The Use of Machine Readable Dictionaries in the Pangloss Project

Machine Readable Dictionaries (MRDs) contain much useful information about lan£uage. Researchers have worked for the last decade on ways to extract this information for language processing systems. But processing dictionaries for use in natural language computation is itself a difficult problem. Transforming information from a version designed for human readers to one usable by computers is a slow process that, for example, may require one to define a "semantics" of font changes on a dictionary printer tape (e.g., that small bold capital letters mean that the word so capitalized is a synonym of the headword). Some fairly straight-forward information, such as word lists, part of speech fists, and orthography indications have been used by various language processing systems, yet the rather obvious use of MRDs -to allow the automatic construction of lexicons for natural language processing systems, and in particular for MT systems -has not been within our reach until recently.