Representing Compounding with OntoLex : An Evaluation of Vocabularies for Word Formation Resources

This paper explores how compounds are represented in resources documenting word formation, and proposes ways to convert them into Linked Open Data using the OntoLex model. The ultimate purpose is to offer a broad empirical evaluation of which of the two OntoLex modules allowing for the representation of compounds {--} Decomp and Morph {--} fits best the different formats and theoretical approaches of the resources we examine. We show that the vocabulary of Decomp alone is rarely sufficient to account for all relevant facts; in almost all cases, it is necessary to resort to the vocabulary of Morph, either to reify the relation between compounds and their constituents or to represent specifically morphological information or other aspects. Special attention is devoted to the format of the Universal Derivations project: the modelling strategy that we propose can be applied to all resources harmonized in that format, potentially allowing for the conversion into Linked Open Data of a large amount of structured data.