This research develops software components for a fully automated system for building knowledge bases about language. The knowledge bases produced are structured to be compatible with computer systems for processing natural language (natural language understanding, natural language generation, and natural language translation systems, for example). The knowledge-acquisition system does not require extensive human intervention or "knowlege engineering," but builds the knowledge base from standard text input, such as conventional dictionaries (many of which are becoming available in machine-readable form. Initial work will be with the Longman's Dictionary of Contemporary English.) The significance of this research is that natural language processing systems, for machine translation, user interfaces, and knowledge based systems, must incorporate extensive knowledge about words and their meanings in the real world. This work is an important exploration of the extent to which the necessary knowledge may be found and built from existing text sources (e.g. dictionaries, encyclopedias). The lexical/semantic knowledge bases generated are intended for widespread utilization, both for applications and for further scientific investigation by the natural language processing research community.
