Towards the automatic derivation of lexical prototypes

In order to model human understanding of natural language, it is necessary to be able to rep-resent the knowledge and processes that humans use. As such, Natural Language Processing (NLP) research needs to take into consideration cognitive theories of language processing; one candidate theory is Prototype Theory. In this paper, we argue for the adoption of a conceptual approach to word meaning in order to achieve a view of lexical semantics consistent with cognitive representation and processes. We detail ongoing research concerned with extending an existing semi-automatic method for deriving patterns of verb usage from corpora. Finally, we argue that this approach has connections with Prototype Theory and detail initial steps to fully-automate this process.