CoreLex is a new type of lexical semantic ontology that is organized around systematic polysemous classes. Lexical items are assigned to such classes instead of being assigned a number of distinct senses. A systematic polysemous class corresponds to an underspecified semantic type that entails a number of related senses, or rather interpretations that are to be generated within context. The underspecified semantic types are represented as qualia structures along the lines of generative lexicon theory.
Underspecified semantic types assign a basic semantic structure to each lexical item in the corresponding systematic polysemous class. On top of this, specific information for each lexical item needs to be defined. This kind of information however is corpus specific and therefore needs to be adapted specifically to and on the basis of a particular corpus of texts by mapping syntactic patterns onto semantic ones. Simultaneously, the acquired knowledge is used to classify unknown lexical items in a probabilistic fashion.
To view an application of CoreLex, try out CoreLex Webber, an interactive tool for creating semantic indexes on Web documents. Or view a ready-made example in the domain of molecular biology.