Formal Concept Analysis, which has been developed at the TH-Darmstadt for more than 10 years now, enables a graphical representation of conceptual hierarchies using mathematical lattice theory (Wille, 1982).
The WordNet project tries to build a lexical database which orders the words corresponding to the mental lexicon. Therefore they collect the words in synonym sets between which semantical relations hold (Miller, 1990). Those relations for nouns are synonymy, antonymy, hyponymy and meronymy where hyponymy and synonymy can be explained with the methods of Formal Concept Analysis as subconcept-superconcept ordering and equal classification.
In this talk I will explain the differences between hyponymy and meronymy, which are both hierarchical classifications of words, and extend the methods of Formal Concept Analysis to combine them with relations like meronymy. I will use that to develop a graphical representation of a word, its neighbours, and its classification in WordNet.
The general aim is to deal with different classifications on one set and to find representations for relations like meronymy, which do not fulfill the strong conditions for a lattice and in addition have more exceptions.
Wille, R.: Restructuring lattice theory: an approach based on hierarchies of concepts. In: Rival, I. (ed.): Ordered Sets. Boston (1982). 445-470.