1. Introduction

One of the more common criticisms about inductive predictive models is that they fail to account for either functional or temporal differences in sites. It is claimed that accounting for these would make more powerful models (Kincaid 1988; Rose and Altschul 1988). While there may be gains in model power, this proposition has not been tested. In the scope of this research, site functionality will not be considered, and only the temporal aspect will be introduced into standard archaeological predictive modelling techniques. However, there has been little work done to prove if either temporal or functional models would actually increase the predictive power of the model. This lack of attention probably stems from both conceptual difficulties in creating temporal or functional models, as well as software limitations of geographic information systems (GIS).

The small amount of work that has been done on temporal modelling can be divided into theoretical and practical approaches. On the theoretical side, there has been discussion of how temporal data could be integrated into archaeological GIS (e.g. see Castleford 1992; Daly and Lock 1999; Harris and Lock 1996). On the practical side, there have been some projects which have attempted to model temporal aspects of archaeological data (e.g. see Holst 2001; Mytum 1996), mostly by employing snapshots of time, or time slices, to represent temporal differences. However, there has been little in the way of application of temporal aspects in predictive modelling.