The theme of Nicholas Negroponte's book, Being Digital (1995), is that the impact of information technology can be summarised in terms of the inexorable change from atoms (the material world) to bits (the digital world). This might equally apply in an archaeological information context — the atoms of the material record are digitised into bits representing that same record. What implications might transpire? One result that has frequently been observed is the apparent reinforcement of what is a supposedly objective material record. Somehow, the act of moving from atoms to bits can seemingly contribute an objective, authoritative character to something. For example, Gidlow comments that:
'At times, the partial nature of the archaeological record can seem erased by the wholeness of a vector database and its mathematical calculation of surfaces.'
Similarly, Miller and Richards argue that:
'… computer models carry more authority than paper images; people expect computers to be right, and the past is therefore presented as a known — and knowable — reality.'
Nor are these problems exclusive to graphical applications. As Lock says, any digital model is a representation of selected elements from the data and theoretical models relevant to the study in question (1995, 14). Likewise, Reilly distinguishes between 'raw' data (which is simply 'out there', lying undiscovered), primary data (which archaeologists abstract and record) and operational data (which are incorporated in a form capable of consultation and use at a later date) (1985, 63-64). There are therefore (at least!) two transformative processes in which computers may at times be implicated. First, although the abstraction of data may not appear to be affected directly by computerisation unless a computer is actually used, nevertheless, the recording structures themselves may be originally derived from computer models. As Reilly (1985, 64) says, there is a qualitative difference between pro-forma sheets and field notebooks — on the one hand, the significant attributes to be recorded have been defined in advance; on the other, the attributes are those the observer chose to record at that point. Secondly, the operationalisation of the data may involve further abstraction, 'cleaning', and, in the process, become physically separated from original recording sheets or notes and hence stripped of the doubt and uncertainty which the observer may well have expressed. This, coupled with a general belief in the neutrality of computers, only serves to enhance the overall illusion of objectivity, whereas there is no immutable truth behind archaeological data (Huggett 1995, 25). At the root of this belief are the same positivist, 'scientific' attitudes held towards computers that were discussed earlier. That said, there is a danger of over-egging the issue — long before computers were brought onto the scene there was a problem with the belief in the objectivity of archaeological recording.
However, the application of computers may well exacerbate the situation. For example, in a paper discussing the anachronism of the term 'computer' in the 21st century, Chesher (2002) argues that the concept of 'computation' carries epistemological assumptions worth questioning:
'The suffix "-puter" is related to the Latin putare, for pruning, cleansing or reckoning. So computing brings things together and cleans them up for use in estimations. It takes data — literally "givens" and extracts them from the contexts out of which they were gathered. Cleansing givens so that they can be brought together erases the process of collection itself. Whenever someone starts reckoning anything, you can bet he has some task in mind. He will decide what is relevant to the task and what is not. That's when the putare happens — the choices behind the computation are discarded, and the results presented as clean and objective.'
In other words, the apparent objectivity is in some senses embedded within the device itself. The operationalisation of data within a computer environment strips out the context of recording — or at the very least, increases the distance from it.
Drawing out one strand of this argument further, to what extent does the organisation of data for computer-based storage affect the nature of that information? If indeed there is a qualitative difference between using pro-forma recording sheets and site notebooks, is there not likewise a qualitative difference between using a paper-based recording system and a computer-based one? Certainly the conceptual models behind the two sets of structures will be somewhat different — a pro-forma sheet will more than likely be represented as several linked tables in a relational database system, for instance, and the process of data modelling involves a more elaborate deconstruction of the data than the design of a record sheet. But do the effects go further than this? There can be choices to be made about the structural models to be employed, although these choices are admittedly limited in the real world. However, hierarchical, network, relational, dimensional, and object-oriented models all conceptualise information in different ways. While a poor design, regardless of its underlying model, will affect the success or otherwise of information retrieval, to what extent does the underlying model impact on the way information is recorded? For example, Andresen and Madsen (1992) show how a typical hierarchical excavation model (where a site consists of a set of structures, a structure consists of a set of features, a feature consists of a set of contexts, a context consists of a set of components …), if implemented as a hierarchical database model, rapidly breaks down in the real archaeological world when problems such as residuality are encountered. In this case, a complex relational model is proposed instead. While an object-oriented model might seem to be more intuitively appropriate, these are not generally commercially available, underlining the presence of a deterministic dynamic in these kinds of technological decisions. In some respects, not a great deal has changed since Smith and Ryan's (1986) critique — for example, databases still model data in an essentially static manner with no historical aspect yet the potential value of retaining a palimpsest-like database, building up data descriptions rather than waiting until they are fully-formed, modelling changing interpretations and the like, is clear (see Ryan, this volume). Indeed, such a facility would be a prerequisite of the kind of multiple hypothesis modelling that Hodder proposes (1999, 120-21).
Such criticisms as there are about databases tend to focus on high-level issues of access rights and security of the data. At a lower level, problems associated with the use of databases are often associated with poor design (for example, see Ryan, this volume). In neither case is the technology itself at issue: only the way in which it has been implemented or the uses made of it. The neutrality of the database tool itself therefore seems to be a taken-for-granted. Yet, if it is such an 'empty vessel' into which data can be poured, why is it that data has to be structured before it can be utilised? And what effect does the process of structuring data for a database have on the way that we think about that data, on the way we go about recording that data, the way in which we retrieve that data, and the way in which we subsequently analyse that data? Such assumptions about the neutrality of technology occupy a middle ground between the pro- and anti-technology positions, but still ignore the social, cultural, and philosophical issues concerned with its use. The technology cannot be entirely separated from issues surrounding its implementation and subsequent use, if only because of the way in which it intermediates between user and data.
Of course, computers may directly influence the way archaeological data is recorded, but the influence can be altogether more indirect. Certain categories of information may not be easily recorded, if at all, in a computerised environment — consequently, if information cannot be included in the digital model, then it cannot be available for analysis. Unfortunately, and inevitably, such categories fall into the areas of interest of post-processualist archaeology, and hence again compounds the positivist impression of computers. For example, Lock (1995, 14) points to cognitive and perceptual spatial data as problem areas, although there has been progress in this area (see contributions in Lock 2000, especially Wheatley and Gillings 2000). Geographical Information Systems are spatially deterministic:
'… data that cannot be captured as a spatial primitive comprising point, line, polygon, or pixel, is essentially excluded from the database … Essentially qualitative or alternative forms of knowledge representation that are crucial to understanding the nature of a place are largely excluded from a GIS.'
(Lock and Harris 2000, xvii).
The representation of symbolic, experiential, and interpretative knowledge within Euclidean space is extremely problematic and hence for the most part avoided. Consequently, as Lock and Harris (2000, xvii) observe, 'GIS "capture" a scientific-based data-driven view of only one "reality"' a remark that might equally be applied to information technology in general. However, the seductive power of the technology means such concerns are often sidelined.
© Internet Archaeology
Last updated: Wed Jan 28 2004