8. The Analysis of Formal Space

I will now turn attention onto another approach I have been recently advocating: spatial modelling of non-geographic structures in GIS through their mapping in relational abstract space (Šmejda 2008). This requires a brief introductory explanation. In one of his influential papers, Spaulding (1960) argued that archaeological description can trace only two generic dimensions of artefacts: their spatial and formal dimensions. Everything else (for instance chronology, social structure etc.) must be derived from these two categories of observation. This is usually done through the identification and interpretation of significant patterns in the data. Later, Binford (1987) noted that this logic of analysis can be applied to the whole domain of the archaeological record, not only to artefacts. These statements have been frequently used and further developed in archaeological methodology (Neustupný 1993, 44-72).

The divide between spatial and formal properties of the archaeological record is of course logical and in many instances useful, because it comfortably follows our everyday experience. On the other hand, we can easily face research problems where this division seems arbitrary, too rigid, and imposing modern preconceptions onto the remains of the past. Naturally, the archaeological record may be structured in accordance with its own frames of reference, not complying with the conventions of modern cartography.

One possible solution for such a situation can be the introduction of a more generic space, consisting of any analytical dimensions, be they the primary spatial and formal dimensions or secondary ones (various indexes, scores etc.). We can call it here a 'fact space'. The fact space is basically a union of spatial and formal dimensions (plus secondary dimensions where appropriate), where any type of recorded observation (fact) we work with can be represented (Šmejda 2008). This step frees our hands (and minds) enough to formulate many old problems in new ways, which can offer a deeper insight into their nature.

The very concept of fact space is nothing revolutionary, as it is automatically created in the course of every analytical description. For example, any descriptive database is one form of its representation. Hence, my present article is not an attempt to introduce a completely new avenue of archaeological research. It is rather an attempt to use present techniques of data synthesis from an unusual mental perspective. In other words, it is not the technology that I seek to develop but an approach to considering data and a way of archaeological reasoning.

As I have already said above, any descriptive database is an example of one particular fact space. The actual database can be very complex and extensive, which means it cannot be reasonably explored and presented directly in any simple way because of its large size and multi-dimensionality. The multi-dimensional fact space can treat a spatial position of individual burials in the same way as any other descriptive variables and the total bulk of data can be explored either by means of statistics and/or GIS. Many tasks can be carried out in both environments; others are handled better in one of them. In principle, however, there is no substantial difference between these two approaches.