oa The Arabic ontology

We overview the Arabic Ontology, an ongoing project at Sina Institute, at Birzeit University, Palestine. The Arabic Ontology is a linguistic ontology that represents the meanings (i.e., concepts) of Arabic terms using formal semantic relationships, such as SubtypeOf and PartOf. In this way, the ontology becomes a tree (i.e., classification) of meanings of the Arabic terms. To build this ontology (see Fig.1), the set of all Arabic terms are collected; then for each term, the set of its concepts (polysemy) are identified using unique numbers and described using glosses. Terms referring to same meaning (called synsets) are given the same concept identifier. These concepts are then classified using Subsumption and Parenthood relationships. The Arabic Ontology follows the same design as WordNet (i.e., network of synsets), thus it can be used as an Arabic WordNet. However, unlike WordNet, the Arabic Ontology is logically and philosophically well-founded, following strict ontological principles. The Subsumption relation is a formal subset relation. The ontological correctness of a relation (e.g., whether "PeriodicTable SubtypeOf Table" is true in reality) in WordNet is based on whether native speakers accept such a claim. However, the ontological correctness of the Arabic Ontology is based on what scientists accept; but if it can't be determined by science, then what philosophers accept; and if philosophy doesn't have an answer then we refer to what linguistics accept. Our classification also follows the OntoClean methodology when dealing with, instances, concepts, types, roles, and parts. As described in the next section, the top levels of the Arabic Ontology are derived from philosophical notions, which further govern the ontological correctness of its lower levels. Moreover, glosses are formulated using strict ontological rules focusing on intrinsic properties. Figure 1. Illustration of terms' concepts and its conceptual relations Why the Arabic Ontology It can be used in many application scenarios such as: (1) information search and retrieval, to enrich queries and improve the results' quality, i.e., meaningful search rather than string-matching search; (2) Machine translation and term disambiguation, by finding the exact mapping of concepts across languages, as the Arabic Ontology is also mapped to the English WordNet; (3) Data Integration and interoperability in which the Arabic Ontology can be used as a semantic reference to several autonomous information systems; (4) Semantic web and web 3.0, by using the Arabic Ontology as a semantic reference to disambiguate meanings used in web sites; (5) Conceptual dictionary, allowing people to easily browse and find meanings and the differences between meanings. The Arabic Ontology Top Levels Figure 2 presents the top levels of the Arabic Ontology, which is a classification of the most abstract concepts (i.e., meanings) of the Arabic terms. Only three levels are presented below for the sake of brevity. All concepts in the Arabic Ontology are classified under these top-levels. We designed these concepts after a deep investigation of the philosophy literature and based on well-recognized upper level ontologies like BFO, DOLCE, SUMO, and KYOTO. Figure 2. The Top three levels of the Arabic Ontology (Alpha Version)