Phraseology in specialized language and its representation in environmental knowledge resources

Zusammenfassung

Phraseology in Specialized Language and its Representation in Environmental Knowledge Resources This research presents the design of a template for encoding and describing phraseological information in the environmental knowledge base, EcoLexicon (ecolexicon.ugr.es). One of the objectives of this study was to formulate a methodology that could also be used to codify phraseological information in other specialized knowledge domains. The focus of our analysis was on verb collocations (verb+noun, noun+verb) because of the need to enrich specialized knowledge resources with information regarding predicates and their arguments. As is well known, verbs are an extremely important part of language since, in a manner of speaking, they set the scene or establish the structure for the rest of the sentence. However, very few specialized knowledge resources include them. The phraseological template proposed in this research takes its theoretical premises from cognitive approaches to Terminology, namely, Sociocognitive Terminology and Frame-based Terminology. In addition, assumptions from linguistic models that deal with predicate argument structure were also included, such as the Lexical Grammar Model, Lexical Constructional Model, Role and Reference Grammar, and FrameNet. The practical guidelines for the design of our template were based on a detailed analysis of the most representative lexicographic and terminographic resources that contain phraseological information. A set of the most important meaning-based resources for verb description was also analyzed. The underlying idea of our research is that verbs and their potential arguments can be classified and structured in a set of conceptual-semantic categories typical of a given specialized domain. In this context, when semantic roles and macroroles are specified as well as the resulting phrase structure, it is then possible to establish templates that represent this meaning for entire frames. In this regard, within the context of a specialized knowledge domain, the range of verbs generally associated with potential arguments can be predicted within the frame of a specialized event. This occurs, of course, because the nature of arguments is constrained by verb meaning. However, it is our assertion that this influence is reciprocal since in specialized language, verbs are also to some extent constrained by their arguments. In this regard, there is an interaction between the meaning components of the entities and processes activated in specialized knowledge event representations. Finally, the results of our analysis have been implemented in EcoLexicon. As shown, the methodology proposed in this thesis for encoding and describing verbal collocations in terminographic resources is useful both for text comprehension and text production.