A computational lexicon for natural language processing needs to deal not only with simple words, but also with lexemes consisting of more than one orthographic word. Multi-word lexemes such as idioms, phrasal verbs, and collocations display different degrees of flexibility concerning their morphology and syntax. In the context of NLP applications, it is vital to know which modifications are possible for each idiom (as this will determine whether both an idiomatic and a literal reading are possible) and to express these acceptable variations in an appropriate way. The present work describes the linguistic specification of English multi-word units in Word Manager, a system for reusable morphological dictionaries developed at the University of Basel. Word Manager makes it possible to express and integrate this description with a morphological lexicon database. The detailed description of the classes of multi-word units in English is the first full-scale application of this part of the Word Manager formalism.
E-mail this message to a friend
