This paper examines linguistic criteria adopted in models of knowledge representation for information retrieval purposes. For that, the model of syntactic-semantic predication proposed by Café (1999) was tested in an attempt to contribute towards the construction of analytical-synthetic systems of knowledge representation. This model of predication was experimentally applied to the construction of schemes of knowledge representation using a corpus taken from the specific terminology used in Plant Biotechnology. The intention was to identify existing conceptual relationships within it and to determine its facets. The research showed that certain rules became constant over the course of the analyses revealing precise scientific standards of classification. In addition, the intense data analysis showed the occurrence of both qualitative and quantitative results in the interpretation of the corpus. It was concluded that by adopting this model it is possible to identify conceptual relationships and determine the basic categories, facets and their indicators, which are essential for the construction of analytical-synthetic constructions. This interpretation of the analyzed corpus indicates the validation of Café's proposal as a model of predication for application in tools of knowledge representation used for information retrieval.
Syntactic-semantic predication; Facetted classification; Complex terminological unit; Analytic-synthetic system
