From relational schemas to subject-specific semantic relations

Compounding is a major word-formation procedure in many languages, and even more so in specialised terminology. The classification of these compound words is a very complex issue due to the large number of semantic relations that can hold between the constituents of the compound. Typologies for different special languages differ considerably from each other and usually combine rather general with highly subject-specific relations. This paper presents a proposal for a two-step classification of these intraterm relations. First, a set of basic relational schemas is worked out, whose purpose is to serve as a tool for the interpretation of semantic relations. These schemas, which are potentially applicable to any domain, are then used to classify the actual compound terms that appear in a corpus of texts from a specific technical field.