Lexical concepts (i.e. semantic units conventionally associated with linguistic forms) are viewed in the article as structures consisting of interrelated facets (i.e. conceptual slots filled with various types of information about the referent) with different structural weight. The paper suggests a way to model the graded structure of lexical concepts by assessing the weight of each constituting facet according to its relevance for defining purposes, frequency of contextual profiling and salience in derivation processes. Thus, the approach taken exploits as many linguistic points of access to the concept as possible and uses three different dimensions to range its facets. The suggested idea is verified with a case study of some common lexical concepts in English (e.g. represented by concrete nouns such as “bird”, “tree”, etc.), which reveals both the advantages and the limitations of the approach taken.