The role of computers to understand the past. The case of archaeological research

Abstract

Abstract As practitioners of an historical discipline, archaeologists are in general, interested in guessing how ancient objects, buildings and/or landscapes were produced and used in the past from the object's visual appearance and material properties. In the pursuit of this general goal, archaeologists act as detectives looking for material cues (mostly visual) that may allow the discovering of social actions that may have happened in the past. This investigation can be made intuitively, but also formally, in which case we need computational tools and techniques to 1) extract necessary information from visual and non-visual data, 2) data processing to evaluate relevant relationships between different items at different spatial and temporal scales, 3) build appropriate models that can be used to provide explanatory hypothesis about the human past. In this paper I review the nature of archaeological problems and suggest computational techniques and methods that can be useful in solving these kinds of questions. Emphasis is made on modern classification and clustering techniques (neural networks, for instance), and computational simulation approaches.