LexInfo: A declarative model for the lexicon-ontology interface

In this paper we motivate why it is crucial to associate linguistic information with ontologies and why more expressive models, beyond the label systems implemented in RDF, OWL and SKOS, are needed to capture the relation between natural language constructs and ontological structures. We argue that in the light of tasks such as ontology-based information extraction (i.e., ontology population) from text, ontology learning from text, knowledge-based question answering and ontology verbalization, currently available models are not sufficient as they only allow us to associate literals as labels to ontology elements. Using literals as labels, however, does not allow us to capture additional linguistic structure or information which is definitely needed as we argue. In this paper we thus present a model for linguistic grounding of ontologies called LexInfo. LexInfo allows us to associate linguistic information with respect to any level of linguistic description and expressivity to elements in an ontology. LexInfo has been implemented as an OWL ontology and is freely available together with an API. Our main contribution is the model itself, but even more importantly a clear motivation why more elaborate models for associating linguistic information with ontologies are needed. We also further discuss the implementation of the LexInfo API, different tools that support the creation of LexInfo lexicons as well as some preliminary applications.