Abstract
Developing a strong and complete lexicon is essential for many Natural Language Processing (NLP) related applications such as Information Retrieval, Question Answering Systems, Natural Language Generation and Machine Translation. Such lexicon should be supported by a strong theoretical framework and implemented using tools that will facilitate its implementation and efficient use both in terms of resources and language preciseness. Recognising the derivative and inflexional nature of the Arabic language, the paper presents a formalism based on mathematics and ontologies of part of the morphological knowledge of the Arabic language based on verbs and their derivatives. In this research, we attempt to use the derivations and their patterns to structure the Arabic language and to strongly link the words’ morphology to their semantics. We first perform a thorough study of part of the Arabic Language and propose a model that is based on set theory and ontologies. We then show how this model can be used for some applications that include NLP applications.