Retrieval of relevant document from a huge set of data is a crucial task particularly in an age when world wide web literally crisscrosses the world. The task becomes all the more difficult if the process involves recovery of fuzzy data. Then, not only does the text comparison between query and document become sufficient, but at the same time concept matching also assumes importance. Our primary aim in this paper is to work out a data extracting system that can handle queries and documents involving fuzzy concepts along with crisp concepts. Here, the keywords from query and document are matched keeping an eye on the meaning of the words. Moreover, the ontology used here is created by treating fuzzy linguistic terms as variables called fuzzy-valued variables, a completely new way of dealing with fuzzy concepts. In our research, we have considered a linguistic variable (a fuzzy set) as a fuzzy-valued variable. We have taken a layered approach and given emphasis on modularity. We take the concerned linguistic variable to sit at the root of a tree with the semantically similar terms situated at the end of the branches of the tree and through this process the meaning of the term is reflected.