Practice-based Ontologies: A New Approach to Address the Challenges of Ontology and Knowledge Representation in History and Archaeology Emad Khazraee 1 and Michael Khoo 1 1 Drexel University, College of Information Science and Technology, 3141 Chestnut St., Philadelphia, PA, 19104, USA {Emad, khoo}@drexel.edu Abstract. Data production in history and archaeology far outpaces data processing. In order to apply computers to this problem, historical data must be converted to machine-readable forms. This process is easy for domains of knowledge that have explicit terminology, but history and archaeology lack these characteristics. This study therefore proposes a phenomenological approach to requirements gathering for knowledge representation and ontology systems for historians and archaeologists. The approach utilizes qualitative and ethnographic research methods to gather data on practitioners’ reasoning and knowledge practices. The design requirements for ontology design can be extracted from the ‘thick description’ produced by this process, and used to build ‘practice-based ontologies.’ This paper presents the theoretical framework and early outcomes of ethnographic research with archaeologists in practice at the University of Pennsylvania. Keywords: ontology, knowledge representation, practice-based ontology, ethnographic method, archaeology 1 Introduction Historical and archaeological data are very diverse. They are dispersed in different institutions, and across different countries. The pace of data production in this field is far higher than the amount of data being processed. This situation leaves large amounts of data to be processed, and many research questions unsolved. Increasing computational power can help to solve more complicated questions, but a precondition to benefit from computers is that data must be converted and formalized into machine-readable formats. Knowledge engineering is “the application of logic and ontology to the task of building computable models” that “analyzes knowledge about some subject and transform it to a computable form for some purpose” [1]. It includes the study of a domain and the mechanics and dynamics of knowledge in that domain, to ascertain whether it is possible to achieve computable forms for that purpose. Davis et al. [18] summarized following five principles for knowledge representation. A knowledge representation: is a surrogate, is a set of ontological commitments, is a fragment