GIS was quickly identified as the ideal medium through which to facilitate comparative survey - a uniform and unifying environment in which to integrate, map and analyse data (Allen et al. 1990). However, no-one has yet created a pan-Mediterranean survey GIS. The scale of the task and the methodological issues outlined earlier arguably present too great a challenge to be achieved in a single step. An initial and more manageable step towards this goal is the collation of survey metadata.
The Mediterranean Archaeology GIS (MAGIS) is an ambitious initiative to collate (formal) metadata about Mediterranean surveys with the aim of disseminating information about individual surveys in order to facilitate comparisons (Foss and Schindler 2007). This project provides a welcome and constructive GIS application with which to advance the comparative survey agenda. By documenting individual surveys using a standard range of metadata, it is possible to begin to understand how they can be compared.
Meanwhile a number of projects that compare and integrate the actual datasets created by different surveys are also underway. These include the Tiber valley project (Kay and Witcher 2005; Patterson et al. 2004) and the Regional Pathways to Complexity project (Attema et al. 1998). In both of these examples, GIS plays a central role in the collation and analysis of legacy data. However, as many of the original surveys did not explicitly document or publish formal metadata, how can their results be rigorously compared?
As stressed previously, we need to build up contextual metadata through a process of data characterisation; that is, to engage in 'source criticism' (see Alcock 1993, 49-53). GIS is an environment in which this understanding can be created through the processes of systematically modelling, visualising and analysing the structure of legacy data. The following case studies demonstrate this approach.
However, before considering these examples, it should be stressed that the digitisation of legacy data does not automatically make surveys comparable. When modelling individual survey results, their idiosyncratic nature can be accommodated with a degree of flexibility. Projects that integrate the results of several surveys demand greater uniformity in order that data can be systematically mapped and analysed (see Gillings 2001, 109). For example, divergent terminologies for specific types of pottery require standardisation. However, this process of lexical harmonisation does not necessarily make the data more comparable; a survey that did not collect a particular type of pottery is no more comparable with a survey which did collect this type of pottery as a result of the standardisation of terminology. Worse still, this methodological difference may be disguised by the process of standardisation (see Miller and Richards 1995); a difference in methodology may be transformed into an apparent difference in past human activity.
In using GIS for comparative survey, there is a danger that survey results are reduced to the lowest common attributes (where, when, what); this may encourage simplistic and inappropriate comparison by disguising real differences in data structures (how have spatial location, date and site interpretation been achieved?). As an initial step, GIS applications should explore the complexity of datasets rather than reducing them to (apparently) comparable basic attributes. For example, we might develop innovative ways of visualising and accommodating variation in data quality (e.g. in spatial coordinates; Boldrini 2007). Such work should be understood to be concerned with data characterisation for the identification of contextual metadata.
In summary, increasing interest in comparative survey has provided new impetus for the use of legacy survey data. Variability of methodologies raises significant issues of data compatibility. GIS is not a simple panacea to these problems; indeed, arguably it could aggravate the situation by masking incompatibilities. However, GIS can play a positive role: the processes of digitisation, visualisation and spatial analysis all assist in the creation of contextual metadata. In turn, this facilitates more informed interpretation and comparison.
© Internet Archaeology/Author(s)
URL: http://intarch.ac.uk/journal/issue24/2/2.4.html
Last updated: Mon Jun 30 2008