PREVIOUS   NEXT   CONTENTS   HOME 

2. The Nature of the Archaeological Evidence

For the rural landscape of Greece, the majority of the archaeological data takes several forms: scattered architecture, terraces, field boundaries, agricultural/industrial equipment (such as presses, mortaria, or kilns), and, predominantly, ceramics (Millett 2000; Pettegrew 2007). The majority of this information has been recovered from intensive archaeological field survey, though there are scattered instances of excavation within the rural landscape (for example, Hjohlman et al. 2005), with the Dema and Vari houses being the most widely known (Jones et al. 1962; Jones et al. 1973). For most surveys, it is fair to say that ceramics have formed the vast majority of recorded archaeological material (Chapman 1999, 65-6). The character of such ceramic assemblages are quite problematic: they are a smear on the landscape, a partial surface reflection of subsurface deposits that elide temporalities, and are only crudely associated spatially. Complicating these problematic assemblages is the fact that the vast majority of what is recorded on the surface is not datable, and most of it is simply counted. Those ceramics that are identifiable are, primarily, only broadly identifiable within wide functional categories: fineware, tableware, coarseware and/or storage and transport vessels (Witcher 2006; 2012, 11-14). The interpretative emphasis has been on identifying specific ceramic wares in order to aid dating of predominant phases within any one artefact scatter — and even the ceramic density that constitutes a 'scatter' or 'site' is problematic.

It needs to be emphasised that this is not meant as a slight on survey archaeology; it is simply the nature of the evidence, and survey archaeologists have long been aware of the problematic nature of the recovered data (compare, for example, the collections: Keller and Rupp 1983; Alcock and Cherry 2004). However, it is still the case that scholars build typologies and categorise entire landscapes based on this small top-slice of identifiable and datable ceramics. We identify key abstractions or aggregations of this problematic data as being of most use in interrogating the broader landscape — for example through a careful comparison of datable artefacts, artefact density, and artefact scatter size (Chapman 1999, 65-7; Renfrew 2003, 312-13) — but the underlying issue remains: the vast bulk of what is recorded in the landscape is not of direct use, and contemporary practice in archaeological visualisation, while a powerful tool in its own right, can obscure some of the underlying qualitative and quantitative failings of the data (Slapšak et al. 2001; Lolos et al. 2007).

Nevertheless, landscape archaeologists tend to further generalise this data by grouping it into larger typological hierarchies, in essence deciding on the relative proportions of farmsteads, villas/estates, hamlets, villages and towns within an area based on these data (see Table 1; key debates on site classification: Gallant 1986; Keay and Millett 1991; Osborne 1992; Foxhall 1997; 2004; Mee and James 2000; Pettegrew 2001; Witcher 2012). However, there are significant discrepancies both between and within projects about how they construct these hierarchies. Not only do surveys employ different descriptors within their site hierarchies, but exactly what density of artefacts constitutes a site and what is simply background 'noise' has never been satisfactorily resolved, irrespective of regional differences in the archaeological signature (or formation) of sites.

Current methodologies of site classification can be quite sophisticated, and indeed, site classification forms a crucial part of the analysis of settlement developments, but as Rob Witcher (2012, 16) has recently stated 'field surveyors have noted that there are problems in defining the archaeological signature of a farm, but there has been little consideration of whether farm is an appropriate category… or the implications of using such a term'. He suggests that we need to develop alternative methods of site classification to stand beside those we already use, as a way to create a more dynamic approach to classification.

Much of the evidence from the rural landscape relates to agricultural production, storage and transport, and the existing evidence for storage and transport in the rural landscape is what we might call 'end-point' evidence: remains of ceramics, recoverable infrastructure, surviving texts. In other words, this evidence represents the end-point of a sequence of behaviours. Despite that, we use this evidence as a proxy for the entire sequence, rather than the stage in the sequence that they represent (Ascher 1961; Chapman 1999; Grosjean et al. 2003; Seetah 2008; Stewart 2013b).

 PREVIOUS   NEXT   CONTENTS   HOME