ABSTRACT
Archaeological data are the basis on which to generate information and knowledge about past events. Existing data are frequently limited to one project's scope or activity and are rarely used for other purposes. Software systems can help archaeologists to obtain relevant information from large amounts of data. To construct these systems, it is necessary to identify and understand the most common methods of inference that occur in archaeology, in order to include them as requirements of knowledge-assisted software systems. To achieve this, a set of archaeological case studies were selected and their discourses analysed, attempting to establish the characteristics of the inferences that were carried out in each case. The discourse analysis method, which has been successfully applied in other disciplines, has allowed us to find the different types of hypotheses that are usually managed in archaeology, and to analyse the potential uses of knowledge extraction techniques for archaeological data.
