Summary
Archaeology constructs new knowledge by interpreting material evidence. Interpretation must be supported by arguments to be persuasive and accepted. Disagreements and conflicts usually arise on how different authors interpret facts or support their discourse. Therefore, new knowledge emerges in a dialogical rather than monological manner. Consequently, the study of argumentation processes is crucial to archaeology, especially when clashes occur.Discourse analysis has been barely applied in archaeology, although some natural processing language techniques have produced promising results. Also, it focuses on the utterances being spoken, paying little attention to the things being referred to. This latter aspect is usually left to be described by ontologies or conceptual models. In this paper, we propose a combined approach to capture both the dialogical argumentation and the things being referred to, so that both kinds of information can be extracted from transcripts and stored in an open-access repository for automatic processing. This is illustrated by a practical experiment in which a group of archaeologists were asked to debate about a controversial topic in archaeological heritage, the dialogue transcribed and analysed, and conclusions extracted.By using this approach, an open corpus of argumentation facts can be employed to obtain a better comprehension of why and how disagreements occur, how solidly each position is backed, and what facts are being used or ignored to construct a discourse. This can help us to design heritage policies, mediate between agents (such as citizens and government agencies) to reconcile diverging interests, or make decisions involving monuments at risk.
Keywords
Argumentation. Disagreement. Discourse. Ontologies. NLP.