Abstract
A necessary component of any natural language processing program is the dictionary, an electronic database containing syntactic and semantic definitions of natural language words. This talk addresses the issue of how one might construct large-scale dictionaries using online resources. Preliminary results are presented on the benefits of using linguistically motivated techniques over brute-force methods for large-scale lexicon development. These techniques have been implemented as part of a set of lexicon-development tools used in multilingual applications such as foreign language tutoring and machine translation.