken.
Much of the survey work around Rome was published in the volumes of Latium vetus and Forma Italiae.The ongoing conceptual changes can be seen in variations in mapping and presentation.Sometimes settlement sites were shown as an area (e.g.Morselli 1980), but sometimes only pre-Roman scatters had spatial extent whereas Roman sites were presented with symbols (e.g.Quilici & Quilici 1980).The standing structures could present the confines of a site (e.g.Quilici 1980).The approaches fluctuated between qualitative and more quantitative.
Students and amateurs often collaborated together in Gruppo Archaeologico Italiano.Surveys carried out reflected the changing ways of recording and defining sites.Salmi (1978) presented his sites as points of diagnostic material, whereas the Narce project (Camilli & Gazzetti 1993) used forms to record scatters with grid reference, size, density and visibility information.Generally, there has been a move to more quantitative recording also covering off-site areas.In addition, new fieldwork practices have been actively promoted (cf. Bernardi 1992;Rendeli 1993;Cambi and Terrenato 1994).Defining a site has often reflected an ideal; sites were expected to have continuous confined distributions of surface material with defined density values (Rendeli 1992: 66-69).
In Italy, as often elsewhere, the way GIS has been used has preserved the concept of a confined site.Although one could present continuous find densities in raster format, scholars still try to present sites as points or areas in vector format.Siteless raster or vector presentations have been propagated (pers.comm.Vince Gaffney), and density maps resulting from international projects have been published (e.g.Lock et al. 1999).Italian standard archaeological maps with bounded site areas have been presented in digital format (e.g.Cattani 1997).Furthermore, rank size analysis and other geographic analyses, practised from early on in Italy (e.g.Guidi 1985), have naturally encouraged defining site areas.
After two decades of quantifying the value of qualitative analysis has been newly recognised.The Roman bias of many of the survey projects (cf. Di Gennaro & Stoddart 1982) has led to the situation where new un-recognised early sites and unknown hidden prehistoric landscapes are emerging when more attention is directed towards low densities of fragile Neolithic and Bronze Age pottery (cf.Attema et al. 2001;Bintliff et al. 1999;Bintliff 2002).The progressive refinement of dating is seen as a future necessity (e.g.Bintliff 2002;Hayes 2000;Malone & Stoddart 2000;Patterson 2000;Patterson & Millett 1998).This has been the opinion of the central Italian archaeologists for a long time (e.g.Cambi & Terrenato 1994;Pacciarelli 2000).Therefore, the project of redating of the pottery from the South Etruria Survey was carried out mainly with the expertise of Italian archaeologists (cf.Patterson et al. 2004).

Conclusions: The Different Scales of Sites
Heritage management has always been interested in maintaining boundaries, and therefore, most archaeologists spend their time looking for site boundaries.As shown here, a site boundary is an essential part of modern conceptualisation of archaeology (cf.Carman 1999).Even if recording practices allow siteless surveys, like in the case of Agro Pontino, sites and their boundaries are still sought.
The concept of a site and the understanding of its characteristics have changed.Well into the 20th century, a site was a monument or a place of an important find.There existed a very clear ideal of a typical site and these predefinitions biased the recognition of different sites.After the Second World War scatters of pottery and tile were recognised as 'proper sites' in central Italian archaeology.Originally, scatters were presented as points defined by a set of qualitative presences.Thus, a site was measured on a qualitative scale.
Gradually, quantitative attributes became increasingly important.At first, the extent of a site was considered essential, but later, find density became a defining attribute.These steps, together with blanket surveying have lead to the total recording of all areas covered and to the realisation of the existence of 'halos' and off-site distributions.Generally, the advances in archaeological theory and method have increased the understanding of the complex nature of a 'site' in the Italian context, and this in return has resulted in further advances.Ever more so now when qualitative find and fabric analysis has again become as important and significant as quantification.