Lexically-based terminology structuring: Some inherent limitations

Terminology structuring has been the subject of much work in the context of terms extracted from corpora: given a set of terms, obtained from an ex-isting resource or extracted from a corpus, identi-fying hierarchical (or other types of) relations be-tween these terms. The present paper focusses on terminology structuring by lexical methods, which match terms on the basis on their content words, taking morphological variants into account. Exper-iments are done on a ‘flat ’ list of terms obtained from an originally hierarchically-structured termi-nology: the French version of the US National Library of Medicine MeSH thesaurus. We com-pare the lexically-induced relations with the original MeSH relations: after a quantitative evaluation of their congruence through recall and precision met-rics, we perform a qualitative, human analysis of the ‘new ’ relations not present in the MeSH. This anal-ysis shows, on the one hand, the limits of the lex-ical structuring method. On the other hand, it also reveals some specific structuring choices and nam-ing conventions made by the MeSH designers, and emphasizes ontological commitments that cannot be left to automatic structuring.