Combining traditional and digital approaches : application of State Sequence Analysis to the study of long-term evolution of settlements

Settlement patterns and networks are traditionally studied as snapshots of a landscape organization. Their structure is described based on the hierarchical levels of each settlement at a given period. These hierarchical levels are attributed considering a number of variables including characteristic artifacts, settlement size and functions, that are studied in a pure archaeological or statistical approach. Some projects already introduced the time-component including chronological variables such as total duration of occupation and date of first implantation in the considered variables when clustering settlements. However, the resulting typologies do not consider chronological variation. This paper explores the way traditional archaeological classifications of settlements can be reemployed in a quantitative approach to study the evolutionary trajectories of settlements in a long-term perspective. Evolutionary trajectories are built compiling hierarchical levels determined for each occupation period of settlements (with either a traditional or a statistical approach), and State Sequence Analysis based on Optimal Matching is proposed as a tool to measure their variability. A cluster analysis finally groups trajectories and detects trends in the long-term development of settlements. Applied to the case-study of the Anapodaris river catchment (Crete, Greece), this novel approach enables to study the trajectories of 6 cities known from the 9th century BC, among which 2 or 3 become independent city-states after the 6th century. Hence, State Sequence Analysis offers a new perspective in the study of the evolution of landscape organization and allows to improve the modelling of city-states territories