Proceedings:
Representation and Acquisition of Lexical Knowledge: Polysemy, Ambiguity, and Generativity
Volume
Issue:
Representation and Acquisition of Lexical Knowledge: Polysemy, Ambiguity, and Generativity
Track:
Contents
Downloads:
Abstract:
Complex nominals present one of the greatest challenges in semantic analysis to date, particularly when viewed from the concerns of polymorphic expressiveness in natural language. From the perspective of the traditional lexicon designer, complex nominals are formed generatively and therefore do not merit explicit listing except when they are clearly non-compositional. In the spectrum of complex nominals, technical terminology occupies a special place, in that it is highly productive yet serves to encapsulate the essential concepts of a particular technical domain, therefore meriting inclusion in a technical glossary for that domain. Interpreting technical complex nominals (henceforth TCNs) -- complex nominals with terminological status-- generatively is not generally feasible because they are often composed of technical words for which conventional dictionaries will not provide coverage. As lexicons, and dictionaries, evolve from being static sources of word definitions to dynamic knowledge bases which function as resources for natural language technologies (such as NL-based information access and retrieval, machine translation, and natural language understanding), it will be increasingly important for them to support processing of technical documentation. Hence, they must serve not just the function of a lexicon but also that of a technical glossary. Given the frequency with which they are coined it is impractical to rely on manual listing of TCNs, and so it becomes critical that on-line lexical resources be able to acquire and interpret TCNs dynamically as text is processed.
Spring
Representation and Acquisition of Lexical Knowledge: Polysemy, Ambiguity, and Generativity